Skip To Content
Cambridge University Science Magazine
By reviewing data from mass spectrometry experiments, literature, and the Protein Data Bank (PDB), they found that protein complexes could be explained by three basic assembly steps: dimerization (2 identical subunits bind), cyclization (3 or 4 identical subunits bind), and heteromeric subunit addition (2 unique subunits bind). Based on these acceptable steps, they then constructed a “periodic table” of all allowable protein complexes, and categorized known existing complexes according to this table. The table is organized by number of unique subunits and number of repeats of those subunits. They used observed frequencies of structures to predict probable structures that have not yet been observed experimentally, and were able to correctly predict at least six recently discovered structures that were not included in the original data set.

Researchers identified three simple assembly steps to account for 96% of known protein complexes, and based on these steps, classified the complexes in a “periodic table” that has been able to accurately predict novel structures.

Based on this classification, the researchers determined that 92% of protein complexes are “bijective”, i.e. protein subunits with the same amino acid sequence have the same structural environment, or the same geometric binding arrangement to other subunits. This elegant result is not at all obvious, and hints that there are simple overarching rules for multi-protein structures. The remaining non-bijective structures fell into two categories: 4% had very high predicted error, suggesting these structures were incorrectly assigned. The other 4%, representing non-bijective structures with low error, present an exception that require further study.

This new categorization is useful for many disciplines: It can provide constraints for discovering and characterizing new protein complexes, and it could be used to create guidelines for protein engineering. The introduction of any new classification is a major step in better understanding the natural world, and it is exciting to see an accurate and simple scheme to comprehend the seemingly innumerable set of protein complexes.

DOI: 10.1126/science.aaa2245

Written by Hannah Wayment-Steele.