Protein-water interactions determined by dielectric methods.

In one of the earliest reports of the physicochemical properties of a protein, Sorensen (1) asked "Does crystallised egg-albumin contain water?" Fol­ lowing a detailed compositional analysis, he found that such samples contain about 0.22 g water per g water-free egg-albumin. In later work, he gives the water content of crystallized hemoglobin as 0.35 gjg (2). We now know that protein molecules can have from 0.20 to 0.70 g strongly associated (bound) water per g protein, and through modern x-ray and neutron diffraction studies, we have for some proteins a detailed knowledge of the location and bonding of much of this water (3-5). By using the latest high-resolution NMR techniques, we can even identify individual molecules of hydration water and characterize their binding sites on the protein molecule (6). Thus, in general and as a convenient method of classification, two kinds of water molecules associated with proteins have been identified: internal water and peripheral water. The internal water molecules, which form an integral part of a protein structure, diminish local charge-charge inter­ actions and reduce destabilizing effects that arise from otherwise unbonded proton donors and acceptors (7, 8). These molecules exchange with bulk water in time scales ranging from tens of seconds to months (9, 1 0). As a rough gauge of the upper limit of internal water, we can consider the small