Multiplexed Chromatin Analysis Using Optical Spectroscopic Statistical Nanosensing

Abstract

In single cells, chromatin packs into organized structures to perform biological functions, such as RNA transcription regulation. Characterizing such structural behaviors, including packing density and mass scaling, is critical in epigenetics research. Partial wave spectroscopic (PWS) microscopy is a label-free, live-cell, high-throughput imaging modality that utilizes optical spectroscopic statistical nanosensing. Rather than resolving the exact chromatin packing structure, PWS extracts statistical packing information from spectroscopic interference signals. In this study, we evaluate its ability to characterize multiplexed chromatin packing density and mass scaling, as well as its spatial confidence interval, using finite difference time domain (FDTD) electromagnetic simulations. We validated the simulation-based analysis algorithm by comparing experimental PWS images against coregistered super-resolution acquisitions, confirming its accuracy in capturing chromatin packing metrics. We then applied this modality to live cells treated with different epigenetic agents, mapping spatial changes in chromatin packing in a high-throughput workflow.

Keywords: chromatin imaging; finite difference time domain simulation; spectroscopic nanosensing.

Figure 1.

(a) FDTD simulation of average backscattering intensity (normalized by backscattering intensity of the glass-nucleus interface) as a function of effective nucleus $\varphi$. (b) Theoretical model and FDTD simulation of $\Sigma$ to $D$ conversion.

Figure 2.

(a) FDTD synthesized PWS 2D $D$ map for one random material with packing scaling $D$ = 2.65. Scale bar: 1 μm. (b) Relative error “e” and confidence interval “C.I.” of spatial average $D$ as it relates to sampling size.

Figure 3.

(a) XY cross section of RI distribution of random materials used for FDTD simulation, with high $D$ region boundaries marked in red. (b) Corresponding simulated PWS $D$ maps from FDTD. (c) Average $D$ value and confidence interval (C.I.) as it relates to the distance from high $D$ region centers. Scale bar: 1 μm for (a) and (b).

Figure 4.

(a) PWS $D$ measurement has an effective Gaussian filter. PWS $D$ map (right) matches the material $D$ distribution (left) convolved with a 2D Gaussian filter (middle) with σ = 0.49 μm. (b) Applying deconvolution on PWS $D$ map (left) reconstructs material $D$ distribution (middle) close to the input configuration (right). Scale bar: 1 μm.

Figure 5.

PWS is capable of measuring $D$ and $\varphi$ simultaneously. (a) Simulation grid setup with 4 regions, each with different $D-\varphi$ combinations. (b) Simulated $\varphi$ and (c) $D$ map, with vertical or horizontal lines showing the expected separation. (d) Multiplexed view of the simulated PWS image with both $D$ and $\varphi$ information. Scale bar: 1 μm.

Figure 6.

Experimental implementation of PWS deconvolution and $D-\varphi$ multiplexing. (a) STORM image of RNAP II labels of a m248 ovarian cancer cell, with the manually drawn dashed line marking the nucleolus zone. (b) Corresponding PWS $\varphi$ and (c) $D$ measurement of the same nucleus. (d) Multiplexed PWS image of the same nucleus. Scale bar: 5 μm.

Figure 7.

Typical experimental PWS image acquiring global change in $D$ values with (a) ActD and (c) MgCl₂ treatment. Scale bars 5 μm. (b, d) $D$ distribution comparison. ***P < 0.001.

Figure 8.

PWS analysis on B-type lamin depletion. (a) PWS $D-\varphi$ multiplexed image of a typical untreated cell. Scale bar: 5 μm. (b) Nuclear periphery $D$ and $\varphi$ distribution comparison between 24 h Auxin treatment and control. ***P < 0.001.