Researchers at the University of Göttingen have built a compressed-air-driven microfluidics system that automates the tedious buffer exchanges required for multiplexed super-resolution microscopy, cutting a major source of human error from one of biology’s most powerful imaging techniques. The platform, designed for DNA-PAINT and Exchange-PAINT workflows, lets scientists image multiple molecular targets in sequence without manually swapping fluids between each round. Because manual fluid handling has long been the weakest link in single-molecule localization microscopy, the system addresses a bottleneck that has kept the technique slow, inconsistent, and difficult to scale.
Why Manual Buffer Swaps Hold Back Super-Resolution
DNA-PAINT works by using short DNA strands that transiently bind to their targets, producing the blinking signals that super-resolution algorithms need to pinpoint molecules far below the diffraction limit of light. Exchange-PAINT extends this idea: after imaging one target, the operator washes out the first set of DNA probes and introduces a new set tuned to a different protein or structure. The foundational demonstration of this approach, led by Ralf Jungmann and colleagues, achieved sub-10-nm resolution in vitro, using microfluidic-style sequential buffer exchanges to cycle imager strands across targets and showed that multiple cellular targets could be imaged in sequence on the same sample. That resolution is extraordinary, but every buffer exchange in a multiplexed experiment introduces a chance for contamination, sample drift, or inconsistent washing, all of which degrade the final image.
The practical consequence is that most labs limit Exchange-PAINT experiments to a handful of targets, not because the chemistry fails but because the manual steps accumulate errors. Each wash cycle can take minutes of careful pipetting, and a five-target experiment might require dozens of individual fluid swaps. Roman Tsukanov, a lead developer of the new system, has noted that automation reduces variability and enables long imaging cycles that would be impractical by hand, especially when working with fragile cells that cannot tolerate rough handling. That observation frames the core problem: the optics and chemistry of DNA-PAINT are already excellent, but the plumbing has not kept pace.
Compressed Air Replaces the Pipette
The Göttingen team’s solution uses compressed air to push tiny, precise volumes of imaging buffer through microfluidic channels, eliminating the need for a human operator to pipette between rounds. Their peer-reviewed report in ACS Nano describes a system purpose-built for Exchange-PAINT and DNA-PAINT, with the explicit goal of improving reproducibility in multiplexed single-molecule localization microscopy. The platform automates the full cycle of introducing imager strands, acquiring data, washing, and loading the next probe set. A super-resolution image of cytoskeleton and focal adhesion proteins produced with the system, shared by the team through the University of Göttingen, illustrates the kind of multi-target detail the approach can deliver.
What makes this more than an incremental engineering improvement is the downstream effect on data quality. When every wash step is identical in volume, timing, and flow rate, the experiment-to-experiment variability that plagues manual protocols shrinks. The research team, which includes Samrat Basak, Kim-Chi Vu, and Jörg Enderlein alongside Tsukanov, has emphasized that consistent fluid exchange also protects delicate biological samples from the mechanical stress of repeated manual interventions. For labs trying to map the spatial organization of ten or more proteins in a single cell, that consistency is the difference between publishable data and weeks of wasted effort.
Microfluidics Already Proved Its Value for Live-Cell Imaging
The Göttingen platform builds on a decade of work showing that microfluidic devices can hold living cells steady enough for super-resolution imaging without killing them. One earlier system used hydrodynamic trapping combined with valves to capture, release, and retrieve individual cells while performing PALM-based imaging at a localization precision below 15 nm. A separate device, designed specifically for fission yeast, reversibly immobilized cells and preserved their growth and division over hours while achieving modal localization accuracy better than approximately 25 nm laterally. Both demonstrations established that microfluidic control of the sample environment does not just stabilize the specimen. It actively enables imaging experiments that would be impossible on a conventional glass slide.
The new compressed-air system extends this logic from sample handling to reagent delivery. Where earlier devices solved the problem of keeping a cell still, the Göttingen platform solves the problem of keeping the chemistry consistent across dozens of sequential imaging rounds. Together, these advances suggest that the entire workflow, from trapping a cell to completing a multiplexed super-resolution dataset, can now be automated on a single microfluidic chip. That prospect matters most for high-throughput applications where dozens or hundreds of cells need identical treatment.
Software and Chemistry Push Speed Even Further
Hardware automation is only one route to faster super-resolution data. Deep learning algorithms have separately demonstrated the ability to accelerate localization microscopy by replacing slow, iterative fitting routines with neural networks that process raw camera frames in near-real time. The combination of automated fluidics and faster computational analysis creates a path toward experiments that would have been prohibitively slow just a few years ago. Most current coverage of the Göttingen system focuses on the hardware, but the real acceleration will likely come from pairing automated buffer exchange with software that can keep up with the data rate.
On the chemistry side, exchangeable oligonucleotide labels offer another speed lever. A recent study showed that carefully tuning the binding kinetics of imager strands can boost blinking rates without sacrificing localization precision, allowing more molecules to be sampled per unit time. In particular, work on programmable DNA binders has demonstrated that off-rates can be engineered to match camera frame rates, so that each fluorophore produces many resolvable events during a single imaging round. When such kinetic optimization is combined with automated buffer delivery, the microscope spends less time waiting for molecules to bind and more time collecting useful information.
Toward Fully Integrated, High-Throughput Super-Resolution
Beyond DNA-PAINT, the compressed-air platform points toward a broader integration of microfluidics with advanced imaging modalities. Earlier work on flow-based sensing has already highlighted how precise control of small volumes can enhance detection sensitivity, as shown by microfluidic designs that couple nanoliter channels to optical readouts in single-molecule assays. The Göttingen system effectively repurposes that level of control from analytical chemistry to structural cell biology: instead of measuring a single analyte, it orchestrates a sequence of labeling and washing steps that together build a multidimensional picture of a cell’s architecture. The same principles could eventually support automated staining, fixation, and even drug dosing protocols upstream of imaging, further compressing what is now a multi-day workflow into a largely hands-off process.
There is also room to integrate more sophisticated feedback between the microscope and the fluidics. For example, algorithms could monitor localization density in real time and trigger buffer exchanges as soon as sufficient data have been collected for a given target, rather than relying on fixed imaging times. This kind of closed-loop control becomes more realistic when both imaging and fluid handling are automated and programmable. Studies on dynamic DNA interactions already hint at how sequence design, binding kinetics, and imaging parameters can be co-optimized; embedding those insights into the control software for systems like Göttingen’s would turn what is currently an expert-driven art into a reproducible protocol that less specialized labs can adopt.
For now, the compressed-air microfluidics platform is best viewed as a crucial missing piece in the super-resolution toolkit. DNA-PAINT and Exchange-PAINT have long promised multiplexed imaging with near-molecular resolution, but manual buffer swaps kept that promise out of reach for routine experiments. By standardizing the fluidic side of the workflow, the Göttingen team has made it far more realistic to perform long, multi-target imaging campaigns without sacrificing data quality or sample viability. As microfluidic control, kinetic probe design, and machine learning-based analysis continue to mature together, the field is moving steadily toward a future in which mapping dozens of molecular species in three dimensions is not a heroic, one-off effort, but a standard, automated procedure.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
