Mining the Proteome for Biomarkers: Top Down & Bottom Up Strategies

Thomas R. Brown Senior Scientist Oxford Biomedical Research, Inc.

Enormous strides have been made in technologies for the separation and identification of proteins from very complex samples such as serum. Advances such as DIGE [1,2] and MUDPIT LC/MS identification of peptides in complex tryptic digests are supporting wide-ranging efforts to identify novel biomarkers for early detection of human diseases [5-7]. However, the complexity of the human serum proteome – comprised of approximately 100,000 proteins whose concentrations span over 12 orders of magnitude – makes the identification of biomarkers akin to identifying a precious diamond in a vast mine. At least 10 fmol of a protein are needed for identification on the latest LC/MS/MS instruments [8], so for top down proteomic identification of very low abundance proteins, one may need to start with over 200 mg of serum proteins. This is well in excess of the sample loading capacity of available separation technologies.

Top Down Proteomics – Removal of High Abundance Proteins: Relatively few high abundance proteins (HAP), including albumin and immunoglobulins, comprise over 90% of the proteome by weight. In contrast, low abundance proteins (LAP) – including disease biomarkers – are very difficult to identify and characterize due to the predominance of HAP. Therefore, removal of HAP prior to 2D-electrophoesis or LC/MS can significantly improve the odds of identifying novel biomarker proteins [9-12]. A number of products have recently been introduced for the removal of HAP in serum, but they tend to be very expensive and must be recycled, which gives rise to concerns regarding sample carryover and decreasing capacity following recycling.

Figure 1. Two dimesional electrophoretic analysis of 150 µg of normal human serum proteins (top) and of the same quantity of proteins following removal of high abundance proteins using ProteoMine^™(bottom).

ProteoMine^™ is a novel, powerful, yet cost-effective approach to the removal of HAP for proteomics research (see Figure 1). Human ProteoMine^™ 1.0 is an immunoaffinity resin comprised of a proprietary mixture of polyclonal antibodies specific for the most abundant proteins in normal human serum^†. Incubation of serum samples with ProteoMine^™ resin removes the most abundant proteins from the sample, with high recovery of low abundance proteins following centrifugation as outlined in Figure 2.

Figure 2. Removal of high abundance serum proteins using a ProteoMine™ disposable spin column. After using a plastic syringe and adapter to wash the column, a diluted serum sample is added to the prepacked ProteoMine™ spin column. Following mixing and incubation to bind high abundance proteins, the ProteoMine™ spin column is centrifuged and the unbound medium and low abundance proteins are captured during two elution spins. These can be analyzed directly or subjected to additional pretreatments to focus on specificf subsets of the proteome.

Available either in kit form with disposable spin tubes and accessories, or as bulk resin for larger samples, Human ProteoMine^™ 1.0 removes > 95% of the HAP from human serum samples. ProteoMine^™ is convenient, fast, and effective, yet inexpensive enough that these columns need not be recycled. ProteoMine^™ pretreatment increases the quantity of medium and low abundance proteins that can be analyzed by 2D electrophoresis or LC/MS by 20 fold. In contrast to alternative immunoaffinity methods for HAP depletion, ProteoMine^™ columns ensure consistent capacity and performance.

The ~5% of the proteins in a serum sample following removal of HAP using ProteoMine^™ 1.0, still span many orders of magnitude and low abundance biomarkers may not be present at a sufficient concentration to permit identification. This major proteomics issue can be addressed by using ProteoMine^™ technology in conjunction with other methods to better focus on subsets of the serum proteome, including glycoproteins [10], oxidatively modified proteins (see below) or phosphoproteins [12].

Interested in Mining the Rodent Proteome? Rodent ProteoMine^™ resin and spin columns have also been developed for analysis of the proteome in rat and mouse model systems (see Figure 3).

Bottom Up Proteomics: Proteins are substrates for hundreds of post-translational modifications that include glycosylation, phosphorylation, and prenylation. These modifications can be used to select for protein subsets based on specific modifications. For example, proteins with specific types of glycosylation can be isolated by lectin affinity chromatography. Similarly, several methods have been reported for the isolation of phosphoproteins prior to proteomic analysis.

Toxicoproteomics: The oxidation of amino acid residues in proteins that results from oxidative stress is associated with over 50 diseases as well as aging [13]. In addition, the toxicity of most xenobiotics involves a direct or an indirect oxidative component – including cytochrome P450/P450 reductase generated reactive oxygen species as well as infiltration of inflammatory cells that generate oxidative damage via neutrophil myeloperoxidase [14] and other enzymes.

Arguments in favor of targeting oxidized proteins for proteomics analysis and biomarker discovery is driven by several characteristics:
1. Oxidative stress, along with chronic inflammation, are now known to be major risk factors for many diseases. In contrast to other oxidative stress biomarkers (8-hydroxydeoxyguanosine, isoprostanes, etc), the oxidation of specific proteins can be used as differential biomarkers for changes occurring in specific tissues [15. 16].

2. Intracellular oxidative stress may lead to protein fragmentation, increased propensity to aggregate and increased susceptibility to proteolysis. In the latter pathway, mildly oxidatively damaged proteins may be turned over rapidly by the proteosome [17,18].

3. Role(s) in pathogenesis: The oxidation of intracellular proteins does not always result in their rapid hydrolysis. Aggregated or misfolded proteins may escape the proteosome pathway and their presence may play a role in a range of human degenerative conditions dubbed “protein misfolding diseases” that include Alzheimer's disease, light-chain amyloidosis and the spongiform encephalopathies [19]. While mutations in certain genes promote the abnormal processing and accumulation of misfolded proteins, oxidative changes in these proteins have been proposed to be a causative factor [20] or, at the very least, an exacerbating factor [21].

4. Protein oxidation and signal transduction: Redox (oxidation/reduction) reactions play critical roles in cellular regulation from embryonic development through senescence. They generally involve more limited and subtle oxidation resulting in less damage and they require a reductive component as a necessary part of the reversible regulatory process. Oxidized proteins and other post translational modifications (PTMs) may participate in certain signaling cascades that often affect key redox transcription factors such as NF-kB, AP-1 and STATS. The main players in protein oxidative signaling are superoxide and nitrating species and recent advances in MS based proteomics have enhanced our capability to map and identify these PTMs [22-24]. Furthermore, the extent of oxidative modification of a protein may affect its ability to participate in signal transduction modulation, e.g. in peroxynitrite nitric oxide-regulated signal transduction cascades [24].

Oxidized amino acids of particular interest: Well over 100 oxidative modifications of proteins have been reported. Of these, a few specific modifications have been shown to be particularly informative.

1. Nitrotyrosine: Biological nitration of protein tyrosine to form 3-nitrotyrosine (NT) is a phenomenon associated with over 50 diseases and inflammatory conditions [15,24,25]. Using a proteomics approach, more than 40 intracellular NT-immunopositive proteins were identified in an in vivo and cell culture model of inflammation [24]. Our nitrotyrosine immunoaffinity resin uses an affinity purified anti-NT and is specifically designed for proteomic identification of this specific set of oxidized protein biomarkers (see figure 4).

2. Protein Carbonyls: Protein carbonyls may be generated by multiple oxidative mechanisms [25], including oxidative cleavage of proteins in which the N-terminal amino acid of the resultant peptide is blocked by an a-ketoacyl derivative [25]. The oxidation of lysine, arginine, proline, and threonine side chains may also generate carbonyl derivatives [25]. Further, products of lipid peroxidation (including 4-hydroxynonenal, and malondialdehyde) as well as glycation and glycoxidation reaction products can react with proteins to introduce carbonyl groups. Our high affinity antibody to DNP has therefore been incorporated into a protocol for the derivitization and immunoaffinity isolation of proteins containing carbonyl modifications. As for ProteoMine^™ and our nitrotyrosine affinity resin, our protein carbonyl immunoaffinity product is available either as bulk resin or in prepacked disposable spin columns.

Figure 4 . Differentially expressed oxidized proteins in rat serum 24 hrs after acute exposure to (B) a liver specific toxin (1 mL/Kg CCl₄ ) vs (D) a lung specifc toxin (2 ppm ozone). Gel images following isolation of oxidized proteins using nitrotyrosine (NT) immunoaffinity spin columns are shown. Differentially expressed spots are circled. Sera from control animals following parallel NT immunoaffinity isolation (A and C) were used to identify oxidant-induced proteomic changes. Mobility of molecular weight markers are shown between the images and approximate isoeletric points are indicated on top of the images.

Additional precautions needed for deep mining of the proteome: Proteins adhere to a variety of surfaces. Indeed, albumin is commonly used as a blocking agent in immunoassay products.

· Don’t lose your biomarker: As one strips away the high abundance proteins in preparation for proteomic analysis, it is critical to minimize losses of low abundance proteins of interest due to adhesion to plastic tubes, pipet tips and, especially, dialysis membranes. Although there are many products available for the rapid centrifugal concentration of dilute proteins through membranes with various MW cutoffs, for dilute protein solutions this approach can result in an unacceptable loss of the low abundance proteins of interest — either by adhesion to the membrane or loss in the flow through fraction (for lower MW proteins). Therefore to avoid these losses, we recommend the use of volatile buffers (included with our ProteoMine^™ spin column kits) for removal of HAP with ProteoMine^™, followed by lyophilization or centrifugal evaporation to concentrate the samples prior to separation and identification. These methods take a little more time than centrifugal ultrafiltration, but ultimately ensure much better recovery of the proteins of interest and further improve odds of biomarker identification.

· Don’t inadvertently add proteins to your sample: Keratin from hair, skin, dust and latex gloves, is a very common protein contaminant that can easily become the predominant protein by weight in a sample that you have taken great pains to pretreat in order to remove irrelevant high abundance proteins. Similarly, plasticizers and other organics can leach from pipet tips, tubes and other surfaces and can add unnecessary complexity to LC/MS peptide analyses. In addition, there are compounds, such as glycerol and PEG, that can be detrimental to proteomic studies using MS. Careful selection of disposables and processing samples in a clean environment such as within a HEPA-filtered laminar flow hood, can help ensure sample integrity.

^†Patents pending

References:
1. Lilley KS and Friedman DB All about DIGE: quantification technology for differential-display 2D-gel proteomics. Expert Rev Proteomics, 2004. 1: 401-9.

2. Wu WW, Wang G, Baek SJ, and Shen RF Comparative study of three proteomic quantitative methods, DIGE, cICAT, and iTRAQ, using 2D gel- or LC-MALDI TOF/TOF. J Proteome Res, 2006. 5: 651-8.

3. Motoyama A, Venable JD, Ruse CI, and Yates JR III Automated ultra-high-pressure multidimensional protein identification technology (UHP-MudPIT) for improved peptide identification of proteomic samples. Anal Chem, 2006. 78: 5109-18.

4. Chen EI, Hewel J, Felding-Habermann B, and Yates JR, III Large Scale Protein Profiling by Combination of Protein Fractionation and Multidimensional Protein Identification Technology (MudPIT) Mol. Cell. Proteomics, 2006. 5: 53 - 56.

5. Petricoin EF, and Liotta LA Clinical Applications of Proteomics J. Nutr., 2003. 133: 2476S - 2484S.

6. Duncan MW, and Hunsucker SW Proteomics as a Tool for Clinically Relevant Biomarker Discovery and Validation Experimental Biology and Medicine 2005. 230: 808 - 817.

7. Hortin GL, Jortani SA, Ritchie CC Jr, Valdes R Jr, and Chan DW (2006) Proteomics: A New Diagnostic Frontier Clin. Chem., 52: 1218 - 1222.

8. Old WM, Meyer-Arendt K, Aveline-Wolf L, Pierce KG, Mendoza A, Sevinsky JR, Resing KA, and Ahn NG Comparison of Label-free Methods for Quantifying Human Proteins by Shotgun Proteomics Mol. Cell. Proteomics, 2005. 4: 1487 - 1502.

9. Liu T, QianW-J, Mottaz HM, Gritsenko MA, Norbeck AD, Moore RJ, Purvine SO, Camp DG II, and Smith RD Evaluation of Multiprotein Immunoaffinity Subtraction for Plasma Proteomics and Candidate Biomarker Discovery Using Mass Spectrometry. Mol. Cell. Proteomics, 2006. 5: 2167 - 2174.

10. Darde VM, Barderas MG, and Vivanco F Depletion of high-abundance proteins in plasma by immunoaffinity subtraction for two-dimensional difference gel electrophoresis analysis. Methods Mol Biol, 2007. 357: 351-64.

11. Gong Y, Li X, Yang B, Ying W, Li D, Zhang Y, Dai S, Cai Y, Wang J, He F, and Qian X Different immunoaffinity fractionation strategies to characterize the human plasma proteome. J Proteome Res, 2006. 5: 1379-87.

12. Yocum AK, Yu K, Oe T, and Blair IA Effect of immunoaffinity depletion of human serum during proteomic investigations. J Proteome Res, 2005. 4: 1722-31.

13. Finkel T, and Holbrook NJ Oxidants, oxidative stress and the biology of ageing. Nature, 2000. 408(6809): p. 239-47.

14. Gaut YG, Tran HD, Byun J, Henderson JP, Richter GM, Brennan ML, Lusis AJ, Belaaouaj A, Hotchkiss RS, Heinecke JW (2001) Neutrophils employ the myeloperoxidase system to generate anti-microbial brominating and chlorinating oxidants during sepsis. Proc Natl Acad Sci, 98: p. 11961-6.

15. Kuo WN, et al., Protein nitration. Mol Cell Biochem, 2000. 214(1-2): p. 121-9.

16. Brown T “Proteomics Based Method for Toxicology Testing” Patent application 10/341,026 (filed 1/05/2003)

17. Grune T, Reinheckel T, and Davies KJ, Degradation of oxidized proteins in mammalian cells. FASEB J, 1997. 11(7): p. 526-34.

18. Davies KJ, Oxidative stress, antioxidant defenses, and damage removal, repair, and replacement systems. IUBMB Life, 2000. 50(4-5): p. 279-89.

19. Bucciantini M, et al., Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature, 2002. 416(6880): p. 507-11.

20. Giasson BI, et al., The relationship between oxidative/nitrative stress and pathological inclusions in Alzheimer's and Parkinson's diseases(1,2). Free Radic Biol Med, 2002. 32: p. 1264-75.

21. Taylor JP, Hardy J, and Fischbeck KH, Toxic proteins in neurodegenerative disease. Science, 2002. 296(5575): p. 1991-5.

22. MacMillan-Crow LA, et al., Tyrosine nitration of c-SRC tyrosine kinase in human pancreatic ductal adenocarcinoma. Arch Biochem Biophys, 2000. 377(2): p. 350-6.

23. Squier T, Oxidative stress and protein aggregation during biological aging. Exp Gerontol, 2001. 36:1539-50.

24. Aulak KS, et al., Proteomic method identifies proteins nitrated in vivo during inflammatory challenge. Proc Natl Acad Sci U S A, 2001. 98(21): p. 12056-61.

25. Berlett BS, and Stadtman ER Protein Oxidation in Aging, Disease, and Oxidative Stress J. Biol. Chem., 1997. 272: 20313

back to top