2.2. Literature review of Saccharomyces cerevisiae proteomic analysis
2.2.4. Proteomics in the study of S. cerevisiae networks
The interactions of proteins in a network can be divided in two main systems. These are the interaction of proteins with metabolites (protein-metabolite), and proteins with proteins
(protein-protein). Protein-metabolite interactions are performed via metabolic pathways, in which each protein catalyses the reaction(s) of one (or more) specific substrate(s). A significant number of metabolic pathways in S. cerevisiae have been discovered, with constant updating underway. The source of these metabolic pathways, as well as the functions of each gene/protein can be accessed free of charge at the KEGG database (http://www.genome.jp/kegg/), BioVyc Query Page (http://biocyc.org/server.html), the Comprehensive Yeast Genome Database (http://mips.gsf.de/projects/fungi/yeast.html) and the Saccharomyces Genome database (http://www.yeastgenome.org/). Moreover, the protein localization databases of S. cerevisiae can be found at the Yeast Protein Localization Database (http://ypl.tugraz.at/pages/home.html) or Yeast GFP Fusion Locolization Database (http://yeastgfp.ucsf.edu/), and interaction networks (or protein- protein interactions) can be found at the Biomolecular Interaction Network Database (http://www.xml.com/pub/r/1290), BioGRID (http://www.thebiogrid.org/), or JWS Online (http://www.jjj.bio.vu.nl/database/index.html). Systems biology studies/groups with a focus on S. cerevisiae are growing (see Yeast System Biology Network (YSBN - http://www.gmm.gu.se/YSBN/), the Manchester Centre for Integrative Systems Biology (MCISB), and one can view and download large-scale yeast data sets from Princeton (http://sgdlite.princeton.edu/cgi-bin/download_dataset).
The study of protein-metabolite interactions networks led to the establishment of metabolic engineering in S. cerevisiae. The aim of metabolic engineering is to increase the cells’
production (such as ethanol [72], and glycerol [73]) via optimization of genetic and regulatory processes. Under certain conditions, expression of metabolic pathways will predominate, and therefore, the reconstruction of these metabolic pathways is necessary to aid in deeper understanding and then metabolic engineering. Proteomics can aid in this. For example, Figure 2.11 shows the reconstruction of glycolysis/gluconeogenesis in the relationship with sucrose metabolism and biosynthesis of amino acids. Protein expression changes are depicted there in relationship with metabolites.
As mentioned before, one of the most important roles of proteomics in S. cerevisiae is to characterise both proteins expressions and their functions on a global scale. The interaction of a protein not only provides the functions of proteins, but also plays a key role in the building of complex systems. In the visual representations of complex networks, each protein is presented as a point (or node), and a line is representative for the physical relationships between proteins. To establish a network of protein-protein interactions, much genetic and proteomic information (such as gene and protein expression, transcription factor binding, and PTM locations and types) is needed on the global scale [74].
Figure 2.11. The reconstruction of central carbon metabolism in relationship to the TCA cycle, starch and sucrose metabolism, and the biosynthesis of amino acids (reproduced from [27]).
Most proteins rarely work by themselves, and they always interact with other biomolecules to express their functions, with these interactions being the basis of life. Therefore, by studying protein interaction networks (‘protein interactome’), we can understand the fundamentals of life of a molecular system. Although the genome sequence of S. cerevisiae was completed in 1996 [75], a complete comprehension of the protein-protein interaction networks in S. cerevisiae (or any organism for that matter) is still a big challenge [76]. The first network model established at the proteome level is known as the two-hybrid system.
This system, developed by Fields in 1989 [77], is based on the modular domain structure of the transcription factor GAL4, consisting of a DNA binding domain and transcription activation. In the two-hybrid system approach, a protein of interest (X) is expressed as a hybrid protein with the GAL4 DNA binding domain, and another protein of interest (Y) is expressed by the GAL4 activation domain. If X interacts with Y (two hybrid proteins), transcription of a gene regulated by UASG occurs [77]. This system has high sensitivity in detecting protein-protein interactions in vivo without information of protein molecules [76].
But this model cannot be applied for the interactions of three or more proteins and those depending on PTMs [76]. This model is also poor at determining the interactions of membrane proteins, moreover, in some cases the inference of the interaction withdrawn from this model might be irrelevant to the physiological of cells [76]. However, this model is still used as a standard technique in biology [76]. Two large-scale yeast two-hybrid screens were used to characterise the protein-protein interactions, whereby 957 putative interactions related to 1,004 proteins were detected [78]. A network of 2,358 interactions established among 1,548 proteins was also characterised, and this network provided connections of global protein-protein interaction patterns based on functional groups or localization assignments, and cross-connections [79].
The application of proteomics for the study of protein interactions requires three essential parts, bait presentation, affinity purification of the complex, and the analysis of the bound proteins [80]. Stable isotopic tag methods have recently been used to investigate these
protein interactions. Examples include the application of ICAT for the identification of specific interacting proteins in Pol II general transcriptional complexes [81], or using N-iso tag to the Rad53p complex to investigate the regulation and function of Rad53p [82].
Therefore, these methods can be extended for quantitative analysis of protein-protein interactions, as well as protein phosphorylation of protein complexes. However, until now, most of the large scale studies of protein-protein interactions have drawn a static picture, whilst the dynamic natures of the real networks in S. cerevisiae are rarely reported [83].
The combination of affinity purification with mass spectrometry has provided a powerful tool [84] for the analysis of proteins complexes, which are both expressed at physiological levels and assembled in vivo [80].
The tandem affinity purification (TAP) of affinity-tagged proteins expressed from their locations of the natural chromosomes coupled with tandem MS/MS offers the best coverage and accuracy [85]. In a recent study, to overcome the missing of protein identifications (potentially due to mass spectrometry problems), two independent mass spectrometry methods (MALDI-TOF and LC-MS/MS) were used to increase the interactome coverage and confidence [86]. Finally, 1,623 and 2,001 purified proteins were detected by MALDI- TOF and LC-MS/MS, respectively, resulting in a total of 2,357 purified proteins identified (from the purification of 4,562 different proteins consisting of all non-membrane proteins) from both methods [86]. These data also suggest that the identification of small proteins using LC-MS/MS was more efficient than that using MALDI-TOF, since the small proteins were not visualized by silver stain as well as these proteins out of the 2-DE gels [86].
Moreover, when compared with western blotting methods, the TAP tag method was more sensitive, since 47% of the proteins were detected whilst these proteins were not detected by western blotting [86].