Brown used this technique in 1998 to study the sexual development of yeast in collaboration with Ira Herskowitz at the University of California, San Francisco. ("Actually, our graduate students, Shelley Chu and Joe DeRisi, did most of the work and thinking," explains Herskowitz.)

When one looks at highly magnified pictures of yeast, these little blobs seem utterly sexless. Yet yeast comes in two mating types—"a" and alpha—as well as a slightly larger diploid cell that carries both "a" and alpha. When a diploid cell runs out of food, it morphs into spores of the two different mating types, which remain enclosed in a tough shell until times improve. Then these spores burst forth and start looking for mates of the opposite type; as soon as they find a suitable mate, they fuse together.

The spores' development is surprisingly similar to the development of human sperm and eggs. "Most of the cells in your body are diploid. They have two complete copies of the genome, one from your mother and one from your father," Brown explains. "But egg or sperm cells carry only one copy of the genome. To produce these sex cells, the diploid cells have to go through meiosis—they divide—and shuffle their chromosomes around.

"Well, yeast does exactly the same thing when its diploid cells form spores," Brown says. "Many of the genetic mechanisms are shared with humans. Many of the genes that are involved in the process, and the gene products, are functionally analogous to one another."

Until the Brown and Herskowitz study, scientists knew of roughly 150 yeast genes involved in making spores. But the microarrays, which show all the genes that are active in a cell at a particular time, gave the team a much richer, more detailed picture of what turns out to be a surprisingly complex process. Instead of 150 genes, they found that "more than 1,000 yeast genes showed significant changes in mRNA levels" during sporulation (the process of making spores). The activity of half of these genes was turned up (as shown in graded shades of red), while that of the other half was turned down (green) in complicated sequences that changed over time.

Around 10 percent of the changes in the genes' activity were in some sense expected, since scientists knew something about the genes involved. "But there was a large bunch of genes about which we knew essentially nothing," Brown says, "and here we suddenly had a really great clue as to what they might be doing."

Guilt by Association
The clue? "Guilt by association," the team called it. "We can guess their function based on the fact that other genes—whose function we do know—are turned on at the same time," Herskowitz says. Genes tend to be expressed just when their products have important roles to play, he explains, so all the genes whose mRNA went up or down at the same time were likely to have similar functions.

These assumptions could be tested very easily in yeast. "We took a set of genes that showed a suggestive pattern of expression during sporulation," says Brown. "Then we knocked these genes out one by one and found that the cells failed to complete their program." The program stopped at various points, depending on which gene was knocked out. Each point revealed the normal function of the knocked-out gene.

"This study is a door opener; our analysis only scratches the surface," says Herskowitz. A searchable database of all the data from this project is now available on Brown's Web site (cmgm.stanford.edu/pbrown), where anyone can apply the information to human sperm and eggs (whose formation, according to Herskowitz, is still "very poorly understood") or to analyze other developmental processes.

— Maya Pines

< Previous  |  Top of page  |  Next >

Yeast Spores and Human Sperm: A Parallel Development

How do yeast cells mate? Ira Herskowitz explains it all...

	A Microarrayer Shows How Genes in a Yeast Cell Respond to Different Types of Stress

© 2012 Howard Hughes Medical Institute. A philanthropy serving society through biomedical research and science education.
4000 Jones Bridge Road Chevy Chase, MD 20815-6789 | (301) 215-8500 | e-mail: webmaster@hhmi.org