High-throughput multicolor 3D localization in live cells by depth-encoding imaging flow cytometry

Abstract

Imaging flow cytometry replaces the canonical point-source detector of flow cytometry with a camera, unveiling subsample details in 2D images while maintaining high-throughput. Here we show that the technique is inherently compatible with 3D localization microscopy by point-spread-function engineering, namely the encoding of emitter depth in the emission pattern captured by a camera. By exploiting the laminar-flow profile in microfluidics, 3D positions can be extracted from cells or other objects of interest by calibrating the depth-dependent response of the imaging system using fluorescent microspheres mixed with the sample buffer. We demonstrate this approach for measuring fluorescently-labeled DNA in vitro and the chromosomal compaction state in large populations of live cells, collecting thousands of samples each minute. Furthermore, our approach is fully compatible with existing commercial apparatus, and can extend the imaging volume of the device, enabling faster flowrates thereby increasing throughput. RESULTS Device and functionalityAn illustration of the device is shown in Figure 1 A. Samples are loaded into the instrument and then directed through a multilaser-illuminated imaging volume. Emission light from the sample is captured by a 60X objective, relayed through a series of lenses, divided into separate spectral channels, and directed onto a camera, whose readout rate is synchronized with the flow speed. As a first demonstration, we show how an astigmatism can be incorporated into the system. By inserting a cylindrical lens between two of the aforementioned relay lenses (see Methods), the axial and lateral focal planes of the instrument are bifurcated, thus an object closer to the objective will appear focused laterally, but defocused axially and vice versa (Figures 1 B, C). The extent of the astigmatism contains information about the depth of the emitter, but is not a direct measurement of the z position. The typical 3D calibration used in microscopy (Huang et al., 2008) is therefore inapplicable and the incorporation of PSF engineering into flow imaging necessitates a novel 3D calibration procedure, which we describe below.