Possible assortment of al and a2 gene segments of human mhc class i molecules

From our aligned nucleotide sequences of human MttC class I rnolecules, we have recently noted that the al sequence of lILA-A6901 is identical to those of A'6801 and A'6602, while their a2 sequences differ by 7 and 16 bases respectively. On the other hand, the a2 sequence of A'6901 is identical to those of A0201, A0206 and A0211, and their al sequences are different by 11.9 and 11 bases respectively. Similarly, HLA-B4102 and B'4501 have the same al sequence, but their a2 sequences differ by 17 bases. The a2 sequence of B'4102 is, however, identical to that of B'4201, but their al sequences are different by 19 bases. There are many other such examples. In addition, pairs of human MHC class I sequences may differ by only l or 2 bases in their al regions, but are different by more than 10 or 15 bases in their a2 regions, or vice versa. These results strongly suggest that the al and a2 gene segments can be assorted, or that the same geue segment can be duplicated or rnultiplicated to be used in different human MIIC class I molecules. Interestingly, this feature is absent from mouse MHC class I molecules.