The Human Genome Project might never have gotten off the ground if it hadn’t been for the automated DNA sequencer. But in the emerging field of proteomics, no single technology drives the industry. Instead, researchers tap a variety of techniques to identify and analyze proteins. Three of the most popular are mass spectrometry, yeast-two hybrid and x-ray crystallography. Each enters the complex world of proteins from a different angle, aiming to illuminate protein composition, interaction and function.
Mass Spectrometry: New Wine, Old Skin
Every protein has its own fingerprint. But unlike DNA, which is assembled from four simple building blocks, proteins are extremely difficult to identify, let alone sequence. While databases of hypothetical protein sequences have been derived from the first draft of the human genome, that information must be verified. Currently, no tool does this better than the mass spectrometer.
Mass spectrometry has been around for more than a century, used by physicists and chemists to identify and characterize small molecules. In a mass spectrometer, the molecules are literally smashed to pieces by an electric charge. The fragments then pass through a magnetic field where each individual mass is measured and recorded. The distribution of all the masses-the “mass spectrum”-is unique to each type of molecule, letting researchers match their readouts against a database of recorded mass spectra.
Until recently, this has meant little for studying larger biological molecules. That’s because biomolecules like proteins could not easily be ionized (given an electrical charge) and converted to a gas-two requirements for mass spectrometry. In 1988, two solutions were proposed. The first employs a laser to impart enough energy to charge the molecule and turn it into a gas. In the second, the molecule passes through a liquid phase before being vaporized.
Steven Gygi of Harvard Medical School uses the second technique, called electrospray, to prepare peptides, a smaller subunit of protein, for analysis. Once vaporized, the peptide ions are brought into contact with helium, and each molecule breaks up. The pattern of fragmentation allows Gygi to determine the precise amino acid sequence of the peptide.
“This technique can tell you other information about the peptide,” Gygi says. “For example, is there a sugar or a phosphate molecule attached, is it the whole peptide or just a piece?”
Yeast Two-Hybrid: Gone Fishing
Proteins usually work in combination with other proteins. So when a protein’s function is unknown, as with roughly 95 percent of human proteins, clues can be sought by rounding up additional suspects. Yeast two-hybrid identifies the company that proteins keep.
Matching proteins is a painstaking process, equivalent to piecing together a complex jigsaw puzzle. Central to the yeast two-hybrid process is a protein catalyst called a “transcription factor.” This catalyst binds to a gene and causes it to produce RNA (which in turn produces functional protein). The part of the catalyst that binds to the gene is called the binding domain; the part that stimulates RNA production is called the activation domain.
The two domains can be split apart, and each fused to a different protein-creating two “hybrid” proteins that can be tested for a match. If there is a fit between these hybrid proteins, the cleaved domains reunite and the catalyst resumes its original function, producing RNA from the bound gene. Yeast cells are the ideal test bed for this process.
Marc Vidal, of Dana Farber Cancer Institute, applies the yeast-two-hybrid method to map interactions among the 19,000 proteins in the worm proteome. “The idea,” he says, “is to generate lots of different maps and test to what extent they can be integrated into a biological atlas. Then you could presumably better locate where [in the genome] you want to do your studies based on the particular biological question you’re asking.”
Although the yeast two-hybrid technique does not directly reveal what a protein does, other technologies, such as x-ray crystallography, can illuminate a protein’s function.
X-Ray Crystallography: In Shape
“The genome is really a linear code encoding three-dimensional information,” says Tom Ellenberger, a researcher at Harvard Medical School. “Everything we need to know about proteins-their interactions with other proteins, the roles they play in particular diseases-is locked up in their intricate 3-D structure.”
It’s not enough just to know a protein’s amino acid sequence or the other proteins it interacts with. What researchers need to know, Ellenberger says, is the protein’s actual shape, including all its contours, indentations and surface irregularities.
For example, deep pockets may indicate the function of an enzyme, a catalyst that regulates a chemical reaction. The deep pockets are where enzymes receive and bind to other molecules.
In this kind of visualization, the most powerful tool is x-ray crystallography, which can determine the shape of a protein down to its atomic structure.
In the laboratory, many millions of copies of a protein molecule can be made to line up and pack together in a precise, crystalline array. The result is a cube on a slide, no larger than half a millimeter, barely visible to the naked eye.
X-rays are directed at the crystal, which scatters the rays. The repeating geometry of the crystalline structure amplifies the diffracted light, yielding a far more intense image than a single protein ever could. A detector measures the position and intensity of each dispersed ray.
Next, the procedure is repeated-only now the crystal is soaked in a heavy metal, such as mercury. By calculating the difference between the x-ray diffraction patterns and applying a mathematical formula, a computer-generated image of the protein structure emerges.
“X-ray diffraction is viewed as this arcane thing that involves complicated math,” says Ellenberger. “Really, it’s pretty simple. The math formula simply acts as a substitute for the lens [in a microscope].”
Ellenberger uses x-ray crystallography to study proteins that copy DNA during cell division. “By knowing the overall physical organization of proteins that function together to do complex tasks, we can learn a lot about their function,” he says. “And other people can then use that info for their own experiments.”