Approximately 1110 non-redundant sequence families and therefore potentially distinct folds for ?- helical integral membrane proteins (MPs) are represented in sequence databases. At the same time less than 100 distinct MP folds have been deposited in the protein data bank (PDB). Thus, for around 1000 ?-helical MP families at least one structure remains to be determined to unambiguously assign a fold. While more than 50% of all drugs target MPs it remains difficult to obtain high-quality crystals of MPs, even though there has been spectacular progress in this area in recent years. Even if a MP can be crystallized, not all functional states might be represented, ensemble states are badly represented by static snapshots, and the model might be perturbed by crystallization aides. Alternative experimental techniques such as NMR and EPR spectroscopy can result in structural restraints for MPs. However, datasets typically remain sparse and are affiliated with an error margin. This sparseness of data leads to an increased demand for computational methods to integrate sparse data from multiple technologies and supplement missing experimental information. Here we propose development of a MP structure determination algorithm integrated in the BioChemical software Library (BCL) ?BCL::MP-Fold? with the following highlights: a) Sparseness of experimental data is counter-balanced by integrating multiple experimental approaches with knowledge-based potentials. b) The absence of suitable templates for many MPs is addressed through a de novo folding algorithm. c) The size and complexity limits of current computational methods are circumvented via initial assembly of disconnected sec- ondary structure elements (SSEs) in the trans-membrane region. d) The computational effort is complemented by a paramagnetic tagging strategy including non-natural amino acids to yield NMR and EPR restraints, that overcome limitations of cysteine labeling strategies, circumvent the tedious and error-prone assignment of NMR signals to side chain atoms, and are especially well suited for use in BCL::MP-Fold. Ultimately, BCL::MP-FOLD will help to determine the structure of protonated EmrE, a homo-dimeric ?-helical MP with four trans-membrane spans per protamer. The protonated state of EmrE is an important interme- diate in the transport cycle of the small multidrug transporter with a yet to be determine three- dimensional structure.
This grant develops computational and experimental methods to determine the structure of membrane proteins. Membrane proteins are targeted by around 40% of small molecule therapeutics yet their structure remains elusive. Ultimately, the methods developed herein will contribute to the structure-based design of nov- el and improved small molecule therapeutics.
Showing the most recent 10 out of 98 publications
