This is the first of two special advertising sections this year on proteomics. The second will appear in the 3 October issue of Science.

Having nurtured the discipline of genomics in the wake of the successful sequencing of the genomes of humans and other organisms, life scientists have now begun to turn their attention to the next challenge: proteomics. Defined loosely as the study of all the proteins in a cell or tissue, the field operates on a larger canvas than genomics. “If genomics resembles the Matterhorn, proteomics is like Mount Everest,” says Nick Roelofs, senior vice president, marketing and sales for Stratagene. “The analogy works in two ways. Like Everest, proteomics is bigger in dimension. It also causes an oxygen problem because it’s higher.”

The “oxygen problem” stems from the sheer magnitude of scientific factors involved in protein research. “A single gene can create multiple protein products,” explains Sharan Pagano, director of business development at Incyte Corporation. “Given the unexpectedly low number of genes in the human genome predicted by both public and Celera genome projects, it is clear that a great deal of complexity is encoded in the proteome” states Steve Ruben, vice president of protein therapeutics for Celera, a business unit owned by Applera Corporation. “Proteins represent a complex set of biologic entities that are composed of 20 different molecules in virtually an infinite number of combinations. They occupy not only two-dimensional but three-dimensional space and form multimeric functional units. Proteins are expressed at various levels in heterologous systems, have differing stabilities, and are modified through mechanisms such as phosphorylation and glycosylation. This makes it extremely challenging to work with proteins in general and as biologics in particular.”

Jim Jersey, senior scientist and president of Charles River Proteomic Services (CRPS), an independent joint venture between Charles River Laboratories and Australian firm Proteome Systems, extends that thought. “Not only is a protein’s function a result of its primary, secondary, tertiary, and quaternary structures,” he points out. “It is also affected by the protein’s interactions with other proteins. Compared with the genome, which is linear and thus one-dimensional, proteins have four additional dimensions of complexity when viewed from this perspective. It is a vastly more complicated endeavor to understand.”

A Difficult Beast
It is the three-dimensional conformation that ultimately determines how a protein reacts chemically and biologically. “The protein is a much more difficult beast to deal with than the genome, given the complexity of the structure and the extensive modifications that happen to it through its function,” explains Brad Crutchfield, vice president and group manager for life sciences at Bio-Rad Laboratories. “It’s very difficult to apply a set standard such as sequencing and the relationship between DNA and RNA. When you’re looking at proteins, you’re looking at various functions.”

Bob Karol, director of new business development, life sciences, for Waters Corporation, puts the problem succinctly. “Proteomics presents very difficult problems because of the complexity and the amounts of proteins found in sera, cells, and other locations,” he says. “They are very complicated molecules. They aren’t static. And they change as they age.”

Proteins perform the functions that account for a cell’s behavior in its environment. That involves several different structural and functional tasks. Thus proteins can act as the stuff that allows a cell to be motile; they can function as receptors for signals from the environment external to the cell; they can even be the molecules that send signals from one location in a cell to another. In addition, proteins often have other chemicals, such as sugars, attached to them. The additives give them properties different from those possessed by unadorned protein molecules.

Those facts present problems to suppliers who want to devise tools for proteomics. “The tools available to genomics aren’t available to proteomics,” says Jersey of Charles River Proteomic Services. “Scientists need more tools. As a supplier of tools, you have to produce an even wider range,” echoes Roelofs of Stratagene. “The other problem is that finding the best tools and communicating that to the customer presents a very difficult situation. It’s pretty easy for DNA. But there are 50 ways to purify proteins, each of which involves different steps. Communicating to customers to help them solve their problems requires a lot of sophistication.”

Problems and Promise
Stefan Löfås, chief scientific officer of Swedish firm Biacore, highlights specific problems. “The general handling is quite straightforward when it comes to water soluble proteins,” he says. “The big hurdle comes when you go into more complicated proteins, such as membrane proteins. It’s a big bottleneck to purify them and make them soluble. Then, when you try to assemble them from solution onto a surface – to create protein arrays, for example – you get a lot of complications, such as losing the proteins’ activity.”

The effort to overcome those problems and to gain a deeper understanding of proteins has an increasing influence on drug discovery. Increasingly, R&D teams in pharmaceutical and biotechnology companies are buying products designed to handle proteins. “Our proportion of clients from the pharma and biotech industries is increasing because we now provide instrumentation and technologies for looking at the binding of drug candidates to receptors,” says Biacore’s Löfås. In addition, says John Gebler, manager of life sciences chemistry for Waters, “Some pharmaceutical groups are now calling proteomics biomarker discovery.” Jersey of CRPS confirms that. “A lot of proteomics is focused on biomarkers – the needle in the haystack problem,” he says.

Suppliers recognize that it will take a long and complex process to develop solutions to the problems that proteins present. “We’re very well aware of the challenges – the temporal nature of the proteome, the vast number of proteins, their biological diversity, and the dynamic range problem,” says Dick Rubin, director of marketing for Ciphergen Biosystems. “We’re all dealing with that, some more successfully than others. But that number of problems means good business, with no single magic bullet that will solve the complete mysteries of the proteome for at least 10 years, and probably longer.”

The Challenge of Separation
In general, protein scientists use familiar tools to carry out their research. However, several new products and technologies for proteomics have emerged in recent years. The remainder of this article will highlight some of those advances and the ways in which they combine to ease the lives of protein chemists.

Separating proteins provides an immediate challenge for proteomics research. Scientists base the separations on such physical properties of proteins as their sizes, shapes, charges, hydrophobicities, and affinities for other molecules.

Research teams have three main technologies available for separating proteins. Most protein scientists feel comfortable with two-dimensional (2-D) gel electrophoresis, a technique introduced in the mid 1970s specifically for separating protein mixtures. More recently, laboratories have started to use high performance liquid chromatography (HPLC) and mass spectrometry for the same purpose.

The two dimensions in 2-D gel electrophoresis refer to specific gels. The first traditionally includes ampholytes that migrate in the gel to form a pH gradient. That gradient helps to separate the proteins based on the isoelectric point of each protein. This part of the process can routinely separate thousands of proteins in a single gel, but it presents significant technical challenges. Scientists find it difficult to prepare gels that produce consistent results. However, suppliers offer immobilized gradient strips that don’t suffer the drifts in the pH gradient often associated with ampholytes.

Once they have divided up a mixture of proteins in the first dimension, scientists take the separated proteins – often on a strip of material – and place them on a vertical gel electrophoresis unit. This second dimension usually consists of a denaturing polyacrylamide gel electrophoresis, or SDS-PAGE, a tool found in most laboratories. That matrix creates an extremely uniform molecular sieve. Since they are denaturing, the gels unwind the proteins, changing their three-dimensional conformation into more or less linear molecules that can then migrate through the sieve; smaller molecules move through the pores of the gel faster than larger ones.

Practical Problems
The technology presents some practical problems. “It is based on a very tried and true method,” says Crutchfield of Bio-Rad. “It works and gives a very recognizable data stream and it doesn’t involve a lot of equipment. The challenge is that there’s a fair amount of art in consistently running experiments when you want to draw comparisons.” Indeed, preparing gels that produce consistent results even in the same laboratory is not easy.

Suppliers have improved the original design by producing immobilized pH gradient strips that are not subject to the drifts in the gradient often associated with ampholytes. Also, precast polyacrylamide gels and/or isoelectric focusing strips provide consistent results and allow scientists to run separations without having to master the art of casting the gels.

Several companies, including Amersham Biosciences, Invitrogen, and Proteome Systems in addition to Bio-Rad, offer complete systems for protein separations. “We try to make separation simple to carry out,” says Crutchfield. “The real key is the validation of all the equipment and reagents. Our 2-D starter kit, part of our ReadyPrep system, allows our customers to ensure that everything is working before the experiment starts.”

From HPLC to Mass Spec
HPLC represents the first alternative to 2-D gel electrophoresis. “Most of the low-hanging fruit has been picked by 2-D gel technology,” says Karol of Waters. “HPLC is orthogonal to it. It separates small peptides and large peptides and proteins.”

In ion exchange chromatography, the method with most value in proteomics, charged particles in a solid matrix held in a stainless steel column bind reversibly to proteins or other molecules. The bound molecules are then detached by increasing the salt concentration or by altering the pH of the mobile phase.

In the past, scientists took pride in preparing their HPLC equipment. “But now only the hard-core R&D people at the cutting edge make their own nanoscale [75 micrometer] columns,” says Gebler of Waters. “And no one now would dream of packing their own analytical HPLC columns.” Instead, most scientists who use HPLC rely on Waters and companies such as Amersham Biosciences and Beckman Coulter to provide the basic tools for their work. “We recently launched 75-micron and 100-micron columns,” says Gebler. “We’re also involved in two-dimensional liquid chromatography in which the proteins are separated by charge in one dimension and by hydrophobicity in the other,” adds his colleague Karol.

Liquid chromatography doesn’t stand alone. Increasingly, protein scientists use it in combination with mass spectrometry. “We’re trying to make the liquid chromatography-mass spectrometry process seamless,” says Karol.

Identifying Individual Proteins
The value of mass spectrometry, supplied by such companies as Applied Biosystems, Thermo Electron, and Waters, stems from its success in identifying individual proteins. “Its ability to analyze proteins is about an order of magnitude more sensitive today than the ability of an investigator to analyze them,” says Ian Jardine, chief technology officer for life and laboratory sciences and vice president of mass spectrometry at Thermo Electron. “The instrument provides very specific information. It can tell you exactly what peptide weights, amino acid sequences, and posttranslational modification you have.”

The key to effective use of mass spectrometry is the input. “One of the main challenges is purifying protein samples before they go into the mass spectrometer,” Jardine explains. In addition to liquid chromatography, scientists can use 2-D gel profiles as the initial separation method prior to mass spectrometry and matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) techniques. Thermo Electron has recently formed an alliance with Amersham Biosciences that links its Finnigan line of mass spectrometers with Amersham’s 2-D gel technology.

New methods of mass spectrometry for use in protein research continue to appear. Thermo Electron recently signed an agreement with Cell Signaling to develop an internal standard for use in tandem mass spectrometry to determine specific quantities of peptides. And in March, says Jardine, “We introduced a hybrid mass spectrometer that uses an ion trap and Fourier transform technology. Peptides are trapped and fragmented in the ion trap and injected into a magnet where Fourier transform mass spectrometry is done.”

Waters has also developed new methods of separating proteins. “We have a series of MALDI devices, including kits for sample preparation, that we’re continuing to expand and work on,” explains Karol. “And we recently introduced the Alliance Bioseparations system. That is a biocompatible, no-compromise system designed to run on the same system as high performance peptide maps and multidimensional protein separation using ion exchange.”

The Impact of Chips
In past years, DNA microarrays revolutionized the field of genomics. Now protein microarrays, also known as protein biochips, are having similar impact on proteomics. While complementary DNA microarrays can measure the levels of messenger RNA expressed in a cell, they cannot directly measure the proteins or protein functionality that those messengers produce. Protein biochips, on the other hand, promise to make direct measurements of the relative levels of proteins and their interactions with other molecules. That makes them ideal for use in searching for pharmaceutically relevant targets and disease-specific marker proteins.

Protein microarrays consist of large numbers of regularly arranged spots of elements that recognize a protein or proteins of interest. The elements might be antibodies or antigens, enzymes or substrates, or membrane receptors and ligands. The technique permits scientists to monitor a cell’s metabolism and response to external stimuli. Any biological protein assay that uses a specific ligand-receptor interaction can be miniaturized into a protein biochip or array format.

When a protein microarray is exposed to a mixture of other proteins, molecules that naturally interact with the proteins fixed on a slide will bind to those protein probes. For identification purposes, the proteins bound to the probes can be labeled and visualized, in much the same way as they are in DNA microarrays. Molecules with strong affinities to the probes represent good candidates for leads in drug discovery, since a drug must bind to its target to be effective.

Making a protein microarray presents much more technical challenges than making a DNA chip, however. The reason: proteins’ lack of stability when compared with DNA. Proteins tend to be bioactive only when in their native state with the correct three-dimensional structure. Changes in pH, temperature, or the ionic strength of a solution can cause native proteins to change shape and denature, and hence become inactive. Controlling those conditions during the creation and use of protein microarrays is no easy task.

Because of those difficulties, the market for protein biochips has developed only slowly. “It is an emerging technology,” explains Löfås of Biacore. “As of now, not much is commercially available owing to the inherent difficulties with proteins and the search for viable applications. There are limited arrays of 20 to 30 spots with antibodies for the detection of, for example, cytokine concentrations. But beyond that, not much is out there yet.”

New Types of Protein Array
Despite the technical challenges, companies are investing heavily in protein arrays, which they regard as the tools that will enable rapid progress in the study of proteomics. Companies that have entered the protein biochip business include Affibody, Agilent Technologies, and Perkin Elmer Life and Biological Sciences. The ready availability of their products means that the days of studying one protein at a time have long gone. As a result, researchers routinely study families of proteins today to understand the complex interactions of protein systems inside cells.

Ciphergen takes a unique approach to protein arraying, called ProteinChip technology and based on its own surface-enhanced laser desorption ionization (SELDI) time-of-flight mass spectrometry platform. “We have strips with 8 or 16 spots on them, formatted like one column of a 96-well microplate,” explains Rubin. “They act as targets in our ProteinChip Reader, which is similar to MALDI instruments.” Indeed, the firm’s approach to protein arrays resembles MALDI technology far more than it does the microarray methodology that applied to most protein biochips.

In June of this year, a new wrinkle to the company’s basic protein array technology emerged. Ciphergen launched a line of surface-enhanced neat desorption (SEND) ProteinChip arrays that, it stated, “will vastly improve the applicability of SELDI-TOF mass spectrometry.” “SELDI is an improvement on MALDI and SEND is the next evolution of the technology,” says Rubin. “The idea is to save a step by embedding the matrix molecule in the surface of the chip so that you don’t have to add it separately.” By reducing the amount of extraneous material on and around arrays, the new technology also makes it possible to separate out proteins in a lower range of molecular weight. “We also increase the signal-to-noise ratio, mostly by reducing the noise, but also by enhancing the signal,” Rubin continues.

Bio-Rad has its own form of protein biochip, a multiple assay product called BioPlex that it introduced early in 2002. “This is a liquid protein chip that gives you the opportunity to look at multiple proteins in a pathway in the same well,” explains Crutchfield. “You can look at real time dynamics in protein pathways.”

The Ties that Bind
In addition to separating proteins, proteomics researchers want to determine the substances to which they bind. Investigations of that type have implications for the understanding of diseases and the development of remedies for them. For example, many processes that underlie the development of cancers result from atypical molecular binding events, such as the linking of carcinogenic compounds and/or their metabolites to DNA and the abnormal activation or suppression of cell signal pathways by aberrant ligand interactions.

Researchers in several areas of cancer research use surface plasmon resonance (SPR) technology to study biomolecular binding events. The method enables scientists not only to isolate the molecular partners that bind with such targets as proteins, carbohydrates, nucleic acids, lipid bilayer vesicles, and even whole cells, but also to gather real-time functional data such as binding specificity, affinity, kinetics, and the active concentration of the molecules involved about the binding event itself. “SPR is a key technology to get kinetic data out in a quantitative format,” says Löfås of Biacore.

Having worked in this field since 1990, Biacore now aims to improve the technology. “We have created new surface chemistry techniques,” Löfås says. “And the technology that we have now monitors in real time not only the assay but also the preparation steps. Users can see how much they have attached to the surface and check the activity of any other control proteins that interact and bind.” So far, he adds, “We have functionality across four spots that provides not only a wealth of functional data but also the capacity to deliver and prepare the sample on MS-MALDI targets. This unique integrated SPR-MS combination provides the user with tools for isolation of analytes from complex mixtures as well as functional analysis, recovery, and identification of proteins and other biomolecules – and all from one sample.”

Pursuing the Holy Grail
Another measurement method applies to a particular form of protein binding – interactions with other proteins – that provides key evidence of disease processes. The yeast two-hybrid method allows researchers to identify related proteins. “It is a really good method for picking out associated proteins in a pathway,” says Stratagene’s Roelofs. “Two-hybrid systems are exquisitely good for looking at which proteins react with other proteins, and hence identifying the switches in the pathway. Pathway detection is the Holy Grail of proteins.”

This technique is based on a transcription factor that binds to a gene and causes it to produce RNA and then protein. Several companies, including BD Biosciences Clontech, Invitrogen, and Stratagene, offer the technology. “It has the advantage of fundamentally being eukaryotic,” says Roelofs. “The negative is that it’s a relatively sophisticated process to grow yeast. For novices entering two-hybrid technology, there’s a learning curve.”

To overcome that problem, Stratagene has developed a bacterial two-hybrid process, called BacteriaMatch, based on research carried out at Harvard University. “We’ve taken the technology from Harvard and made it into a product that moves the two-hybrid interaction concept into a bacterial system. It’s very easy to grow bacteria,” Roelofs explains. “We’ve made the expression system we sell as close to a eukaryotic environment as can be had while staying bacterial.” The company is about to release an upgraded version of the system that it first marketed in late 2001.

Data Deluge
Like any facet of modern life science, protein studies generate large amounts of sequence data and other molecular information, including 3-D structure and protein function. Pagano of Incyte highlights a particular problem that protein scientists encounter. “The redundancy of scientific nomenclature causes complications,” she says. “Names are given because of function. The redundancy adds to the difficulty when you’re doing analysis and want to identify a protein.”

Plainly, proteomics researchers need software capable of storing large volumes of data, comparing the data, and recognizing relationships between sequences for the same organism or even for a group of organisms. Companies that offer this type of software include Compugen and Hitachi Genetic Systems as well as Incyte.

Incyte’s BioKnowledge Library is a proteomics database product that contains information about more than 60,000 scientific references to over 77,000 proteins. “One can look at the underlying sequence of any protein, look to see if there’s a clone, and use it in the analysis,” Pagano says. The company based its original protein database on the yeast proteome. In 2000 it launched databases devoted to the proteomes of humans and of mice and rats.

Bio-Rad contributes to this area with a product called WorkSpace. “This is a bioinformatic model designed to standardize the whole experiment leading up to the database and making it easier for data to be handled,” explains Crutchfield. “You can see the before and after views – which proteins have arrived and which have left in, say, a drug interaction.”

CRPS provides another type of service to protein scientists, providing proteomic testing and research on a fee-for-service basis. “We work a lot on sample preparation techniques in terms of enrichments, depletions, and the like,” says Jersey. “What attracted us to Proteome Systems was its recognition of the criticality of sample preparation techniques, regardless of the subsequent separation technology to be used. Proteomics is an exercise in reducing complexity before separation and ultimately detection by mass spectrometry. This was one of Proteome Systems’ core philosophies with which we agreed.”

Bioinformatics Meets Proteomics
Celera has taken a far-reaching approach to drug discovery based on applying bioinformatics to proteomics. Like many informatics based firms, the company best known for its contribution to sequencing the human genome has decided to leave the pure information business. Rather, Celera has begun to focus on proteins as a route to understanding, diagnosing, and treating diseases.

The focus on drug design and other forms of preclinical research started in late 2001, with Celera’s acquisition of Axys Pharmaceuticals. The company’s plan, refined late last year, “supports increased investment in clinical programs, greater efficiency, and economy in target discovery,” according to Ruben. “The information generated from this proteomic analysis requires huge amounts of computational power and storage capacity and Celera has one of the largest commercial supercomputing centers in the world. It also has the advantage of leveraging the genomic and SNP databases that are a product of the Human Genome Project here. This has allowed identification of unique candidates as potential targets in disease,” he adds.

Celera is evaluating cell surface proteins from normal and diseased cells to identify likely therapeutic targets for antibodies. Many scientists believe that those proteins represent the most promising targets for near-term drug candidates via antibodies. The company applies proprietary techniques to capture cell surface proteins, which it then identifies and quantifies, using a program based on liquid chromatography and mass spectrometry that allows for simpler automation and higher throughput and efficiency. Celera scientists have validated this discovery platform by identifying differentially expressed proteins on the surface of pancreatic cancer cells; the same proteins are undergoing further validation to confirm their viability as targets. The company now focuses its proteomics effort on lung cancer and colon cancer. “In a sense, the proteins are the real actors in the biological systems that define health and disease states”, Ruben says. “By exploring differences in the proteomic levels of expression, we can understand not only the potential mechanisms by which the disease state is manifested but may also identify better, more specific ways of diagnosing and treating the diseases.”

Some of the largest challenges in the emerging field of proteomics involve the need to amplify low-abundance proteins to produce quantities for biochemical analysis and the development of robust bioinformatics systems that can collect, manage, and analyze the data from those projects. Proteomics continues its development at a rapid pace, challenging both suppliers and researchers to invent new tools and methods. “We believe that proteomics will become increasingly important to the drug development and discovery process,” says Jersey. “It’s a high growth opportunity.”

WEBLINKS ADVERTISERS American Society for Biochemistry and Molecular Biology (ASBMB) advancing the science of biochemistry and molecular biology through publication of scientific and educational journals, scientific meetings, and advocacy for funding of research and education 301-530-7145 www.asbmb.org Ardais Corporation provides clinical genomics products and related services to facilitate clinically based discovery programs 781-274-6420 www.ardais.com BD Biosciences Pharmingen products for immunology, apoptosis, cell biology, neurobiology, and molecular biology research 858-812-8800 www.pharmingen.com Bio-Rad Laboratories instruments and reagents for life science research including genomics and proteomics 510-741-1000 www.bio-rad.com Bruker Optics, Inc. Fourier Transform infrared, near infrared, and Raman spectrometers for various industries and applications 978-439-9899 www.brukeroptics.com Ciphergen Biosystems, Inc. protein microarray systems and related products that discover, characterize, and assay proteins from native biological samples 510-505-2100 www.ciphergen.com MDL Information Systems discovery informatics and software solutions for the life sciences 510-895-1313 www.mdl.com Pierce Biotechnology, Inc. biochemicals, kits, and systems for genomics and proteomics research 815-968-0747 www.piercenet.com Roche Applied Sciences biochemicals, kits, and systems for genomics and proteomics research 317-845-2000 www.biochem.roche.com SANYO Sales & Marketing Corporation / SANYO Electric Biomedical Co., Ltd. constant temperature equipment (incubators, ovens, refrigerators, and freezers) www.sanyo-biomedical.co.jp FEATURED COMPANIES Affibody AB protein microarrays www.affibody.com Agilent Technologies, Inc. microfluidics/lab-on-a-chip systems www.agilent.com Amersham Biosciences kits and reagents for protein separation www.amershambiosciences.com Applied Biosystems mass spectrometry systems www.appliedbiosystems.com BD Biosciences Clontech proteomic kits and reagents www.clontech.com Beckman Coulter, Inc. chromatography products www.beckmancoulter.com Biacore International AB systems to study molecular binding www.biacore.com Bio-Rad Laboratories electrophoresis equipment and supplies www.bio-rad.com Celera Genomics drug discovery www.celera.com Cell Signaling Technology signal transduction reagents www.cellsignal.com Charles River Laboratories International, Inc. drug discovery www.criver.com Charles River Proteomic Services drug discovery / proteomics services www.criver.com/products/dd/proteomics.html Ciphergen Biosystems, Inc. protein microarrays and analytical instruments www.ciphergen.com Compugen, Inc. bioinformatics www.cgen.com Harvard University university www.harvard.edu Hitachi Genetic Systems bioinformatics www.hitachi-soft.com Incyte Corporation gene expression databases www.incyte.com Invitrogen Corporation kits and reagents for protein separation www.invitrogen.com PerkinElmer Life and Analytical Sciences research instruments, kits, and reagents www.perkinelmer.com Proteome Systems, Ltd. proteomic technologies for drug discovery / bioinformatics www.proteomesystems.com Stratagene proteomic kits and reagents www.stratagene.com Thermo Electron Corporation mass spectrometry systems www.thermo.com Waters Corporation chromatography products www.waters.com

Note: Readers can find out more about the companies and organizations listed by accessing their sites on the World Wide Web (WWW). If the listed organization does not have a site on the WWW or if it is under construction, we have substituted its main telephone number. Every effort has been made to ensure the accuracy of this information. The companies and organizations in this article were selected at random. Their inclusion in this article does not indicate endorsement by either AAAS or Science nor is it meant to imply that their products or services are superior to those of other companies.