The Solution Structure of [Cu(aq)]2+ and Its Implications for Rack-Induced Bonding in Blue Copper Protein Active Sites

Abstract

The structure of [Cu(aq)]²⁺ has been investigated by using full multiple-scattering theoretical (MXAN) analysis of the copper K-edge X-ray absorption (XAS) spectrum and density functional theory (DFT) to test both ideal T_d and square-planar four-coordinate, five-coordinate square-pyramidal, and six-coordinate octahedral [Cu(aq)]²⁺models. The best fit was an elongated five-coordinate square pyramid with four Cu−O_eq bonds (2× 1.98 ± 0.03 Å and 2× 1.95 ± 0.03 Å) and a long Cu−O_ax bond (2.35 ± 0.05 Å). The four equatorial ligands were D_2d-distorted from the mean equatorial plane by ±(17 ± 4)°, so that the overall symmetry of [Cu(H₂O)₅]²⁺ is C_2v. The four-coordinate MXAN fit was nearly as good, but the water ligands (4× 1.96 ± 0.02 Å) migrated ±(13 ± 4)° from the mean equatorial plane, making the [Cu(H₂O)₄]²⁺ model again D_2d-distorted. Spectroscopically calibrated DFT calculations were carried out on the C_2v elongate square-pyramidal and D_2d-distorted four-coordinate MXAN copper models, providing comparative electronic structures of the experimentally observed geometries. These calculations showed 0.85e spin on Cu^II and 0.03e electron spin on each of the four equatorial water oxygens. All covalent bonding was restricted to the equatorial plane. In the square-pyramidal model, the electrostatic Cu−O_ax bond was worth only 96.8 kJ mol^-1, compared to 304.6 kJ mol^-1 for each Cu−O_eq bond. Both MXAN and DFT showed the potential well of the axial bond to be broad and flat, allowing large low-energy excursions. The irregular geometry and D_2d-distorted equatorial ligand set sustained by unconstrained [Cu(H₂O)₅]²⁺ warrants caution in drawing conclusions regarding structural preferences from small molecule crystal structures and raises questions about the site-structural basis of the rack-induced bonding hypothesis of blue copper proteins. Further, previously neglected protein folding thermodynamic consequences of the rack-bonding hypothesis indicate an experimental disconfirmation.