The heart of the platform is a periodic array of Y-shaped gold nanocavities, each unit cell consisting of a symmetric cavity with three identical arms arranged at 120 degrees. This geometry was chosen deliberately. Compared with conventional nanohole arrays, Y-shaped cavities concentrate electromagnetic energy at their sharp inner vertices, producing intense localized hotspots where the Raman scattering cross-section of nearby molecules is dramatically amplified. The researchers fabricated the arrays on indium tin oxide coated BK7 glass using electron-beam lithography, spinning a thin ZEP520A resist film, exposing the Y-shaped pattern in a hexagonal lattice with periodicities of 350 and 347.3 nanometers, and then depositing a 50-nanometer-thick gold layer by electron-beam evaporation. Scanning electron microscopy confirmed that the finished structures faithfully reproduced the designed geometry with uniform orientation and periodicity across a 150 by 150 micrometer sensing area.

Optical characterization revealed that the metasurface supports plasmonic resonances near 670 nanometers, arising from coupling between localized surface plasmon modes and propagating plasmon modes within the periodic array. Numerical simulations based on Rigorous Coupled-Wave Analysis predicted two resonances, at approximately 670 and 742 nanometers, in good agreement with the experimental transmission spectra. The small discrepancies are attributable to nanofabrication tolerances, particularly slight rounding of the Y-shaped cavity vertices during resist development, which modifies the effective plasmonic mode volume, and to minor differences in the refractive index of the polymeric adhesion layer. Importantly, although the dominant extinction maximum sits near 670 nanometers, the optical response is broad and multimodal, extending into the near-infrared where it overlaps substantially with the 785-nanometer excitation laser used for SERS. The authors emphasize that maximum far-field extinction does not necessarily coincide with maximum local field enhancement, so efficient Raman amplification can still be achieved through these broader near-infrared modes. The 785-nanometer choice also suppresses fluorescence background, a common nuisance in biological samples, improving signal-to-noise during viral measurements.

To quantify the enhancement mechanism, the team ran finite-difference time-domain simulations of the local electric field at 785 nanometers and estimated the electromagnetic SERS enhancement factor using the standard fourth-power approximation, which accounts for simultaneous amplification of both the excitation field and the Raman-scattered emission. The calculated maximum enhancement was approximately 6.4 × 10³, concentrated at the sharp inner vertices of the Y-shaped cavities. Two physical mechanisms drive this amplification: the lightning rod effect, in which high curvature at the arm tips accumulates charge and boosts the local field, and inter-cavity plasmonic coupling, in which neighboring cavities in the periodic lattice further intensify the field distribution. While higher enhancement factors have been reported elsewhere, including values around 1.3 × 10⁶ for pyramidal nanohole metasurfaces previously developed by the same group, the modest figure here conceals a crucial design advantage: the hotspots are distributed uniformly across the entire metasurface. That homogeneity dramatically reduces spectral variability from point to point, which the authors identify as the key to the tight spectral clustering observed in their statistical analysis.

For the sensing experiments, the team used non-infectious, inactivated samples of human norovirus (HNoV) and murine norovirus (MNV-1), both members of the Caliciviridae family of single-stranded RNA viruses, supplied by the Istituto Zooprofilattico Sperimentale del Mezzogiorno. No viral propagation or handling of infectious material took place during the study. Droplets of virus solution at a concentration of 10⁵ tissue culture infectious dose 50 percent per milliliter were deposited directly onto the plasmonic sensors, incubated for one hour in a humid chamber, rinsed seven times with bidistilled water, and dried under nitrogen before spectral acquisition. SERS measurements employed a custom-built automated Raman platform with a 785-nanometer excitation laser, a motorized precision stage for spatial mapping, and a high-sensitivity spectrometer covering 200 to 2,500 inverse centimeters, with 20 milliwatts of laser power and 10-second integration times per spectrum.

The experimental design was deliberately rigorous. Four independently fabricated metasurfaces were used, two dedicated to human norovirus and two to murine norovirus, with 25 spectra acquired from different spatial locations on each sensor, yielding 100 spectra in total. Measurements were performed in separate sessions by different operators, allowing the team to probe substrate-to-substrate, spatial, temporal, and operator-dependent reproducibility all at once. The resulting spectra showed high consistency and well-defined vibrational features. Both viruses produced fingerprints containing bands associated with disulfide bonds, aromatic amino acids such as phenylalanine, tryptophan, and tyrosine, protein backbone amide modes, and nucleic-acid-related structures, consistent with their close taxonomic relationship. The authors caution that because multiple biomolecular components can contribute to the same spectral region, individual band assignments should be regarded as putative rather than definitive molecular identifications.

Despite the broad similarity, distinctive differences emerged. The human norovirus spectra displayed features that were absent or weaker in the murine dataset, such as a disulfide stretching band near 524 inverse centimeters, an adenine band near 730, and a valine contribution near 970, while the murine spectra showed their own characteristic peaks, including phenylalanine at 605, adenine at 746, histidine at 779, tyrosine at 841, and an amide III band near 1,312 associated with alpha-helical structure. These differences likely reflect subtle variations in protein composition, capsid organization, amino-acid exposure, and nucleic-acid vibrational contributions. The researchers are careful not to over-interpret: given the overlapping nature of Raman signatures in complex biological systems, no single band can be treated as a unique biomarker. What matters is that the reproducible differences carry enough information for multivariate statistics to exploit.

That statistics came in the form of principal component analysis, applied in the R environment after baseline correction and normalization of all spectra. The two-dimensional score plot built from the first two principal components revealed two cleanly separated clusters, one for each virus, with tight grouping within each cluster indicating high spectral reproducibility. Quantitatively, the first principal component accounted for 87.1 percent of the total spectral variance and the second for an additional 11.6 percent, meaning that more than 98 percent of the captured variance was explained by just two dimensions. Analysis of the PCA loading vectors showed that the discrimination arises from distributed contributions across multiple Raman regions rather than from any single dominant feature. Crucially, the clustering pattern persisted across independently fabricated sensors, different spatial positions, separate sessions, and different operators, indicating that the separation is driven by intrinsic spectral differences between the viral capsids rather than by measurement artifacts or substrate variability.

The authors position the work as a proof of concept rather than a fully optimized diagnostic assay, noting that the study was not designed to determine an analytical limit of detection. Even so, the tested concentration of 10⁵ TCID₅₀ per milliliter falls within the range commonly associated with active norovirus infections, making it a clinically relevant benchmark. What sets the platform apart from most previously reported SERS-based viral sensing strategies is its complete freedom from antibodies, aptamers, Raman reporter tags, molecular amplification, or any target-specific functionalization. Earlier approaches, such as a dual SERS nanotag immunoassay for norovirus built on molybdenum trioxide nanocubes, depended heavily on the specificity of antigen-antibody sandwich complexes, with all the reagent costs and preparation complexity that entails. The Y-shaped metasurface instead reads the virus’s own biochemical identity directly.

The research group, which has previously applied similar PCA-assisted plasmonic platforms to discriminate Shiga toxin variants and to detect rotavirus in wastewater with supervised learning, sees a clear path forward. Future studies will evaluate sensor performance in clinically and environmentally relevant matrices, assess interference effects, expand the spectral datasets, and implement supervised machine-learning classifiers such as linear discriminant analysis, support vector machines, and random forests, which the team has already shown can boost classification performance in related SERS applications. If those steps succeed, a chip of gold Y-shaped cavities, a near-infrared laser, and a statistical model could one day offer rapid, label-free identification of closely related viral pathogens in clinics, food safety laboratories, and environmental monitoring networks, no molecular labels required.

Subject of Research: Label-free discrimination of human and murine noroviruses using plasmonic metasurface-enhanced Raman spectroscopy

Article Title: Plasmonic metasurface-enhanced Raman spectroscopy for label-free discrimination of human and murine noroviruses

Article References: Guilcapi, B., Milano, A., D’Avino, A., Sagnelli, D., Rippa, M., Marchesano, V., Fusco, G., & Petti, L. (2026). Plasmonic metasurface-enhanced Raman spectroscopy for label-free discrimination of human and murine noroviruses. Frontiers in Photonics, 7, Article 1864447. https://doi.org/10.3389/fphot.2026.1864447

Image Credits: AI Generated

DOI: 10.3389/fphot.2026.1864447

Keywords: norovirus, SERS, plasmonic metasurface, label-free biosensing, Raman spectroscopy, principal component analysis, nanocavities, viral diagnostics, nanophotonics, Caliciviridae, hotspots, biosensor

The study, led by Mingqian Zhang, Ziyi Li, Lei Liu and Hai-Chen Wu of the Institute of Chemistry at the Chinese Academy of Sciences together with colleagues, reconstitutes the complete catecholamine biosynthetic pathway in a test tube using purified enzymes and their required cofactors. The researchers then coupled this miniature chemical factory to an engineered protein nanopore — a nanoscale pore embedded in a membrane through which an ionic current flows. Whenever a metabolite from the pathway interacts with the pore, it briefly perturbs that current, producing an electrical fingerprint specific to that molecule. By reading these fingerprints as they accumulate, the team could follow the transformation of each intermediate into the next, second by second.

The technical heart of the work lies in how the researchers achieved specificity. Small molecules are notoriously difficult for nanopores to distinguish because they zip through the pore too quickly and too subtly to leave clean, separable signals. The team solved this by integrating two orthogonal molecular recognition modalities within a single nanopore platform. In essence, different chemical strategies were used to make different members of the pathway register at the pore: some metabolites, such as the catecholamines themselves, generate characteristic blockade patterns as they translocate, while the amino acids phenylalanine and tyrosine were detected through a tailored probe chemistry involving a copper-coordinated molecular adaptor and a cucurbituril host that captures the amino acid and presents it to the pore as a distinct complex.

With this dual recognition scheme in place, the researchers demonstrated highly specific detection of all six key players in the pathway: phenylalanine, tyrosine, levodopa, dopamine, noradrenaline and adrenaline. Extended data traces show characteristic current blockades and dwell-time distributions for each analyte at 100 micromolar concentration, recorded in a high-salt buffer at a transmembrane potential of +100 millivolts. Gaussian fits to the blockade histograms and single-exponential fits to the dwell times confirm that each metabolite produces a statistically distinct signature, which is the prerequisite for unambiguous identification in a mixed sample.

Having validated each molecule individually, the team turned to the full cascade. They initiated the enzymatic sequence and sampled the reaction mixture every five minutes, feeding each sample to the nanopore. The resulting single-channel recordings amount to a time-lapse film of neurotransmitter synthesis: phenylalanine signals fade as tyrosine rises, tyrosine gives way to levodopa, and levodopa is converted onward to dopamine, noradrenaline and finally adrenaline. Because the readout is label-free and single-molecule, nothing was added to the reaction to make it visible, and no amplification step could distort the quantitative picture of pathway dynamics.

This time-resolved view matters because the catecholamine pathway is not a simple conveyor belt. It is a branched network whose flux is governed by enzyme kinetics, cofactor availability and regulatory feedback, and disruption at any node is implicated in conditions ranging from Parkinson’s disease and attention-deficit hyperactivity disorder to orthostatic hypotension and stress-related endocrine disorders. Conventional analytical methods — liquid chromatography, mass spectrometry, electrochemistry, ELISAs — can measure metabolite concentrations at chosen endpoints, but they are generally snapshot techniques that require aliquots, labels or lengthy separation steps, making continuous, coupled monitoring of a multi-enzyme cascade cumbersome.

The platform also opened a window on pathology. A striking application involves iodotyrosine dehalogenase 1, or DEHAL1, an enzyme best known for recycling iodide in the thyroid, where it removes iodine from iodinated tyrosine residues. Defects in the DEHAL1 gene cause hypothyroidism, and 3-iodo-L-tyrosine, the substrate of this enzyme, has long been known to inhibit tyrosine hydroxylase, the rate-limiting enzyme that converts tyrosine to levodopa. The researchers used their nanopore assay to examine what happens to the catecholamine cascade when dehalogenase activity is compromised. They found that 6-methylisothiocyanate, a compound that disrupts the dehalogenase, perturbs catecholamine biosynthesis in a way that offers a mechanistic explanation for how DEHAL1 deficiency may compromise neurotransmitter production — potentially linking an inherited thyroid disorder to impaired catecholamine output.

Figure 5 of the study, titled ‘Effect of thyroiditis-like disorders on catecholamine metabolism,’ presents this perturbation experiment, with extended recordings showing how the cascade changes when monoiodotyrosine is present. The single-molecule records reveal shifts in the relative abundance of the pathway intermediates, illustrating the value of the method: rather than inferring a block from diminished end-product levels, researchers can see exactly where flux stalls and which intermediates accumulate. That level of mechanistic detail is precisely what drug developers and enzymologists need when trying to dissect regulatory nodes in metabolic networks.

Broader implications extend well beyond catecholamines. The authors describe their integrated enzymology-and-nanopore-sensing platform as a general framework for studying complex biochemical pathways with single-molecule resolution. Because protein nanopores can be engineered with different recognition elements, the same design philosophy could in principle be applied to other branched metabolic cascades, to drug-metabolism studies, or to screening enzyme inhibitors in real time. The approach builds on two decades of stochastic sensing, in which engineered pores have been taught to identify analytes as diverse as metal ions, nucleotides, amino acids, peptides, proteins and microRNAs, and it extends that lineage from static identification to dynamic pathway monitoring.

The team has deposited its source data publicly on Zenodo, and the paper — received in December 2025, accepted in August 2026 and published on 1 September 2026 — carries no reported competing interests. What the work ultimately delivers is a new kind of laboratory instrument: a single pore through which the chemistry of a neurotransmitter can be watched being born. If the framework generalizes as its authors hope, metabolic biochemists may soon spend far less time taking snapshots of pathways and far more time watching them move.

Subject of Research: Time-resolved single-molecule monitoring of catecholamine biosynthesis from phenylalanine using an engineered nanopore sensor

Article Title: Nanopore-enabled time-resolved monitoring of catecholamine-related phenylalanine metabolism

Article References: Zhang, M., Li, Z., Hao, W., Yi, Y., Zhou, K., Liu, L., & Wu, H.-C. (2026). Nanopore-enabled time-resolved monitoring of catecholamine-related phenylalanine metabolism. Nature Nanotechnology. https://doi.org/10.1038/s41565-026-02273-3

Image Credits: AI Generated

DOI: 10.1038/s41565-026-02273-3

Keywords: nanopore sensing, catecholamines, dopamine, adrenaline, phenylalanine metabolism, single-molecule detection, neurotransmitter biosynthesis, DEHAL1, iodotyrosine dehalogenase, label-free biosensing, Nature Nanotechnology, enzymology