Electrostatic properties can play a significant role in affecting the properties and activities of proteins, for example influencing how and where various substrates, inhibitors, cofactors, and other proteins bind. If these themselves have a large net charge or dipole, this effect might be particularly significant. While the precise electrostatic potential about a protein involves a detailed and complex calculation and interpretation, one can often get a first clue by examining two very simple overal properties, the net charge and the dipole moment, and how the latter lines up in comparison with key structural features of the protein.

Actually, since proteins in general are not electrical neutral, one should really speak in terms of a position-dependent first moment of charge distribution. However, we chose to refer to this quantity as a 'dipole moment', because this is how the concept is best recognized by the general scientific community. Although this calculated dipole moment is only a rough approximation, due to several simplifying assumptions made, comparisons of calculated dipoles among different proteins can be meaninfful if they are all calculated the same way, eg. if the same degree of ionization of the residue sidechains, the same atom charges, and same centering of the protein within the coordinate system are used. Further, if the protein consists of more than one peptide strand, only those strands comprising the unique set are included in the calculation, because some PDB files contain complete biological units, while many other contain only partial units.

This project was inspired by a recent study of electrostatic properties of cholinesterases and related proteins (see Felder, Botti, Lifson, Silman and Sussman 1998, J. Molec. Graphics and Modelling, "External and Internal Electrostatic Potentials of Cholinesterase Models"; and Botti, Felder, Sussman and Silman 1998, Protein Engineering, "Electrotactins: A Class of Adhesion Proteins with Conserved Electrostatic and Structural Motifs". In this study we noted that cholinesterases appear to have a large dipole moment oriented rougly parallel to a long gorge through which substrate is thought to travel to the active site.

We wanted to know if such dipole moments are actually unusually large. We therefore developed a program to calculate net charge and dipole moment for a large representative set of the PDB, the June 1998 pdb-select list of Hobyuhm and Sander of EMBL-Heidelberg to 45% sequence identity, and assembled these into a database. From the averages and standard deviations of these data, one can then compare the charge dipole moment of any given protein, and see if it is unusually large or small, both absolutely and for a protein of that size (eg. from the ratio of charge or dipole to the number of atoms). Assuming that the distributions are approximately Gaussian, we measure the number of standard deviation units the given charge or dipole is above or below the average. Values near +/- 1 or more are significantly larger or smaller, if +/- 2-3 or more, very much so.

Two additional features were added to this server. First, if the protein consists of several peptide strands, an option is given to calculate the dipole moment for each strand separately. It is often interesting to compare how the dipoles of the individual constituents compare with that of the overal structure. Second, one can specify up to 20 residues, to examine the angle the dipole moment vector makes with the Beta carbon (Alfa-carbon for GLY) of that residue, to get an appreciation of how the dipole lines up against structural features of the protein. For example, for Torpedo Acetylcholinesterase, residue D72, which sits near the top of the catalytic gorge, the dipole makes an angle of about 150^o, which is roughly alligned along it, but in the opposite direction.