Probe-Seq: Method for RNA Sequencing of Specific Cell Types from Animal Tissue.

Most organs and tissues are composed of many types of cells. To characterize cellular state, various transcription profiling approaches are currently available, including whole-tissue bulk RNA sequencing, single cell RNA sequencing (scRNA-Seq), and cell type-specific RNA sequencing. What is missing in this repertoire is a simple, versatile method for bulk transcriptional profiling of cell types for which cell type-specific genetic markers or antibodies are not readily available. We therefore developed Probe-Seq, which uses hybridization of gene-specific probes to RNA markers for isolation of specific types of cells, to enable downstream FACS isolation and bulk RNA sequencing. We show that this method can enable isolation and profiling of specific cell types from mouse retina, frozen human retina, Drosophila midgut, and developing chick retina, suggesting that it is likely useful for most organisms.

To make a 50% solution, dissolve 5 g of powdered dextran sulfate in 8 ml of UltraPure water overnight, then bring total volume to 10 ml. Store 50% dextran sulfate at -20 °C 8. RNasin Plus (Promega, catalog number: N2615). Store at -20 °C 9. 5 ml round-bottom polysterene test tube with 35 µm cell strainer cap (Corning, catalog number: 352235). Store at RT Depending on the tissue and cell type of interest, other strainer sizes are available (i.e., 70 µm). 10. Oligo probes and hairpin (see section on "Oligo and hairpin ordering") 11. Fluorescent oligos (see section on "Probe Hybridization Recipes") 12. Permeabilization Buffer (see Recipes) 13. FACS Collection Solution (see Recipes) 14. 40% wHyb (see Recipes) 15. Hyb1 (see Recipes) 16. Probe Mix (see Recipes) 17. Fluorescent Oligo Mix (see Recipes)  c. Another probe design software is available (Passaro et al., 2020). The website can be accessed here: http://oligominerapp.org/home. However, probe design for Probe-Seq did not use these websitse because they were not available at the time. Therefore, we have not tested the probes generated by these software.

Reagents for Library Preparation
Before you start: • Get access to your institution's computing cluster.
• Transfer the OligoMiner-master folder in the SABER_Probe_Design folder on the cluster. For example, use the rsync command in the Terminal (Mac) to transfer files between the cluster and the local computer (example shown below).
• Download the "Complete Genome" with the "Balance" setting. Unzip the file and transfer the folder in the SABER_Probe_Design folder on the cluster. 7 www.bio-protocol.org/e3749 Note: If the target RNA is short and does not produce sufficient number of oligos, including introns may help. This is especially true if using nuclei as most of the RNA will be nascent. 9. Click "get BED". This will download your BED file. 10. Open the BED file. We use a software called Brackets, which is free to download at: http://brackets.io/.
11. If there are multiple isoforms listed (i.e., multiple NM_# with different #), choose one of the isoforms to keep and manually delete the others. Save the file.
12. Note the 2 features in the BED file as you will need them later: a. The chromosome number. b. Strand orientation "+" or "-".
13. Transfer the BED file to the SABER_Probe_Design folder on the cluster [i.e., type "rsync 14. Log into the cluster.
Note: How you log into your cluster will be different institution to institution.

Note: The versions will be different for every cluster.
17. Run intersectBed between the genome-wide probe set (using the correct chromosome BED file) and your BED file with the option of-f 1. (i.e., Type in "intersectBed -a Notes: a. Make sure you're running this command in the folder with the genome-wide probe set (i.e., mm10b) and your gene-specific BED file (i.e., Grik1_Mm).
b. If you have enough computing power, it's also possible to run this command locally.
18. If your gene-specific BED file had a "+" orientation, you need to reverse complement. Use the probeRC.py file from the OligoMiner-master folder (i.e., Type in "python OligoMiner- 23. Under "DNA oligos," click on "Order Now" and "All ordering options." 24. Under "Ordering," click on "Plates" and order 25 nmole DNA Plate oligo in a 96 well format. 9 www.bio-protocol.org/e3749      11. Transfer homogenate to a 5 ml polypropylene tube (Thermo Fisher Scientific). Make sure to seal the cap by pushing in.

Note: If a temperature-controlled centrifuge is not available, RT spin is also acceptable. However,
there may be slightly more RNA degradation. 14. Spin 2,000 x g for 5 min at 4 °C.
Note: The centrifugation speed has changed to 2,000 x g after fixation.

Note: If an in situ hybridization oven is not available (i.e., Bellco), a water bath should suffice
although it has not been tested in our lab.
18. Spin 2,000 x g for 5 min at RT.  Coverslip with mounting media.

FACS (~30 min for 1 million cells)
34. Gate first based on Hoechst histogram (Figure 3). This step will ensure that you get 2N cells. Make sure to NOT include doublets. You can also gate based on FSC/SSC for cells vs. debris, but this step is likely unnecessary as debris will be negative for Hoechst.

Note: If your tissue contains polyploid cells (4N+ cells), it will be difficult to distinguish them from
doublets or other cell clumps. In this case, it may be best to use the nuclei Probe-Seq protocol.
35. Gate using a 2D plot with appropriate wavelengths. The negative population will likely run diagonal.
The positive population will be left-or right-shifted compared to the diagonal events (i.e., the Vsx2 + cells). Often, the separation will not be as obvious as GFP or well characterized cell surface markers.
Thus, it's a good idea to run a negative control (fluorescent oligo only control). to determine the events that have high fluorescence in all channels. * Do not collect these events even if they have a high fluorescence intensity in the Vsx2 channel. These events are also highly fluorescent for the empty autofluorescence (561 nm) channel. If using a histogram or a plot without an autofluorescence channel, these events will look like they are Vsx2 + even though they are unlikely to be the population of interest.
Note: It is advisable to keep one laser channel unused as this can be used to determine autofluorescence (i.e., the 561 nm channel in Figure 3)   • Check the ordered oligo sequences that the "TTT" linker and the hairpin primer sequences have been added.
• During probe synthesis, make sure that the hairpin primer on the oligos and the hairpins match (i.e., Grik1.25 and 25.25 hairpin).
There's no separation between gene-positive and gene-negative events • Make sure that the marker gene is expressed highly.
• Check the number of oligos. If less than 15-20, there might not be enough tiling oligos, and a new marker gene may be necessary.
• Make sure that the temperature of the solution does not deviate from the appropriate temperature (i.e., 43 °C) during solution changes.

Cells are clumped together
• Make sure you do not use Tween-20 in your wHyb or Probe Mix.
• Dissociate the tissue for longer.
• There may be other dissociation methods for your tissue of interest.
The signal is not strong enough to separate the populations • If the number of oligos is low, you can try using both Exons and Introns when generating your probe set. This has not been tested in our hands, but it should increase signal.
• Incubate for longer time (24 h instead of 16 h).