Figure 2. Correlative confocal-STORM... | Rockefeller University Press

Figure 2.

$Figure 2. Correlative confocal-STORM imaging reveals the internal structure and size of RDs. (A) Cartoon of a single-color–labeled CT. A subset of RDs is tagged by co-replicative labeling (green circles). Each RD consists of several replicons, which are simultaneously activated during S-phase and co-replicate their DNA synchronously. At confocal resolution (∼300 nm), RDs appear as subdiffraction-sized spots, whereas at a resolution of ∼20 nm by using STORM imaging, it is possible to resolve the individual replicons or co-replicating stretches. (B) Experimental pipeline for correlative confocal-STORM imaging. NRKs labeled with ATTO 633-dUTP were plated onto gridded coverslips and screened in 2D to identify cells of interest containing a few CTs and to be imaged in 3D at confocal resolution. The coverslips were then put in switching buffer, and the cells of interest imaged in 2D with the STORM microscope at super-resolution. The correlative confocal stacks and 2D-STORM images were analyzed to find the optimal confocal substack-STORM overlay and to discard out-of-focus regions of the STORM images. Finally, DBSCAN clustering analysis was performed on the cropped STORM images to obtain estimates of number of co-replicating stretches/RD, RD size, and RD diameter. Bars: (overview images) 5 µm; (lower panels) 500 nm. (C) ATTO 633–labeled CTs imaged by correlative confocal-STORM imaging. Panels from left to right: optimal confocal z-substack projection (gray), STORM image (gray), overlay (confocal in magenta, STORM in green). Bars: 500 nm; (magnified portion) 50 nm. Small groups of co-replicating stretches can be resolved in the STORM images and are not discernible from the confocal images. (D) Representative confocal-STORM overlay showing qualitatively the correspondence between diffraction-limited RDs and super-resolved groups of co-replicating stretches. Clustering analysis: example STORM image after filtering and intensity thresholding; density-based clustering of detected co-replicating stretches to RD (color-coding shows forks belonging to the same cluster, and the red lines denote the convex hull around the centers of detected co-replicating stretches; yellow squares mark unclustered, solitary co-replicating stretches). Bars: 500 nm; (magnified portion) 100 nm. (E–G) DBSCAN: histograms of the number of co-replicating stretches counted per cluster/RD and median value; n = 87 CTs from 37 cells (E), the NND between co-replicating stretches and median value (F), and the horizontal Feret-diameter of RD and median value (median including clusters of two co-replicating stretches = 105 nm; G).$

Correlative confocal-STORM imaging reveals the internal structure and size of RDs. (A) Cartoon of a single-color–labeled CT. A subset of RDs is tagged by co-replicative labeling (green circles). Each RD consists of several replicons, which are simultaneously activated during S-phase and co-replicate their DNA synchronously. At confocal resolution (∼300 nm), RDs appear as subdiffraction-sized spots, whereas at a resolution of ∼20 nm by using STORM imaging, it is possible to resolve the individual replicons or co-replicating stretches. (B) Experimental pipeline for correlative confocal-STORM imaging. NRKs labeled with ATTO 633-dUTP were plated onto gridded coverslips and screened in 2D to identify cells of interest containing a few CTs and to be imaged in 3D at confocal resolution. The coverslips were then put in switching buffer, and the cells of interest imaged in 2D with the STORM microscope at super-resolution. The correlative confocal stacks and 2D-STORM images were analyzed to find the optimal confocal substack-STORM overlay and to discard out-of-focus regions of the STORM images. Finally, DBSCAN clustering analysis was performed on the cropped STORM images to obtain estimates of number of co-replicating stretches/RD, RD size, and RD diameter. Bars: (overview images) 5 µm; (lower panels) 500 nm. (C) ATTO 633–labeled CTs imaged by correlative confocal-STORM imaging. Panels from left to right: optimal confocal z-substack projection (gray), STORM image (gray), overlay (confocal in magenta, STORM in green). Bars: 500 nm; (magnified portion) 50 nm. Small groups of co-replicating stretches can be resolved in the STORM images and are not discernible from the confocal images. (D) Representative confocal-STORM overlay showing qualitatively the correspondence between diffraction-limited RDs and super-resolved groups of co-replicating stretches. Clustering analysis: example STORM image after filtering and intensity thresholding; density-based clustering of detected co-replicating stretches to RD (color-coding shows forks belonging to the same cluster, and the red lines denote the convex hull around the centers of detected co-replicating stretches; yellow squares mark unclustered, solitary co-replicating stretches). Bars: 500 nm; (magnified portion) 100 nm. (E–G) DBSCAN: histograms of the number of co-replicating stretches counted per cluster/RD and median value; n = 87 CTs from 37 cells (E), the NND between co-replicating stretches and median value (F), and the horizontal Feret-diameter of RD and median value (median including clusters of two co-replicating stretches = 105 nm; G).

This Feature Is Available To Subscribers Only