A "minimal essential Mhc" and an "unrecognized Mhc": two extremes in selection for polymorphism.

The high polymorphism of classical Mhc molecules found in mammals is not simply the result of strong selection for pathogen resistance in the recent past, since there are virtually no examples of diseases caused by infectious pathogens for which resistance is determined by particular Mhc haplotypes, and in the best-studied case, a particular aspect of malaria in humans, the selection is remarkably weak. We discuss three possibilities to explain high polymorphism in mammals: accumulating, merging and boosting. The mammalian Mhc is complicated and redundant, so that every Mhc haplotype may give some level of resistance due to multiple classical Mhc genes as well as other disease resistance genes; this frustrates the attempts to demonstrate selection for disease resistance. We have looked at two vertebrate groups that may represent two extreme examples of selection for Mhc polymorphism. Birds, like mammals, have highly a polymorphic Mhc that determines strong allograft rejection. However, chickens have a much smaller, compact and simpler Mhc than mammals, as though the Mhc has been stripped down to the essentials during evolution. The selection on a single Mhc gene should be much stronger than on a large multigene family and, in fact, there are a number of viral diseases for which resistance and susceptibility are determined by particular chicken Mhc haplotypes. We have determined the peptide motifs for the chicken class I molecules from a number of haplotypes, which may explain some disease associations quite simply. On the other hand, salamanders have very low Mhc polymorphism and slow allograft rejection. We have isolated axolotl Mhc molecules and shown that they cosegregate with the locus that determines graft rejection in the axolotl, have only a few alleles and only weakly stimulate axolotl T lymphocytes in mixed lymphocyte culture. We believe that salamanders have classical Mhc molecules but most T cells do not recognize them, so that there is no strong selection for polymorphism.