FAQ

Q: What is an enzyme interaction site?

A: The broadly defined binding or interaction site of an enzyme means the collection of all amino acid residues that interact with any ligand related to enzyme function, including substrates, products, cofactors, and inhibitors. The expressions of “functional site” and “recognition site” are also used in the literature.

Q: What is the difference between the catalytic site and the interaction site of an enzyme?

A: Catalytic sites generally only consist of a few, highly conserved amino acid residues that are responsible for the enzyme action. By contrast, the interaction site usually includes 10 to 20 amino acid positions that interact with various ligands but do not necessarily participate in the catalytic mechanism?

Q: What is a consensus interaction site?

A: The consensus interaction site of an enzyme is based on all available structures of the enzyme and its close homologues, and includes all residue positions that contribute to ligand binding in any of the structures.

Q: There are excellent enzyme databases such as Brenda (http://www.brenda.uni-koeln.de/). Why do I need another enzyme database?

A: BRENDA contains a very large collection of facts related to enzymes, including reaction specificity, functional parameters, substrates, products, and inhibitors. IntEnz (http://www.ebi.ac.uk/intenz/index.html) is a relational database integrating enzyme data from a number of sources, including BRENDA and the ENZYME database (http://www.expasy.org/enzyme/). Interestingly, these extensive databases provide no information on enzyme-ligand interactions, and no resource has been developed that permits the systematic study of all residues involved in ligand binding.

Q: Why should I use PRECISE rather than the Catalytic Site Atlas or CSA database (http://www.ebi.ac.uk/thornton-srv/databases/CSA/)?

A: CSA provides annotation only for catalytic residues, and does not contain any information on the rest of the binding site.

Q: Is there any database similar to PRECISE?

A: The database that comes closest is the Sequence Annotated by Structure (SAS) facility of the PDBsum website (http://www.ebi.ac.uk/thornton-srv/databases/pdbsum/). SAS generates lists of hydrogen bonds and non-bonded contacts from the 3D coordinates in a PDB file. However, results are presented for each enzyme-ligand complex separately, whereas PRECISE finds all relevant structures and collects all interactions for a summary which provides the consensus characterization of the binding site in a homologous family of enzymes.

Q: Why use both PDBsum and PRECISE?

A: PRECISE focuses only on the binding site, whereas PDBsum provides much more information on both the protein and its ligand. As a matter of fact, whenever information on a specific complex is required, we contact PDBsum.

Q: How can I search for interaction data in PRECISE?

A: Search is currently by the PDB ID or by the EC number. We will add a keyword search option, similar to the SearchLite option of the RCSB Protein Data Bank (http://www.rcsb.org/pdb/).

Q: What happens if an incomplete EC number is entered?

A: PRECISE returns a list of all enzymes that are compatible with the query. Clicking on any member of the list, it further expands until a unique EC number is found.

Q: A query with a unique EC number frequently returns several PDB IDs with a additional character. E.g., entering EC number 1.1.1.1 (alcohol dehydrogenase), we get the list as follows: 1B16_A, 1J5R_A, 1JVB_A, and 3BTO_A. What are these?

A: Each of these PDB IDs represents a cluster of dehydrogenases that have the same catalytic function and hence the same EC number, but show little similarity to each other. The fifth character in the ID is the chain identifier. For example, the cluster of 1B16 includes different structures of the alcohol dehydrogenase from drosophila lebanonensis, 1J5R from the bacterium thermotoga maritima, 1JVB from the archaeon sulfolobus solfataricus, and 3BTO represents a large group of dehydrogenases from various mammals. While these enzymes have the same function, they diverged to a degree such that their binding sites should be characterized separately.

Q: What is the result of a query by PDB ID?

A: The result is a list of non-homologous chains that are present in the query PDB file and participate in enzyme action. The user can select any of these chains for analysis.

Q: My query PDB ID does not exist in the PRECISE database. What can be the reason?

A: Most likely the specified protein is not an enzyme. Another possibility is the structure was entered into the PDB after the last update of PRECISE.

Q: How are the homologous enzymes selected?

A: All enzymes in the same cluster of homologues have the same EC number and pairwise BLAST p-value of 10^-40 or less.

Q: What is the main result of a query?

A: The main output page shows the color-coded sequence of the representative of the cluster. The colors indicate the residues that belong to the binding site.

Q: What do the colors mean?

A: The colors represent the total number of interactions found at each amino acid position in all chains of the cluster. The scheme is blue (few interactions) to red (many interactions). The number of interactions by color is indicated at the bottom of the page.

Q: My query did not find any interactions. What happened?

A: The interactions in PRECISE are based on the enzyme-ligand complexes in the PDB, and hence no interactions are found if such structures are not available.

Q: Are ions such as sulfate or phosphate considered as ligands?

A: Yes, if they are present in the PDB. However, these can be separated from other ligand types by the filtering the results.

Q: Why is the database called “Predicted and Consensus Interaction Sites in Enzymes”?

A: We already described the consensus interactions sites, extracted from the relevant structures in the PDB. Since for a large fraction of enzymes no or limited interaction data are available, future releases of the database will also include interactions predicted by solvent mapping, a novel method for determining protein binding sites based on their structure. For reference see Silberstein, M., Dennis, S., Brown, L., Kortvelyesi, T., Clodfelter, K. and Vajda, S. (2003) Identification of substrate binding sites in enzymes by computational solvent mapping. J. Mol Biol. 332, 1095-1113.

Q: How can I visualize enzyme-ligand interactions?

A: At present each interaction in the detailed list of interactions is linked to the relevant file in PDBsum which provides a 2D LIGPLOT representation of the ligand and the surrounding side chains of the protein. While we will retain this link, we plan to add a rotatable 3D representation of the binding site using the Java-based Jmol molecular viewer.

Q: I found an error in PRECISE. What should I do?

A: Based on diverse and continuously updated information, errors in the database are essentially inevitable. Some information may be incomplete or incorrect in the PDB. For example, EC numbers are frequently missing. It is more difficult to decide which chain(s) of a protein belong to a given EC number. Other decision may be debatable. For example, it is far from trivial to decide whether two homologous clusters of enzymes with the same EC number have binding sites similar enough to permit their merger. Finally, we can obviously err when creating the database. For example, some ligands such as infrequently occurring ions may be misclassified. Peptide ligands are also difficult to identify. Therefore we will appreciate any comment concerning the correctness, interpretation, and usefulness of information in the PRECISE database. The contact persons are listed on the webpage.