Single-protein optical holography

The study addresses a central challenge in optical microscopy: generating detectable image contrast from micro- and nanoscale objects. Phase-contrast concepts (originating with Zernike) and their successors exploit interference between incident and scattered light to render phase objects visible. While numerous variations—including phase-shifting interferometry, digital holography, quantitative phase imaging, and implementations optimized for nanoparticle tracking—have advanced sensitivity down to tens of nanometres, they have struggled to match the sensitivity of non-holographic, common-path interferometric approaches at interfaces (e.g., iSCAT and mass photometry). Common-path schemes achieve single-protein sensitivity but inherently couple the reference and scattered fields, limiting independent control and obscuring direct access to phase information. The research question is whether a non-common-path, inherently phase-stable holographic microscope can achieve single-molecule sensitivity while decoupling amplitude and phase to enable quantitative measurements (including mass and polarizability) of single biomolecules.

Classical phase-contrast microscopy and differential interference contrast introduced quantitative phase sensitivity for transparent samples. Phase-shifting interferometry, digital holography, and quantitative phase imaging extended capabilities to measure amplitude and phase, with applications to nanoparticle imaging and sizing. Holographic and dark-field schemes have detected metallic nanoparticles down to ~20 nm in imaging and ~15 nm with differential interference. Common-path interferometry (e.g., iSCAT) with laser illumination improved sensitivity to single proteins and enabled mass photometry via contrast scaling with polarizability, but phase information remains ill-defined due to inseparable reference and scattered fields and Gouy phase effects. Prior holographic methods required image sequences of the same particle for phase retrieval and were limited by static background of glass substrates. There is thus a gap for a holographic platform combining high sensitivity, independent phase control, and robust phase retrieval for single biomolecules.

The authors implement total-internal-reflection optical quadrature microscopy using a non-common-path, Mach–Zehnder-like interferometer. A single-emitter 450 nm laser is split: one arm illuminates the sample via prism-coupled total internal reflection; the other serves as a mode-filtered reference. Polarization optics impose controlled phase shifts: a quarter-wave plate in the reference arm yields orthogonal polarizations with a π/2 phase difference; a half-wave plate in the scattering arm imposes a π shift between orthogonal polarizations. After recombination at a non-polarizing beamsplitter and separation by downstream polarizing beamsplitters, four synchronized cameras record interferograms with relative phase shifts of 0, π/2, π, and 3π/2. Total internal reflection enhances the field at the interface and suppresses background from out-of-plane scatterers; illumination leakage to the detector is minimized except for interface imperfections. The setup is engineered for phase stability via short beam paths, rigid cage mounting, environmental covers, and thermal stabilization. A separate focusing channel maintains axial focus via feedback on the reflected 520 nm alignment beam. Each camera j records I_cam = |E_ref|^2 + |E_scat|^2 + 2|E_ref||E_scat| cos(Δφ + Δφ_j). Reference-only and scatter-only images (obtained by blocking arms) are used to subtract non-interferometric terms and normalize by |E_ref|, isolating |E_scat| cos(Δφ + Δφ_j). Complex reconstruction multiplies each camera’s corrected image by exp(iΔφ_j) and averages across the four cameras to retrieve the complex scattered field E_scat, yielding amplitude and relative phase maps. A 150 µm laser coherence length ensures consistent interference over the field of view. Data are normalized to exposure time and laser power. For validation, AuNPs (20, 40, 60 nm) are immobilized and recorded under controlled phase ramps using a piezo-driven mirror. For protein detection, a protein solution (DynΔPRD) is introduced; landing events are identified by sliding-window differences (ten-frame windows at 6.25 ms/frame) applied to reconstructed complex videos after temporal phase correction. Events are localized and fitted using a complex-valued PSF model (Bessel J0-based with complex phase) derived from averaged experimental PSFs to obtain amplitude (contrast) and relative phase per landing. Holographically reconstructed intensities are cross-validated against dark-field acquisitions (reference arm blocked). Calibration of scattering cross-section uses Mie-theory values for AuNPs, regressing measured contrast to derive a mapping to σ_scat; protein σ_scat then yields specific excess polarizability via the Rayleigh relation.

• The non-common-path, four-channel optical quadrature microscope achieves single-molecule sensitivity comparable to common-path interferometry while enabling independent amplitude and phase retrieval. • The method provides a five orders of magnitude improvement in sensitivity in terms of minimum detectable scattering cross-section over prior holographic approaches, with a dynamic range beyond 10^3 in holographic contrast. • Amplitude–phase decoupling is demonstrated: under a controlled phase ramp, single 40 nm AuNPs show sinusoidal intensity modulations across the four cameras but reconstruct to a constant amplitude with relative standard deviation ~0.31%; phase stability without ramping is a few milliradians per second. • Holographically reconstructed scattered intensity matches dark-field intensity to within ~20% (attributed to imperfect coherence), enabling quantitative amplitude extraction after calibration. • Single-protein holography: individual proteins (DynΔPRD) landing on glass are detected in both amplitude and phase via differential analysis, despite no visible changes in raw amplitude images due to glass roughness. • Contrast histograms resolve oligomeric states corresponding to ~90, 180, 360, and 540 kDa (monomer, dimer, tetramer, hexamer), with linear scaling of contrast with molecular mass and an average mass resolution of 28 kDa. • Relative phase of landing events, after correcting for the TIR-induced spatial phase gradient, collapses onto a line versus contrast as expected for non-absorbing dielectric particles; heavier (larger) oligomers exhibit slightly larger relative phase due to increased separation from the interface. • Post-acquisition refocusing using the complex field (numerical propagation) optimizes contrast and aids robust quantification. • Calibration with AuNPs shows that the cubic root of holographic contrast scales linearly with particle radius (Rayleigh regime), and the sixth root of dark-field contrast does likewise; dark-field intensities are ~20% larger than holographic estimates, consistent with coherence considerations. • Using Mie-calibrated AuNP scattering cross-sections and Rayleigh theory, the specific excess polarizability of protein is determined as 239 ± 4 Å^3 per kDa, implying an estimated protein refractive index n ≈ 1.47; these values fall within literature ranges.

By separating and independently controlling reference and scattered fields in a phase-stable, non-common-path geometry, the method overcomes a key limitation of common-path interferometric microscopy: ill-defined phase. Accurate complex-field retrieval enables quantitative amplitude and phase imaging at the single-protein level, supports post-processing focus optimization, and provides a route to intrinsic biophysical properties such as polarizability. The linear contrast–mass relationship validates mass measurement and oligomer resolution without labels. The observed phase–contrast dependence aligns with expectations for dielectric nanoparticles near an interface. Compared with dark-field, the interferometric approach yields well-defined events on rough glass surfaces by avoiding interference with uncontrolled substrate-scattered wavefronts. Although current mass precision trails state-of-the-art mass photometry, the platform demonstrates sensitivity sufficient for sub-100 kDa proteins and offers unique phase information inaccessible to common-path schemes. Anticipated improvements in illumination power density, optimized TIR field enhancement, phase control, and background suppression should further improve sensitivity and precision, potentially surpassing common-path methods for single-molecule mass metrology and enabling nanometre-scale distance and identity measurements.

The study introduces total-internal-reflection optical quadrature microscopy that achieves holographic imaging of individual biomolecules with decoupled amplitude and phase. The approach detects and quantifies single proteins below 100 kDa, resolves oligomeric states with an average mass resolution of 28 kDa, and directly measures the specific excess polarizability of proteins (239 ± 4 Å^3 kDa⁻¹). Calibration with AuNPs validates quantitative amplitude retrieval and theoretical scaling. Access to the complex field enables post-acquisition refocusing and phase-sensitive analyses, expanding holography to sub-20 nm scales in biological and materials applications. Future work should increase illumination power density, optimize TIR incidence for maximal field enhancement, implement active phase modulation and structured illumination to stabilize illumination phase, add a bleed-through channel for absolute phase measurement, and employ atomically flat substrates to minimize background, with the prospect of surpassing common-path geometries in single-molecule sensitivity and mass resolution.

Current experiments are limited by an illumination power density of ~50 kW cm⁻² (about an order of magnitude below mass photometry), reducing shot-noise-limited sensitivity. The TIR incidence angle was not optimized, yielding only ~10% field enhancement versus a potential threefold intensity increase. Absolute scattering phase is not measured due to unknown reference–illumination phase; only relative phase is retrieved. For low-contrast (small) particles, phase precision deteriorates due to finite localization precision. Imperfect coherence between reference and scattered fields produces a ~20% intensity discrepancy between holography and dark-field. The roughness of standard glass substrates contributes background speckle, and while differential analysis mitigates this, atomically flat substrates could further improve precision. Present mass resolution (28 kDa) remains above that of state-of-the-art mass photometry.