How we approach the maintenance and upkeep of our bodies is astonishing. We almost never look under the hood unless we are feeling unwell, then we need to go to a specialized facility in the hopes of getting tests performed, these tests require expensive lab equipment and trained personnel, and based on single measurements using these tests you will either be sent home or potentially diagnosed with a disease.  There are virtually no other complex machines we treat this way. Indeed, if the only time we ever looked under the hood of an airplane was when it started to wobble in flight, we would never fly!  We don’t treat planes (or trains or automobiles) in this way. Instead they are chock-full of sensors that allow us to monitor their operation continuously in real-time and detect deviations from normal operating parameters ideally before they become problems. And yet for the one machine our lives literally depend on, our own bodies, we are flying blind.

The burgeoning field of wearable health devices promises to solve this problem through low cost portable computing devices bristling with sensors that can monitor our health and our environment. The real challenge is the fact that current sensors for biology are extremely limited. Most existing solutions are not portable, continuous or real-time. Indeed, today there is only one commercially successful, portable, continuous biosensor available: the continuous glucose monitor.

My colleagues, Catherine Klapperich, Allison Dennis, and Mark Grinstaff, and I set out to address this critical lack of biosensors. And it turns out a solution is all around us. What is really good at sensing biology is other biology.  Microbes in particular are an enormous and untapped reservoir for portable, inexpensive and diverse sensing proteins. And microbes have evolved over 3 billion years to detect and respond to virtually all stimuli relevant to human biology, biotechnology, and our environment.  Remarkably, microbes are also the cornerstone of the continuous glucose monitor: this device senses glucose using the protein glucose oxidase derived from microbes.  

Inspired by this success, we have developed a method to identify the many other sensing proteins produced by microbes, isolate and characterize these proteins, and engineer them into biosensor devices. 

Our paper in Nature Communications (doi:10.1038/s41467-020-14942-5) reports the first proof-of-principle of this novel approach. This focus of the paper, with lead authors Chloe Grazon and R Baer, is a class of proteins that that are new to sensor devices: bacterial allosteric transcription factors (aTFs).  In the paper we describe a method for mining isolated bacteria for aTFs for specific analytes, and use it to discover and characterize the first known bacterial aTF that senses progesterone.  We also developed a method for transducing analyte recognition by any aTF into an optical signal, and used this method to develop a real-time, optical progesterone sensor. The progesterone sensor, validated in artificial urine, is sensitive enough for clinical use and is compatible with an inexpensive and portable electronic for point-of-care applications.

The paper provides a foundation for developing an array of sensors by identifying bacterial aTFs for specific target molecules.  And we are developing additional technologies to transduce, immobilize, and deploy these sensors into biosensor devices to provide solutions appropriate to diverse applications.

Since our initial success, we have also substantially expanded our screening platform.  Our initial platform was limited to identified and culturable microbes.  But these represent only a small fraction of the total microbial diversity.  Our next-generation approach enables metagenomic screening directly from DNA isolated out of complex microbial communities.  And the screen directly targets not only aTFs, but also redox proteins like the glucose oxidase that powers the continuous glucose monitor.  Moreover, because we identify both the protein and its genetic blueprint, the gene, we are able to deploy both rational protein engineering and directed evolution to optimize the proteins parts for sensing applications.

Our approach thus has the potential to mine the full diversity of microbes for sensors for a range of applications in healthcare, biotechnology, drug development, environmental management, agriculture, food and water safety, consumer technology, biological threat detection, and the military.

And one of the most exciting aspects of this work is the remarkable interdisciplinary team of investigators that have come together to help realize this vision. This team includes our colleagues Karen Allen and Doug Densmore who are helping us develop the next generation of engineered proteins and lab automation that will be leveraged for sensor development.  And it includes an extraordinary team of undergraduates, graduate students, and post-docs working together across disciplines to learn from each other and engineer cutting edge technology.  Our project meetings are a remarkable sharing of ideas spanning molecular biology, chemistry, materials science, computational biology, device development, and electrical engineering.  They are also the most fun I have every week.  If you are interested in joining us, or just learning more, please feel free to reach out to us.   

Oh, and one more thing: if we and our colleagues in the biosensor community are successful in developing the next generation of real-time continuous biosensors, this will not just change the way we monitor our health.  It will also enable fundamentally new methods for treating diseases based on engineering principles.  Excitingly, we already have an example of this – the Bionic Pancreas, developed by our colleague Ed Damiano using the continuous glucose monitor.  To learn more, check out his Ted Talk.  And be prepared to be inspired.