A computer search was carried out for homologies between HLA‐A and HLA‐B antigen sequences and the sequences of constant and variable regions of immunoglobulins and of all other sequenced proteins. Searches were made both with relatively short peptide sequences from the HLA antigens and with those longer peptide sequences which were available in 1978. Significant homology of HLA antigen sequences to immunoglobulin constant region sequences was found in two cases: (1) a short decapeptide sequence which includes the fourth cysteine residue of HLA‐B7 and (2) an 89‐amino‐acid residue (Ac‐2) C‐terminal fragment of the papain‐solubilized HLA‐B7 molecule. The difficulty of establishing statistically significant sequence homology with relatively short peptide sequences is emphasized by computer‐based comparisons of the decapeptide sequence with randomly generated peptide sequences. It is concluded that statistically significant homology with short sequences can be assured only when extraordinarily high degrees of homology are present and additional constraints are included in the matches, for example, matches at relatively rare amino acid residues such as Cys, His and Trp. The homology of the 89‐amino‐acid residue sequence to constant region sequences of immunoglobulins is as great as or greater than that of β2‐microglobulin. These findings and the unique domain structure involving a disulphide loop of comparable size strongly favour a common evolutionary origin for this region of HLA‐A and ‐B, β2‐microglobulin and immunoglobulin constant regions.
|Original language||English (US)|
|Number of pages||20|
|Journal||Scandinavian Journal of Immunology|
|State||Published - Jun 1980|