The 20S proteasome is a multicatalytic protein complex that plays an important role in intracellular protein degradation from archaebacteria to eukaryotes. This complex is made up of two copies each of seven different alpha (alpha) and seven different beta (beta) subunits arranged into four stacked rings (alpha7beta7beta7alpha7). Although the proteasome's cylindrical structure is conserved, the subunit composition of the 20S protein complex varies during the evolution, and the number of subunits increases from archaebacteria to mammals. To fully characterize the 20S proteasome subunit composition and understand the subunit functions, we, the authors of this chapter, have developed and employed various mass spectrometry (MS)-based approaches to generate a comprehensive profile of the 20S proteasomes from rat liver and Tropanosoma brucei. We have identified 7 alpha and 10 beta subunits, including 7 essential and 3 nonessential beta subunits from rat 20S proteasome complex using two-dimensional (2-D) gel electrophoresis and tandem MS (MS/MS). In addition, multiple isoforms of most of the subunits were determined; indicating the composition of rat 20S proteasome complex was much more complicated than expected. Further analysis of the intact protein molecular weight of each subunit using LC-MS confirmed the heterogeneous population of the 20S proteasome and revealed that many of the experimental measured molecular weights do not correspond well with the theoretical values deduced from the sequences in protein databases. This finding is mostly due to the sequence errors in the protein databases and possible posttranslational modifications. Although the protein sequences of rat 20S proteasome are present in the databases, the sequences of the 20S proteasome from T. brucei were not available at the time when the analysis was carried out. To determine the subunit composition of the 20S proteasome from T. brucei, we developed a homology-based database searching tool to identify unknown proteins based on the novel sequences determined by de novo sequencing using MS/MS. As a result, 14 subunits (7 alpha and 7 beta) were identified on the 2-D gel, which was later confirmed by the full-length sequences. Using the same approach, we also identified and characterized an activator protein, PA26, from T. brucei. The purified recombinant PA26 self-assembles into a heptamer ring, which can bind and activate the 20S proteasome from T. brucei as well as rat. Compared to the human PA28 complex, PA26 may be the prototype activator protein involved in proteasomal protein degradation. Therefore, the MS-based strategy developed here for identification of the known and unknown protein complexes can be generalized for the study of other protein complexes.