Aromatic Residues in Proteins: Re-Evaluating the Geometry and Energetics of π-π, Cation-π, and CH-π Interactions

Abstract

Aromatic residues can participate in various biomolecular interactions, such as π-π, cation-π, and CH-π interactions, which are essential for protein structure and function. Here, we re-evaluate the geometry and energetics of these interactions using quantum mechanical (QM) calculations, focusing on pairwise interactions involving the aromatic amino acids Phe, Tyr, and Trp and the cationic amino acids Arg and Lys. Our findings reveal that π-π interactions, while energetically favorable, are less abundant in structured proteins than commonly assumed and are often overshadowed by previously underappreciated, yet prevalent, CH-π interactions. Cation-π interactions, particularly those involving Arg, show strong binding energies and a specific geometric preference toward stacked conformations, despite the global QM minimum, suggesting that a rather perpendicular T-shape conformation should be more favorable. Our results support a more nuanced understanding of protein stabilization via interactions involving aromatic residues. On the one hand, our results challenge the traditional emphasis on π-π interactions in structured proteins by showing that CH-π and cation-π interactions contribute significantly to their structure. On the other hand, π-π interactions appear to be key stabilizers in solvated regions and thus may be particularly important to the stabilization of intrinsically disordered proteins.

Figure 1

Selected geometric parameters to represent aromatic–aromatic and cationic–aromatic interactions. (A, B) Definition of parameters for an aromatic–aromatic pair, Phe–Tyr, where D is the distance between the centroids of the Phe and Tyr rings, Tθ₁ is the angle of elevation of the Tyr centroid relative to Phe, Tθ₂ is the angle of elevation of Phe relative to the Tyr centroid, and P is the angle between vectors normal to the two rings and yields P ≤ 90°. (C, D) Definition of parameters for cationic–aromatic amino acid pairs involving Lys, Arg, or Tyr. (C) Parameters for a Tyr–Lys pair, where θ₁ describes the angle between the NZ atom of Lys and the normal of the Tyr ring’s plane (represented by the n₁ vector), where D is the distance of the NZ atom from the ring centroid. (D) Parameters for the Tyr–Arg pair are as for panel (C), with the addition of parameter θ₂, which denotes the angle between a vector normal to Arg’s plane and vector D.

Figure 2

π–π interactions between Phe, Tyr, and Trp. Binding energies (rainbow colorbar) from QM calculations of pairwise interactions in aqueous solution between Phe–Phe (A), Phe–Tyr (B), Tyr–Tyr (C), Phe–Trp (D), Tyr–Trp (E), and Trp–Trp (F) pairs projected onto density contour gradients (white–brown colorbar) created by counting the frequency of each pair geometry as sampled from 6535 high-resolution PDB structures. The geometries of pairwise interactions are mapped in terms of angular parameters P and Tθ₂, these being two of the four geometric measures required to represent all pairwise interactions (see Figure 1). For the symmetric cases (i.e., Phe–Phe, Tyr–Tyr, and Trp–Trp), Tθ₂ density values also include Tθ₁ to account for randomly choosing one residue to calculate Tθ₁ rather than the other. The quantum calculations were performed on selected pairs from those found in the sampled database. We note that each selected pair underwent energetic optimization, and consequently, its final geometry may deviate from its original starting structure. The pairwise interactions are categorized as π-stacked or CH−π based on distance and angle cutoffs (see Methods for further details). All other geometries are classified as “other”. The size of the symbol represents the D geometric parameter (i.e., the distance D between the centroids of the two aromatic rings). (G) Selected pairwise geometries and their corresponding binding energies for Phe–Tyr pairwise interactions (geometries 1–4; see panel B). These pairwise interactions are minimal energy geometries of Phe–Tyr characterized by CH−π or π–π interactions. Interaction-type categorization follows Figure S1.

Figure 3

Aromatic–aromatic pairs interacting through π-stacking. A configuration for each pairwise interaction between Phe, Tyr, and Trp is shown together with its corresponding binding energy, thereby illustrating the relationship between π–π and CH−π interactions along with their minimal binding energy geometries, categorized according to Figure S1.

Figure 4

Cation−π interactions between Phe, Tyr, or Trp and Arg or Lys. The binding energies (rainbow colorbar) of pairwise interactions in aqueous solvent were calculated for pairs selected from geometries obtained from high-resolution protein structures and projected onto density contour maps (white–brown colorbar) calculated as described in Figure 2. The cation−π interactions are shown for interactions between (A, B) Tyr, (C, D) Phe, or (E, F) Trp π-systems and (left column) an Arg or (right column) Lys cation. The pairwise interactions are categorized geometrically as H-bonding, CH−π, π-stacked, and other. (G) Three interacting Tyr–Arg pairs possess H-bonding, π-stacked and CH−π geometries (geometries 5–7).

Figure 5

Geometry of Arg–Tyr pairs engaged in hydrogen bonding. (A) Mapping of the NH1 atom of Arg relative to Tyr, where the circles represent configurations identified as hydrogen-bonded interacting pairs. The QM energy of these pairs is in agreement with those of the most populated geometries found in the PDB. (B) Lowest energy conformation of Tyr–Arg pairs was identified as H-bonded.

Figure 6

Geometries of cation−π interactions involving Arg cations with Phe, Tyr, and Trp π-systems. (A–C) Maps of the position of the NH1 atom of Arg (shown by circles) relative to (A) Phe, (B) Tyr, and (C) Trp π-systems. The circles correspond to selected pairwise geometries, categorized as cation−π, whose binding energies (rainbow colorbar) were calculated using QM and whose population is shown on the contour map (white–brown colorbar). (D–F) Conformations with the lowest energy are for the Phe–Arg, Tyr–Arg, and Trp–Arg cation−π pairs, respectively.

Figure 7

Summary of binding energies for pairwise interactions involving Phe, Tyr, Trp, Arg, and Lys residues. Average binding energies of pairwise interactions are categorized into three groups (based on geometric parameters): (A) π-stacked (purple) and CH−π (red) motifs involve solely aromatic residues Phe, Tyr, and Trp. (B) Cation−π (blue) and CH−π (red) interactions occur between the π electrons of a Phe, Tyr, or Trp residue and a cation or CH group from a basic Lys or Arg residue. (C) H-bonding interactions between Tyr and a Tyr, Trp, Lys, or Arg residue.

Figure 8

Synergism between pairwise interactions in proteins. QM calculations of pairwise interactions may experience some structural deviation when modeled in isolation compared with modeling in the context of the whole protein. Three examples are shown for pairwise calculations. (A) Pairwise interactions between Tyr189 and Arg293 (colored yellow) extracted from PDB ID 2R24 chain A. The yellow dashed line represents the cation−π interaction, identified by PyMOL, for the QM optimized pair with a distance of 4.7 Å. The calculated QM binding energy of this interaction is −4.3 kcal/mol. (B) Pairwise interactions between Tyr23 and Arg19 extracted from PDB ID 5ZN0 chain A, where the yellow dashed lines represent the CH−π interaction. The calculated QM binding energy of this interaction is −2.8 kcal/mol. (C) Pairwise interactions between Arg252 and Tyr552 extracted from PDB ID 7WNO chain X, where the yellow dashed line represents the simultaneous cation−π interactions of the original structure. The calculated QM binding energy of this interaction is −4.0 kcal/mol.