The specificity of guinea pig liver transglutamlnase was Investigated by the determination of modification sites within polypeptides and proteins of known sequence and (In some cases) folded structure. It was shown that some globular proteins have substrate properties for transglutamlnase In conformations which resemble their native states. Novel substrate sites for transglutamlnase were determined within the following proteins: 1) bovine B-Iactoglobulln. 2) the Hls3~ Gln388 mutant form of yeast phosphoglycerate kinase. 3) bovine B-caseln. 4) porcine pepsin. Despite the high exposure of many glutamlnyl residues within these proteins only a small fraction of these residues were observed to be reactive towards transglutamlnase. This Is taken to Indicate that features such as the chemical nature of the amino acid side chains In the local vicinity of unreactive glutamlnyl residues strongly determine the specificity of transglutamlnase. When structural models were available for substrates of transglutamlnase, the local secondary structure associated with substrate sites could be assessed. When no such models were available computer based methods were used to predict the local secondary structures associated with these sites. This approach allows substrate sites to be classified according to their local conformational preference Into conformationally flexl ble (type A substrates) and more conformationally restricted (type B substrates). Since diverse amino acid sequences are observed to surround the reactive glutamlnyl residues of many of the non-physiological substrates of transglutamlnase, It was assumed that the glutamlnyl residues within these sequences were probably reactive due to having; 1) a favourable stereochemistry during modification by transglutaminase 2) a lack of Inhibitory features. In order to determine why some exposed glutamlnyl residues were reactive and others not, It was necessary to find features which were present In unreactive sequences but absent In reactive ones. Through the use of this approach an "anti-consensus sequence" motif was Identified. This was based on the observation that exposed glutamlnyl residues which were unreactive towards transglutamlnase often have charged residues within their surrounding sequences. The distribution of allowed/disallowed residues within substrate sequences, together with what Is known concerning the conformational preference of transglutamlnase for its substrates was built Into a preliminary set of "rules'·. These rules may provide a basis for understanding the observed specificity of transglutamlnase. The application of these rules to a number of model systems has resulted In the correct prediction of both reactive and unreactive glutamlnyl side chains within a number of proteins. The demonstration of the substrate properties of the Hls3~ln388 mutant of phosphoglycerate kinase Illustrates the feasibility of Introducing a novel substrate site for transglutamlnase Into a protein using recombinant DNA technology.

published_or_final_version / Physiology / Doctoral / Doctor of Philosophy