Proteomics arrived on the biotech scene a decade ago, promising a new discovery route to disease markers. Much more complex than genomics, it represents the industrial-scale analysis of the vast array of proteins in their myriad configurations. It was assumed that because of their direct relationship to the organism’s phenotype, proteomics would yield a wealth of new targets, immediately exploitable for revolutionary new pharmaceuticals. But there have been many disappointments in the interim, and the technology has been slow to deliver.
The instrumentation on which proteomic discoveries depend are older and still cumbersome, expensive, and complex to operate in the 1990s, yet they are only recently undergoing revolutionary improvements.
Liquid chromatography combined with mass spectrometry is a widely adopted means of studying proteins and has been the focus of much innovation in the past few years. These include redesigned mass spectrometry instrumentation combined with stable isotopic labeling and novel 2-D gel technologies fused with new methods of differential staining. With the introduction of more user-friendly software and affordable pricing, mass spectrometry is being widely pursued for target identification in drug discovery.
However, this rapid expansion has brought concern over the quality of studies published in peer-reviewed journals. A number of scientists at academic and commercial institutions have complained of poor experimental design, use of inappropriate statistical analysis, and erroneous protein identification, which appeared in the published proteomics database. These problems are compounded by an insufficient number of samples when quantitative differences in peptides are correlated with disease states. The questionable validity of many studies has led to calls for standardization of procedures in order to avoid wasting time and resources pursuing false leads.
Focus on Phosphoproteins
Mark McDowall, Ph.D., strategic development manager, MS, and James Langridge, Ph.D., senior manager proteomics business, of Waters (www.waters.com), say the characterization of serum and plasma is particularly fraught with peril given the vast span (a factor of 1011) over which protein concentrations range.
“We’ve looked at a number of depletion strategies,” said Dr. Langridge, “most are quite successful in removing the most abundant, targeted proteins, but it doesn’t get you too much further into the serum proteome; depletion is not a perfect solution. Therefore, we approach the problem from the other direction, developing enrichment strategies for the less abundant proteins that are the targets of your investigations.”
Drs. Langridge and McDowall developed a kit based on their investigations that has just been launched by Waters for the capture of phosphopeptides from biological samples. “If you look in the literature, there are a number of groups doing peptide analysis of serum proteins in which the results are biased toward the high-abundance proteins,” states Dr. McDowall. “You sample the same proteins over and over again, so the peptides you detect are overrepresented, resulting in an inability to detect peptides of interest—a situation we refer to as undersampling.”
As they explained, the classic separation of complex protein mixtures in large proteomic studies was most commonly carried out using 2D-PAGE. But in really complex protein mixtures, such as a tumor cell lysate, there maybe several thousand individual proteins and 2-D gel analysis would be overwhelmed due to the complexity and dynamic range.
Waters introduced the Protein Expression System, which substitutes microcapillary liquid chromatography for 2-D gel separation, with a direct connection to the mass spectrometer. This allows identification of the peptide ions with a high degree of quantitative and qualitative accuracy, according to the company. Protein identification is then achieved via database search using the exact mass of the tryptic peptide and fragment ion information.
The Waters Protein Expression System is also designed to perform as a parallel mass analyzer, in which the system alternates between low and elevated collision energy. This allows the acquisition of two data functions from a single LC-MS analysis. The spectra within each function can then be processed yielding highly specific mass data, providing more accurate identification of the target peptides. The system also provides simultaneous qualitative and quantitative proteomics within a single run, so the sample can be mined exhaustively in silico, and up/down regulated proteins and biomarkers can be unambiguously identified.
Although effective sample depletion is demanding, in some situations the Waters group has successfully removed overabundant proteins from complex mixtures. Dr. Langridge and his colleagues depleted proteins, including albumin, IgG, anti-trypsin, IgA, transferrin, and haptoglobin, from serum of patients with Gaucher’s disease. They used a commercial multiaffinity removal system column (Agilent) for this purpose. They analyzed samples of serum from patients using the Waters Protein Expression System and were able to demonstrate a number of marker proteins that may be useful in identifying affected individuals.
Abundant Protein Concerns
Other companies supplying the mass spec market are concerned with the challenges offered by abundant proteins. “Traditional labeling of proteins in quantitative expression studies use amine- or thiol-reactive coupling chemistries, but if you have proteins overrepresented in your mixture, they soak up all the label,” warns Peter Banks, Ph.D., technology leader at PerkinElmer (www.perkinelmer.com).
“There are simple procedures for removing much of the albumin and IgG from serum or plasma and more sophisticated approaches for getting rid of the 20 most frequent proteins,” he continues, “but if you really want to go after the rarer, more interesting proteins, you need to use selective enrichment technology.”
PerkinElmer recently released the Immuno-catch™ Kit, an immunocapture technology for mass spectrometry. The kit contains streptavidin-coated 96-well microplates along with the buffers for binding and elution of the target proteins. The streptavidin-coating procedure provides high binding capacity for biotinylated capture agents, including antibodies. Moreover, the coating is designed to permit easy elution of captured analytes. These properties provide a high-throughput, multiplex format for the enrichment of specific target proteins and peptides and selective downstream analysis and detection using standard MS and gel technologies, says Dr. Banks.
The Immuno-catch Kit is part of PerkinElmer’s new Affinity-tools™ group of protein enrichment and detection offerings, based upon conventional immuno- and biotin-binding protein capture strategies. Other items include Phos-tools™ for the analysis of protein phosphorylation, comprised of Phos-trap™ and Phos-tag™ for the selective enrichment and detection of protein phosphorylation, respectively and ExacTag™, highly multiplexed isobaric mass tag reagents for the quantification of protein expression.
A number of other companies have introduced products for protein fractionation. Among these is GenWay Biotech (www.genwaybio.com), which developed an immunoaffinity separation procedure that can specifically capture an additional 207 proteins after the top-14 abundant proteins are removed from plasma or serum samples.
According to Sergey Sikora, Ph.D., senior director of business development, the GenWay SuperMix column uses a proprietary sample-preparation solution to further reduce the dynamic range of plasma for detecting the low-abundance proteins. With some manipulation, after removal of the top-14 highly abundant proteins, IgY-SuperMix separates the next concentration level of moderately abundant proteins.
Avian polyclonal IgY antibodies have properties that allow highly specific fractionation of complex protein mixtures. While the HAP partitioning columns remove the most abundant proteins from plasma, the real issue is the next level of moderately abundant proteins. These block access to the effective catch of low abundance proteins orders of magnitude less frequent and the most relevant for therapeutic exploitation.
“To tackle this challenge, we further developed our IgY microbeads system (Seppro® technology) by establishing an IgY-SuperMix column to separate the moderately abundant from the low-abundance proteins,” Dr. Sikora explains. “The SuperMix column was developed by immunizing chickens with a flow-through fraction of the IgY-12 column and constructing the column with affinity-purified IgY antibodies reactive against the moderately abundant fraction.”
Through a collaboration between Guy Poirier (Centre de Recherche du CHUL, Laval University, Québec) and Weijun Qian (Pacific Northwest National Laboratory), the SuperMix columns were applied to fractionating the flow-through fraction of plasma samples by IgY-12 or IgY-14 column, which resulted in a bound/eluted fraction (designated MAP fraction, containing over 200 proteins) and the flow-through fraction, containing the low-abundance proteins. SDS-PAGE profiles of the fractions coupled with liquid chromatography and mass spectrometry demonstrate that IgY-SuperMix is able to identify over 500 moderate- and low-abundance proteins in a one-step process.
Fulfilling Its Therapeutic Promise
Early enthusiasm over proteomics as a discovery tool for new therapeutic strategies has declined over the years. It was initially believed that analyzing the proteome would zero in on the phenotype and its variation in pathologic conditions and this would build new drug models and diagnostics.
However, the real-world situation turned out be much more complex and demanding, and it has proven difficult to establish relationships between variation in the proteome and particular disease states. In fact, proteomics did not immediately yield a cornucopia of therapies in the same manner that the sequencing of the human genome did not open up a vast array of treatment options.
The failure of proteomics to live up to expectations may have been the result of overhype, but it is more likely due to the difficulties involved in generating clean and unambiguous data. This situation is changing now, with the availability of much better instrumentation and an improved understanding of the pitfalls of experimental design. The next few years should establish the application of mass spec and other methodologies at a new level of proteomic investigation.