Estimating the number of bound waters in Gd(III) complexes revisited. Improved methods for the prediction of q-values

Two literature computational methods for the prediction of the number of inner-sphere aqua ligands, q, have been applied to a test set of seven Gd(aminocarboxylate) complexes. The first method is based on the hypothesis that q should be proportional to the solvent-accessible surface area of the ligand-complexed Gd ion (Castonguay, L. A., Treasurywala, A. M., Caulfield, T. J., Jaeger, E. P., and Kellar, K. E. (1999) Prediction of q-Values and Conformations of Gadolinium Chelates for Magnetic Resonance Imaging. Bioconjugate Chem. 10, 958). The second method is based on the hypothesis that q-values can be deduced by examining series of steric energy versus ionic radii plots as a function of coordination number (Reichert, D. E., Hancock, R. D., and Welch, M. J. (1996) Molecular Mechanics Investigation of Gadolinium(III) Complexes. Inorg. Chem. 35, 7013). This study identifies deficiencies in these methods and, with respect to the first method, describes some apparent errors. Although neither method was reliable at predicting q, two alternate approaches based on either molecular mechanics strain thresholds or exposed Gd surface area thresholds are shown to predict observed q-values for all Gd aminocarboxylate complexes in the test set.