The classical approach in proteomics couples two-dimensional gel electrophoresis (2-DE) with post-gel identification by mass spectrometry. While this technique has proved quite efficient, its limitations are now well described (Wilkins et al., 1998). Large-scale proteomics studies indicate that complete analysis of whole cells or tissues is too complex for available technology. There is currently no single proteome analysis strategy that can sufficiently address all levels of the organization of the proteome (Gygi et al., 2000; Santoni et al., 2000). Therefore, to achieve a more complete analysis of a proteome it is desirable to focus on subcellular proteomes (Rabilloud et al., 1998). Cells are compartmentalized, thus providing distinct environments for biochemical processes such as protein synthesis and degradation, provision of energy-rich metabolites, protein glycosylation and DNA replication. The compartmentalized structure of a cell is supported by subsets of gene products that are specifically targeted to particular subcellular structures. Therefore, protein localization is linked to cellular function that requires proteome analysis with subcellular resolution (Dreger, 2003a,b). Although many years have passed since most cellular organelles were initially characterized by microscopy and subcellular fractionation, a complete catalogue of the proteins in each organelle has yet to be obtained. The complexity of eukaryotic cells hinders a single step characterization of the complete proteome which necessitates alternative approaches. While the classical proteomics approach using 2-DE was successful for analyzing the proteome of different organisms, it was also evident that the number of proteins expressed in complex eukaryotic cells largely exceeds the resolving power of 2-DE. To overcome this limitation, subcellular fractionation is of choice and is used by a number of researchers (Jung et al., 2000). Subcellular or organelle proteomics have been recently reviewed extensively by a number of authors (Jung et al., 2000; Schirmer and Gerace, 2002; Brunet et al., 2003; Dreger, 2003a,b; Huber et al., 2003; Taylor et al., 2003; Warnock et al., 2004). The approach has been used for the analysis of synaptic proteins (Walikonis et al., 2000; Phillips et al., 2001), synaptic vesicles (Hartinger et al., 1996) and yeast plasma membranes (Navarre et al., 2002). This also led to a number of comprehensive global organellar proteomics studies. In all those cases, either one or two-dimension gel electrophoresis separation of proteins was used prior to mass spectrometric protein identification. However, for analysis of membrane proteins, gel-based separation was not fully successful and alternative techniques were devised. Whilst proteomics have the potential to define the composition of organelles, it is limited by organellar cross contamination that can arise during subcellular fractionation. Thus the precise localization of proteins can be hindered by difficulties in preparing pure organelles (Brunet et al., 2003; Dunkley et al., 2004). However, comparative proteomics of organellar subfractions can mitigate these problems, as demonstrated by a recent study involving the nuclear envelope (Schirmer and Gerace, 2002). There are also analytical difficulties associated with the monitoring of dynamic changes in the proteome at the subcellular level. This is because the organelles are not fixed entities but rather dynamic structures interacting with each other and remodeling themselves in response to various stimuli. Therefore, analysis of cell organelles in various conditions is required to understand the dynamic nature of integrated cell function (Brunet et al., 2003). Although organelles are thought to be discrete entities with particular cellular functions, complex mechanisms of intracellular communication and contact sites between the organelles make it difficult to evaluate the biological significance of proteins that are usually associated with one organelle but are detected in the proteome of another organelle. The field of genomics provides a list of potential proteins encoded by an organism's genome, while data derived from proteomic analysis can provide further information that allows the assignment of specific proteins to different subcellular structures. In recent years, organellar proteomics has profiled mitochondrial, chloroplast, nucleolar proteomes, uncovered minor Golgi proteins (Taylor et al., 2000), and compared functional states of the Golgi complex (Wu et al., 2000). Future comparative proteomic studies can provide a better insight towards a complete map of all the cellular proteins in each organelle, in each tissue, at each stage of development.