In 2009, artists Alexandra Daisy Ginsberg and James King teamed up with synthetic biology students at the University of Cambridge to create the Scatalogue. The students engineered E. coli  to respond to a stimulus by changing color, creating a signal robust enough to be seen with the naked eye. The team envisioned a future in which these bacteria could detect problems in the gut, turning different colors to indicate different diseases; the artists brought the concept to life in the Scatalogue art project, a collection of cheerily colored (fake) feces to represent this hoped for advance in diagnostics.

While the Scatalogue isn’t yet a reality, it might not be far off. Advances in synthetic biology and expanded understanding of the gut microbiome enable researchers to create increasingly more sophisticated bacterial sensors that can recognize and record conditions in the complex environment of the human gut. While fecal analysis gives researchers a good idea of what’s happening at the far end of the gastrointestinal tract, it doesn’t necessarily indicate what’s happening in the several other feet of intestines between mouth and rectum (1). 

Scientists hope that one day their genetically engineered bacterial systems will enable early diagnosis and continuous monitoring for gut conditions such as inflammatory bowel disease and cancer.

The first steps

Harvard University biologist Pamela Silver spent the first part of her career studying the movements of proteins and RNAs within cells; her early work formed the basis for a drug that is currently approved to treat some blood cancers. In the early 2000s, she was ready for a change. She linked up with the Synthetic Biology Working Group at the Massachusetts Institute of Technology (MIT) and decided to pivot her team to work in a new field.

She started thinking about engineering bacteria to sense specific things in their environments. Applying these biosensors in the gut, she said, “seemed kind of obvious. The gut is a key environment where bacteria live, including E. coli. One of the tenets of synthetic biology is that we know a whole lot about engineering certain organisms, and one of them is E. coli.”

The beauty of the lambda switch is that once it's on, it stably stays on. 
- Pamela Silver, Harvard University

Bacteria evolved many ways to sense their environments; in one well characterized system, a specific promoter drives expression of a downstream gene in response to a drug called anhydrotetracycline. Silver decided to use this system in the first proof of concept study to create a bacterium able to sense conditions in the gut.

She didn’t only need the bacteria to respond to chemicals in the gut; she also needed them to “remember” the information. To endow the bacteria with memory, Silver borrowed a genetic switch from another model organism: bacteriophage lambda. The switch consists of two elements, cI and cro. Expression of one of these genes strongly represses transcription of the other, so the system is stable in the cI state (“off”) and in the cro state (“on”). “The beauty of the lambda switch is that once it's on, it stably stays on,” said Silver.

Silver engineered the bacteria so that the promoter would drive expression of cro, flipping the lamba switch to the cro, or “on,” state. She also added in a cro-driven reporter gene that would tell them when the switch had been flipped on.

Then came the test. Silver gave the engineered bacteria, all of which began in the “off” or cI state, to two groups of mice. One of the groups also received a low dose of anhydrotetracycline. When the researchers checked the bacteria coming out the other end, they found that they stayed in the “off” state in mice that hadn’t received the drug. However, in the other group of mice, the engineered bacteria reliably sensed the presence of the drug, flipping the switch to the “on” state, demonstrating that these bacteria could sense, remember, and report on conditions in the gut (2).

Trust your gut

In the years since that landmark study in 2014, research on bacterial biosensors has blossomed. Researchers moved from proof of concept studies in which bacteria were triggered by a drug to engineering bacteria that could sense molecules relevant for human disease.

Scientists aren’t engineering these sensors from scratch, though. For the most part, they’re borrowing two-component systems that already exist in nature. Two-component systems, explained Jeffrey Tabor, a bioengineer at Rice University, “are the primary means that bacteria use to sense their environments. They’re bacteria’s eyes and ears for their world.”

These systems generally consist of a sensor protein in the cell membrane, which activates in response to a specific stimulus. This protein transfers the signal to a second protein, which turns gene expression on or off, allowing the cell to respond appropriately to the stimulus.

Even so, finding a two-component system that responds to a stimulus that is relevant for human disease isn’t necessarily straightforward. In fact, Tabor’s discovery of a sensor for a molecule involved in gut inflammation happened almost by accident.

For this project, he said, “we were originally looking for bacterial sensors of environmental pollutants.” They noticed that one of the marine bacteria they were studying had a novel sensor that likely controlled an enzyme for using thiosulfate, which some bacteria can “breathe” when oxygen is absent. They guessed, correctly it turned out, that the sensor allowed bacteria to sense thiosulfate in the environment.

This wasn’t particularly useful for their research on pollutants, but Tabor knew that thiosulfate also associated with inflammation in the gut. This sensor then, might be useful for monitoring inflammatory conditions in the mammalian gut.

Tabor borrowed this sensor from the marine bacteria and engineered it into a strain of probiotic mammalian gut adapted E. coli  known as Nissle 1917. He also added a green fluorescent protein reporter.

Sure enough, when researchers administered these engineered bacteria to mice with experimentally induced colitis, the bacteria’s trip through this inflammatory environment made them glow green (3).

While this experiment was successful, the technique needs some finetuning before it’s ready for human use. One of the big challenges is making the signal easier to see. If the diagnostic is administered as a sort of bacteria laden milkshake, Tabor said, the first problem is that, “99 percent of those bacteria will die in the stomach, give or take. The ones that don’t die get spread out along the 28-foot tube that is your gastrointestinal tract.” This can result in a weak signal once the bacteria reach the far end. In the mouse study, scientists had to use flow cytometry, a single cell analysis technique, to identify the signal from the bioengineered bacteria.

Patients, and probably many general practitioners, don’t have access to flow cytometry, so the team is currently working on ways to make the signal more visible. Tabor and his team are developing a strategy to stick the bacteria together using hydrogels, which will hopefully help more of them survive their journey through the stomach and keep them together so that the signal is concentrated enough to be visible to the naked eye once the bacteria are excreted.

Tabor imagines a product like this one day serving as an early warning system. “Say you’re a Crohn's disease patient and your doctor asks you to drink this bacterial solution every morning at your house. And if your toilet water turns blue, you are having an inflammatory flare, and you need to go see the doctor.”

Human gut disorders are complex and involve many more factors than thiosulfate. “We are pretty sure we're going to need to find more sensors from human gut bacteria to reliably sense human inflammation,” said Tabor. “What we're doing is getting clinical samples — fecal samples from healthy people and sick people — and we're screening them against lots and lots of two-component systems from human gut bacteria in order to discover clinically translatable human inflammation biosensors.”

Cell biographies

Although bioengineered sensors that can detect one or even a few different biomarkers could be useful as diagnostics or early warning devices, they can’t really provide scientists with a complete picture of everything that’s happening in the incredibly complex gut environment.

Randall Platt, a bioengineer at the Swiss Federal Institute of Technology, thinks he may have a potential solution. Platt had always been interested in CRISPR and gene editing, but when he founded his own lab in Zürich, he wanted to explore CRISPR’s recording capabilities.

In nature, CRISPR-Cas systems take genetic information from an attacker, like a bacteriophage, and integrate it into the bacterial genome to provide the bacterium with a sort of immune “memory” of the attacker. Platt and postdoctoral researcher Florian Schmidt wanted to see if they could modify this system to record other types of “memories.”

We can create biographies inside of cells in any environment. 
- Randall Platt, the Swiss Federal Institute of Technology

In 2018, they successfully used this system to grab a snippet of the cell’s own RNA, convert it back into DNA, and integrate it into the genome, preserving this piece of transcriptional history (4). While most of the cells grab only a single snippet, by using a population of cells, they created a more detailed record of gene transcription, providing information about how cells responded to different environments. Even though this isn’t a direct recording of the environment, a lot can be inferred about the environment by measuring how the bacteria respond.

This project started as a theoretical exercise; Platt’s team just wanted to see if they could do it. “Once we succeeded, we realized how powerful it could be. We can create biographies inside of cells in any environment,” said Platt. “My interest is in human health and trying to help people and patients. And so, we wanted to see if we could use this to gain some really valuable clinical information about what's happening inside of the gastrointestinal tract.”

The research team found that in monocolonized mice (mice colonized with just the reporter bacteria), the bacteria recorded different patterns of gene expression based on the diets the mice ate, or whether they had experimentally induced colitis (5).

Platt said that the diet analysis, “is where the technology really shines, because E. coli  have evolved to be experts in adapting to different carbon sources.” By analyzing transcriptional memory, researchers got clues about what carbon sources the bacteria used, which wasn’t necessarily constrained to what the mouse ate.

“We could tell which diets were good for the mice to be eating,” said Platt. “If the bacteria didn't like the diet, in the sense that it wasn't rich in a range of carbon sources and it was only enriched in bad things like starch or fat, the E. coli  would start to eat mucus carbon sources. So essentially, if you don't feed the E. coli, they feed off of the mucus and deplete that barrier.” The protective layer of mucus lining the gastrointestinal tract is crucial for maintaining a healthy gut, so if bacteria consume it, this could lead to problems.

In the near future, said Platt, “we see these bacteria as a great way to make nutritional assessments in individual people, to figure out which diets are good for individuals to eat in a therapeutic context in which patients need appropriate nutrition to manage their other health conditions like cancer or diabetes.”

Further in the future, Platt hopes that these reporter bacteria will help unravel some of the mysteries surrounding irritable bowel syndrome (IBS), the many varied causes and mechanisms of which are poorly understood.

“IBS is more of an umbrella term for a whole bunch of different conditions related to things like intolerances or malabsorption.… Part of the difficulty in labeling these as proper disorders might be that there's just no way to measure what's actually going on in the intestine,” said Platt. “We believe if we put our tool in, we can start to understand the mechanistic basis of what makes one patient different from another and then know how to take action in each case.”

A perfect partnership

Tabor’s and Platt’s approaches have their own advantages and specific use cases, but both techniques require the bacteria to exit the body before they can report on what they found inside. Timothy Lu, an MIT bioengineer, might have a way to get answers faster.

Lu originally trained as a computer scientist, but developments in genetics and genetic engineering, including the Human Genome Project, prompted him to take his career in a new direction.

Bacteria aren't good at communicating over long distances. So, we had to figure out how to combine these together with some other modality, and electronics made a lot of sense. 
- Timothy Lu, MIT

“The idea that you could think about DNA as a kind of code was really appealing,” said Lu. “So, I switched fields and jumped into the field of synthetic biology. And it's been fun ever since trying to figure out how to program cells to do different things.”

During his early work on bacterial biosensors for the gut, Lu recognized some inherent shortcomings. While bacteria are easy to administer, said Lu, “the downside is that it's hard to get more granular detail about where you are in the gut and what’s going on in real time. You have to wait for the bacteria to come out the other end. Bacteria aren't good at communicating over long distances. So, we had to figure out how to combine these together with some other modality, and electronics made a lot of sense.”

Lu teamed up with MIT electrical engineer Anantha Chandrakasan to see what a partnership between bacteria and electronics could achieve. In 2018, this pair debuted a bioelectronic sensor that detects bleeding in the gut (6). Lu and his team engineered a bacterium to produce luminescence in response to heme, which is present in blood. They packaged these bacteria with an electronic sensor to measure the luminescence output and wirelessly transmit this information to a computer or smartphone. In tests, this bioelectronic package detected blood in pigs’ gastrointestinal tracts in near real time.

Next, Lu wanted to make a few upgrades. First of all, at more than 3 centimeters long, the original device was too big for human use. With the help of fellow engineers Rabia Yazicigil of Boston University and Giovanni Traverso of MIT, he shrunk the capsule down to less than 1.4 cm in length, smaller than a typical pill camera. In work described in a preprint, the researchers also created a multifunction sensor with different types of bacteria that report on four different markers associated with inflammatory bowel disease (7). In the future, Lu envisions a sensor with a dozen or more types of engineered bacteria to sense different chemicals in the gut for even greater precision when diagnosing and monitoring disease.

While many researchers, including Tabor and Lu, focused on markers of gut inflammation, this certainly isn’t the only application. Jeff Hasty’s team at the University of California, San Diego showed in a preprint that engineered bacteria also detect colorectal cancer in mice (8).

Platt said that unraveling the mystery of what’s happening in the gut is likely important for a slew of diseases. “There's so much we don't know about the human intestine, but we know that it's really, really, important. It's connected to all sorts of diseases from diabetes to inflammatory bowel disease. It can even change your behavior.”

Whether packaged in hydrogels or working in concert with miniature electronics, whether sensing one environmental factor or many, scientists hope that their approaches will not only enable us to understand gut processes more fully, but also provide diagnostics and monitoring to help inform management of life altering diseases.

References

1. Tanna, T., Ramachanderan, R. & Platt, R. J. Engineered bacteria to report gut function: technologies and implementation. Current Opinion in Microbiology  59, 24–33 (2021).
2. Kotula, J. W. et al. Programmable bacteria detect and record an environmental signal in the mammalian gut. Proc Natl Acad Sci U S A  111, 4838–4843 (2014).
3. Daeffler, K. N.-M. et al. Engineering bacterial thiosulfate and tetrathionate sensors for detecting gut inflammation. Mol Syst Biol  13, 923 (2017). 
4. Schmidt, F., Cherepkova, M. Y. & Platt, R. J. Transcriptional recording by CRISPR spacer acquisition from RNA. Nature  562, 380–385 (2018).
5. Schmidt, F. et al. Noninvasive assessment of gut function using transcriptional recording sentinel cells. Science  376, eabm6038 (2022).
6. Mimee, M. et al. An ingestible bacterial-electronic system to monitor gastrointestinal health. Science  360, 915–918 (2018).
7. Inda, M. E. et al. Ingestible capsule for detecting labile inflammatory biomarkers in situ. 2022.02.16.480562 Preprint at https://doi.org/10.1101/2022.02.16.480562 (2022)
8. Cooper, R. M. et al. Engineered bacteria detect tumor DNA. 2021.09.10.459858 Preprint at https://doi.org/10.1101/2021.09.10.459858 (2022)