New insights into subcomplex assembly and modifications of centrosomal proteins

Recent work has provided a wealth of new molecular and functional information on the centrosome in a wide variety of organisms. Several large-scale RNA interference (RNAi) screens [1-4], proteomics [5-16] and comparative genomics studies [17-20] have identified the molecular components of the centrosome and related structures such as the spindle pole body, cilia and flagella. However, the exact number of centrosomal proteins is difficult to determine since especially the mass spectrometry (MS)-based studies suffer from the fact that true components can be lost and contaminants are included when isolating organelles such as the centrosome. Furthermore, many proteins only transiently associate with centrosomes and less abundant proteins may escape identification due to the limited sensitivity of MS techniques. Nevertheless the list of identified centrosome proteins continuously grows. Centrosome:db, a database devoted to the human centrosome proteome, currently contains a total of 464 genes encoding proteins that stably or transiently localize to the centrosome [21]. Genes were added on the basis of different kinds of evidence, including high-throughput proteomics datasets and Gene Ontology (GO) annotation in Ensembl and the Human Protein Reference Database (HPRD). A second collection of centrosome proteins can be found in the MiCroKit database [22], which extracts proteins from scientific literature that have been shown to localize to the centrosome, midbody and/or kinetochore by fluorescence microscopy in 7 different model organisms. MiCroKit currently lists 445 human centrosome proteins, of which 371 are also present in Centrosome:db. Andersen and colleagues contributed a large fraction of centrosomal proteins (108) to these databases by examining the human interphase centrosome by MS. Their initial analysis resulted in the identification of roughly 500 proteins. In order to discriminate between contaminants and true centrosome components, they used protein correla