Combining Quantitative Structure Activity Relationship (QSAR) Analysis With Experimental Approaches Permits The Rapid Screening Of T Cell Epitopes From Defined Antigens

The study of relationship between the antigen structure and its antigenicity makes progress in molecular immunology and structural immunology, and the latter needs understanding the complex immunological phenomena on the molecular level. It is known that protein antigen exerts its function not through the whole molecule, but exhibits its speciality through its epitopes. So it becomes an important task to screen the epitopes to fasten the vaccine development, and one of the bottlenecks in vaccine research. For a certain protein antigen, there exist B cell epitopes, Th cell epitopes, CTL epitopes and so on, which are related to immunlogical recognition. Among them, there exist dominant, subdominant and cryptic epitopes, protective and neutralizing epitopes, toxic, and pathological and self-antigen cross reactivity epitopes and other epitopes, which play different roles in protective immunity. On the other hand, because the immunological reaction caused by natural antigen does not meet the need to prevent infection, it is essential to screen and modify epitopes to enhance the protective immunity, and these are based on epitope screening and identification. But for a certain protein antigen, if only by using the experimental methods to identify epitopes, it needs to synthesize many overlapping peptides or to split the whole gene to many small segments, and to transfect and express, then to screen and identify epitopes by using peptide binding experiment or immunological function experiments. These approaches take much time and money. Whereas it will be get twice the result with half the effort, if reducing the number of the peptide segments by using epitope prediction in advance, and then identifying the epitopes using the corresponding experiments.In this thesis, we used quantitative structure activity relationship (QSAR) analysis combining with experimental approaches to screen the Th cell epitopes and CTL epitopes, and tried to establish a new technical platform to identify T cell epitopes. Th cell epitopes play an important function in Th cell immunological reaction. When endocytosed by the antigen presence cells (APCs), the foreign antigens are lysised as peptide segments by the lysase in the endosome, and Th cell epitopes bind with the MHC class II molecules. Then the complex is transferred to the surface of the APC to be recognized by the T cell receptor (TCR), and the Th cell is activated to exert its immune regulation function. Generally, Th cell epitopes are composed of 9-25 amino acids, and it is useful for design and development of subunit vaccine to understand the sequence characteristic of Th cell epitopes. So it is of theoretical importance and practical significance to predict and modify the Th cell epitope structure.In order to develop an automated and computerized method for the identification of helper T-cell epitopes, a new set of descriptors called electronic-topological property vector (ζ,in short) was developed from the primary structures of peptides to describe the its structure, based on the distance and the electric property of each amino acid side chain, looking on each amino acid side chain as a pseudo atom. We classified MHC class II restricted helper T-cell epitopes belonging to diverse alleles as two kinds, one included the Th epitopes being predicted and assigned active "+" with value "1", the rests were assigned non-active "-" with value "0". Then a semi-quantitative model was proposed based on the training set of 28 immunogenic peptides including 14 Ad and non-Ad restricted peptide sequences, assigned active and non-active respectively. We used the semi-quantitative model to estimate the activity of each peptide sequence. If the calculated value of the peptide sequence is close to one, it belongs to the MHC II restricted Th cell epitopes to be predicted, and if the calculated value of the peptide sequence is close to zero, then it doesn't belong to the kind of MHC II restricted Th cell epitopes to be predicted. In order to explain the accuracy of the prediction syste...