The bacterium E. coli may be the most-studied organism on Earth, but it continues to give up secrets. A Boston University engineer has, for the first time, mapped more than 1,000 interactions in the cell’s regulatory network, potentially opening the window to understanding how bacteria mutate and develop resistance to antibiotics.
Timothy Gardner, an assistant professor of biomedical engineering, said that bacteria have strong survival skills. “They respond to varying conditions, nutrients, toxins, stresses and competition from different organisms in constantly varying environments by basically changing their metabolic program,” he said. “The system that orchestrates these responses is not a trivial system. It is actually hundreds of genes that are connected in an interdependent network.”
In a study published in the journal PLoS Biology, Gardner and his research team looked at existing models of metabolism in E. coli, then came up with an algorithm to make predictions about what would happen if they tweaked a cell, akin to flipping a switch.
They collected data sets on about 2 million data points detailing E. coli genes and regulatory interactions between them, then ran the data through the algorithm. “It’s like as if there were a black box that you can see some things coming out of,” said Gardner, who was recently elected to the governing council for the Institute for Biological Engineering. “By studying precisely all the things that come out, you can infer what’s actually inside the black box.”
They found that their algorithm exceeded the accuracy of existing models by about 40 percent. “It could predict regulatory interactions from the data with about 80 percent confidence, so about 8 out of 10 predictions were correct,” said Gardner. “For computational work in biology, that’s an extremely good performance.”
Having proved with E. coli that the algorithm could calibrate the control system, the researchers are now expanding their data set to try to map microbial interactions quickly and thoroughly in other bacteria. As Gardner explains it, their approach is like using a satellite to map a coastline rather than sailing a boat around it.
Their strategy so far has been to target pathways of virulence or infection, coordinated group action and antibiotic tolerance. Initially, the researchers disturb the physiological mechanisms to induce group responses. They then try to examine virulence expression and responses, and then mix those responses with different types of stresses and environmental conditions to see if there is a relationship between them. They also plan to do antibiotic studies where they start to treat the virulence with combinations of antibiotics to see how those responses interact with the virulence pathways.
While Gardner is mapping the regulatory network in the shewanella oneidensis bacterium, wherein metabolic reactions produce electricity, his group is applying the same methods to a practical problem half-way around the world. They are collaborating with researchers in South Korea who are trying to develop cold-tolerant plants for the country’s agricultural system. The Korean team had been treating model plants with cold temperatures and then measuring the plant’s genetic response.
“They were overwhelmed by the data,” said Gardner, “so we set up an experiment where they did cold permutations and then they turned off about 40 different genes, each at a separate time, and measured the network response.” At BU, Gardner’s team then applied their algorithm to the data and identified all six of the currently known regulatory circuits, and predicted about 50 more.