Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Dec 9;137(48):15152-60.
doi: 10.1021/jacs.5b08424. Epub 2015 Nov 30.

Carbohydrate-Aromatic Interactions in Proteins

Affiliations
Free PMC article

Carbohydrate-Aromatic Interactions in Proteins

Kieran L Hudson et al. J Am Chem Soc. .
Free PMC article

Abstract

Protein-carbohydrate interactions play pivotal roles in health and disease. However, defining and manipulating these interactions has been hindered by an incomplete understanding of the underlying fundamental forces. To elucidate common and discriminating features in carbohydrate recognition, we have analyzed quantitatively X-ray crystal structures of proteins with noncovalently bound carbohydrates. Within the carbohydrate-binding pockets, aliphatic hydrophobic residues are disfavored, whereas aromatic side chains are enriched. The greatest preference is for tryptophan with an increased prevalence of 9-fold. Variations in the spatial orientation of amino acids around different monosaccharides indicate specific carbohydrate C-H bonds interact preferentially with aromatic residues. These preferences are consistent with the electronic properties of both the carbohydrate C-H bonds and the aromatic residues. Those carbohydrates that present patches of electropositive saccharide C-H bonds engage more often in CH-π interactions involving electron-rich aromatic partners. These electronic effects are also manifested when carbohydrate-aromatic interactions are monitored in solution: NMR analysis indicates that indole favorably binds to electron-poor C-H bonds of model carbohydrates, and a clear linear free energy relationships with substituted indoles supports the importance of complementary electronic effects in driving protein-carbohydrate interactions. Together, our data indicate that electrostatic and electronic complementarity between carbohydrates and aromatic residues play key roles in driving protein-carbohydrate complexation. Moreover, these weak noncovalent interactions influence which saccharide residues bind to proteins, and how they are positioned within carbohydrate-binding sites.

Figures

Figure 1
Figure 1
Amino acids proximal to carbohydrates in X-ray crystal structures of protein–carbohydrate complexes. Propensities of amino acids (in order of increasing hydrophobicity) in carbohydrate-binding sites from the data set compared to the distribution of amino acids across all proteins in Uniprot. Alternative methods for normalization are given in Figure S1; however, the overall trends shown here are preserved. Color code: white, hydrogen-bonding side chains; gray, aliphatic hydrophobic side chains, including Gly, Pro, Cys and Met; beige, aromatic side chains.
Figure 2
Figure 2
Distribution of aromatic and aliphatic amino acids around carbohydrates. (A–C) β-d-Glc, and (D–F) β-d-Gal. (A, D) α- and β-faces and ring C–H bonds. (B, E) Centers, represented as spheres, of aromatic and aliphatic side chains interacting with the faces of the carbohydrates (i.e., within 6 Å of any carbohydrate carbon or the ring oxygen). (C, F) Proportions of aromatic and aliphatic side chains interacting with the α- and β-faces reported to the nearest carbon atom of the pyranose ring. See Figure S2 for the analyses for all monosaccharides.
Figure 3
Figure 3
Definition of parameters for CH−π interactions and participating amino acids. (A) Parameters used to identify CH−π interactions: CH−π angle (θ, ≤ 40°), CH−π distance (C-X, ≤ 4.5 Å), C-projection distance (Cp–X, ≤ 1.6 Å for His and TrpA; ≤ 2.0 Å for Phe, TrpB, Tyr). (B) Raw-count distribution of aromatic side chains identified making CH−π interactions with carbohydrates. For Trp, CH−π interactions were identified for cases where either the five- or six-membered ring interacts with a CH proton, TrpA and TrpB, respectively, and where the two rings both interact with separate CH protons, TrpA+B. (C) Structure of proteinogenic aromatic amino acids, with corresponding electrostatic surface potentials for the π-systems (highlighted in beige) of the side-chain moieties: indole (Trp); phenol (Tyr); benzene (Phe); imidazole (His). For indole and phenol, the forms as hydrogen-bond donors (H-bonded to water) are shown, as these are predominant in protein X-ray crystal structures. To show the differences in the π-systems, the scale is shown from ≥130 kJ mol–1 (electropositive, blue) through neutral (green) to ≤ −130 kJ mol–1 (electronegative, red).
Figure 4
Figure 4
Relationship between carbohydrate electrostatic surface potential and formation of CH−π interactions. (A) Orthogonal views of a minimized conformation of β-d-Gal, representative of the majority of those found in the database, which has the ω-angle favored by Gal in solution and in protein crystal structures, in stick-model representation with C–H protons numbered systematically. (B) ESP calculated for the minimized conformation. To show the differences in the C–H bonds, the scale is shown from ≥260 kJ mol–1 (electropositive, blue) through neutral (green) to ≤ −260 kJ mol–1 (electronegative, red). This is double that used for the aromatic systems; i.e., similar changes in color here signify bigger differences than in Figure 3C. (C) Juxtaposed aromatic moieties of amino acids engaged in CH−π interactions with β-d-Gal.
Figure 5
Figure 5
Hydroxyl group stereochemistry influences carbohydrate electrostatics and CH−π interactions. (A) β-d-Gal, (B) β-d-Glc, and (C) α-d-Glc. Column 1: Stick models for representative minimized conformations viewed from the α-faces with C–H protons numbered. Column 2: Normalized calculated ESPs for the same orientation of the minimized conformation. The scale is shown from ≥260 kJ mol–1 (electropositive, blue) through neutral (green) to ≤ −260 kJ mol–1 (electronegative, red); as with Figure 4B this is double that used for the aromatic systems in Figure 3C. Column 3: The distributions of aromatic side chains that form CH−π interactions with the monosaccharides. Column 4: Average frequency of involvement of the monosaccharide C–H protons in the CH−π interactions. For complete analyses for all monosaccharides see Figures S6 and S7.
Figure 6
Figure 6
1H NMR chemical shift perturbations in carbohydrate–aromatic interactions in solution. (A) Interactions between methyl glycosides and 7.5 mM indole in D2O. The circle color and size is scaled to represent the chemical-shift change relative to indole-free solutions (Δδ = δindole – δindole-free). From left to right: β-d-Gal, β-d-Glc, and β-d-Man. (B) Δδ shift for H5 and methyl C–H protons of methyl-β-d-Gal versus the Hammett σp parameter of the 5-substituent in a series of substituted indoles. To allow for solubility limitations, all perturbations were normalized to 7.5 mM indole using the linear dependence of chemical-shift perturbation on indole concentration, Figure S9. Linear fits of the data are shown for H5 (gradient = 5.7, R2 = 0.86) and Me (gradient = 2.1, R2 = 0.63). Δδ values were independent of glycoside concentration. ppb = parts per billion.

Similar articles

See all similar articles

Cited by 47 articles

See all "Cited by" articles

References

    1. Transforming Glycoscience: A Roadmap for the Future; National Academies Press: Washington, D.C., 2012.
    1. Ernst B.; Magnani J. L. Nat. Rev. Drug Discovery 2009, 8, 661.10.1038/nrd2852. - DOI - PubMed
    1. Kiessling L. L.; Splain R. A. Annu. Rev. Biochem. 2010, 79, 619.10.1146/annurev.biochem.77.070606.100917. - DOI - PubMed
    1. Solís D.; Bovin N. V.; Davis A. P.; Jiménez-Barbero J.; Romero A.; Roy R.; Smetana K.; Gabius H.-J. Biochim. Biophys. Acta, Gen. Subj. 2015, 18501186.10.1016/j.bbagen.2014.03.016. - DOI - PubMed
    1. Quiocho F. Annu. Rev. Biochem. 1986, 55, 287.10.1146/annurev.bi.55.070186.001443. - DOI - PubMed

Publication types

Feedback