The Application Of The Artificial Neural Network To The Study Of Proteasome Cleavage

The proteasome plays an essential role in the cleavage in the process of the antigen peptide processing and presenting. In order to further study the specificity of the proteasome cleavage, an artificial neural network is used to build the model of predicting proteasomal cleavage sites. We train the neural network using a ten fold cross-validation and get the prediction accuracy 80.21%. Compared to other models with the same test set, the performance of this model is more satisfying. The specificities of the cleavage sites and its adjacent positions come from the contribution of the amino acids of the samples to the cleavage sites, and it shows the information of proteasome interacting with antigen protein. It demonstrates that the proteasome cleavage of the target protein is selective, but not random.In the eukaryote cell, the ubiquintin-proteasome system plays a pivotal role in the cleavage of the antigen peptide in the process of the antigen peptide processing and presenting in the MHC class I pathway. The ubiquintination was formed by multiple ubiquitin molecules to the substrate with covalent attachment. And then the substrate was transported and degraded in the 26S proteasome.The mammal 26S proteasome is ATP-dependent protein hydrolyzing enzyme complex with 20S catalytic particle in the middle and two 19S regulatory particles in the two sides. The 20S catalytic core of the 26S proteasome has the shape of a barrel made of four rings which are composed of seven differentαsubunitsα₁-α₇ at two outer rings and seven differentβsubunitsβ₁-β₇ at two inner rings at which the catalytic sites locate. Three of theβsubunits contain proteolyses sites, which are sequestered in the hollow interior of the 20S particle. Substrates enter the 20S through a narrow channel formed by theβsubunits, whose N-termini, depending on their conformation, can either obstruct or allow substrate entry and thus function as a gate. This entry channel is narrow and only permits passage of unfolded, linearized polypeptides. The combination of the 20S CP and 19S RP results to the conformation changing of a subunit and opens the entry channel. And the substrate protein enters the 20S catalytic particle so that it can be degraded by the proteasome. In eukaryotes, 20S CP has two regulatory particles, 19S and US. The 20S core proteasome can associate with two regulatory particles. Association with 19S produces the ATP-depending 26S proteasome particles, which is able to recognize and degrade the ubiquintin-conjunctated proteins. 19S RP is responsible for the substrate deubiquinating, unfolding and translocation into the 26S proteasome. The association between the 11S and 20S produced the football proteasome which is not able to catalytize and degrade the substrate protein, but can active the proteasome and contribute to the production of the antigen peptides. The 20S proteasome associates the 19S RP at one side and 11S at another side to form the hybrid proteasome, whose cleavage capability is strengthened, compared to other proteasomes. Furthermore, in vertebrates, three catalytic activities were identified, each associated with distinct subunits of the proteasome. These are chymotryptic-like (ChTL), tryptic-like (TL) and postglutamyl hydrolase (PGPH) activities. This form of the proteasome is called the constitutive proteasome. The stimulation withγ-interferon replaces these three catalytically active sites of the proteasome by alternative subunits. This form of the proteasome is often referred to as the immunoproteasome.The theoretical prediction depending of the cleavage products of the proteasome in vitro have been reported, such as PAProC, MAPPP and NetChop. However, their predicting accuracy and robustness have more potential to improve. The study makes good use of the artificial neural network to construct the model which predicts the specificity of the proteasome cleavage. The model will make great contribution to the further understanding of the MHC class I process of the antigen peptide processing and presenting. It also instructs the the designing and developing of the vaccines of the cancer or tumor.