The State of Extrinsic Protein Labeling

Proteins govern life. They carry oxygen, catalyse reactions, transmit signals, and form the structural fabric of every cell. Understanding how individual proteins behave, how they fold, interact, and function is fundamental to medicine, diagnostics, and drug discovery. Yet most tools available today require attaching a synthetic fluorescent dye to a protein before it can be detected. This labelling step, while effective, is not neutral: it can alter the protein's shape, disrupt its binding interactions, and distort the very behaviour it was designed to reveal.

Intrinsic UV Autofluorescence Long-Standing Challenges

Single-molecule fluorescence techniques like FCS, FRET, FLIM, and STED have transformed our understanding of molecular biology. However, they almost universally depend on extrinsic fluorescent labels. The field has long sought an alternative rooted in a simple observation: proteins are not dark. Aromatic amino acids like tryptophan, tyrosine, and phenylalanine emit an autofluorescence under ultraviolet (UV) illumination. More than 90% of all proteins contain at least one such residue.

Exploiting this intrinsic UV autofluorescence for single-molecule detection has been a long-standing challenge. Proteins are orders of magnitude dimmer than conventional dyes, and UV microscopy introduces specific difficulties: rapid photobeaching, high background from buffer and optics, and the absence of nanophotonic platforms optimised for the UV spectral range. Early efforts were limited to large proteins carrying dozens of tryptophan residues, a restriction incompatible with most biologically relevant targets.

Progress in Nanophotonics and Plasmonic Nanopores

Recent years have seen decisive progress. The Wenger group at Institut Fresnel demonstrated that UV-optimised plasmonic nanoantennas, including aluminium nanoapertures, optical horn antennas, and rhodium nanocube dimers, can enhance UV autofluorescence sufficiently to resolve single proteins containing as few as one tryptophan residue. Zero-mode waveguides (ZMWs), nanoscale apertures that confine light below the diffraction limit, have also been shown to boost brightness and enable operation at physiologically relevant micromolar concentrations. In parallel, plasmonic nanopores have emerged as powerful single-molecule sensors capable of monitoring the translocation of individual biomolecules with optical readout, moving beyond the limitations of conventional ionic current detection.

Project Research Intent

Despite these advances, combining UV autofluorescence with nanophotonic enhancement and nanopore-based translocation in a unified, label-free platform remains an open frontier and is precisely what this project addresses.

My project under DYNAMO is centred on UV autofluorescence correlation spectroscopy (UV-FCS), a technique that measures the natural UV emission of individual proteins as they diffuse through a tiny illuminated volume, revealing information about their concentration, size, and dynamics. The intrinsic signal is weak, so two complementary strategies are deployed: reducing noise and amplifying the signal.

First is suppressing the background with FLCS. Fluorescence Lifetime Correlation Spectroscopy (FLCS) adds a time dimension to the measurement and can mathematically separate genuine signal from scattered laser light and other parasitic contributions. This lifetime-gating approach substantially improves the signal-to-background ratio without any modification to the protein or sample.

The second strategy is amplifying the signal with nanophotonics. Zero-mode waveguides (ZMWs) are nanoscale apertures fabricated in metal films confining light to volumes below the diffraction limit. When a protein enters this ultrasmall hotspot, its fluorescence is dramatically enhanced. This makes it possible to detect individual proteins that would otherwise be invisible, and to work at the physiologically relevant concentrations found in living systems.

Further, we plan to extend the detection platform to plasmonic nanopores: nanoscale channels that force proteins to pass through one at a time, driven by an electric field, and give electrical readouts. But our focus is on the optical readout strategy. Each translocation event produces a distinct optical burst, carrying information not only about the protein's presence but also about its dynamics and identity. Combined with the UV autofluorescence readout, this approach opens the possibility of characterising proteins with single-molecule resolution, label-free, in real time.

Together, these tools form a coherent label-free platform for single-protein analysis. The integration of UV autofluorescence with nanopore translocation could enable a new class of biosensor, one capable of identifying proteins by their intrinsic optical fingerprint as they thread through a nanoaperture. This would have direct implications for medical diagnostics, proteomics, and the study of protein misfolding diseases such as Alzheimer's and Parkinson's.

Ultraviolet Nanophotonics Core Objectives

The central ambition of this project is to detect individual protein molecules without attaching any fluorescent label to them and using only the faint ultraviolet light that proteins naturally emit[cite: 5]. Proteins glow under UV illumination because of their aromatic amino acids, but this intrinsic signal is orders of magnitude weaker than conventional synthetic dyes, and UV microscopy comes with its own set of challenges: rapid photobeaching, high background from buffers and optics, and a near-total absence of nanophotonic tools designed for the deep-UV spectral range[cite: 5]. The project tackles these challenges step by step, building from visible-light calibration experiments all the way to plasmonic nanoantennas and solid-state nanopores[cite: 5].

Visible-Light Methodological Foundations

Before entering the UV, it was essential to validate the experimental toolkit on a well-controlled visible-light system[cite: 5]. Low-emitting nanoprobes like gold nanoclusters and lanthanide nanoparticles were chosen as model systems because their weak, complex emission closely mimics the signal-to-background conditions expected in the UV[cite: 5]. Using Fluorescence Correlation Spectroscopy (FCS), the photophysics of these emitters were characterised at the single-particle level: their blinking behaviour, per-emitter brightness, and response to excitation power were systematically studied[cite: 5]. The effect of surface ligands on gold nanocluster brightness was also explored, along with the use of zero-mode waveguides (ZMWs) for brightness enhancement[cite: 5]. These results form the methodological foundation for all subsequent UV work[cite: 5].

Background Suppression and Signal Optimization

With the FCS toolkit established, the focus turned to the most pressing obstacle for single-molecule detection: background noise[cite: 5]. Fluorescence Lifetime Correlation Spectroscopy (FLCS) extends standard FCS by adding a time dimension[cite: 5]. Because different emitters and, crucially, genuine signal versus scattered laser light decay at different rates after excitation, FLCS can mathematically isolate the contribution of each species[cite: 5]. The result is a dramatic improvement in signal-to-background ratio without touching the sample in any way[cite: 5].

A first major result was the extension of FLCS to a ternary mixture for the first time[cite: 5]. Three emitters with nearly identical emission spectra were resolved simultaneously[cite: 5]. The system combined gold nanoclusters with two organic dyes, Alexa Fluor 647 and CF640R, pushing multiplexing well beyond what conventional spectral separation can achieve[cite: 5].

A second key contribution was a rigorous investigation of the FLCS detection limit, a question that, surprisingly, had never been systematically addressed[cite: 5]. An in-depth study identified the microscope parameters and figures of merit that govern sensitivity, and led to a modified confocal configuration delivering a tenfold improvement in the limit of detection[cite: 5]. This advance was immediately applied to measure the association and dissociation dynamics of the biotin–streptavidin interaction, one of the strongest non-covalent bonds in biology and a widely used benchmark in biotechnology[cite: 5].

Deep-UV Protein Autofluorescence Optimization

Armed with optimised detection strategies, the project moved into deep-UV spectroscopy, the domain where proteins emit their intrinsic fluorescence[cite: 5]. Two intertwined problems had to be solved simultaneously: photobeaching (proteins are fragile under UV illumination) and background (the buffer, the glass optics, and the laser scatter all contribute unwanted signal)[cite: 5]. We could not apply FLCS here due to the similar fluorescence lifetime of the protein and the background[cite: 5].

A systematic study of buffer composition revealed that pH stabilisation using phosphate-based buffers, combined with the antioxidant glutathione, significantly extended the photostability of protein autofluorescence[cite: 5]. Active sample stirring was found to further reduce photobeaching by continuously refreshing the detection volume with unbleached protein[cite: 5]. These seemingly simple steps turned out to be decisive: together with an optimised confocal configuration, they yielded a tenfold improvement in the detection limit for label-free streptavidin, surpassing the current state of the art and enabling the observation of streptavidin–biotin binding cooperativity at the single-molecule level[cite: 5].

DNA-Origami-Assembled UV Nanoantennas

Optimised buffer conditions improve the signal, but to reach reliable single-protein sensitivity, the brightness of individual molecules must be actively boosted[cite: 5]. Plasmonic nanoantennas that concentrate light into nanoscale volumes far below the diffraction limit offer exactly this capability[cite: 5]. The challenge in the UV is that most plasmonic metals (gold, silver) do not support resonances in this spectral range[cite: 5]. This project pioneered the use of rhodium (Rh) as a UV-compatible plasmonic material, assembled into nanoantennas using DNA origami scaffolds[cite: 5]. Single-streptavidin autofluorescence measurements in the rhodium nanoantenna hotspot showed up to 22× brightness enhancement (average 6.6×), shorter autofluorescence lifetimes, and a more than tenfold increase in total photon budget[cite: 5].

Nanopore Integration and Photo Clogging Challenges

The final frontier is to combine all of the above—UV autofluorescence, plasmonic enhancement, and nanophotonic confinement with solid-state nanopores[cite: 5]. In a nanopore experiment, individual molecules are driven through a nanoscale aperture by an electric field; each passage generates a detectable signal[cite: 5]. The goal here is to replace the conventional ionic current readout with an optical one, using fluorescence bursts as proteins transit through a ZMW-integrated nanopore[cite: 5].

Preliminary experiments in the visible range used a 635 nm laser to distinguish free Atto643 dye from Atto643-labelled streptavidin passing through the nanopore, with single-molecule bursts successfully detected[cite: 5]. However, protein samples showed far fewer translocation events than expected, pointing to non-specific adsorption of proteins to the nanopore surface, a known challenge that surface passivation strategies are expected to resolve[cite: 5].

In the UV, polystyrene beads were used as test particles with a 266 nm laser[cite: 5]. Here, a different obstacle emerged: UV illumination caused the nanopore to become clogged[cite: 5]. While the blockage could be temporarily cleared by incubating in tetrahydrofuran (THF), it recurred immediately upon renewed UV exposure[cite: 5]. Understanding and eliminating this photo clogging phenomenon is the key open challenge for the next phase of the project[cite: 5].

Selected Output

• M. Kayyil Veedu, J. Osmólska, A. Hajda, J. Olesiak-Bańska, and J. Wenger, “Unveiling the photoluminescence dynamics of gold nanoclusters with fluorescence correlation spectroscopy,” Nanoscale Adv., vol. 6, no. 2, pp. 570–577, 2024[cite: 5].
• M. Kayyil Veedu, G. Lavilley, M. Sy, J. Goetz, L. J. Charbonnière, and J. Wenger, “Watching lanthanide nanoparticles one at a time: characterization of their photoluminescence dynamics at the single nanoparticle level,” Nanoscale, vol. 17, no. 6, pp. 3270–3276, 2025[cite: 5].
• M. Kayyil Veedu, A. Hajda, J. Olesiak-Bańska, and J. Wenger, “Three species multiplexing of fluorescent dyes and gold nanoclusters recovered with fluorescence lifetime correlation spectroscopy,” Biochim. Biophys. Acta BBA - Gen. Subj., vol. 1868, no. 6, p. 130611, Jun. 2024[cite: 5].
• M. Kayyil Veedu and J. Wenger, “Breaking the Low Concentration Barrier of Single‐Molecule Fluorescence Quantification to the Sub‐Picomolar Range,” Small Methods, p. 2401695, Feb. 2025[cite: 5].
• N. Corduri et al., “DNA‐Origami‐Assembled Rhodium Nanoantennas for Deep‐UV Label‐Free Single‐Protein Detection,” Adv. Funct. Mater., p. e32006, Mar. 2026[cite: 5].