Into the Future

---

By Lorin Charlton, PhD and Ryan Leskiw

Although major advances have been made over the past five years, technology challenges continue to limit the field of proteomics. The promise — and the value — that proteins hold for new diagnostics, therapeutic treatments and medicines is predicated on technologies that will not only support the research, but that will also lead researchers to ask questions that are, at present, unanswerable.

The discovery and analysis of proteins and — more importantly — the application of the results of proteomic research is a daunting undertaking, and there is a long way to go before wide-scale protein identification is complete, or even possible. Proteins are far more numerous, labile and changeable than genes, and they can be much more difficult to locate and measure. In spite of this, new and creative enhancements to existing technologies — as well as exciting new technologies — are being developed to enhance our understanding of the myriad roles that proteins play in human health. This article provides a very brief overview of the current state-of-the-art proteomics technologies as well as a description of some of the fascinating new research that will undoubtedly pave the way for future proteomics research.

Two-dimensional gel electrophoresis continues to be the workhorse tool in proteomics, and it has become increasingly useful. But 2-D gels do not always provide reliable quantitative results, and they cannot identify certain classes of proteins, including membrane, highly acidic or basic, and low-abundance proteins. ICAT (isotope-coded affinity tags) methodology has helped to address some of the problems with 2-D gels, and provides a new method for quantification and identification of proteins from complex samples using mass spectrometric analysis. ICAT could theoretically make any separation technology quantitative, including high-performance liquid chromatography and capillary electrophoresis. Another new development is the protein biochip, also called a protein or antibody array. Protein biochips can help measure protein-protein interactions, protein-small molecule interactions and enzyme-substrate reactions. They may also allow for differential profiling, such as distinguishing the proteins of a healthy cell from those of a diseased cell and for subsequently identifying potentially relevant biomarkers.

Mass spectrometry (MS) methods, however, remain the cornerstone proteomics tool for characterizing proteins. MS identifies large numbers of proteins and peptides, and as the technology continues to be refined, it holds the promise of increasingly rapid identification and characterization of proteins. MALDI-TOF, released in 1990, was the first system to provide rapid protein identification by using peptide mass fingerprinting. Newer technologies include the MALDI-quadrupole time-of-flight mass spectrometer (MALDI-QTOF-MS). This instrument allows a novel kind of data-dependent analysis to be performed on ICAT samples. By setting the instrument to select ions for collision-induced-dissociation (CID) based on the ICAT ratio, all peptide pairs that present a “biologically” uninteresting ICAT ratio (e.g., 1:1) can be ignored, while those that display a ratio of interest (e.g., 10:1) can be automatically targeted for CID. In this way, the data can be more rapidly and efficiently mined. Other new MS systems include the MALDI-time-of-flight/time-of-flight mass spectrometer (MALDI-TOF/TOF-MS), micromass ULTIMA QTOF (mLC-ESI-QTOF-MS) and the MALDI-quadrupole ion trap-time-of-flight mass spectrometer (MALDI-QIT-TOF-MS). Each of these has specific applications and enhanced specifications.

The Protein Engineering Network of Centres of Excellence (PENCE) has funded a number of proteomics technology development research projects as part of its mandate to advance the technological innovations that are critical for supporting such research. These networked, interdisciplinary projects include state-of-the-art technologies such as protein chips, high throughput mass spectrometry and bioinformatics in an effort to advance our knowledge of technologies, allowing us to ask questions previously impossible to address. Some of the more innovative projects and technologies are described below.

NMR-BASED TECHNOLOGIES
Structural proteomics (or structural genomics, as it is sometimes called) has become a cornerstone in proteomics research. Delineation of the structure of proteins can assist the determination of their function, activity and probable interactions. Structural proteomics can also aid in the rational design and identification of potential drug targets. Much structural research has historically been carried out using X-ray crystallography. However, NMR (nuclear magnetic resonance) spectroscopy is playing an increasingly important role. Recent improvements in NMR technology (e.g., cryoprobes, improved coil designs, 900 MHz instruments) along with advances in NMR methodology (e.g., TROSY (transverse relaxation-optimized spectroscopy) selective labelling) are making it possible to determine the structures of increasingly larger proteins more accurately and more rapidly than ever before.

PENCE researchers have been actively developing methods to not only extend the limits of NMR protein structure analysis, but to also greatly accelerate the process of assigning, calculating and refining protein structures. These efforts are already having a significant impact on structural proteomics across Canada and around the world.

For instance, Lewis Kay, PhD of the University of Toronto (U of T) (Toronto, ON) has exploited advanced heteronuclear TROSY techniques to fully assign proteins as large as 723 residues. This new approach shattered the previous NMR world record by nearly 400 residues.

Mitsuhiko Ikura, PhD, also with U of T, has developed a number of high throughput ‘protein-conditioning’ technologies that allow NMR spectroscopists to rapidly screen protein conditions or manipulate proteins to generate optimal NMR spectra.

Brian Sykes, PhD of the University of Alberta (U of A) (Edmonton, AB) has developed the NMR Smart-Notebook — a semi-automated NMR assignment tool that uses artificial intelligence and built-in NMR expertise to greatly accelerate NMR assignments.

Several new software tools developed by David Wishart, PhD at U of A allow proteins to be automatically assigned in less than one minute, using only a single NMR spectrum. His research group has also developed programs that permit the rapid generation (< 30 seconds) of protein structures using only chemical shift data.

Other members of the PENCE NMR technology team (Lawrence McIntosh, PhD at U of A and Julie Forman-Kay, PhD at U of T) have worked to refine and apply the techniques developed by their colleagues to a number of important protein-folding and structural proteomics problems.

MASS SPECTROMETRY
In the post-genomic era, the field of proteomics faces two major challenges. The first is assigning structure and function to the tens of thousands of proteins encoded by prokaryotic and eukaryotic genomes. The second challenge is the accurate quantitative analysis of changes in protein levels/activities that occur within a proteome as a response to biological perturbations. Such perturbations can occur in a dynamic cellular system either as normal developmental and metabolic changes or as abnormalities associated with disease. To tackle these challenges, new technologies are required that permit the facile activity-based analysis of many proteins in parallel within samples of high complexity.

One such strategy, being developed by David Perrin, PhD and Stephen Withers, PhD at the University of British Columbia (UBC) (Vancouver, BC), combines quantitative mass spectrometry-based detection and the use of specific biotinylated mechanism-based inactivators or affinity labels to profile all enzymes of a particular reaction mechanism/specificity within a proteome (Scheme 1, page 7). These reagents target specific enzymes such as glycosidases or cysteine proteases within a proteome and react with catalytic residues in the active sites of these enzymes forming covalent adducts. The proteome is then proteolyzed, resulting in a mixture of tagged and untagged peptides. The tagged active-site peptides are then affinity-isolated using biotin/avidin technology and detected by mass LC/MS. The sequence of these peptides can then be determined via MS/MS. Since the sequence of an active-site peptide is highly conserved within an enzyme family, a database search can then identify the enzyme family from which the tagged peptide originated.

PROTEIN EXPRESSION
Most high throughput structural genomics initiatives have exploited bacterial over-expression systems to produce proteins for structural analysis, because of the rapid production time. However, many targets of interest are not amenable to bacterial systems because of post-translational modifications or the need for higher-level folding components. To start bridging this gap, PENCE researchers, led by David Rose, PhD of U of T, are developing Drosophila cell over-expression systems for structural analysis. While slower than bacteria, these cells are much closer to mammalian systems for protein synthesis and folding. This approach is particularly appealing when studying Drosophila proteins as models for their mammalian counterparts, as these can then be expressed in their native cell types. The research team has made modifications to the expression system to allow for protein secretion or localization within the cell, antibiotic selection of transformed cells, and the incorporation of selenomethionine for X-ray crystallographic phasing.

OTHER PROJECTS
Many of the PENCE researchers are also working on advances in proteomics technologies outside of the network-funded projects. For example, the laboratory of Andrew Emili, PhD at U of T has developed an integrated series of computer-based algorithms, statistical methods and software applications that allow for reliable isotope- and chemistry-free quantitative profiling of complex protein mixtures using basic liquid chromatography-tandem MS procedures. The tool kit automatically quantifies the peak intensities of hundreds of peptides, compensates for residual signal noise and fluctuations in peptide retention times, and correctly aligns and matches related peaks in different proteomic data sets. These advances greatly facilitate the routine examination and comparison of the relative abundance of hundreds of distinct proteins in different biological samples without the need for tedious and expensive labelling procedures. This enables researchers to quickly probe large numbers of clinical samples with sufficient precision so as to maximize the opportunities for biomarker discovery.

Another recent breakthrough that PENCE researchers have been involved in is the development of the UBC CLIP CHIP, the first dedicated complete human protease and inhibitor DNA microarray chip. Two critical genomic and proteomic challenges are the profiling of the transcriptome and expression pattern of every active protease, and the identification of all the substrates of every protease in a particular cell, tissue or organism at any particular time. Temporal and spatial localization of a protease with its substrate is essential for substrate recognition and cleavage in vivo. Determination of the relevance of potential substrates can be obtained by DNA microarrays, which can reveal the co-ordinate expression of a protease with its substrate in tissues.

In addition to the new ICAT approaches that have been adopted to profile substrates in cell culture, PENCE researcher Dr. Chris Overall of UBC has approached the genomic side of the equation by making the first complete human protease and inhibitor chip, termed the CLIP CHIP. In collaboration with Dr. Carlos Lopez-Otin of the University of Oviedo (Oviedo, Spain), Overall annotated and compared the entire human and mouse protease genomes (Puente et al. 2003 Nature Reviews Genetics 4, 544-558). The CLIP CHIP has now been designed and version 1.1 has been printed on glass slides. In January 2004, version 1.2 will be made with a redesigned layout that features two sets of the array spotted per slide, with each protease and inhibitor spotted in triplicate. In addition, a secondary custom array is located adjacent to each protease array to enable custom spotting of cDNAs of interest for each intended project. For example, the first analyses of human breast carcinomas also have key oncogenes and tumour suppressors specific to this disease spotted in order to relate their expression with individual proteases. Updates on the CLIP CHIP, along with tables of the complete human and murine protease degradomes, and the papers cited here are downloadable from www.clip.ubc.ca.

While there is much to do to perfect the methods and technologies that will allow us to rapidly and effectively study the entire human proteome, research projects such as those outlined here are providing a strong foundation to build upon. Based on the strength and innovation of its research community, Canada is well positioned to be at the forefront of the development of new technology platforms for proteomics. By continuing to push the envelope on several different fronts, these researchers and their teams are helping the promise of proteomics to become a reality.

Lorin Charlton, PhD is manager of Scientific Affairs at PENCE. Ryan Leskiw is manager of Logistics and Communications at PENCE.