Real-time protein sensor uses reversible Fab fragments

Congratulations to Claire Michielsen and her team at Eindhoven University of Technology for their exciting new publication in ACS Sensors, titled “Continuous Protein Sensing Using Fast-Dissociating Antibody Fragments in Competition-Based Biosensing by Particle Motion”. This innovative work explores how engineered antibody fragments can be used to build a protein sensor that works in real-time, continuously tracking protein concentrations, even in complex fluids like milk.

Why does this matter? Protein sensors are becoming key components in health, diagnostics, and industrial biotech. From monitoring biomarker levels in blood to tracking inflammation or controlling bioprocesses, these molecular sensors offer new ways to detect proteins and follow their changes over time. With this study, Michielsen and her team have taken a significant step forward in demonstrating a new method for continuous protein monitoring, a challenge many in the field have been working to solve.

From signal to insight: Real-time protein detection made simple

Modern protein sensors are transforming how we measure and understand complex biological processes — from diagnosis and monitoring of disease, to real-time tracking of inflammation and biomarker levels. But how do these molecular sensors actually work?

At the heart of most protein sensors is molecular recognition: the ability of a sensor protein to bind specifically to a target molecule. This often relies on engineered domains or Fc-tagged antibodies that recognize particular protein structures. Upon binding, many sensors undergo conformational changes or structural transformations, producing a readable signal — whether it’s a change in light (as in a fluorescent protein or bioluminescent sensor), electric current (nanopore sensors), or particle movement, like in this study.

In the work by Michielsen et al., the authors engineered a modular biosensor capable of continuous protein detection without needing chemical regeneration. Using custom antibody fragments, they developed a sensor chip surface by immobilizing ligands via biotinylated tethers — much like what you’d find in a protein A sensor kit. These sensor domains were specifically designed to bind and unbind rapidly, enabling a reversible signal that could be monitored over time.

This kind of design is only possible through advanced protein engineering methods, including directed evolution, combinatorial screening, and in some cases, de novo design. Tools like protein structure prediction and amino acid sequence analysis are also essential for fine-tuning binding affinities and optimizing sensor performance.

Today, there are many types of protein-based biosensors, including:

• Genetically encoded sensors for cellular studies
• Optical devices like FRET-based sensors
• Active-reset sensors for continuous workflows
• Ultrasensitive optical sensors for low-concentration targets

The sensor featured in this article falls into the category of a real-time, reversible biosensor, a design especially suited for continuous measurement in both the lab and future clinical or wearable applications. It highlights how combining engineered proteins, modular components, and precise fluid handling can produce highly specific detection tools that are compact, robust, and ready for real-world use.

Methods used in the study

Before diving into the details, here’s a quick summary of the sensor setup used in this innovative study on continuous protein sensing:

Sensor Setup at a Glance:

• Target analyte: Lactoferrin, a protein involved in inflammation and immune response.
• Binder: Custom-designed antibody fragments (Fabs) with fast dissociation kinetics.
• Sensor design: Competition-based biosensor, where tethered particles bind and unbind in real-time.
• Surface: A low-fouling polymer layer functionalized with analyte analogues.
• Detection: Video microscopy and single-particle tracking to detect binding state changes.
• Fluid handling: Sample delivery was automated using AMF’s LSPone laboratory syringe pump, integrated with a 12-port rotary valve for precise and repeatable sample switching.
• Validation methods:
  • Free Particle Motion (f-BPM) for kinetic screening
  • Surface Plasmon Resonance (SPR) for kon/koff analysis
    Long-term, real-time testing over 40+ hours

How the methods work

To develop a real-time protein sensor with reversible behavior, the research team combined modular and tunable biosensor design, protein engineering, and advanced fluid handling using AMF’s LSPone system. The goal: to create a sensitive detection platform capable of tracking protein concentrations dynamically in real-time, even in complex biological samples like milk and serum.

Fab screening and selection

Thirteen synthetic antibody fragments (Fabs) were selected using phage display from a combinatorial library — a widely used method in engineering sensor proteins. These fragments underwent screening to identify those with optimal binding kinetics, especially high dissociation rates, which are essential for a reversible biosensor that can operate continuously without requiring a reset step. This reflects a broader protein design trend: building biosensors that are modular, adaptive, and responsive to changes in analyte concentration.

Free particle motion (f-BPM)

In this lab-based analysis, Fabs were immobilized on microbeads, which were then suspended over a sensor surface coated with polyclonal antibodies, a typical sensor chip protein A configuration. When the target protein (lactoferrin) was introduced, the particles experienced a shift in motion due to molecular recognition. This shift, driven by conformational changes and binding events, was captured via video microscopy, providing a quantitative signal. The method enabled dose–response characterization and sensitive detection at nanomolar levels.

Surface plasmon resonance (SPR)

To further investigate the sensor proteins, the team used SPR to map the association/dissociation curves of each Fab on a functionalized sensor chip. This real-time analysis in lab environment allowed precise ranking based on kon/koff values. SPR also reflects the kind of protein sensor kit workflow where researchers create a sensor surface and evaluate ligand binding using optical methods.

Sensor assembly and control via LSPone

The final modular biosensor was built using dsDNA tethers to link particles to a polymer-coated chip. Analogue molecules (ssDNA-lactoferrin conjugates) were immobilized as ligands to simulate target binding. This is similar in concept to commercial protein A sensor kits, where the user builds a detection surface from engineered components. AMF’s LSPone syringe pump, paired with a multi-port valve, provided automated control over sample delivery, flushing, and cycling, a crucial feature for stable signal detection over extended periods.

Long-term monitoring

To prove the system’s suitability for continuous measurement, the biosensor was operated over 40 hours with multiple cycles of protein detection in both buffer and milk matrices. The results demonstrated consistent binding and unbinding, confirming that the sensor could detect signals and adapt dynamically to changing protein levels. This represents a major step forward in sensitive, real-time monitoring tools, the kind that could support next-generation diagnosis and monitoring platforms in both research and clinical settings.

This work not only demonstrated a reversible, high-performance protein sensing model but also showed how tightly integrated systems, from sensor domains to fluidics, are unlocking the full potential of biosensing technologies in real-world scenarios.

AMF’s LSPone at the core of the fluidic setup

In this protein sensing study, AMF’s LSPone programmable syringe pump played a key role in enabling precise, automated fluid handling throughout the experiments.

In any continuous monitoring system, especially one dealing with biomolecular recognition, consistent fluid exchange is essential. The researchers needed to cycle between different solutions — from blank buffers, to samples spiked with varying lactoferrin concentrations, to washing steps, all without introducing bubbles, dead volumes, or timing errors that could compromise the signal stability of the sensor.

To achieve this, they integrated the LSPone pump with a 12-port rotary valve, creating a miniaturized fluidic control system that matched the precision of the single-molecule biosensor. This setup enabled:

• Smooth, low-pulsation delivery of fluids
• Predictable and reliable flushing and switching between samples
• Tight control of flow rates, timing, and direction
• Automation with consistent measurement cycles over 40+ hours without manual intervention

Thanks to this integration, the team could demonstrate the performance and reversibility of their engineered protein sensor under real-world conditions, even in challenging media like milk. The reproducibility of the data across multiple cycles (as shown in supplementary figures) was only possible because the fluidic transitions were stable and free of disturbance, a credit to the mechanical precision of the LSPone and its microfluidic compatibility.

By automating fluid control, the system also moved closer to the kind of lab-on-chip, wearable, or implantable biosensor envisioned for future patient monitoring or in vivo diagnostics. Whether used for diabetes, inflammation, or other disease-related biomarkers, platforms like this one benefit from plug-and-play tools like AMF’s LSPone to reduce complexity and improve throughput in the lab.

This study is a great example of how a sensor device, molecular biology, and fluidic engineering can come together to produce powerful, user-friendly, and real-time monitoring technologies for tracking protein levels with high sensitivity and minimal maintenance.

What is the LSPone?

The LSPone is AMF’s high-precision laboratory syringe pump, specifically designed for microfluidic and analytical applications that demand low flow rates, repeatability, and plug-and-play integration. Built with flexibility in mind, the LSPone simplifies fluid handling for researchers working on cutting-edge protein detection, biosensor design, and continuous measurement workflows.

Whether you’re building a custom biosensing platform, designing a modular biochip, or validating new engineered proteins in a controlled setup, the LSPone brings unmatched control over:

• Flow rate: From a few nanoliters to several microliters per second
• Plunger speed: Adjustable in micrometers per second, ensuring high-resolution control
• Fluid switching: When paired with our multi-port rotary valves, it allows rapid and seamless transition between samples, washes, and reagents
• Scripting & automation: Full control via serial commands or integration into your lab automation suite (Python, MATLAB, etc.)

Its compact form factor, low dead volume, and ease of use make it ideal for applications where signal drift, dead volumes, or inconsistent sample delivery would compromise the performance of a sensor or analytical method. In the case of protein sensors like the one described in this study, these features are not just helpful, they are essential.

The LSPone supports applications ranging from glucose monitoring to inflammation tracking, from cell culture media exchange to sample prep in diagnostics and chemical biology workflows. By integrating cleanly into existing systems and reducing manual intervention, it enhances data reliability, increases throughput, and allows researchers to focus on their science, not their plumbing.