How MIT Built Reconfigurable Circuits From Engineered Bacteria

The printing process began with a planned layout, followed by preparation and growth of the engineered strains before they were deposited onto the agar. The experimental protocol describes growing the strains on selective agar and preparing them for printing in a multiwell format.

Reconfiguring a circuit consequently meant changing the spatial pattern: which strain was printed at each position and which colonies were placed next to one another. The underlying DNA did not need to be rewritten for each new circuit. This shifts part of the programming task from genetic engineering to physical assembly.

What the living circuits demonstrated

The researchers tested the architecture with increasingly complex operations, including:

• OR logic, in which any of several inputs can produce an output.
• Implication-type operations, which combine inputs according to a conditional logical relationship.
• A 1-to-4 demultiplexer, which routes one input toward one of four outputs according to control signals.
• Arithmetic circuits, including half-adder and full-adder arrangements.

The full-adder demonstration used 24 printed colonies. That result showed how computation could be distributed across a cooperating population rather than confined to one genetically overloaded cell.

Eight hours is slow for computers—but not necessarily for biology

A calculation took roughly eight hours. That is extraordinarily slow compared with silicon logic, where operations occur on electronic timescales many orders of magnitude shorter. These bacterial circuits are therefore not intended to replace processors or perform conventional digital computing.

Their potential advantage is different: they can sense biological or chemical conditions locally and respond without relying on a conventional electronic system. Processes such as plant drought stress, pest damage, or disease development unfold over hours or days, making an eight-hour response potentially relevant to biological control applications.

The trade-off is thus speed for biological integration. A living circuit could, in principle, detect a combination of signals directly where the response is needed and then activate a biological output.

Possible applications on plant roots and leaves

The researchers point to plant-associated environments as one possible use. A circuit placed on a root or leaf could combine signals associated with drought, insect attack, or pathogens and respond only when the appropriate pattern is detected.

A future system might, for example, trigger production of a fungicidal compound or another plant-protective output. These applications remain a research direction rather than a demonstrated agricultural product, but they illustrate why chemical computation could be useful even when it is much slower than electronics.

What the approach changes about synthetic biology

Traditional genetic-circuit design often asks how many functions can be packed into one cell. MIT’s architecture asks a different question: which functions should be separated across different cell populations, and how should those populations be arranged?

That separation can reduce the burden on individual cells while making the physical layout an additional programming layer. Genetic engineering defines what each strain can do; printing determines how those capabilities are connected. The result is a reconfigurable biological platform built from a small library of specialized parts rather than a wholly new strain for every logic design.

The work was funded in part by the Defense Advanced Research Projects Agency and the Intelligence Advanced Research Projects Activity.