It's always puzzling when one person turns to a life of crime and his sibling curbs his criminal instincts. It's also puzzling that certain skin cells, called melanocytes, can become either deadly melanomas or harmless moles even though they have acquired the same mutation. So when Michael Green was musing about his work while fishing one day, he thought up a way to identify the genetic changes that send a melanocyte down the pathway to malignancy. The results were surprising, and they suggest a new way to fight skin cancer.

Scientists are like artists, Green says, because they try to create something different from what has gone before. That's why he chose an academic career rather than becoming a physician. "Most people who come out of an M.D./Ph.D. program do additional training in medicine," he says. "But that would have been a detour for me, because I knew I wanted to do research."

Green got his start in research as a graduate student at Washington University in St. Louis in the lab of Robert Roeder, who was studying eukaryotic gene transcription. This process extracts information from genes by transcribing the sequence of "letters" in DNA into a corresponding sequence in messenger RNA (mRNA), the template for protein synthesis. "There wasn't a lot known [in the late 1970s]," Green recalls. "Cloning technologies were just becoming available, and there was a lot of excitement about eukaryotic gene regulation."

Regulation of gene transcription, which helps determine when or how much of a particular mRNA is made inside a cell, has remained a major focus in Green's lab. One important regulator is a stretch of DNA called the promoter, a row of switches outside the coding region of the gene that can be flipped on or off or turned up or down by a suite of proteins called transcription factors. These factors determine the rate at which a gene is transcribed—or even if it is transcribed at all. Improper gene regulation can result in too little or too much of the protein encoded by that gene. And it can sabotage development or trigger cancer and other medical conditions.

One of Green's major contributions to this field was to elucidate how a certain class of transcription factors, called activators, stimulate gene transcription. Activators bind to specific DNA sequences in promoters. Green showed how activators, such as the yeast Gal4 protein, target the general transcription machinery and recruit it to the promoter, enabling transcription to occur.

Green attributes part of his success to knowing which problems to pursue. "You have to risk picking those questions that you think are going to give the most important insights and answers," he says. "You also have to be willing to abandon certain ideas and projects if they don't look as if they are going to be fruitful."

Another interest in the lab stems from Green's postgraduate work with Thomas Maniatis at Harvard University. It concerns the process by which pre-mRNAs—the intermediates between DNA and mRNA—are edited into mRNAs. Green's group identified an important part of the mechanism that cuts out noncoding parts of a pre-mRNA (called introns) and splices the cut ends together. "We showed that during splicing the intron forms a little loop called a lariat," Green says. "That was unexpected because both the [pre-RNA molecules and the mRNA molecules that are formed] are linear."

The luxury to pursue such work, even though it veers from the expected path, appeals strongly to Green. "I like the freedom to chase problems and ask questions pretty much on my own," he explains.

As it turned out, he didn't abandon his interest in medicine when he became a scientist, because his work on gene regulation eventually led to him to study cancer, which arises when certain genes fail to restrain cell division. For example, his research group recently determined why certain cells either commit suicide or hibernate if they acquire a cancer-causing mutation in a gene called BRAF. This mutation is found in both moles and melanoma cells, but the melanocytes in moles respond by hibernating, whereas those in melanomas proliferate and spread throughout the body, causing metastatic melanoma. This dreaded disease kills 8,000 people in the United States each year.

The group's genome-wide scans revealed a network of 17 genes that prevent a melanocyte with a BRAF mutation from progressing to a melanoma. One of these genes encodes a secreted protein called insulin-like growth factor–binding protein 7 (IGFBP7). This discovery was a big surprise, because researchers had expected to find a protein that acts inside the cells that produce it. But cells that secrete anticancer proteins have an advantage, because they can control neighboring wayward cells from their home base.

When the group exposed cultured melanocytes to IGFBP7, the cells went into a state of hibernation, much like the cells in moles do. However, when they treated cultured melanoma cells carrying the BRAF mutation with IGFBP7, the cells committed suicide. Moreover, IGFBP7 stopped the growth of human melanomas implanted into mice. "We think that IGFBP7 is a potential treatment for melanoma, which is currently untreatable when it becomes metastatic," Green says. "IGFBP7 may also have implications for other types of cancers that have the same BRAF mutation."

The cancer studies in Green's lab are not confined to BRAF, IGFBP7, or melanoma. In fact, the group has used similar genome-wide scanning strategies to address a number of interesting questions in cancer biology, such as how another cancer-causing gene, RAS, silences gene expression, and to identify cancer-suppressing genes and genes that regulate metastasis. Through such basic research, the group might uncover more leads to anticancer drugs. "We are driven by fundamental questions in cancer biology," Green says.