Many simple, sin.-le-domain proteins fold via first order kinetics, indicative of a single, dominant free energy barrier. Because folding usually involves the burial of hydrophobic side chains, the acquisition of native structure may be associated with a decrease in the heat capacity of the system. If the transition state ensemble involves the burial of hydrophobic residues, the folding rates show a well-known concave downward dependence on temperature, exhibiting a maximum folding rate with respect to temperature. Within the framework of transition state theory, the maximum folding rate for a specific native structure depends simply on the entropic barrier as well as the heat capacity of activation. The latter is related to the mean hydrophobicity when the protein is largely unfrustrated with regard to its stabilizing interactions. As an example, here we show that the maximum folding rate of the three-helix bundle structure of 1prb7-53, the GA module of an albumin binding domain, can indeed be fine-tuned using computational, methods to identify and design structurally consistent mutations that modulate its hydrophobic content. Specifically, we find design that the logarithm of the maximal folding rate depends linearly on the mean hydrophobic content of the designed sequences, where faster folding correlates with higher mean hydrophobicity. (C) 2004 Elsevier B.V. All rights reserved.