Proteasomes, major proteolytic sites in eukaryotic cells, play an important
part in major histocompatibility class I (MHC I) ligand generation and thu
s in the regulation of specific immune responses. Their cleavage specificit
y is of outstanding interest for this process.
In order to generalize previously determined cleavage motifs of 20 S protea
somes, we developed network-based model proteasomes trained by an evolution
ary algorithm with experimental cleavage data of yeast and human 20 S prote
asomes. A window of ten flanking amino acid residues proved sufficient for
the model proteasomes to reproduce the experimental results with 98-100% ac
curacy. Actual experimental data were reproduced significantly better than
randomly selected cleavage sites, suggesting that our model proteasomes wer
e able to extract rules inherent to proteasomal cleavage data. The affinity
parameters of the model, which decide for or against cleavage, correspond
with the cleavage motifs determined experimentally. The predictive power of
the model was verified for unknown (to the program) test conditions: the p
rediction of cleavage numbers in proteins and the generation of MHC I ligan
ds from short peptides.
In summary, our model proteasomes reproduce and predict proteasomal cleavag
es with high degree of accuracy. They present a promising approach for pred
icting proteasomal cleavage products in future attempts and, in combination
with existing algorithms for MHC I ligand prediction, will be tested to im
prove cytotoxic T lymphocyte epitope prediction. (C) 2000 Academic Press.