Flexible Molecular Computer Functions Inside a Cell
A newly created molecular computer works in human cells and offers the flexibility of a general-purpose circuit. The advance, described in Nature Biotechnology in May, brings closer the eventual possibility of placing bio-based computers inside cells to diagnose and treat disease on a cellular level.
“In theory, there is no limit to the decision-making complexity” that this system can handle, says bioengineer Yaakov Benenson, PhD, a Bauer Fellow at Harvard University’s Center for Systems Biology. Until now, molecular computers have mostly been test-tube prototypes that tackled just one specific task, such as tic-tac-toe. Benenson, along with Ron Weiss, PhD, assistant professor of electrical engineering at Princeton University, devised a way to engineer a general purpose circuit by taking advantage of the cell’s cell-regulation pathways.
The scientists first got their machinery to work inside a human kidney cell by mimicking a virus. They transfected the cell with genes that code for the circuit. The cell then took up the genes and created the computer network for them.
The network itself is made up of engineered mRNA strands that encode a chosen protein and smaller RNA strands that interfere with the translation of the mRNAs. Scientists can engineer these small, interfering RNAs (siRNAs) to bind any number of possible disease markers in a cell. (Weiss and Benenson did not experimentally validate sensing disease markers in this work.) In the simplest scenario, once an siRNA binds a disease marker, that siRNA can’t interfere with translation of the mRNA, and the protein is made. The protein can be whatever the designer likes: a fluorescent tag or therapy for the diseased cell, for example.
By adding interacting pairs of mRNAs and siRNAs, the researchers can individualize the network to handle any problem that can be expressed as a Boolean logic formula—equivalent to Boolean operations run on traditional silicon-based computers. The formula could be simple: “If marker A or marker B is present, then make the protein.” Or it could be much more complex: “If marker A is present and marker B is absent, or if marker A is present and marker C is present, or if marker D and E are both absent, then make the protein.”
Weiss and Benenson tested their system using a network of five siRNAs and two mRNAs. Complex functions, Benenson says, are limited by the scalability of the molecular components.
According to Darko Stefanovic, PhD, associate professor of computer science at the University of New Mexico, many functions “will require unacceptably complex forms.” Yet, Stefanovic comments, “the paper presents an innovative way of accomplishing logic computation using transcriptional networks.” It’s a promising direction, he says, for synthetic biology.