Three-dimensional Reconstruction and Quantification of Proteins and mRNAs at the Single-cell Level in Cultured Cells.

Gene expression is often regulated by the abundance, localization, and translation of mRNAs in both space and time. Being able to visualize mRNAs and protein products in single cells is critical to understand this regulatory process. The development of single-molecule RNA fluorescence in situ hybridization (smFISH) allows the detection of individual RNA molecules at the single-molecule and single-cell levels. When combined with immunofluorescence (IF), both mRNAs and proteins in individual cells can be analyzed simultaneously. However, a precise and streamlined quantification method for the smFISH and IF combined dataset is scarce, as existing workflows mostly focus on quantifying the smFISH data alone. Here we detail a method for performing sequential IF and smFISH in cultured cells (as described in Sepulveda et al., 2018 ) and the subsequent statistical analysis of the smFISH and IF data via three-dimensional (3D) reconstruction in a semi-automatic image processing workflow. Although our method is based on analyzing centrosomally enriched mRNAs and proteins, the workflow can be readily adapted for performing and analyzing smFISH and IF data in other biological contexts.

intergrade the analysis of the IF data in the pipeline. Here, we detail a streamlined workflow for performing, acquiring, and analyzing sequential IF and smFISH data via 3D reconstruction in Imaris software, followed by quantifications using MATLAB and R scripts. We use co-translational targeting of pericentrin (PCNT) polysomes to the centrosome during mitosis in adherent cultured cells as an example (Sepulveda et al., 2018) to demonstrate how to use Imaris software to reconstruct the PCNT mRNAs and proteins in 3D confocal z-stacks and to apply MATLAB and R scripts to quantify their intensities, volumes, and relations (e.g., spatial distribution of molecules and overlapping between signals) in a semi-automatic manner. Our protocol can be readily applied to other sequential IF and smFISH experiments to precisely quantify RNAs and proteins in 3D space.   RT (e.g., 1:1,000 dilution of rabbit anti-PCNT antibody in 1x PBS in our example). 5. Perform three 5-min washes with 1x PBS and incubate the cells with 70 µl of secondary antibody solution overnight in the dark at 4 °C (e.g., 1:500 diluted anti-rabbit Alexa Fluor 488).
Note: The dilution factor and incubation time of the primary and secondary antibodies need to be empirically determined; the sample condition described is for detecting PCNT proteins. To perform 5-min washes, gently transfer the coverslip to a well in a 24-well plate containing 1 ml of 1x PBS. Incubate for 5 min at RT without shaking the plate.

Imaging
Note: When using multiple fluorophores, it is important to carefully select the wavelength ranges, emission filters, and dichroic mirrors to avoid signal bleed-through. Use settings to minimize photobleaching but also maximize the signal-to-noise ratio. A camera with high sensitivity is preferred to obtain good smFISH signals (e.g., an electron multiplying CCD camera). We recommend acquiring images sequentially, channel by channel, starting from the longest wavelength. Below are the general IF and smFISH acquisition settings using the Dragonfly spinning confocal system in the Jao lab.    Figure 2D). e. Click 'Finish' to complete the rendering of protein surface ( Figure 2D). Examples are in    Set 'voxels outside surface' to '0' and inside surface to the highest voxel values of your imaging system (e.g., '30000' in this example). This will generate a masked channel of protein surface (e.g., Channel number 8 in Figure 5F). b. Click the RNA spot object generated in step 4. Add Filter ( Figure 5C)   b. Define 'fpath' as the location of the images (e.g., line 14 in Figure 6). c. Define protein and RNA of interest for quantification. These are the 'protein_surface_number' in line 18 and 'spot_number' in line 19 in Figure 6.
Note: For example, in Figure 5F, 'Protein surface number 2' is the Surface object created for the protein of interest. You will thus define 'protein_surface_number = 2' in line 18 in Figure 6. 'RNA spot number 1' is the Spot object created for the RNA of interest. You will thus define 'spot_number = 1' in line 19 in Figure 6.    the distance relative to the centrosome, from 0 µm (first row) to 20 µm (last row) with 0.5 µm intervals (Figure 9). Each column represents the amount of mRNA at each distance relative to the nearest centrosome (Figure 9). h. Specify the working directory and the location of the .csv files (e.g., in lines 11 and 12, respectively, in Figure 10).
i. Run the R script to obtain two .csv files within the new folder: (1) combined original dataset; (2) combined normalized dataset. b. Define 'fpath' as the location of the images (e.g., line 14 in Figure 11). c. Define centrosome protein surface number, overlapping RNA spot number, and nonoverlapping RNA spot number (e.g., in lines 18, 19, and 20, respectively, in Figure 11). Figure 5F, RNA spot number 2 and 3 are the Spot objects created for overlapping and non-overlapping RNA spots, respectively. You will thus define 'overlapping_RNA_spot_number = 2' in line 19 and 'nonoverlapping_RNA_spot_number = 3' in line 20 in Figure 11.

Note: In
d. Define the deconvolved channel number for RNA (e.g., line 23 in Figure 11) e. Run the script to obtain .csv files in the image folder. The first column represents the distance relative to the centrosome, from 0 µm (first row) to 20 µm (last row) with 0.5 µm intervals. The second column represents the percentage of RNA signals that overlaps with the protein surfaces within a given distance to the centrosome.   2. If MATLAB indicates errors about ImarisReader, run the script of 'CellsReader.m' to make sure that ImarisReader is functional before performing data analyses.