What's the most frequent knockout phenotype? Death.

It's a wry joke among geneticists, but it's also true. And it limits the types of questions that can be answered using knockout technology.

One response to the problem has been to make inducible knockouts, which will produce a certain gene product until the gene is disabled by an extraneous signal. Another has been the construction of tissue-specific knockouts, which delete a gene from all or most cells of a given type.

In the May 5, 2005, issue of Cell, scientists from Stanford University debuted a yet more fine-grained solution to the same problem: mosaic mice.

Creating mosaic animals has long been a standard and powerful genetic analysis tool in Drosophila. But of course, if mice can be poor models in drug discovery, fruit flies are further removed still from humans.

Homologous Drosophila chromosomes line up during cell division, making recombination, and the creation of mosaics, comparatively easy. Creating a mosaic mammal proved much more difficult.

Liqun Luo, assistant professor of biological sciences at Stanford University and senior author of the study, and his colleagues succeeded by developing a technique they dubbed Molecular Analysis with Double Markers, or MADM. The technique engineers three separate mutations onto the sister chromosomes: two marker protein fragments and the mutation of interest.

At the outset, one chromosome contains half of the gene for one of each of two different marker proteins, separated by a recombination site; the chromosome allele has the halves of the two proteins switched. Downstream from the marker protein fragments is the gene of interest, which is present at the outset in one functional and one nonfunctional copy.

When recombination occurs, which it does in a few percent of cells during mitosis, one-quarter of recombined daughter cells will be homozygous for the nonfunctional gene variant. Because the marker proteins, as well as the alleles of interest recombine, all of the cells also will express one marker protein (in the case of the experiments described in Cell, green fluorescent protein). Expression of the other marker protein shows that the cell now is homozygous for the functional gene, while heterozygous cells express either no marker or both. Post-mitotic recombination, which does not alter the cell's genotype, also will lead to double labeling.

One strength of the MADM technique is the 1-to-1 coupling of knockout and labeling. For green fluorescent protein, which is one of the markers the scientists used in their paper, "the background rate is zero," Luo told BioWorld Today. "So even if only one cell turns green, we know it has to contain the mutated gene of interest." Luo said that "possibly, others have succeeded before us [in creating mosaic mammals], but their marker techniques may have been insufficient to really demonstrate it."

Current conditional knockouts require the insertion of flanking sequences for each gene separately. Once a given chromosome has the marker protein fragments engineered in, "with our method, as long as a gene is knocked out, we can convert it to a conditional knockout by crossing," Luo said. "If we just make those 20 pairs, it will really streamline conditional knockouts," he added.

"The technique does not replace flox, in the sense that our frequency is lower," Luo said. While flox knockouts, depending on the strength of the promoter used, can lead to a population in which most or all cells are knockouts, MADM will lead to a cell population with a few percent of knockout cells. Whether or not that is preferable depends on the experimental question being addressed.

The scientists showed proof of principle for the technology by creating such a knock-in on chromosome 19, and using it to study the relationship between lineage and wiring of one type of brain cell, cerebellar granule.

Besides lineage studies, such as the one reported in Cell, other uses of the technology include the analysis of complex genetic diseases; Luo specifically pointed out that one very useful application could be in studying cancer, to visualize and study cells in which loss of heterozygosity - one cellular cause of cancer - has occurred.

The technique is patented, and currently is being licensed to different companies, Luo said. His lab also is in the process of establishing collaboration, as well as improving the technique and working on inserting the marker protein fragments into other chromosomes besides chromosome 19.