Abstract
Cryoelectron tomography (cryo-ET) enables three-dimensional visualization of molecular structures within tissue and intact cells, providing a powerful tool for studying the spatial organization of biological components at nanometer resolution. Realizing this potential, particularly for submegadalton complexes, is facilitated by fiducial-based labeling. Gold nanoparticles (AuNPs) have emerged as powerful electron-dense labels for cryo-ET, but multiplexing using only conventional monomeric AuNPs is challenging, limiting their application in multitarget studies. Here, we describe functionalized dimeric AuNPs with a precisely defined size range applied to bioimaging, enabling multiplexed labeling by allowing reliable discrimination between monomeric and dimeric AuNPs, thereby supporting identification of distinct molecular targets within the same cryotomogram. Each dimer consists of two covalently linked gold particles of approximately 2.6 nm diameter separated by a defined spacing of about 1.4 nm, with high structural homogeneity validated by small-angle X-ray scattering (SAXS) and electron microscopy. The anti-GluN1 Fab, which targets the N-methyl-D-aspartate receptor (NMDAR), was site specifically conjugated to a dimeric AuNP. A deep learning classifier enabled reliable discrimination between monomeric and dimeric AuNPs in tomograms. We confirmed that the dimeric AuNP-Fab conjugates bind robustly to recombinant GluN1/GluN2A receptors, validating their use for structural labeling. In situ cryo-ET of brain tissue further confirms that the dimeric labels reach NMDARs within the glutamatergic synaptic cleft. Combined with monomeric AuNPs, this dimeric AuNP platform establishes a generalizable approach for distinguishable labels optimized for cryo-ET. The compact size and structural uniformity of monomeric and dimeric AuNPs make them ideally suited for nanoscale molecular mapping in crowded cellular environments.