The team, from the University of Utah and the Howard Hughes Medical Institute (HHMI), outline their new technique to analyse DNA sequences in this week's online edition of Nature Genetics. They claim to have developed a method for deleting any piece of DNA from the mouse genetic blueprint very efficiently.
"Diseases are known to occur as a consequence of deleting non-gene DNA sequences, and this new method allows us to evaluate what these sequences do," said Professor Mario Capecchi, co-chair of human genetics at the University of Utah.
The new method "is significant because it makes it practical to do this for a vast amount of the total genome," he added.
Capecchi explained it currently costs about $10,000 (€7,500) to genetically engineer a mouse with a particular gene or other DNA sequence knocked out. So using existing technology to knock out the estimated 20,000 mouse genes would cost $200 million, and knocking out the roughly 300,000 non-gene DNA sequences in mice would cost $3 billion, he adds.
The new method is a faster, cheaper way of mutating DNA, costing $200 to create each mouse with a mutant gene or other DNA sequence, although the savings are not proportionate because more mutant mice must be bred to obtain desired mutants
Capecchi said the new technique could speed up a US National Institutes of Health (NIH) effort to mutate all mouse genes in cell culture and create 900 new lines of mutant mice by 2010 - an effort NIH said "will be extremely useful for the study of human disease".
To quickly and cheaply mutate DNA, Capecchi and his colleague Dr Sen Wu, used short pieces of DNA known as loxP. Pieces of loxP act like signposts that say 'cut the DNA between these two signposts'. In the new method, loxP is inserted in a chromosome on one mouse, and in a different position on the same chromosome in a second mouse, which also has a gene named Cre in its cells. When the mice are bred, an offspring mouse has loxP DNA on two sites on the same chromosome and also has the Cre gene. The protein made by Cre acts like a knife, cutting the DNA wherever loxP is found.
The offspring mouse, in turn, is bred with a normal mouse. In a "remarkably high" average of 10 percent of the offspring, the desired large stretch of DNA is either deleted or duplicated so scientists can see what goes wrong in the mutant mouse and thus learn the normal job of that piece of DNA.
Capecchi said the same method also can be used to make two chromosomes break and recombine, with each new chromosome containing a piece of each of the two old ones. The process, "translocation," also can create new genes from two other genes. Many cancers - certain sarcomas, leukaemia and lymphomas - start with translocation.
Existing methods result in the desired translocation of two chromosomes in one of every 1 million to 10 million cells cultured, Capecchi said, while the new method produces the desired result in one of every 100 cells, a factor of 10,000-100,000 times better.
That is important, because if translocation of two chromosomes only occurs in one of every million cells, not enough cells are present in which subsequent steps can occur so that cancer develops.
"If we now improve that 10,000-fold, the pool of cells [with translocated chromosomes] may be large enough to create a mouse that develops the cancer," so the new method makes it easier to breed mice with human cancers, Capecchi said.
Wu and Capecchi then improved their new method by using it in combination with a "transposon" or so-called "jumping gene" named piggyBac, which comes from a moth. They used piggyBac to insert pieces of loxP DNA randomly into numerous spots on the mouse genome. So scientists easily can breed multiple generations of mice with loxP surrounding various large stretches of DNA.
The mice are bred with mice that carry the Cre gene so any large stretch of DNA that is located between two short pieces of loxP DNA can be mutated.
Because the whole mouse genome has been sequenced - the order and location of base pairs determined in detail - scientists can identify locations in individual mice where the jumping genes have carried pieces of loxP DNA. Then, they can select a mouse with loxP on both sides of the large piece of DNA they want to mutate and study.
If the jumping gene carrying loxP hops into the middle of a gene, then the gene will be mutated. So while the method began as a way to delete large stretches of non-gene DNA, it has the bonus of also being able to disable genes more easily and inexpensively than existing knockout techniques, Wu said.
