Making bacteria behave
Electrical engineer programs cells to do his bidding
By Steven Schultz
Princeton NJ -- In Ron Weiss' lab, bacteria do as they are told. Weiss, after all, is not a biologist -- not a student of nature trying to understand how life works in all its complexity. He is an electrical engineer, someone who builds machines that work as he wants them to, when he wants them to.
"My motivation is to be able to program cells," said Weiss. His idea is that cells could function as computers, even robots, fully programmable and able to respond to cues from each other, their environment and human operators. Just as natural cells, already programmed by their own genes, build a bacterium's flagellum, a snail's shell or a person's brain, so an engineered cell might carry out equally intricate jobs at a human's command.
"Say I want to do some kind of complicated task, such as building myself a house or sensing whether I have some toxic agent in the environment and then going and attacking it, or trying to fix the tissues and organs in my body because they are failing on me can we program cells to do all these various tasks?" Weiss asked.
Although his work to answer that question is at an early stage, he already has begun to create a toolkit of basic information-processing and communications devices -- like the chips and network cards of computers -- that can be inserted into cells.
His research makes him a pioneer in a specialty that is barely five years old. While biologists are adept at editing DNA and engineering cells to create new medicines and crops, few have tried to recruit cells as general-purpose machines, said Tom Knight, a professor of electrical engineering at the Massachusetts Institute of Technology and one of Weiss' doctoral advisers.
"Ron is one of the few people in the world who have actually signed up to this agenda and are acting on it," said Knight. "He is in there doing it and is training a set of students to do it. That is impressive."
A risky decision
Weiss' specialty is so new that one of its key developments was his own Ph.D. dissertation, which he completed last year. Just five years ago, as an electrical engineering graduate student at MIT, he was on a more conventional track. His project was to study how computing would change if engineers could program billions of devices at once, rather than restricting their thinking to the silicon chips and hard-plastic cases of today's machines. As part of this research, he thought about how biology tackles the very same problem. Eventually, he tired of merely thinking about it and decided to actually try to program organisms.
"My advisers warned me against it," Weiss said of his change in plan. His original project was close to completion, and the new one would require learning the entirely foreign field of experimental biology. With his advisers' eventual approval, he set about building a working biology lab, signed up for biology courses and began reading lab manuals. It was a nerve-wracking six months before he accomplished the most basic feat of genetic engineering, mastering a technique that is second nature to many molecular biology graduate students.
It was another year before he used the technique to accomplish one of his first goals: making the biological equivalent of a device called an inverter. An inverter is one of the most basic elements of a computer chip; it takes a positive signal and flips it to a negative, a one to a zero. Instead of the electrical signals of conventional computers, the information currency in Weiss' bacterial circuits would be proteins. A high concentration of a certain protein would be a "one," a low concentration a "zero."
In the fall of 2000, after several roadblocks, Weiss found the combination of six genes that processed the proteins in the way he wanted and turned his E. coli cells into little digital inverters. The cells sensed the concentration of one input protein and pumped out high or low concentrations of another in response.
That day is etched in his mind. "I ran to my adviser and said, 'Look, my cells are working!'" Weiss recalled. "It was a rush -- just the whole notion of me being able to make these cells perform digital logic and then seeing it work with my own eyes was very exciting."
Weiss received his Ph.D. in 2001 and came to Princeton as an assistant professor of electrical engineering that fall. He is teaching a graduate seminar in cellular and biological computing and has again built a biology lab from scratch, this time in Princeton's E-Quad. He also has begun to establish collaborations with biologists, physicists and others interested in the intersection of biology and engineering.
"The nice thing about Princeton being a small community is that people are not confined to working only in their own departments," said Weiss. "Collaborating across departments is standard behavior at Princeton, and that is very important for the work I am doing."
So far, Weiss and his students have assembled some of the standardized elements needed to make digital circuits within cells. In one petri dish, cells in group A send a signal to group B, which signal group C. He also is continuing his original project to make a set of logic-performing switches, like his first inverter.
"We have to figure out what are the primitive components in building a digital circuit inside a single cell, so it can do the same things over and over, reliably and predictably," Weiss said.
While most of that work involves applying the language of engineering to biology, sometimes the engineering effort draws on concepts from biology. In collaboration with colleagues at the California Institute of Technology, Weiss is using a technique called "directed evolution," an accelerated version of natural selection, to mutate his gene circuits until they have the desired properties.
In another project, Weiss programmed cells to emit different signals depending on where they are in the petri dish -- a primitive version of what an embryo does when it begins sprouting arms, legs and internal organs from a mass of undifferentiated cells.
In some sense, making cells do even some of the most outlandish things Weiss imagines is more a matter of tweaking than inventing. "That's exactly why this sort of thing is potentially doable, because we're not reinventing chemistry from scratch," he said. "We're reusing existing mechanisms in the cell.
"Cells already do all these things," he continued. "They coordinate their activity. They grow. They self heal. They communicate. They produce materials. We're materials, right? They make us. They make bones, so they can produce various other structures that have integrity to them."
As his work has advanced, Weiss has become increasingly intrigued by the mystery of those natural processes and hopes that the lessons he learns in building genetic circuits will help biologists understand principles behind nature's own engineering. "The more I do this work, the more I am fascinated by those questions," he said. "As our ability to modify things increases, so does our ability to understand them."
Editor: Ruth Stevens