The pH dependence of predictive models relating electrophoretic mobility to peptide chemico-physical properties in capillary zone electrophoresis.

We applied best fitting procedures to capillary electrophoresis (CE) mobility values, measured at varying acidic pH, of a set of 21 peptides with a molecular mass ranging from about 350 to 1850 Da. This method allowed the contemporary measurements of C-terminus and carboxylic group of the side-chain of aspartic and glutamic acid dissociation constants and of peptide Stokes radius at different protonation stages. Stokes radius was related to peptide molecular mass M at the power of a fractional coefficient, and best correlation was found at pH 2.25, the fractional coefficient being equal to 0.68. This value is close to that proposed by R. E. Offord (Nature 1966, 211, 591-593), who suggested a proportionality between the polymer Stokes radius and M(2/3). The coefficient value decreases at higher pH, reaching a value of 0.58 at pH 4.25, corresponding to a mean peptide conformational transition towards more compact structures as a consequence of C-terminus dissociation. The measurement of the dissociation constants of each peptide allowed us to determine the percentage error on peptide charge predictions performed utilizing mean dissociation constants. Even for the charge, the best predictive performance is obtained at the most acidic edge of the range of the pH studied, mainly at pH 2.25. Conclusively, this study shows that the best performance of predictive models for peptide CE mobility is obtainable in the very acidic pH range (2.25-2.50) and in the absence of electroosmotic flow, and that a satisfactory predictive equation of peptide electrophoretic mobility (m2V(-1)s(-1) is given by mu = 85.4(Z/M(0.68))10(-8).