Molecular Mimicry - Probability of Mimicry Events

Probability of Mimicry Events

The prerequisite for molecular mimicry to occur is thus the sharing of the immunodominant epitope between the pathogen and the immunodominant self sequence that is generated by a cell or tissue. However, due to the amino acid variation between different proteins, molecular mimicry should not happen from a probability standpoint. Assuming five to six amino acid residues are used to induce a monoclonal antibody response, the probability of 20 amino acids occurring in six identical residues between two proteins is 206 or 1 in 64,000,000. However, there has been evidence shown and documented of many molecular mimicry events.

To determine which epitopes are shared between pathogen and self, large protein databases are used. The largest protein database in the world, known as the UniProt database (formerly SwissProt), has shown reports of molecular mimicry becoming more common with expansion of the database. The database currently contains 1.5 X 107 residues. The probability of finding a perfect match with a motif of 5 amino acids in length is 1 in 3.7 X 10−7 (0.055). Therefore, within the SwissProt database, one would expect to find 1.5 X 107 X 3.7 X 10−7 = 5 matches. However, there are sequence motifs within the database that are overrepresented and are found more than 5 times. For example, the QKRAA sequence is an amino acid motif in the third hypervariable region of HLA-DRB1*0401. This motif is also expressed on numerous other proteins, such as on gp110 of the Epstein-Barr virus and in E. coli. This motif occurs 37 times in the database. This would suggest that the linear amino acid sequence may not be an underlying cause of molecular mimicry since it can be found numerous times within the database. The possibility exists, then, for variability within amino acid sequence, but similarity in three-dimensional structure between two peptides can be recognized by T cell clones. This, therefore, uncovers a flaw of such large databases. They may be able to give a hint to relationships between epitopes, but the important three-dimensional structure cannot yet be searched for in such a database.