Every day in her lab, Carol Gross plots the steps to one of the

most intricate dances in nature's repertoire. Gross, vice chair of the UCSF's department of microbiology and immunology with a joint appointment in the School of Dentistry's department of stomatology, studies the bacterium E. coli for clues to how living cells maintain themselves against environmental stresses, particularly heat. At the business end of her microscope, somewhere deep in her petri dishes, are hints to the rules governing life itself.

The processes Gross has discovered are bogglingly complex and startlingly efficient. "This biological system does something over and over again," says Gross. "It adapts, and responds, and responds in a graded way to stress it senses. It's a very complex feedback system, and we study E.coli because we can understand it in bacteria." When you heat an

E.coli cell, DNA transcription of certain genes shifts into high gear - and that, biologically speaking, makes perfect sense. Sudden stress surely requires that a cell defend itself, probably with proteins whose manufacture is directed from the genome. But how exactly does a cell learn of its own distress? What cascade of chemical reactions manages E.coli's response to heat?

It's not a purely academic question: Though Gross works with harmless strains of E.coli, her findings may point the way toward disabling the pathogenic bacteria that so regularly make headlines. Even more importantly, the homeostatic mechanism at work here is highly conserved from one species to the next. The methods E.coli uses to maintain itself are not so different from those employed by the cells in our own bodies.

Compared to human cells, E.coli is a simple environment. Each cell is composed of an inner cytoplasm and an outer periplasm: a sphere within a sphere. Both compartments are surrounded by membranes (the double membranes help account for the bacterium's resistance to antibiotics). The cell's DNA, of course, is contained in its cytoplasmic core.

When the E.coli cell is unstressed, DNA transcription of genes involved in self-protection moves at a relatively slow pace. This, Gross has shown, is the result of the inhibition of at least two co-factors, sigma E and sigma-32, required by an all-important enzyme, RNA polymerase, to complete the transcription of genes related to defense. Sigma E is bound in the cytoplasm by another protein, RseA, which straddles the inner membrane, lying partly in the cytoplasm, partly in the periplasm. Sigma 32 is held in check by a mysterious protease whose functions Gross's lab is just now delineating.

All that changes when E.coli begins to suffer. At 40 degrees centigrade and higher, the cell launches a two-pronged defense. Creation of proteins to protect the cytoplasm depends on activation of sigma 32, Gross has found, while preservation of the periplasm depends on activation of sigma E.

That the cell has a double-fisted approach to defending itself is interesting enough. But Gross' lab has also mapped the cascade involving sigma E in fascinating detail.

The most abundant proteins in the cell, called porins, are the primary ingredient of its outer membrane. Porins must travel from the cytoplasm through the periplasm, then fold precisely for integration into the outer membrane. As heat rises, Gross has found, newly manufactured porins begin folding more slowly. As a result, part of the porin's molecular structure, the C terminus, remains exposed in the periplasm when normally it would be hidden, and it binds to a periplasmic protease called DegS.

The protease, normally moribund, suddenly targets and slices that strand of RseA that lies in the periplasm. Then a second protease, YaeL, swings into action, cleaving from the inner membrane that part of RseA that lies in the cytoplasm. Now disabled (and presumably on its way to being degraded by other proteases), RseA can no longer bind sigma E, which becomes available again in the cytoplasm for use by RNA polymerase. Transcription of the genome accelerates, and soon the cell is making proteins with a variety of functions related to defense of the periplasm: fortifying the outer membrane with lipids, for instance, and marshaling more porins toward the surface.

It has taken years for Gross and her colleagues to describe this regulatory response in E.coli, a cell that contains seven known sigma factors and nearly 400 transcription factors. The work is valuable not just because it connects the dots between molecules as they interact, but because it tells us something about the rules governing these interactions - the music to which this dance is choreographed.

"Not every regulatory system has to be studied in this level of detail," Gross says (perhaps with some relief). "You just want to see the repertoire of principles working here."

In E.coli there are intriguing glimmers into how human cells sustain themselves - and how they may eventually be manipulated. The cellular protease DegS, for instance, is held inactive in the bacterium's periplasm by a part of its own molecular structure called the PDZ domain. The same inhibitory strategy may be used by human mitochondrial proteins involved in apoptosis (cell death). Detailing how the PDZ domain switches E.coli's protease on and off, then, conceivably could yield a new way to induce self-destruction, or prevent it, in human cells.

Indeed, Gross and her colleagues, part of the first wave of Mission Bay colonists to UCSF Genentech Hall, are now investigating how transcription of the genome itself varies in response to environmental stress. "The strategies we're finding are universally used," she says. "But we couldn't find them without the genomic resources available at UCSF. It's a great time to be doing this kind of research."