Abstract: MATH/CHEM/COMP 2002, Dubrovnik, June 24-29, 2002

 

 

Prediction of the pore domain of

P-segment-containing ion channels

 

Ana Jeroncic and Davor Juretic

 

Department of Physics, Faculty of Natural Sciences, Mathematics and Education, University of Split, N. Tesle 12, HR-21000 Split, Croatia

 

 

 

In the last few years a new word, channelopathy, has entered the medical and scientific vocabulary. It describes majority of well-known human and animal neurological disorders that result from defects in ion channel function.

A subclass of these proteins, the so-called P-segment-containing superfamily of ion channels includes K+, Na+, Ca2+ and cyclic nucleotide gated channels. The group is characterized by the presence of 2TMH+P-segment pore-forming domain, with the P-segment being the membrane-reentering loop between the two transmembrane helices (TMH). Structurally, P-segment is shown to contain a short α-helix and extended structure in which selectivity filter function resides.

We introduce a new method for identification of P-segment superfamily members based on prediction of structural core location (2TMH+P) in the protein primary sequence. Preference functions, originally derived for determination of transmembrane segment location and implemented in our program SPLIT, are used for this purpose. Depending on its hydrophobic local environment, every amino acid residue is joined with corresponding preference in order to predict membrane-associated helix or β-sheet secondary structure, or to form either turn or undefined structure. We show that this approach is also valid for prediction of membrane unilayer-spanning secondary structure such as P-segment helix, as well as for pinpointing some local aspects of structure such as the flexibility of selectivity filter. The preference profiles together with training-databases derived set of rules are used to determine presence of 2TMH+P motif in the sequence.

Compared with other methods in the field our approach is somewhat unique. Aside from HMM based Pfam model for ion transport proteins, this is the only method so far, used to identify members of the whole superfamily. Also, unlike others, we do not use primary sequence similarity as input information. Potential gain of this approach is isolation of yet
undetermined ion channels with no sequence similarity with known families. Of particular interest are prokaryotic ion channels, since these proteins have shown to be good models for investigation of ion channel structure and function.