In 2004, when Nancy Craig's collaborators at the National Institutes of Health made an image of an enzyme that breaks and joins DNA, she couldn't wait to get to their lab. "It was so exciting to see this protein that we had been working on for several years," says Craig, who had become fascinated with DNA in college. "I remember a day when a chemistry class, physics class, and biology class were all talking about DNA. The fact that one molecule could come up in all those contexts piqued my interest."
Part of a jumping gene called Hermes encodes the enzyme that sparked Craig's excitement. Jumping genes, or transposons, snip themselves out of one site in a genome and insert themselves into another. If they blunder into a gene—or its regulator—they can mutate it, causing diseases such as hemophilia. Craig studies transposons because she's interested in how they wander around and what they might do other than create havoc. To improve gene therapy, she also wants to tame transposons.
In the early 1980s, Craig was studying the cutting and rejoining of DNA that occurs as a certain bacterial virus inserts itself into its host's genome, when she read an article about Tn7, a piece of mobile DNA that clearly used a different mechanism than the virus she was studying. This transposon uses a reserved parking spot in the genome instead of settling down at random. The location is safe for the cell, because it lies between genes, rather than within a gene. One of the most exciting moments in Craig's career happened when her group was able to insert Tn7 into the correct part of DNA in a test tube. "That meant we could take the system apart and put it back together to figure out how it works," she says.
The group studied the proteins and biochemical reactions that insert or remove transposons and the codes that serve as parking signs. They detected the DNA sequence recognized by Tn7 in almost all the organisms they studied—even humans. "So picking the neighborhood is something that transposons do," Craig says. "It's as if organisms are saying, 'Put this transposon here, and it won't be harmful.' So it makes sense to have a treaty between the host and the transposon."
A lot of treaties have been made, it seems, because transposon DNA accounts for about half the human genome.
With funding from HHMI, Craig was able to study additional transposons, including Hermes, which is found in house flies. "We keep going after the fundamental question: How do all the proteins that are involved in transposition interact with each other?" Craig says.
Knowledge of these proteins' structures can help answer that question, and x-ray crystallography is one way to obtain structural information. Fred Dyda and Alison Hickman at the National Institutes of Health were the collaborators who deduced the crystal structure of the enzyme that integrates Hermes into host DNA. After Craig rushed to their lab in 2004, she was shocked to see so many similarities between that enzyme and the one that inserts retroviruses, such as HIV, into chromosomes. "In retrospect, since we knew they had the same mechanism, we shouldn't have been so surprised," Craig says.
Because of the similarity, Craig hopes her studies might indirectly help AIDS patients. "The things we learn about the fundamental mechanism of transposition should contribute to our understanding of how HIV works," she says.
Craig's interest in gene therapy might improve medical care more directly. One challenge is to introduce a normal version of a defective gene into a safe location in the genome. The hazards of using a retrovirus that inserts into many different positions in the genome were revealed in a 2003 gene therapy trial in France in which 10 infants with X-linked severe combined immunodeficiency ("bubble boy syndrome") were treated with a retrovirus containing a normal version of their defective gene. Nine of the 10 were cured of their immunodeficiency, but four developed leukemias, because the retrovirus inserted its DNA next to a cancer-promoting gene, switching it on. "So we are trying to steer transposons to particular places," Craig says.
One promising transposon comes from bats. After one of her colleagues noticed a sequence in bat DNA that looked suspiciously like a jumping gene, Craig isolated it and showed that it was an active transposon. This was the first cut-and-paste transposon found in mammals. "We are working on it right now to see if it might be a better vector for gene therapy than other transposons," Craig says.
Bats account for nearly one-fourth of mammalian species. So Craig and others have speculated that the knack of maintaining active transposons in the genome might contribute to a group's diversity. "It is clear that transposons have had enormous effects on assembling new genes and changing regulatory elements [during evolution]," Craig explains.
The connections between transposons and evolution, gene therapy, and many other topics make Craig's work particularly stimulating. "I find it an enormously interesting process, because I never do the same thing twice in a day," she says. "It's better than the best crossword puzzle."