Our Closest Relative Among Model Organisms: The Mouse
    A  large group of scientists have been studying our closest relative among model organisms: the mouse Mus musculus. Not just a vertebrate, but a mammal, the mouse has about the same number of genes as we do—the usual estimate is between 80,000 and 100,000 genes. Its DNA is so similar to ours that from the perspective of his work on yeast, Ira Herskowitz of the University of California, San Francisco, say; "I don't consider the mouse a model organism. The mouse is just a cuter version of a human, a pocket-sized human."
    Whole segments of mouse chromosomes are arranged in the same sequence as ours, making it easy to identify the human equivalents of many mouse genes. And new methods of manipulating mouse genes have turned this rodent into a source of increasingly important findings.
    In a dramatic experiment that opened up the field in 1982, HHMI investigator Richard Palmiter at the University of Washington and Ralph Brinster of the University of Pennsylvania showed that it was not necessary to wait for nature to produce mouse mutants—such mutants could be created to order. When they injected a rat's growth hormone gene into mouse eggs, the resulting mice grew up to twice as large as their littermates. A picture of one of their giant mice, next to its normal sibling, ran in newspapers across the world.
    Their technique was limited, however, since the researchers could not specify where the foreign gene would land in the mouse genome. Nor could they foretell how many copies of the injected gene would be taken up by any particular egg. Another drawback was that while they could add genes to a mouse's DNA, they could not eliminate genes from its genome—and therefore could not infer a gene's function from the effects of the gene's absence.
    A new era in mouse genetics began in the early 1980s, when Mario Capecchi, an HHMI investigator at the University of Utah, developed a method of "gene targeting" by homologous recombination (the exchange of two similar DNA fragments), first in cultured mammalian cells and later in living mice. It enables scientists to create strains of mice with mutations in virtually any gene—most importantly, "knockout" mice in which a particular gene is missing. Gene targeting has proved extremely useful in determining the functions of specific genes. In the past five years, the roles of more than 500 genes have been deciphered with this approach, many of them by HHMI investigators.

    Capecchi has used the method to analyze genes that regulate the development of mice—particularly the so-called Hox genes, mammalian versions of the homeotic genes that Edward Lewis first discovered in fruit flies. Whether from flies, worms, mice, or humans, all homeotic genes contain a segment called a homeobox, made up of 180 base pairs of DNA, which codes for an important domain of a transcription factor—a protein that regulates the activity of other genes.
    Similar genes, in the same order, control the development of the front and back part of the bodies of flies and mice. These homeobox-containing genes lie on a single chromosome in the fly (top row of colored squares) and on four separate chromosomes in mammals (lower rows of squares). The genes are color coded to match the parts of the body in which they are expressed.
    While flies have just one cluster of homeobox-containing genes that lies on a single chromosome, mammals have four similar clusters lying on four separate chromosomes. "These clusters arose by duplication, in the course of evolution," Capecchi explains. In each cluster, the genes located at one end direct the development of the anterior part of the body, while the genes at the other end control the formation of the posterior part. The genes in the four clusters work together, "talking" to each other to produce a more complex creature than the fly, says Capecchi. He hopes to disentangle their functions and work out the logic of how mammals are built.