Lectin-coated gold nanoparticles tag bacterial sugars for Raman detection

Researchers have built a simple gold nanoparticle probe that recognizes two common sugars on bacterial surfaces and boosts their Raman signals enough to generate distinct spectral fingerprints for three different bacterial species. The system, described in an open-access paper published on 16 July 2025 in the journal Gold Bulletin, pairs citrate-stabilized gold nanoparticles with Concanavalin A, a plant lectin that binds D-glucoside and D-mannoside, terminal sugars frequently found on microbial glycans. The team, led by Ravichandran Manisekaran and Luz M. López-Marín, proposes the construct as a low-cost, label-free route toward presumptive microbial detection, an area where current glycan identification typically relies on expensive techniques such as mass spectrometry or antibody-based labeling.

The design exploits a well-known property of nanoscale gold: when molecules sit within a few nanometers of a corrugated metal surface, their Raman scattering can be exponentially enhanced, an effect known as surface-enhanced Raman spectroscopy, or SERS. Gold was chosen over silver or copper because it is far more inert in biological media, resisting the redox chemistry that would degrade other metals. The lectin serves as the targeting arm, while the nanoparticle serves as the optical amplifier. Because ConA recognizes a shared structural motif of mannose and glucose rather than a single unique glycan, the authors argue that one functionalized particle can in principle tag any glycan displaying those sugars, whether attached to proteins, lipids, or other molecules on a microbial surface.

Building the probe was deliberately straightforward. The team synthesized spherical gold nanoparticles by the Turkevich–Frens method, reducing a gold precursor with trisodium citrate under boiling conditions until the solution turned wine red. Ultraviolet–visible spectroscopy showed a well-defined plasmon band at 518 nanometers, consistent with spherical, aggregation-free particles of roughly 15 to 20 nanometers. X-ray diffraction confirmed a face-centered cubic gold crystal structure, and applying the Scherrer equation to the most intense reflection gave an average crystallite size of 18 nanometers with a standard deviation of 1.2 nanometers. Transmission electron microscopy of 270 particles yielded an average diameter of 18.6 ± 2.3 nanometers, slightly larger than the crystallographic estimate because individual particles are composed of multiple crystal grains. Dynamic light scattering measured a hydrodynamic diameter of 25.47 ± 1.2 nanometers, and the particle concentration was calculated at 3.8 × 10^13 particles per milliliter using the Haiss equation.

Concanavalin A was then attached through physisorption rather than covalent chemistry. The lectin was incubated with the nanoparticles, and because gold surfaces spontaneously adsorb proteins through noncovalent forces, a protein corona formed around each particle. The researchers distinguished between a soft corona of loosely bound protein and a hard corona of tightly bound protein, and removed the soft fraction by repeated centrifugation and washing. This step mattered for the optical function of the probe: SERS enhancement requires the analyte to come within about 10 nanometers of the gold surface, so a thick, loosely attached protein layer would push the target sugars out of range. After washing, dynamic light scattering showed a major particle population with a hydrodynamic diameter of 32 ± 0.8 nanometers, an increase of roughly 7 to 8 nanometers over the citrate-coated particles, consistent with replacing a thin citrate shell with a single lectin layer about 4 nanometers thick.

Several analyses confirmed that the lectin had indeed formed a stable, functional coating. Fluorescence spectroscopy took advantage of the four tryptophan residues in ConA, whose emission is quenched by gold when the protein lies within 10 nanometers of the surface. Free ConA showed dose-dependent fluorescence with a maximum at 335 nanometers; particles still carrying a soft corona showed weaker but still dose-dependent signals, while particles with only a hard corona showed clear quenching and a blue shift of the emission maximum to below 314 nanometers, a pattern typical of static quenching. Negative staining and transmission electron microscopy visualized the transition directly: unwashed particles were surrounded by abundant electron-semitransparent material, whereas washed particles carried a tight, thin protein layer. Zeta potential measurements traced the same story, moving from −40 millivolts for citrate-capped particles to −16.8 millivolts with a soft corona and −28.3 millivolts for the more stable hard-corona particles. Based on the steric geometry of the lectin, the authors estimated about 61 protein molecules adsorbed per particle, and found that a ConA-to-gold ratio of 35:1 gave reproducible functionalization with little residual free protein. The hard corona contained 40.5 micrograms of ConA per milliliter of colloid, and SDS-PAGE confirmed the released protein at the expected monomeric size of roughly 23 kilodaltons. As a control, the team prepared parallel assemblies coated with bovine serum albumin, a protein chosen for its inertness.

The critical test was whether the lectin retained its sugar-binding activity after adsorption onto gold, since protein deposition can sometimes denature adsorbed proteins. The researchers used methyl-α-D-mannopyranoside, a simple ConA ligand, at concentrations of 5, 10, 20, and 40 millimolars. After incubation with the functionalized particles and washing away unbound sugar, SERS measurements with a 785-nanometer excitation laser showed enhanced Raman bands whose intensities rose with sugar concentration. Bands at 878 and 1442 reciprocal centimeters were attributed to bending vibrations of C–OH and CH2 groups respectively, with additional enhanced signals between 755 and 1150 reciprocal centimeters assigned to α-pyranose ring vibrations. Crucially, no enhanced signals appeared when the same sugar solutions were analyzed with the albumin-coated control particles, demonstrating that the signal depended on the specific lectin–sugar interaction rather than on nonspecific adsorption.

With the model ligand established, the team turned to whole bacteria. They grew three species to logarithmic phase: the Gram-positive Bacillus subtilis, the Gram-negative Escherichia coli, and the acid-fast Mycobacterium fortuitum, organisms from distinct taxa that are known to cause opportunistic infections. After washing the cells and incubating them with the nanotag, the researchers acquired SERS spectra that showed strong signals in all three samples, indicating that the probe enhanced Raman signals associated with mannose- or glucose-containing surface structures. Although the complexity of whole bacterial cells made precise band assignment difficult, the three species produced clearly differential spectra. A band at 735 reciprocal centimeters in the B. subtilis spectrum was assigned to vibrational modes of carbohydrates in Gram-positive bacteria, while a band near 1002 reciprocal centimeters seen in the E. coli and M. fortuitum spectra was tentatively attributed to phenylalanine in proximity to the tagged sugars.

To strengthen the assignment of sugar-related signals, the team isolated a mycoloyl trehalose-rich fraction from M. fortuitum, a lipidated glycan characteristic of mycobacteria, and analyzed it separately. Strong bands at 1298 and 1466 reciprocal centimeters found in the whole-bacterium spectrum also appeared in the enriched glycolipid fraction, supporting the conclusion that the lectin-based nanotag enhances signals from molecular fragments attached to its targeted glycans. One band at 1120 reciprocal centimeters, previously assigned to mycolic moieties, appeared in the whole-bacterium spectrum but not in the enriched fraction, an observation the authors attribute to different conformational arrangements of acyl trehaloses in the two sample types.

Transmission electron microscopy of the tagged bacteria showed where the nanotags landed. On E. coli, particles appeared at the cell surface, while for B. subtilis and M. fortuitum significant numbers of particles associated with released material outside the cells. This pattern matches known glycan ultrastructure: B. subtilis produces glucose-containing strands up to 5 micrometers long, longer than the cell itself, and mycobacterial capsules feature secreted or loosely bound mannose- and glucose-containing glycans. The authors note that the construct can tag both cell-surface and secreted glycans, which extends its potential reach beyond intact cells.

The study is explicitly a preliminary proof of concept. The bacterial analysis involved three species, spectra were acquired in triplicate, and the authors state that further work, including principal component or artificial intelligence approaches, will be needed to assign discriminatory signals to specific glycans and to establish limits of detection. They also caution that microbial cell surfaces are extremely complex and variable over time, which complicates detection analyses. Still, the combination of a simple physisorption-based preparation, a low-cost spectroscopic readout, and differential spectral profiles across three bacterial genera supports the authors’ case that a single lectin-functionalized nanotag could serve as a tool for rapid, presumptive microbial identification, pending validation with larger and more diverse microbial samples.

Subject of Research: A lectin-functionalized gold nanoparticle nanotag for detecting bacterial surface glycans by surface-enhanced Raman spectroscopy

Article Title: Concanavalin A-functionalized nanogold – a glycan tag for surface enhanced Raman spectroscopy

Article References: Manisekaran, R., Becerra, A. H., Vázquez-Ibarra, M., Hernández-Padrón, G., & López-Marín, L. M. (2025). Concanavalin A-functionalized nanogold – a glycan tag for surface enhanced Raman spectroscopy. Gold Bulletin, 58(2), Article 14. https://doi.org/10.1007/s13404-025-00367-9

Image Credits: AI Generated

DOI: 10.1007/s13404-025-00367-9

Keywords: gold nanoparticles, Concanavalin A, surface-enhanced Raman spectroscopy, bacterial glycans, microbial detection, lectins, protein corona, nanotechnology, diagnostics, mycoloyl trehalose, Raman spectroscopy, Gold Bulletin