| Abstract: | A model was constructed that predicts the electric charge of a protein and its isoelectric point from its primary and quaternary structures. By using two different patterns of mutation and purifying selection, four schemes of nucleotide substitution were simulated. In the absence of selection for a specific value of pI, proteins are expected to evolve toward a mildly basic pI. Thus, the selection for maintaining extreme values of pI must be stringent, and proteins with extreme pI's will evolve very slowly. This prediction is consistent with observations on the evolution of histones and ubiquitin. The mean charge change is expected to be about 0.005 units pI per nucleotide substitution. The amount of electrophoretically hidden variation is expected to be considerable even for large degrees of divergence at the nucleotide and amino acid levels. Electrophoretic detectability depends on the size of the protein. The longer the protein the larger the amount of variation at the amino acid level that is undetectable by isoelectric focusing. This property may be partially responsible for the imperfect correlation between molecular weight and gene diversity observed for electrophoretic data. Very basic and very acidic proteins are expected to generate less electrophoretic variability than proteins with intermediate pI's. Unequal rates of mutation between nucleotides and asymmetrical patterns of purifying selection have almost no effect on the equilibrium pI of proteins, but affect the rates of change in pI, and increase the amount of electrophoretically hidden variation in comparison to the expectations derived from random patterns of mutation and constant selection. Comparison of detectability of protein differences among four electrophoretical techniques suggests that the best performance is obtained by the sequential electrophoresis method. |
