A new sensing architecture to detect small molecules

You have probably performed a sandwich assay before, even if you didn’t know it. Sandwich assays are the molecular sensing technology behind lateral flow tests like the at-home COVID test. After a sample is dripped onto the test strip, it combines with detector antibodies attached to tiny gold particles, which bind to any viral proteins present. When the sample reaches the test line, a second antibody anchored to the strip binds to a different site on the viral protein, creating a molecular ‘sandwich’. As the gold particles accumulate, the test line turns dark red, becoming visible to the naked eye.

Sandwich assays are also used for home pregnancy tests, which target the hormone human chorionic gonadotropin (hCG). Like COVID antigens, hCG is a large molecule, with plenty of space for two antibodies to bind to its surface. But smaller molecules like testosterone or the neurotransmitter serotonin don’t have room for two different binding sites, and require much more sophisticated detection technology than the sandwich assay’s visible readout.

To get around this problem, researchers in the Laboratory of Chemical Nanotechnology (CHEMINA) in EPFL’s School of Engineering have used short, synthetic strands of DNA called aptamers. Instead of requiring two different binding sites, only one ‘primary’ aptamer needs to bind to a small molecule. A ‘secondary’ aptamer is then engineered to recognize the shape of the primary aptamer in its bound state, rather than the target molecule itself. The work has been published in Nature Chemistry.

“Aptamer-based technologies have gained momentum in recent years, with a growing number of startups relying on them for human health applications. Instead of simply refining sensors, we wanted to rethink the molecular recognition architecture itself to address a long-standing gap in small-molecule sensing,” says CHEMINA lab head Nako Nakatsuka.

A distinct mode of DNA recognition

Until now, most aptamer-based sensors have relied on a single aptamer to do all the work of recognizing a target molecule and signaling its presence. The signal readout can be especially complex, often requiring specialized equipment to detect tiny fluorescent or electrical changes that flag binding events.

In the new approach, two aptamers share the work. The primary aptamer captures the target, and the secondary aptamer reports that capture by binding to the shape the primary aptamer adopts. Because DNA strands normally identify one another by nucleotide sequence, rather than 3D structure, the work introduces a distinct mode of DNA recognition that adds a second layer on top of target binding.

“A sandwich test for small molecules is just one application. More broadly, this two-layer aptamer design lets us combine separate recognition steps rather than asking a single aptamer to do everything,” Nakatsuka says.

A rapid test for serotonin

For their study, the team used existing aptamers designed to detect four small molecules: serotonin, dopamine, testosterone, and the steroid deoxycorticosterone 21-glucoside (DOG). First, they allowed the four primary aptamers to bind to their targets. Then, they screened billions of DNA sequences to identify those that would bind, not to the target molecule, but to the shape of the primary aptamer in its bound state. They identified secondary aptamers for serotonin, testosterone, and DOG, and performed additional tests to confirm that these sequences recognized and bound specifically to primary aptamers that had already captured their targets. As a proof-of-concept, they developed a lateral flow test that successfully detected serotonin, with a result visible to the naked eye.

Going forward, the team is working on improving test sensitivity, as well as the strength of the aptamers’ binding interactions. They are also investigating why a secondary aptamer for dopamine could not be identified. Beyond sensing, two-layer recognition could also be harnessed in molecular computing systems, where binding events act like binary switches to perform simple logic functions.

“By allowing aptamers to recognize a molecule’s state rather than simply its identity, our work expands what DNA-based systems can detect, creating a foundation for more complex layers of molecular recognition,” Nakatsuka says.