13 day chick embryo
13 day chick embryo

There are a bunch of really interesting articles I want to talk about, but most will have to wait until after the holidays; with one sister and a brother visiting along with their assorted families, and my own Christmas responsibilities (including a three-year-old who confidently expects a motorcycle for Christmas: “But Santa SAID!”), I don’t have much time for detailed commentary.

One paper I do want to comment on, though, is
Koch, M., Camp, S., Collen, T., Avila, D., Salomonsen, J., Wallny, H.-J. et al. (2007). Structures of an MHC Class I Molecule from B21 Chickens Illustrate Promiscuous Peptide Binding. Immunity, 27(6), 885-899.

I may be mistaken, but I believe this is the first non-mammalian MHC class I molecule for which a crystal structure is available. (Actually, it’s not available, because the information they deposited in the Protein Data Bank is not yet released, even though the paper is out.) There are a number of interesting aspects about this molecule, but one is that it apparently violates one of the aspects of MHC that I considered (without thinking very much about it) to be a fairly fundamental characteristic. As such, it raises a question about MHC evolution that hadn’t previously occurred to me.

MHC class I binds to peptides, so that cytotoxic T lymphocytes can examine them. Mammalian MHC class I alleles (that have been tested) all bind to peptides with a broadly similar mechanism — much of the strength of the interaction comes from generic features that any peptide would have (the main chain) but there are also features that confer a certain amount of specificity to the interaction. Typically, a mammalian MHC class I molecule will bind to peptides that are about 9 amino acids long; two or three of those amino acids will snap, Lego-like, into an appropriate notch in the MHC. (See this post for more detail.) As a result, each MHC class I molecule can bind to many different peptide sequences, but only a limited subset of the large universe of 9-amino-acid long peptides. It’s this specificity that makes CTL epitope prediction possible. (However, it’s unusual but not extraordinary to identify peptides that bind, but don’t match the defined motif. For example, the exhaustive study of LCMV by Kotturi et al1 picked up two non-canonical peptides, out of the the 27 total known peptides for LCMV.)

Blogging on Peer-Reviewed ResearchThe interesting thing about this particular chicken MHC class I allele2 is that it seems to be much sloppier (“promiscuous”) than the mammalian alleles that have been looked at so far: “This molecule has a novel mode of peptide binding that allows peptides with completely different sequences to be presented to T lymphocytes.” Several other chicken alleles seem to also show promiscuous peptide binding, though other alleles probably are more conventional (or at least mammal-like) and have clear motifs.3

As a side note, this MHC class I allele is strongly associated with resistance to Marek’s Disease (a herpesvirus of chickens),4 and the authors here suggest that the promiscuous peptide-binding of B21 allows this allele to present more peptides from Marek’s disease, conferring greater resistance:

.. it is clear that many more peptides from representative Marek’s disease virus (MDV) genes are predicted to bind the BF2*2101 molecule than the MHC class I molecules from haplotypes such as B4, B12, and B15 that do not confer strong resistance to Marek’s disease. On this basis, it is likely that the promiscuous BF2*2101 molecule would have a greater chance of binding key protective peptide(s) than the fastidious MHC class I molecules from the other haplotypes.

Koch et al Fig 2
Chicken MHC BF2*2101

I am rather skeptical about this as the cause of resistance to Marek’s Disease. In the case of rapidly-mutating viruses like HIV, having a broad CTL response may well be correlated with resistance, but for herpesviruses, which are very stable genomically, even a single peptide epitope response can certainly confer protection (at least in mice); and it would be really surprising (even for a very “fastidious” allele) that there would be nothing at all (herpesviruses are big viruses, with lots of proteins) that binds. It’s possible, but I would look at other things first. 5

Back to my original issue. I’ve talked (ad nauseum!) about the evolution of MHC diversity. Very briefly6 MHC genes have by far the most alleles of any genes, and that is true for virtually every species that’s been examined — fish, birds, reptiles, mammals. There are two widely-accepted explanations for this diversity: Overdominance, or frequency-dependent selection. (Frequency-dependent selection seems to have a little more evidence on its side.) Both explanations depend on different MHC alleles presenting different peptide subsets from pathogens.

Now here’s my question, and in light of the chicken BF2*2101 data I don’t have an obvious answer for it. (This may be because of my tinsel-addled state, and maybe once the holidays are over the explanation will be obvious to me.) Why do MHC alleles bind to peptide subsets? What is the advantage of binding to subsets, compared to binding to all peptides? If MHC can bind promiscuously to many different peptides, then pathogens would not be able to escape from the MHC. For example, we would have much greater resistance to HIV, because there would be no escape mutants.

The explanation I used to have is that if MHC binds to all peptides, then there would be too many self peptides presented, and the process of negative selection in the thymus would eliminate too many T cells during their maturation. In other words, MHC alleles have to trade off the ability to present many pathogen peptides, in order to leave gaps in the self-reactive repertoire for T cells to actually develop.

But these B21 chickens can apparently get away with promiscuous peptide binding. How does that work? I am suddenly very curious about thymic selection in these chickens. What kind of TcR repertoire do they have? I wonder if they’ve looked.


  1. Kotturi, M. F., Peters, B., Buendia-Laysa, F. J., Sidney, J., Oseroff, C., Botten, J. et al. (2007). The CD8+ T-cell response to lymphocytic choriomeningitis virus involves the L antigen: uncovering new tricks for an old virus. J Virol, 81(10), 4928-4940. []
  2. BF2*2101[]
  3. Wallny H, Avila D, Hunt LG, Powell TJ, Riegert P, Salomonsen J, et al. Peptide motifs of the single dominantly expressed class I molecule explain the striking MHC-determined response to Rous sarcoma virus in chickens. Proceedings of the National Academy of Sciences of the United States of America. 2006 31;103(5):1434-9.
    and
    Kaufman J, Milne S, Gobel TWF, Walker BA, Jacob JP, Auffray C, et al. The chicken B locus is a minimal essential major histocompatibility complex. Nature. 1999 Oct 28;401(6756):923-925.[]
  4. Briles WE, Stone HA, Cole RK. Marek’s disease: effects of B histocompatibility alloalleles in resistant and susceptible chicken lines. Science (New York, N.Y.). 1977 14;195(4274):193-5.[]
  5. There is a little precedent for the suggestion, in that it’s been proposed that Epstein-Barr virus, a human herpesvirus, evolved to escape binding by a prevalent MHC class I allele in a particular population: de Campos-Lima, P. O., Gavioli, R., Zhang, Q. J., Wallace, L. E., Dolcetti, R., Rowe, M. et al. (1993). HLA-A11 epitope loss isolates of Epstein-Barr virus from a highly A11+ population. Science, 260(5104), 98-100 and several more recent studies. I am not overwhelmed by those studies either, although the latest (Midgley, R. S., Bell, A. I., McGeoch, D. J., & Rickinson, A. B. (2003). Latent gene sequencing reveals familial relationships among Chinese Epstein-Barr virus strains and evidence for positive selection of A11 epitope changes. J Virol, 77(21), 11517-11530. ) is the most convincing.[]
  6. See here, here, here, and here for more detail, if you care[]