The Main-Chain Oxygen: Unappreciated Effects on Peptide and Protein Structure

Current limitations in protein structure prediction and design suggest an incomplete understanding of the forces governing protein folding. As such, noncovalent interactions in proteins, particularly hydrogen bonds, have received great attention [1,2]. In common secondary structure patterns like the α-helix and β-sheet, main-chain N–H hydrogen bond donors approach their carbonyl acceptors approximately along the carbonyl bond axis [3], despite conventional wisdom that hydrogen bond energies are maximized when donors approach at 120° to the carbonyl bond axis [4]. This observation can be rationalized using a modern, quantummechanically based model of the carbonyl lone pairs that indicates that the two orbitals differ from the sp-hybridized VSEPR “rabbit ears” assumed commonly. Specifically, one lone pair, approximately sp-hybridized, is oriented along the carbonyl bond axis, while the second, purely p-orbital orients orthogonally (Figure 1). Canonical hydrogen bonds in protein secondary structure therefore often employ the s-rich lone pair; however, the role of the p-type lone pair is less clear. We have previously noted that backbone n→π* interactions are well poised to exploit this p-type lone pair [5]. In an n→π* interaction, the filled p-type lone pair of a carbonyl oxygen interacts with the empty π* orbital of an adjacent carbonyl group, and the mixing of these orbitals releases energy. These interactions have energies generally greater than 0.27 kcal/mol each [6], and are ubiquitous in folded proteins [7,8], particularly in the α-helix [9]. Yet, no analogous role for the p-type carbonyl lone pair has been identified in β-sheets. We now posit that a previously unappreciated hydrogen bond occurs within the backbone of individual residues in β-sheets.