Where to attach dye molecules to a protein: lessons from the computer program WHAT IF

Genomic and proteomic projects are producing a flood of data that all requi re interpretation which often is best performed based on a three dimensiona l structure of the molecule(s) involved. These structures can be determined experimentally, or modelled by homology. Because of the complexity of the questions and the heterogeneity of the data, the software used for modellin g proteins must become even more versatile. We describe several case studie s in which the questions asked, the data, and the requirements on the softw are all are very different. It is shown how structural knowledge about a pr otein helps to determine the best place to bind a fluorescent dye. Such dye s are needed to determine protein-protein, protein-DNA interactions or intr insic fluorescence microscopy. Further, using dyes you can trace molecules in the cell and thus get a handle on subcellular localisation. The first ex ample (OCT-1) involves the search for free amino groups in a protein-DNA co mplex. The second example (BPTI) is a case, in which the amino acid distrib ution shows that amino groups are spread all over the structure, so that th e natural structure has to be modified to get an answer. The third example (HFE) involves a model built by homology. In this case the amino group dist ribution can also be predicted. All these studies were performed using the WHAT IF software package. This p ackage is available including source code, documentation, etc. See http://w ww.cmbi.kun.nl/whatif/ (C) 2001 Elsevier Science B.V. All rights reserved.