Concept 41 DNA is only the beginning for understanding the human genome.
Mario Capecchi describes proteomics; the large-scale study of protein structure and function. Brian Sauer explains gene knock outs.
Hi, I'm Mario Capecchi. I came up with a method that is being used to analyze the functions of proteins. This will help us make sense of the tens of thousands of genes discovered by the Human Genome Project. Some call this new field of large-scale studies of protein structure and function "proteomics." This amounts to carrying Beadle and Tatum's "one gene-one protein" experiments to a massive scale. A particularly good way to study the function of a human protein is to manipulate its homologous protein in mice. Our chromosomes are highly similar in structure. . . . . . and it is estimated that mouse genes are about 99% identical to humans. In the mid-1980s, biologists began to insert new genes into mouse embryonic stem (ES) cells. These cells, which are derived from the inner cell mass of a developing blastocyst, can be grown in culture, like bacteria or yeast. However, they retain the ability to develop and differentiate like normal embryonic tissue. A plasmid carrying a gene to be transferred, or transgene, is cut once with a restriction enzyme. Then, the linearized plasmid is added to the culture. Upon exposure to a short pulse of electrical current, about 10% of the cells take up the new gene. This method of DNA uptake is called electroporation. Blastocysts are then harvested from a pregnant mouse. A number of electroporated ES cells are injected into a host blastocyst. The chimeric blastocyst is then inserted into a surrogate mother, where the transgenic ES cells develop along with the host blastocyst cells. The chimeric pup is, in fact, derived from four parents and visibly shows its mixed heritage. The ES cells are usually derived from mating black parents, and the blastocyst is from a mating of two albino parents. So the transgenic offspring has black and white patches. Initially, there was no way to predict where the transgene incorporated in the host genome. In some cases, it might insert harmlessly into an intergenic region. In other cases, it might disrupt a functioning gene. In rare cases, it recombines with its homologous sequence. My method precisely "targets" a transgene to a particular position on a chromosome by homologous recombination. In one early experiment, I "knocked out" the int-2 gene, which produces a growth factor involved in early mouse development. First, I constructed a "targeting vector" from the first three exons of the int-2 gene. I inserted a neomycin-resistance gene into one of the coding exons, disabling the int-2 gene. I also added a nonhomologous DNA sequence and the thymidine kinase (tk) gene from the herpes simplex virus (HSV). Then I used electroporation to introduce the targeting vector into cultured ES cells. One of two scenarios occurs during integration of the targeting vector in the mouse chromosome. In the first scenario, during homologous recombination, int-2 sequences on the targeting vector and the host chromosome align and exchange. Because its sequences are different, the thymidine kinase gene is excluded during the exchange. In the first scenario, during homologous recombination, int-2 sequences on the targeting vector and the host chromosome align and exchange. Because its sequences are different, the thymidine kinase gene is excluded during the exchange. However, if random integration occurs, the entire targeting vector is inserted, including the thymidine kinase gene. Because of the way the targeting vector was constructed, we can easily distinguish between the two scenarios. First, neomycin selection kills any cells that have not integrated the transgene. Then, gancyclovir, an antiviral drug, kills cells that have integrated the thymidine kinase gene during nonhomologous recombination. Only cells that have homologously integrated the targeted transgene survive the double selection and reproduce on the culture plate. These cells are inserted into a blastocyst, and the blastocyst is integrated into a surrogate mother that will carry the chimeric embryo to term. Producing a chimeric pup is only the first step of creating a gene knockout. Some of the chimeric pups have sex cells derived from the ES cells. These are from black mice and so will have a gene encoding black coat color. Breeding the chimeras to albino mice produces some black mice with the ES genome in their germ plasm. Producing a chimeric pup is only the first step of creating a gene knockout. Some of the chimeric pups have sex cells derived from the ES cells. These are from black mice and so will have a gene encoding black coat color. Breeding the chimeras to albino mice produces some black mice with the ES genome in their germ plasm. Because the chimeric mouse produces two types of sex cells with the black coat gene – one with the transgene and one without – black mice are then screened to show which carry the neomycin gene insert. Although I used Southern blots in my original experiment, PCR is more widely used today. PCR primers are made so that in mice with the integrated transgene, PCR yields a 400-base pair fragment. In wild-type mice, PCR would not produce a fragment because the neo gene isn't present. When the results are run on a gel, the mice with the targeted mutation show one band, represented by +. Those without a band are represented by -. Now, back to Mendelian genetics. Each of the positive mice is heterozygous for the mutation, so about 25% of the offspring of two heterozygotes will be homozygous for the knockout genes. These are called null mutants; they don't have a functioning int-2 gene. These mice can be screened for developmental, anatomical, biochemical, or behavioral differences. Sometimes a phenotype can be the result of partial gene function. By using homologous recombination, knockouts or other types of very specific mutations can be made. For example, later work with the homeotic (Hox) genes of mice produced an extremely striking illustration of how homeotic genes work together to determine the mammalian body plan. We studied the effect of essentially duplicate copies of a Hox gene found on different chromosomes – Hoxa-11 and Hoxd-11. Null mutants for either gene showed only subtle differences in the anatomy of the forelimb. One limitation of my knockout system is that the gene is knocked out in all cells right from the beginning of development. Some null mutants simply don't survive. Hi, I'm Brian Sauer. I refined Mario's method to get around the problem of specifying the timing and location of gene knockouts. The refinement takes advantage of the Cre recombinase system in P1 bacteriophage. Cre is a protein that eliminates DNA between two sites called loxP. Each loxP contains a 13-base pair sequence at the 5' end, and an inverted version of this sequence at the 3' end. There are eight base pairs in the middle. One molecule of Cre binds to each of the 13-base pair sequences. The four Cre molecules then form a tetramer that removes the DNA between the loxP sites. One loxP site is left in the chromosome. I discovered that the Cre/lox recombination system also works in mammalian cells! When this system is combined with Capecchi's homologous recombination system, we can knock out genes after the mice grow into adults. Researchers interested in learning and memory did this when they knocked out the mouse NMDA receptor in one type of brain cell called CA1. We believe that the receptors in these cells are crucial for memory formation. First, using homologous recombination, they created a mouse strain with one of its NMDA receptors flanked by two loxP sites. When these sites surround a gene, we say the gene is "floxed." Another mouse was created that not only had a floxed NMDA gene, it also carried the cre gene on another chromosome. The cre gene was under control of a promoter that is only active in CA1 brain cells. When the two strains were crossed, some of the progeny received two floxed genes and one cre gene. Southern blotting detected the floxed genes, and PCR detected the cre gene. In these mice, the NMDA gene was not excised by Cre until the CA1 cells had formed and turned on the promoter – about three weeks after birth. Only these cells and no others were missing NMDR receptors. Without NMDA receptors in their CA1 cells, the knockout mice could not use landmarks to remember where they were in space. They couldn't remember the location of an underwater platform – that they had been placed on earlier – based on surrounding landmarks. The mutant's siblings, who had NMDA receptors, located the platform after lining up the surrounding landmarks. Together, gene targeting by homologous recombination and the Cre-lox recombination system allow us to make essentially any change in the mouse genome — from point mutations to large-scale chromosome deletions. These changes can be activated in essentially any tissue at any time during development.
The first time Mario Capecchi submitted his ideas on gene targeting to NIH for funding, he was rejected. Four years and several successful experiments later, NIH funded Capecchi's project and issued an apology.
The Cre/lox recombinase system was first found in phage. Why would such a system have evolved?