Hydration water and bulk water in proteins have distinct properties in radial distributions calculated from 105 atomic resolution crystal structures

Abstract

Water plays a critical role in the structure and function of proteins, although the experimental properties of water around protein structures are not well understood. The water can be classified by the separation from the protein surface into bulk water and hydration water. Hydration water interacts closely with the protein and contributes to protein folding, stability, and dynamics, as well as interacting with the bulk water. Water potential functions are often parametrized to fit bulk water properties because of the limited experimental data for hydration water. Therefore, the structural and energetic properties of the hydration water were assessed for 105 atomic resolution (<or=1.0 A) protein crystal structures with a high level of hydration water by calculating the experimental water-protein radial distribution function or surface distribution function (SDF) and water radial distribution function (RDF). Two maxima are observed in SDF: the first maximum at a radius of 2.75 A reflects first shell and hydrogen bond interactions between protein and water, and the second maximum at 3.65 A reflects second shell and van der Waals interactions between water and nonpolar atoms of protein-forming clathrate-hydrate-like structures. Thus, the two shells do not overlap. The RDF showed the features of liquid water rather than solid ice. The first and second maxima of RDF at 2.75 and 4.5 A, respectively, are the same as for bulk water, but the peaks are sharper, indicating hydration water is more stable than bulk water. Both distribution functions are inversely correlated with the distribution of B factors (atomic thermal factors) for the waters, suggesting that the maxima reflect stable positions. Therefore, the average water structure near the protein surface has experimentally observable differences from bulk water. This analysis will help improve the accuracy for models of water on the protein surface by providing rigorous data for the effects of the apparent chemical potential of the water near a protein surface.

Fig 1

a. The ideal model of water distributed around a protein molecule. b. Representation of terms. The RDF is the density of water as a function of distance from a particular water molecule w. The actual density of water in the orange shell (UDV) is the number of waters (w1, w2 and w3 represent three of the water molecules) divided by the volume of the UDV. The small cells (e1, e2, e3 etc) are expected to have different water density because they have different distances from the protein surface. The observed density of water in the UDV is the average of the expected density of each small cell. c. The first maximum of an RDF. The sharpness of a maximum of an RDF is defined as HD/HW, where HD=H-D, HW (dashed line) is the width of the peak at height W and W= HD/2 + D.

Fig 1

a. The ideal model of water distributed around a protein molecule. b. Representation of terms. The RDF is the density of water as a function of distance from a particular water molecule w. The actual density of water in the orange shell (UDV) is the number of waters (w1, w2 and w3 represent three of the water molecules) divided by the volume of the UDV. The small cells (e1, e2, e3 etc) are expected to have different water density because they have different distances from the protein surface. The observed density of water in the UDV is the average of the expected density of each small cell. c. The first maximum of an RDF. The sharpness of a maximum of an RDF is defined as HD/HW, where HD=H-D, HW (dashed line) is the width of the peak at height W and W= HD/2 + D.

Fig 1

a. The ideal model of water distributed around a protein molecule. b. Representation of terms. The RDF is the density of water as a function of distance from a particular water molecule w. The actual density of water in the orange shell (UDV) is the number of waters (w1, w2 and w3 represent three of the water molecules) divided by the volume of the UDV. The small cells (e1, e2, e3 etc) are expected to have different water density because they have different distances from the protein surface. The observed density of water in the UDV is the average of the expected density of each small cell. c. The first maximum of an RDF. The sharpness of a maximum of an RDF is defined as HD/HW, where HD=H-D, HW (dashed line) is the width of the peak at height W and W= HD/2 + D.

Fig 2

a. The SDF (blue), water-polar atom radial distribution function (SDFpol, red), and water-non polar atom radial distribution function (SDFnon, green). The number of waters is averaged over 105 crystal structures. The plots are calculated from equation (1). The inset shows the enlarged second shell. b. The model of the first and second water shells on the protein surface. Red and blue beads indicate the polar (or charged) and non-polar atoms on protein surface, respectively. The red and blue arcs represent the first and second shell, respectively. The two shells do not overlap.

Fig 2

a. The SDF (blue), water-polar atom radial distribution function (SDFpol, red), and water-non polar atom radial distribution function (SDFnon, green). The number of waters is averaged over 105 crystal structures. The plots are calculated from equation (1). The inset shows the enlarged second shell. b. The model of the first and second water shells on the protein surface. Red and blue beads indicate the polar (or charged) and non-polar atoms on protein surface, respectively. The red and blue arcs represent the first and second shell, respectively. The two shells do not overlap.

Fig 3

a. The raw RDF (blue) from equation (2) and the water density distribution for normalization (pink) from equation (3) among 105 crystal structures. b. The normalized RDF (red) of water in crystal structures and the RDF of pure bulk water at 298K (green dashed line) .

Fig 3

a. The raw RDF (blue) from equation (2) and the water density distribution for normalization (pink) from equation (3) among 105 crystal structures. b. The normalized RDF (red) of water in crystal structures and the RDF of pure bulk water at 298K (green dashed line) .

Fig 4

(a) The normalized RDF (red line) and RDFB (dotted blue line) calculated from equation (5). (b). The SDF (red line) and SDFB (dotted blue line) calculated from equation (6).