· Bioengineering
MIT engineers just printed a circuit board — except the transistors are alive, and the wiring is a smell.
Bacteria, arranged like components on a breadboard, can now run logic operations. Voigt says there's nothing your iPhone does that these circuits couldn't do too.
Researchers at MIT, led by Christopher Voigt and postdoc Hamid Doosthosseini, have engineered bacteria to behave like transistors, then printed them onto Petri dishes as living "circuit boards." The system uses five engineered strains of Pantoea agglomerans — two that act as transistors and three that act as relays — arranged in patterns that let signaling molecules carry information between colonies, the same way electrical current carries information between transistors. The team built circuits that add numbers, process multiple signals at once, and route signals to different outputs, including one 24-colony circuit built to add two inputs together. Each calculation takes about eight hours, but the goal isn't speed — it's eventually placing these circuits on plant roots to detect stress and trigger a biological response. The work is published in Nature Chemical Biology.
Each colony on the plate acts like a component on a circuit board, spaced precisely to control how signals travel between them.
MIT Turned Bacteria Into Living Transistors
MIT researchers have engineered bacteria to behave like transistors, making it possible to build living "circuit boards" that can be printed directly onto growth material in a Petri dish.
The project is led by Christopher Voigt, head of MIT's Department of Biological Engineering, with Hamid Doosthosseini, an MIT postdoc, as lead author. Former postdoc Haorong Chen is also a co-author.
Doosthosseini frames the potential simply: "any operation can be built with these five strains."
That's a striking claim for a system built entirely out of living cells, and it's backed by circuits the team actually demonstrated — not just theorized about.
The research is published in Nature Chemical Biology, and it reframes what "computing" can mean when the components are alive and growing rather than etched into silicon.
How a Transistor Works — And How Bacteria Can Copy It
In conventional electronics, a transistor is essentially a switch. It controls whether electrical current flows through a circuit, based on some other input signal.
The MIT team built a biological version of that same idea, just swapping electrical current for something else entirely: small signaling molecules that carry information between parts of a circuit.
| Electronic Circuit | Living Bacterial Circuit |
|---|---|
| Silicon transistors | Engineered bacterial cells |
| Electrical current carries signal | Small signaling molecules carry signal |
| Wires connect components | Spacing between colonies controls signal path |
| Operates in nanoseconds | Operates over roughly 8 hours |
In this system, engineered bacteria regulate the movement of those signaling molecules, effectively deciding whether a "signal" gets passed along to the next part of the circuit.
That's the core reframing at the heart of the project: a transistor doesn't have to be made of silicon. It just has to reliably control whether a signal moves forward.
Why Cramming a Whole Circuit Into One Cell Doesn't Work
Synthetic biology circuits aren't new. Researchers usually build them by engineering cells to produce proteins and transcription factors that interact with each other in specific, planned ways.
A typical circuit might detect a target molecule and then activate production of a particular output — a workable approach, but one with a real structural limit.
To avoid unwanted interactions inside the circuit, researchers generally need a different transcription factor for each distinct function they want the cell to perform.
Only a limited number of suitable transcription factors actually exist, which puts a hard ceiling on how complex a circuit can get while still living inside a single cell.
Packing too many functions into one cell creates a second problem too — it strains the cellular machinery the cell needs just to keep making proteins at all.
The MIT team's answer was to stop trying to fit the whole circuit inside one cell, and instead spread it across many cells working together.
Five Strains, One Flexible Toolkit
For the transistor role, the team used Pantoea agglomerans, a bacterium that commonly grows on surfaces, including plants — a detail that matters later.
They engineered two forms of this bacterial transistor, both controlled by a molecule called OC-6. One version switches on in the presence of OC-6, and the other switches off.
Each transistor can also detect a second molecule, OC-12. Depending on whether OC-12 is present, and whether the transistor is currently switched on, the cell produces a third signaling molecule: OHC-14.
Alongside the two transistor types, the team engineered three additional strains that act purely as relays, converting an incoming OHC-14 signal into a new output that a downstream transistor can then read.
Together, those five strains form a flexible toolkit — five interchangeable building blocks that can be arranged and rearranged into a wide range of different circuit designs.
Printing a Circuit Board Onto a Petri Dish
Having the right biological components is only half the challenge. The team also had to physically arrange them into something that behaves like an actual circuit.
Their solution was refreshingly literal: they printed bacterial colonies directly onto plates of agar, a standard microbiology growth medium.
Neighboring colonies were spaced about 5 millimeters apart — a distance the researchers chose deliberately, not arbitrarily.
That spacing limits exactly how far each chemical signal can travel, so a signal reaches only the nearest colony rather than washing indiscriminately across the whole plate.
The nearest colony then passes information along to the next component in the chain, creating a controlled, directional path for information to move through the circuit.
In effect, physical distance is doing the job copper wiring does in an electronic circuit board — defining exactly which components can talk to which.
What These Circuits Can Actually Do
With the toolkit and the spacing method in place, the team tested what kinds of computation these living circuits could actually perform.
A single transistor, depending on where it sat within a circuit, could carry out several distinct logic operations, including "multi-input," "or," and "imply" gates.
Combining transistors into larger systems unlocked more advanced behavior. The team built circuits able to add two signals together and process several signals simultaneously.
They also built a working demultiplexer — a circuit that takes one incoming signal and routes it to one of several possible outputs, based on a separate control signal.
The largest circuit demonstrated in the study contained 24 connected bacterial colonies, all working together to add two inputs.
Voigt summed up the underlying philosophy directly: "Computationally, there's nothing that your iPhone can do that these circuits couldn't do."
Living Circuits on Plant Roots — Slow, But Fast Enough
Here's the part that separates this from a purely academic exercise: these circuits are dramatically slower than any electronic computer, and the team is fine with that.
Voigt was clear about the actual goal: the aim isn't replacing computers, but putting computational control directly into biology itself.
If bacteria are living on a plant's roots — or if the plant itself is effectively doing the computing — running a simple overnight calculation is more than fast enough relative to an entire growth season.
That's the specific application the team has in mind: circuits placed on plant roots that detect different forms of environmental stress, like drought or a pest attack.
When a particular stress signal is detected, the circuit could then trigger an appropriate biological response — one example the researchers point to is producing a fungicide directly at the site of an infection.
Choosing Pantoea agglomerans as the base organism makes more sense in this light too — it's a bacterium that already naturally grows on plant surfaces.
Manohar Insight
What I find most clever about this project isn't the biology itself — it's the decision to stop competing with electronics on electronics' own terms.
Eight hours per calculation sounds like a weakness right up until you remember these circuits are meant to live on a plant root for an entire growing season, not a data center.
That reframing — computing at the speed biology actually needs, not the speed silicon happens to offer — is the real insight underneath the "living transistor" headline.
The five-strain toolkit is the other detail I'd flag. Distributing a circuit across many simple cells, instead of cramming complexity into one, sidesteps a real synthetic biology bottleneck.
It's a very engineering way of thinking about biology — treat cells like modular parts on a breadboard, not as one impossibly overloaded chip.
I'd watch the plant-root application closely. A circuit that can sense drought stress and trigger its own fungicide response, without any external sensor or controller, is genuinely useful agricultural technology.
The real breakthrough isn't that bacteria can compute. It's that they can now be wired together on purpose, like parts on a shelf, rather than engineered one impossible cell at a time.
Join the Signal
If a plant could sense drought stress and trigger its own defense overnight using circuits like these, would you trust that response more or less than a sensor-and-software system? Share your take in the comments.
Discussion