Proteomics in its broadest sense intends to monitor simultaneously the dynamic behavior of all functional aspects of all proteins in a cell, tissue or organism as the study model undergoing a change in conditions. However, at the moment, there is not a single protein-analytical technique available that can generate such a complete overview. The dynamic aspects of protein functioning, including protein quantity, turnover, post-translational modification, complex formation, (trans)location and 3D structure dynamics, are so diverse that near-complete information might only be obtained using different experimental approaches. In addition, the number of different functional proteins in the proteome is estimated to be between 200,000 and 1,000,000. The broad range of biochemical characteristics of such a complex set of proteins requires a number of complementary techniques to be able to view a considerable part of the proteome. As a consequence, a proteomics experiment needs to be cleverly designed with an a priori choice of what information is wanted, and with a sophisticated selection of the techniques that will provide this information in the most optimal way.
One of the popular experimental approaches is protein profiling to find the differential proteins in a study model under different conditions. It is the determination of the relative abundance of a predefined set of proteins by antibody-based multiplex techniques, or of as many proteins as possible using nondirective ‘fishing’ approaches. One of the most popular nondirective methods is the separation of isolated proteins by 2DE followed by the relative quantification of visualized proteins, and by the mass spectrometric identification of the differential proteins.
A number of recent reports have presented novel information with respect to the detected differential proteins. Petrak et al. analyzed the results of 186 2DE experiments published in recent issues of the journal Proteomics and came to the conclusion that some differential proteins are repeatedly detected regardless of the experiment and the tissue or species Citation[1]. A list of the top 15 generally detected proteins and protein families was presented for both humans and rodents. In a separate study, we performed bioinformatic and statistical analysis of the results from 66 studies using 2DE with MALDI-TOF mass spectrometry (MS) covering additional journals and additional species Citation[2]. We reached an almost identical conclusion and presented the top-44 list of generally detected proteins and top-28 list of generally detected protein families. Using a text-mining approach, Ponomarenko et al. searched proteomics reports for frequently occurring protein names Citation[3]. In this way, they again reached a similar conclusion, although they did not select for differentially expressed proteins. shows the frequently detected/reported proteins in at least two of these studies.
Looking for an explanation for this phenomenon, we came to the conclusion that it may partly be due to the use of the 2DE technique, which preferably visualizes the hydrophilic, middle-sized (10–100 kDa) and abundant proteins. However, this preference cannot explain why those proteins change their relative abundance and why this is irrespective of the species, tissue and experimental conditions. It suggests that there should be a plausible biological explanation. DNA microrarray analysis of yeast cells has revealed a set of approximately 900 genes whose expression changes drastically in response to various environmental alterations Citation[4]. This general stress response appears to be evolutionary conserved, and has been proposed to protect cells against macromolecular damage by sudden detrimental changes in the environment. Such a nonspecific stress response might give the cell sufficient time to allow stress-specific adaptation to take place, increasing the survival chance Citation[5]. The corresponding so-called minimal cellular stress proteome consists of a few hundred evolutionarily highly conserved proteins that are involved in sensing membrane lipid, protein and DNA damage, redox sensing and regulation, cell cycle control, macromolecular stabilization and repair, and control of energy metabolism Citation[6]. Since most of the generally detected proteins have a function that complies with one of these cell-survival processes, it is very likely that they also belong to the minimal cellular stress proteome Citation[2]. We noticed that in different experiments, the same protein responded to a change in conditions but not necessarily in the same direction. In some experiments, a protein may increase in abundance, while in an another experiment it may decrease. Apparently, a change in conditions can either increase or reduce cellular stress, and it is possible that the generally detected proteins are markers for the stress level. Cellular stress may be a factor involved in chronic disorders, suggesting that this set of proteins might be used to search for conditions promoting a ‘relaxed’ performance of cells and tissues, thereby promoting health. A total of 20% of the generally detected proteins reported by us are molecular chaperones, indicating that the endoplasmic reticulum (ER) stress proteins form a significant part of the minimal cellular stress proteome Citation[2]. ER stress has already been shown to play a role in the association between obesity and Type 2 diabetes Citation[7]. First attempts with ER stress-reducing compounds in obese mice have been performed Citation[8].
Based on proteomics studies, almost all of the generally detected proteins have been put forward as potential biomarkers for diverse medical conditions, not least in recent years. Just a few examples of this are: non-small-cell lung carcinoma (enolase 1), hepatitis C virus-related hepatocellular carcinoma (enolase 1, HSP70), osteoarthritis (triosephosphate isomerase), nasopharyngeal carcinoma and uveal melanoma (cathepsin D). Since we are probably dealing with nonspecific stress-response proteins, it is obvious that the identification of those proteins as biomarkers specific for an experimental or disease condition should be conducted with extreme care. Rather, specific biomarkers need to be found beyond the listed proteins, which probably will require the use of more profound proteomics approaches.
In examining the protein changes by 2DE analysis of models for neurodegenerative and non-neurodegenerative diseases, Zabel et al. noticed that up to 36% of variant proteins were shared Citation[9,10]. To investigate this further, they compared the proteome of murine embryonic stem cells carrying gene dosage modifications with that of the parental cell lines. They observed that the shared proteins in comparison to cell line-specific proteins are of higher abundance, have fewer interaction partners, and are more polymorphic, probably as a consequence of lower evolutionary constraint Citation[11]. They hypothesized that the shared proteins might belong to a special class of cellular proteins, which have the common function of rebalancing or buffering the dosage effects of many other proteins in protein networks as a consequence of a gene-dosage alteration. As observed for the generally detected proteins, most of these ‘balancer proteins’ do not always change in the same direction of abundance under different conditions. Furthermore, they are enriched for metabolic and stress proteins, suggesting that those proteins are dealing with a form of cellular stress that is induced by altered gene dosage. Comparing the list of 33 ‘balancer proteins’ with our list of 44 ‘generally detected proteins’ shows a rather limited overlap, but when protein families are considered this becomes more obvious , while at the same time pointing out clear differences between the two protein categories. For instance, no molecular chaperones were present among the balancer proteins, suggesting that ER stress is not their target. The interesting conclusion that may be drawn from this is that cells have distinct systems for dealing with environmentally induced stress and genetically induced stress.
Although the data reviewed here all come from 2DE analyses, it can be anticipated that alternative methods of protein separation and quantification, such as liquid chromatography MS/MS with stable isotope-labeled samples, will also produce a list of proteins shared between different experiments. It will be interesting to see whether such proteins also comply with the minimal stress proteome or the balancer concept.
Table 1. Generally detected proteins reported by at least two studies.
Table 2. Overlapping protein families.
Acknowledgements
I thank P Wang from Maastricht University and P Verhaert from the Technical University Delft for helpful suggestions.
Financial & competing interests disclosure
The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
Related Research Data
References
- Petrak J, Ivanek R, Toman O et al. Déjà vu in proteomics. A hit parade of repeatedly identified differentially expressed proteins. Proteomics8, 1744–1749 (2008).
- Wang P, Bouwman F, Mariman E. Generally detected proteins in comparative proteomics – a matter of cellular stress? Proteomics9(11), 2955–2966 (2009).
- Ponomarenko EA, Lisitsina AV, Petrak I et al. Identification of differentially expressed proteins using automatic meta-analysis of proteomics-related articles. Biomed. Khim.55, 5–14 (2009).
- Gasch AP, Spellman PT, Kao CM et al. Genomic expression programs in the response of yeast cells to environmental change. Mol. Biol. Cell11, 4241–4257 (2000).
- Kültz D. Evolution of the cellular stress proteome: from monophyletic origin to ubiquitous function. J. Exp. Biol.206, 3119–3124 (2003).
- Kültz D. Molecular and evolutionary basis of the cellular stress response. Annu. Rev. Physiol.67, 225–257 (2005).
- Özcan U, Cao Q, Yilmaz E et al. Endoplasmic reticulum stress links obesity, insulin action, and Type 2 diabetes. Science306, 457–461 (2004).
- Ozcan L, Ergin AS, Lu A et al. Endoplasmic reticulum stress plays a central role in development of leptin resistance. Cell Metab.9, 35–51 (2009).
- Zabel C, Sagi D, Kaindl AM et al. Comparative proteomics in neurodegenerative and non-neurodegenerative diseases suggests nodal point proteins in regulatory networking. J. Proteome Res.5, 1948–1958 (2006).
- Zabel C. Proteins are specifically altered in disease – really? Expert Rev. Proteomics3, 469–470 (2006).
- Mao L, Zabel C, Herrmann T et al. Proteomic shifts in embryonic stem cells with gene dose modifications suggest the presence of balancer proteins in protein regulatory networks. PLoS ONE2(11), E1218 (2007).