Isolation of murine large intestinal crypt cell populations with ow sorting

Amber N. Habowski (  Habowski@uci.edu ) University of California Irvine https://orcid.org/0000-0003-1107-3208 Jennifer M. Bates University of California Irvine Jessica L. Flesher University of California Irvine https://orcid.org/0000-0002-3808-7661 Robert A. Edwards University of California Irvine Marian L. Waterman (  marian.waterman@uci.edu ) University of California Irvine https://orcid.org/0000-0003-4823-4968


Introduction
Multiple cell sorting protocols have been optimized to isolate intestinal stem cells, but each lack the resolution to purify daughter cells and differentiated progeny populations 1,2 . For example, the transgenic stem cell lineage marker Lgr5-EGFP enables puri cation of GFP-bright stem cells, but a mosaic expression pattern of the transgene in the intestine has made it di cult to con dently separate daughter cells from GFP-negative stem cells and differentiated cells 3,4 . Single cell RNA sequencing (scRNA-seq) captures cellular diversity when analyzing mixed cell populations and has been useful for de ning intestinal lineage trajectories and diversity of mature cells (for example enterocytes and enteroendocrine cells) [5][6][7][8][9] . However, the low sequencing depth of scRNA-seq misses moderate-to-lowly expressed transcripts and is not compatible with other downstream analysis or applications.
The sorting protocol detailed here enables sorting of colonic crypt cell populations from the large intestine of the mouse, independent of transgene markers. This protocol is compatible with a variety of downstream applications including bulk RNA-seq and mass spectrometry. Importantly, our analysis has validated the identity of the isolated populations, enabling others to use this protocol for FACS analysis of their intestinal system to chart changes in crypt dynamics and populations. For FACS analysis <1/3 mouse colon is more than su cient for a snapshot of crypt populations, although for sorting, several mice may need to be pooled depending on the downstream application. This protocol is also compatible with additional antibody markers or mice of any strain/gender (including transgenic mice -with compatible uorophore). We recommend additional markers use FITC-EGFP channel.  7. Gently scoop out the small intestine to the left side. Cut through the pelvis just to the right of the rectum, then nd the rectum and cut at the end near the junction with the skin. Avoid cutting blood vessels. 8. Cut the small intestine where it meets the cecum. Gently tug the top of the colon and if cut correctly at the rectum, it will slowly pull free with minimal mesentery tissue. 14. Remove colon tissue from dissociation solution #1. Cut tissue into small ~3-5 mm pieces, using forceps to dangle the tissue above dissociation tube #2 with edge of tissue resting on tube rim to pull taut and cutting with scissors. Make sure all pieces are immersed in the dissociation solution #2. Place back on rotator at 4 °C for an additional 30 min rotation.

Reagents
15. Turn on centrifuge (swinging bucket preferred) during this period to allow time to chill to 4 °C. Adjust settings: 500 xg, 5 min, 4 °C, decrease deceleration speed to low setting.
16. After the 30 min dissociation step is completed (1 hr in total), collect conical tubes and shake for 3 minutes very aggressively and rapidly (up and down motion). Solution should quickly become cloudy with an observable abundance of oating cells.
17. Pour suspension through 100 µm lter into new conical tube (100 µm lter #1 tube -example label M#1 -100f #1). Rinse lter with cold PBS ( nal volume of 40-50 mL). Collect tissue chunks trapped in the lter and place back in the dissociation tube. Re-use this lter for the next 100 µm lter for this sample (move the lter to the conical tube labeled 100 µm lter #2).
18. Add ~20 mL of PBS to the dissociation tube that contains tissue chunks and store on ice. It is important to not allow the tissue chunks to get dry.
19. Repeat steps 17-18 for all samples and then immediately spin down all 100 µm lter #1 tubes at 500 xg, 5 min, 4 °C, with decreased deceleration speed. NOTE: All 4 collection tubes will need to be processed rapidly in sequence so that centrifugation steps are done together. The best rates of cell survival depend on minimizing the time cell suspensions are sitting and ensuring that when they are sitting it is always on ice.
20. Manually shake tissue in the dissociation solution tubes (now with 20 mL PBS) again for 3 minutes rapidly.
22. Gently pour off supernatant from the 100 µm lter #1 tube and resuspend in 1 mL of FACS buffer. Mix well with a pipet to achieve a homogenous suspension. Filter the suspension through a 40 µm lter into a new tube (40 µm lter #1 tube). Rinse the lter with PBS lling to 50 mL. Repeat for all samples and save the 40 µm lters for the next step.
23. Repeat step #22 using the suspension from the second shake (100 µm lter #2 tubes) and lter into 40 µm lter #2 tube using the same 40 µm lter saved from step #22 for each sample.
24. Centrifuge the suspension in both 40 µm lter tubes (#1 and #2) at 500 xg, 5 min, 4 °C, with a decreased deceleration speed to protect cell viability. 28. Carefully remove and discard supernatant from all tubes using a pipet. Dispense 500 µL of FACS buffer to each tube -if more than one tube was collected per sample, merge the contents of these into one tube with 500 µL FACS buffer total. Add 50 µL of DNAse and mix. Mechanically mix the DNAse and cell suspension up and down 5-10 times with a P-1000 pipet. Incubate at room temperature for 5 min. For sorting followed by RNA isolation we recommend sorting directly into TRIzol -a step that preserves RNA integrity. Collection tubes can be FACS tubes or Eppendorf tubes.

Collect the cells by centrifugation
For sorting followed by mass spectrometry we changed the sorting machine sheath uid to 100 mM ammonium bicarbonate and sorted directly into 50 µL of 100 mM ammonium bicarbonate -a step that preserves protein integrity and prevents salt contaminants. We recommend sorting into PCR tubes (that can be tted inside of Eppendorf tubes) -although this depends on the set-up of downstream mass spectrometry equipment.
Regardless of downstream applications store collection tubes on ice/chilled prior to, during, and after sorting.

Populations to collect (live cells):
Stem Cells
Solution: This is likely caused by not being aggressive enough with the manual shaking steps. To test how aggressive and effective the shaking is an additional (3 rd ) round of shaking the tissue in PBS can inform on whether additional cells are recovered. If the cell pellet is much larger than the rst two pellets, there is a clear need to shake harder starting in the beginning 2. Large/decent pellets during 100 µm lter, but small after 40 µm lter.
Solution: This is likely due to cells not being released into a true single cell suspension. To address this, one can shake for longer and/or more aggressively during the initial shaking or add some additional gentle shaking of the PBS resuspension prior to using the 40 µm lter. Alternatively, if the problem persists, do not use the 40 µm lter following the 100 µm lter step. Instead, continue with the protocolincluding the important DNAse treatment step -and prior to adding Live/Dead stain, lter the cell suspension into a FACS tube with a 40 µm lter cap (depending on the single cell suspension several lter cap might be needed -this can be painstakingly slow but will improve yield).
3. During Eppendorf centrifugation the cell pellet is poor and a bit uffy.
Solution: If the pellet is small and/or of poor quality always repeat the centrifugation step. Gently ick/vortex the tube to release the pellet and centrifuge again. If possible, a swinging bucket centrifuge can improve quality of pellets.

Low cell viability.
Solution: Because of the long duration of this protocol and the fragility of mature epithelial cell types there is an innately low cell viability. Some important things to implement to improve viability include keeping cells on ice/chilled at all times unless protocol speci es otherwise. Additionally, working quickly (immediately after sacri ce) and smoothly during the dissection and linearization and ensuring that the tissue does not dry out are important features of the protocol. Maximally active Rock inhibitor and high quality FBS in the FACS buffer increases viability.

Total cell yield during sorts is very low.
Solution: This protocol is not designed for an optimized yield of cells, but rather for a high-quality separation that can distinguish cell populations. Please see section below on improving yield if this feature is speci cally important for downstream applications.
6. During the FACS procedure, cells are clumpy, clog the machine, or do not run at a constant ow rate.
Solution: Thoroughly mix the sample with a pipet and/or gently vortex. Dilute the sample with FACS buffer, and pass through a 40 µm lter cap. Since extracellular, extruded DNA from lysed cells is a major reason for cell clumping, make sure the DNAse concentration and treatment time is su cient for the number of cells. In addition, also make sure the FACS buffer contains FBS to help prevent clumping.
Additional Steps to Improve Cell Yield: 1. Instead of using 100 µm lter followed by 40 µm lter during the initial centrifugation steps, use only the 100 µm lter (or a 70 µm lter instead). Immediately prior to adding the Live/Dead stain, lter the cell suspension into a FACS tube with a 40 µm lter cap (depending on the volume and density of the single cell suspension you might need several caps -this can be painstakingly slow. It will nevertheless improve yield).
2. For all ltering steps (50 mL conical) swirl the pipet tip along the lter and pipet up and down, to help solution pass through. Be sure to add additional PBS/FACS buffer to rinse the lter which will collect additional cells.
3. IMPORTANT: Pre-wet pipet tip with FACS buffer before resuspending any cell solution to prevent cellular adherence to the walls of the tip.       FACS gating strategies that de ne six colonic crypt cell populations. As a rst step, standard gating is performed to select single, live cells based on forward and side scatter (Step 1). A dump channel then removes dead cells along with immune cells (Cd45+) and endothelial cells (Cd31+) (Step 2). Epcam+ cells (Step 3) are then gated using Cd44, Cd24, and cKit to isolate six distinct populations (Steps 4-7).