When yeast are starving, they do what any rational microbe would do: they stop dividing and hunker down, trying to conserve resources until the good times return. But you can trick yeast with what could be called 'unnatural starvation': you don't starve yeast by withholding natural environmental nutrients like phosphate, sulfate, or nitrogen; you starve them of nutrients that yeast in the wild normally make themselves, like the amino acid leucine. To do this, you need a mutant yeast that is an auxotroph - a yeast missing a gene essential for making leucine, for example. Normal yeast don't ordinarily starve for leucine; they can make it themselves from other nutrient building blocks. In contrast, a leucine auxotroph can't make its own leucine because a leucine-making gene is missing - it needs to get leucine from the environment. This means you can starve this mutant yeast in two different ways - the 'natural' way, by limiting ordinary nutrients like phosphate and nitrogen, and an 'unnatural way', by removing leucine from its environment. As it turns out, these two different methods of starvation produce two different starvation responses - natural starvation causes an auxotroph to almost immediately stop reproducing itself and conserve available sugar (the main source of yeast food); under unnatural starvation conditions, yeast will only gradually slow their division rate, and they keep metabolizing sugar like there's no tomorrow. This unnatural response to unnatural starvation provides a clue that a regulatory system somewhere is out of whack. A Princeton research group led by David Botstein has followed up on this clue, and they suggest that the regulatory problems in this unnaturally starved mutant yeast resemble the kinds of regulatory problems you find in cancer cells, whose infamous hallmark is that they keep dividing under conditions that would make a normal cell stop. To track down why auxotrophic yeast behave differently under unnatural starvation conditions, Botstein's group looked for mutations that made the yeast behave naturally again. They found the mutations they were looking for in a critical regulatory pathway that controls how cells change their behavior in response to stress, called the TOR pathway. The TOR pathway does essentially the same thing in yeast cells as it does in human cells - based on environmental signals, it controls the cell's decisions about growth, metabolism, and division, exactly the kinds of decisions that go wrong in cancer cells. As tumors grow, cancer cells are starved for nutrients but they don't act like starving cells - they keep dividing relentlessly, and they burn through their energy stores fast and then seek out more. We know that the TOR pathway plays a role in human cancers. The Princeton researchers suggest that some of the mutants they uncovered in their study might be worth following up in humans - perhaps we can use the TOR pathway to manipulate tumor cells the way we can manipulate starving yeast.