FREQUENCY-ANALYSIS OF AMINO-ACIDS IN THE RECOGNITION REGIONS OF T-CELL RECEPTORS

Citation
F. Laraochoa et al., FREQUENCY-ANALYSIS OF AMINO-ACIDS IN THE RECOGNITION REGIONS OF T-CELL RECEPTORS, Biosystems, 39(1), 1996, pp. 77-86
Citations number
47
Categorie Soggetti
Biology
Journal title
ISSN journal
03032647
Volume
39
Issue
1
Year of publication
1996
Pages
77 - 86
Database
ISI
SICI code
0303-2647(1996)39:1<77:FOAITR>2.0.ZU;2-F
Abstract
In immunoglobulins (Igs), key amino acids in the Complementarity Deter mining Regions (CDR) are responsible for maintaining specific conforma tions called canonical structures, In T-cell receptors (TCRs), protein members of the Ig superfamily, the corresponding residues for maintai ning these canonical structures have not been found, In previous studi es we have found in Igs that the frequency of use of amino acids in so me positions of the CDRs follows an inverse power law distribution, wh ile the frequency of amino acids in the rest of the positions of the C DRs follows an exponential law distribution, The positions that follow the inverse power law distribution are precisely those involved in ma intaining the canonical structures, while those positions for which th e distribution fits the exponential distribution are those that should be properly involved in the recognition mechanism, In this paper, whe n the same analysis is applied to the use frequency of amino acids on the CDRs of TCRs, it is found that some positions that have been previ ously identified as having a structural role are those fitting the inv erse power law. That finding combined with the cooperative or long-ran ge interaction properties of systems that follow the inverse power law leads us to propose that the lack of determined key residues in certa in positions is compensated by 'equivalent' residues in other position s within the CDRs in order to maintain the canonical structures. Other positions that follow the exponential distribution are those which ca n be involved in the recognition process, These results coincide with a computer-generated model of TCR/peptide/MHC interaction previously p ublished by the authors.