Separate, sort, and identify

To begin the identification and characterization, cells are broken apart. Hundreds of incumbent proteins are separated by mass using gel electrophoresis, and a special soap allows further, improved separation using liquid chromatography. The proteins are sorted into groups of about five to 15 similarly sized particles. Proteins that are expressed from the same section of the genetic sequence but that differ due to post-translational modifications tend to fall in the same group.

Electrophoresis, which separates molecules using electrical current, and liquid chromatography, which separates molecules by passing them through sticky pores, are nothing special for chemists and biologists. Accusations of ordinariness, however, should probably end there.

Gel electrophoresis of yeast proteins. This process separates proteins by mass after a cell has been broken down. (click here to enlarge picture)

Once the proteins are separated into like-sized sets, they are fed into a Fourier-transform mass spectrometer with a 9.4 tesla superconducting magnet. There are fewer than six research labs in the world with this type of instrument, according to Kelleher, and only the Kelleher team is pursuing the top-down approach to identifying and characterizing proteins.

Schematic highlighting the differences between the Kelleher team's top-down approach to protein characterization and the more traditional bottom-up approach. (click here to enlarge picture)

The mass spectrometer ionizes the proteins and sets them spinning in the instrument's magnetic field. With a bit of computational analysis, the instrument's read on the spin can be converted into a mass measurement because the ionized protein's spin in the gas phase is a function of its mass. The faster the protein ion spins, the lower its mass. Masses, in turn, can be compared to a database and equated to a particular type of protein.

Of traditional character

Identification of proteins, therefore, is a straightforward enterprise that can be carried out using a more ho-hum mass spectrometer. Characterizing the post-translational modifications--tracing the steps between the protein initially expressed by the gene and the protein at work, or at times wreaking havoc, in your body--is a far more daunting prospect.

In a traditional characterization scheme, a protein is broken into many small pieces, typically less than 20 amino acids long. Various enzymes chew the protein into these pieces after separation but before the sample enters the mass spectrometer. Researchers predict the type and size of the resulting fragments based on the make-up of the protein and the types of enzymes used. They compare these predicted values to the values that result from the mass spectrometer. If the values match, that section of the protein is assumed to be unmodified. If they do not, it is assumed a modification has taken place.

This approach is powerful and time-tested. But since many fragments go missing, fingering modifications is difficult. The smaller the piece, the less likely it is to be detected and properly identified. The method also tends to go south if numerous fragments have been modified because it becomes more difficult to determine which modified fragment is really the cousin of the predicted fragment.