ROSALIND Part 1 – A cell-free biosensor platform for the rapid detection of water contaminants.

Regulating in vitro transcription of a fluorescence activating RNA creates a powerful new platform for the rapid, on-site detection of waterborne contaminants.
ROSALIND Part 1 – A cell-free biosensor platform for the rapid detection of water contaminants.
Like
Day 1 of the Lucks Lab at Northwestern University, August 1 2016. I would officially begin my postdoctoral fellowship two weeks later.

In the summer of 2016 I arrived in Evanston, Illinois to work as a postdoctoral fellow. I had no intention of doing a postdoc, but The Lucks Lab had been recruited to Northwestern University’s Center for Synthetic Biology and Julius Lucks had asked me to join his lab with a simple request to “bring your ideas.” It was an offer of intellectual freedom that was hard to refuse and a chance to embed myself within the synthetic biology (“SynBio”) community that was being grown in Chicago. However, the full story behind our paper, “Cell-free biosensors for rapid detection of water contaminants,” starts a few years earlier.

My first exposure to SynBio occurred as graduate student at the University of Missouri. A take-home exam for my second semester biochemistry course prompted me to “build on work by the Synthetic Biology Group of Prof. Christina Smolke.” We were tasked to devise an RNA sensor for intracellular magnesium concentration and as I dove into the literature I quickly became enamored by the abstraction of biology as a programmable platform. Despite not working in a SynBio lab, I geared my dissertation work on RNA aptamers towards the field of synthetic biology and stayed on top of what was happening in the field, in large part due to the community of synthetic biologists on Twitter.

The exam prompt that started it all.

In 2014, Keith Pardee et alia published their seminal manuscript “Paper-based Synthetic Gene Networks” in which they described the use of cell-free systems (capable of performing transcription and translation) as a “stable, sterile, and abiotic” platform that “extends laboratory capabilities out of the lab and into the field” through freeze-drying. They demonstrated a variety of sensors, but perhaps most impressive was their work in developing strain-specific sensors for Ebolavirus using programmed RNA toehold switches. My fascination with this work led me to apply and gain acceptance to the Cold Spring Harbor Laboratory (CSHL) SynBio Summer Course, which offered a module on cell-free systems instructed by the course co-founder, Julius Lucks. Truthfully, I had no idea who Julius was, but I was intrigued by his group’s creation of small transcription activating RNAs (STARs) and their development of SHAPE-Seq technology to probe RNA structure and function in high-throughput. An RNA synthetic biologist seemed like an interesting person to learn from.

Unbeknownst to me at the time, Julius was using the CSHL SynBio summer course as a test bed for new ideas using cell-free systems – the inaugural course (2013) even managed to pull together enough data for a publication characterizing the dynamics of RNA circuits in cell-free systems. Our 2015 summer cohort was tasked with pushing these ideas forward by implementing thresholding strategies that would allow us to tune the sensitivity of cell-free diagnostics. Although these early attempts were unsuccessful, the environment at the course actively encouraged and facilitated open-ended experimentation. Within days of my arrival at CSHL, the team approved and paid for me to have my plasmids shipped from my graduate dissertation lab to CSHL. This allowed me to perform the initial experiments that led to the demonstration that an engineered split fluorescent aptamer can be used to directly visualize RNA toehold hybridization – dissertation research that I would later publish in ACS SynBio and bring with me to the Lucks Lab. 

Exploring thresholding ideas at the CSHL SynBio summer course, August 2015.

When I arrived in Evanston in 2016, I brought with me—as requested—a few ideas. This included a CRISPR project de jour and freeze-dried transcription-only sensors, the latter of which was motivated by my experience with fluorescence activating RNAs. We reasoned that transcription-only sensors would be cheaper, faster, and more predictable than cell-free transcription-translation systems. We knew that a fluorescent RNA could replace a fluorescent protein as a sensor “output” in a cell-free system, all we needed was a way to sense the “input” and convert the molecular recognition event into the “output” through activation of transcription. What came next is what you don’t see published in the paper – dozens of failed attempts. As an RNA-centric group, we focused initially on using STARs and riboswitches as the sensing modality, which likely didn’t work due to kinetic incompatibility with the fast T7 RNA polymerase. We switched over to using E. coli RNA polymerase holoenzyme; however, the low signal-to-noise ratio ruled it out for use in transcriptional biosensing.

It wasn’t until a year after starting the project that we demonstrated that an allosteric transcription factor protein spiked into a T7 transcription reaction—with an appropriately designed template—could tightly regulate the production of the RNA output. I’d love to take full credit for the idea as a stroke of genius, but the reality is that innovation doesn’t occur in a vacuum. There was an enormous body of literature that we drew inspiration from and an environment in the Lucks Lab that made it conducive to the free and open exchange of ideas. If there was such a moment, it would have likely happened en route to grab a burrito with my colleague Eric Strobel, who helped developed cotranscriptional SHAPE-Seq in the lab along with Kyle Watters. Cotranscriptional SHAPE-Seq relies on doping a catalytically dead restriction enzyme (EcoRI E111Q) into a transcription reaction to create a transcriptional “roadblock.” As Eric further developed roadblocking strategies for RNA structure mapping and discussed them with me over our frequent lunch burritos, these ideas for stalling transcription almost certainly permeated into the development of the ROSALIND platform and the “regulated in vitro transcription” concept it’s built upon.

Collection site of the first real-world surface water samples tested in the ROSALIND platform. Lake Michigan, Evanston, Illinois. 

After proving out the concept with the TetR model transcription factor we immediately began to ask several questions centered around field-deployment and application. We turned to the Jim Collins Lab to help us brainstorm applications and develop a freeze-drying process and formulation that would allow us to break the cold-chain limitations of our technology. To further address whether this technology could be used in the field we looked in our backyard—Lake Michigan—and demonstrated functionality in surface freshwater. We also started to work on a low-cost handheld fluorescence illuminator to visualize the reactions, which started with me cutting up a cardboard box and micropipette tip racks from the lab, soldering a copper wireframe with a handful of resistors and LEDs, and gluing on a piece of stage light film to serve as an optical filter. With the help of Northwestern’s research shop, we were able to iterate on that early design and produce a beautiful 3D-printed device complete with a printed circuit board.

An early prototype of the ROSALIND handheld fluorescence illuminator.

Around this same time, Kirsten Jung joined the lab as a PhD student. During her rotation we tasked her with reproducing the work and upon joining the lab she dived into our pursuit for additional transcription factors to prove out the modularity of the concept. As I began to transition out of my postdoc, she quickly took the lead on the project and powered us through the revision – collecting thousands of data points on sensitivity, specificity, longevity, and leading our team through field-testing in Paradise, California. There is, of course, a lot more to the story than can be covered in a single post (see Kirsten’s behind the paper post: "ROSALIND Part 2 – Field-deployable Biosensors for Detecting Waterborne Contaminants with RNA Engineering"). For more background I’ll point to this pre-submission tweet thread from when we posted our manuscript as a pre-print and our latest tweet threads from myself, from Kirsten, and from Julius.

So what’s next? With the support of the Chain Reaction Innovations program at Argonne National Laboratory we’ve launched Stemloop, Inc. – a company developing cell-free biosensors for public health applications. We’re making it our mission to ensure that anyone, anywhere, can sense the world around them with cell-free biosensor technology.

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in