ONE | CRYSTALLOGRAPHIC RENDERINGS

Diane is the head of a protein crystallography laboratory. She is a tenured faculty member based in the departments of chemistry and biology at research institute on the East Coast of the United States. Her laboratory hums with activity. It takes up a number of rooms on the fourth floor of a large building that also houses chemistry laboratories and several other academic units, including the anthropology department. The groups' dozen graduate students and postdocs occupy a large open space lined with laboratory benches. Stacks of used petri dishes, test tubes, Styrofoam containers, used pipette tips, and beakers are scattered across the surfaces of their black laboratory benches. Some benches have apparatuses to purify proteins or automated PCR (polymerase chain reaction) machines for amplifying specific strands of DNA. Microscopes, water baths, hot plates, and shakers take up the remaining surfaces. Shelves lining the walls are stacked high with darkened glass jars full of chemicals, fresh supplies wrapped in plastic, and gleaming glassware. The fume hoods that line the walls are filled with instruments that must be carefully calibrated. Personal workspaces are decorated with pictures of family, friends, comic strips and cartoons, and one or two of the desks sport glamour shots of celebrities.

The machines scattered through the lab are part of a larger assemblage of attendant refrigerators, incubators, and centrifuges that take up space in the halls and adjoining rooms. Vapors billow forth from liquid nitrogen canisters and -80°C freezers. A massive X-ray diffraction machine is housed in its own set of rooms. A protective glass wall separates this machine and its radiation from users and the computers they use to process diffraction data. Down the hall a darkened computer room equipped with several workstations feels like a quiet sanctuary. It is here that lab members render their long-sought-after crystallographic data into three-dimensional computer graphic models.

Just down the hall from the computer lab is Diane's office. When I arrived for our first interview, her aging dog, Max, greeted me with sweet, mournful eyes. His arthritic gait told me he wasn't long for this world. He was Diane's constant companion in the lab, and her grad students would come by throughout the day to take him out for walks. His presence sometimes made the lab seem like a family, especially when he came to the weekly lab meetings. These group meetings gave Diane and her students opportunities to present their progress on current research projects, as well as seek support to navigate difficult problems. At the same time this was also their opportunity to negotiate who was going to take on routine chores, tasks that involved cleaning communal workspaces and maintaining equipment. If some took up these tasks with a sense of collegial duty, others had to mute their disdain for washing glassware or cleaning out the refrigerators.

In our first interview, I learned a lot from Diane about what motivated her to pursue a career in the field of protein crystallography. Diane studied chemistry as an undergraduate student. One of the questions that fascinated her was how enzymes and their substrates interacted to catalyze biochemical processes. She explained: "I really liked the idea of trying to understand enzymes. What did they do? How did they catalyze this reaction? What was the detailed mechanism involved?" She wanted to be able to "see" biochemical reactions unfold: "I figured out that I couldn't think of working on a project if I didn't understand what it looked like. I needed that first.... You have to have some kind of concrete thing to start with, even if it's just one picture. At least there is something more tangible involved there. It was just that it seemed to me that that was the starting place for science. The first thing you ask is, 'What does it look like?'" This desire to see what proteins "look like" is evocative of the efforts of nineteenth-century natural historians' efforts to describe "nature's panopoly" through drawings, images, and models. It is only once they could see what proteins "look like," that modelers could begin to compare and contrast these remarkably diverse forms. What Diane was proposing seemed at first like a high-tech, high-resolution natural history of enzymes.

Protein crystallography was the technique that could give her visual access to these unseen dimensions of cellular life. Yet, this was more than a descriptive project. It was only with a three-dimensional model of the precise atomic configuration of a molecule that she would be able to discern the fine details of a protein's "active site," the area in the molecule that reacts chemically with other molecules. This was crucial knowledge if she was going to be able to figure out how that protein interacted with other molecules and how it catalyzed chemical reactions in the cell. Crystallographic models were "tangible" objects that could give her "concrete" access to these molecular interactions. The models she built would enable her to design experiments to intervene in an enzymatic reaction, and these interventions would help her interpret what particular proteins were up to inside the cell.

When Diane was ready to begin her graduate work in the 1990s, however, many of her mentors saw protein crystallography as a dead-end discipline. This was a time when rapid advancements in molecular genetic techniques and recombinant DNA technologies were shaping the direction of research questions and funding. Researchers in the biological sciences had already diverted their interest and investments from atomic-scale models of proteins to the automation of genetic sequence analysis. Informatic models of the genetic determinants of life held sway and molecular genetic studies dominated the life sciences. Diane was warned not to pursue training in protein crystallography. In some ways, she was getting sound advice. It was hard to build models of protein structure. It took Nobel Prize laureate Max Perutz twenty-two years to solve an atomic resolution crystallographic structure of just one protein molecule. His crystallographic model of hemoglobin, published in 1967, was at that time only the second high-resolution protein structure to be determined. Twenty years later faster computers and better equipment helped to speed things up. But by 1990 the Protein Data Bank housed just 486 protein structures determined by this technique. The future of the field did not look bright.

Yet, Diane's desire to gain visual access to the molecular realm was so strong that she ignored the sage advice, and began graduate work in protein crystallography. Her perseverance, it seems, paid off. By the time she began her tenure-track academic position in 1999,...