Mapping transcriptome

1This protocol contains many centrifugation steps where the supernatant needs to be discarded and the pellet saved. Be careful not to dislodge pellet fragments when pipetting the supernatant to avoid losing nuclei over the course of the protocol. Monitor pellet integrity during each pipetting step. We find that the 2-step centrifugation scheme significantly helps to retain nuclei during supernatant removal steps. The swinging-bucket rotor centrifugation helps form a tight pellet at the bottom of tube, with no accumulation of cells on the side of the tube. The fixed-angle rotor centrifugation shifts the pellet to one side of the tube, which makes it easier to pipette out the supernatant without touching the pellet. If a swinging-bucket rotor is not available (in which case, perform a 3 min centrifugation at 2,500×g in a fixed angle rotor centrifuge instead of the recommended 2-step centrifugation), or if the pellet appears unstable, re-centrifuge the sample for an additional 1 min or spin the tube for ~10 sec in a minicentrifuge before removing the remainder of the solution with a P200 pipet. When starting with as low as 1-2 million mammalian cells, a clear pellet should be visible throughout the protocol.

2Alternatively, prepare the mix in a PCR tube and run the following program in a thermocycler: 3 min at 94 °C, then cool to 4 °C at 0.1 °C/s.

3Enzyme concentrations and bridge concentration may need to be optimized when starting with a different cell number or a different cell type. We have successfully prepared libraries starting from 5-20 million cells or with various human cell lines without altering the protocol, but have observed some variation in the yield of final “useable” sequenced molecules and library complexity.

4Do not allow the tube to become warm to avoid protease activity.

5For the wash step after the cell lysis, we remove Triton from the buffer, as leftover Triton during the subsequent step with 0.5% SDS at 37 °C may lyse the nuclear membranes. Protease inhibitors and RNaseOUT are not necessary for this short wash step.

6The appearance of the pellet is usually different after the nuclei preparation at centrifugation step 3.3.5. and after the SDS treatment at centrifugation step 3.4.1. At the former step, the pellet appears white and resuspends easily by pipetting. At the later step, the pellet looks more transparent, and is significantly harder to resuspend. Pipette up and down vigorously until the pellet is fully resuspended.

7Throughout this protocol, when the pellet is washed, we perform the first wash in PBS for simplicity, and the second wash in a buffer similar to the buffer used in the subsequent enzymatic step to equilibrate the ionic conditions.

8RNA fragmentation is performed by heating up the samples to 70 °C in the presence of magnesium (2.5 mM). The resulting RNA fragment length distribution is sharply dependent on the incubation time at 70 °C. If necessary, one can optimize the fragmentation by titrating the heating time and quantifying the RNA length distribution after decrosslinking the sample as in step 3.13., extracting the RNA using standard RNA extraction protocols, and running the RNA on a gel or bioanalyzer. We found that a fragmentation between 2-5 min works well for various cell types.

9The PEG solution is very viscous and should be added last to be able to mix all the other components more efficiently.

10The genomic digestion can be shortened or lengthened as desired. While we have prepared successful ChAR-seq libraries with 3 h digestion, a longer digestion may be required to completely digest the genome for a given cell type and cell number. Alternatively, the digestion can be done overnight, which also provides a convenient break point for the library preparation.

11Ligation can be done overnight. With the incubation times in this version of the protocol, one proceeds from fixed cells on Day 1 to precipitated DNA in the morning of Day 3. However, Day 2 is long and extends over 12 h. One can easily split Day 2 into two shorter days by allowing either the genomic digestion or bridge-DNA ligation to proceed overnight. We have found that these modifications do not dramatically affect the final results.

12The second 70% ethanol wash is important to get rid of any residual SDS. Presence of residual SDS can create foaming during acoustic shearing and result in incomplete shearing.

13It is key to verify the DNA was sheared to the proper fragment size distribution before proceeding. If the fragment size distribution reveals incomplete shearing, it is possible to repeat step 3.15. using the same incident power, duty factor, and number of cycles per burst. The additional shearing time may need to be optimized but can be estimated by referring to the table of recommended shearing time vs. target peak size in the acoustic shearer manual. More specifically, compare the recommended shearing time for the obtained fragment peak and the recommended shearing time for the target peak size. An additional shearing time corresponding to the difference of these two values provides a good starting point.

14Agitation can be performed using a thermomixer at 900 rpm or a rotator and is necessary to prevent beads from settling.

15The goal of the PacI digest is to eliminate molecules where the bridge was not DpnII digested, and that therefore cannot have any ligated DNA. A PacI site is present between the DpnII site and the 3’ end of the top-strand of the bridge (Figure 1B). The sequencing adaptor will be removed from any molecule containing a bridge that was not cut by DpnII, and thus these incomplete molecules will fail to amplify during library preparation PCR.

16The on-bead amplification at step 3.20. is limited to a low number of PCR cycles (5-7×) to ensure that the library is not overamplified. It is key to avoid overamplification, which can lead to sequence representation bias in the final library or a high PCR duplication rate. However, the library concentration after the on-bead amplification is usually too low for direct sequencing. Additional rounds of amplification are necessary, which are done off-beads in step 3.23. To estimate how many additional PCR cycles can be done without overamplifying the library, a “side qPCR” is performed at step 3.22. The side qPCR provides a direct measurement of the PCR efficiency vs cycle number, and allows one to determine the range of PCR cycles where the amplification remains exponential (Figure 3).

17After on-bead PCR amplification, the amplified library is in the supernatant. However, the beads still contain the single-stranded ChAR-seq library with sequencing adapters. It is thus possible to re-amplify the library off the beads, by repeating the protocol starting from step 3.20. This can be useful if any of the downstream step fails, or if more of the same library needs to be prepared. Therefore, after collecting the supernatant, the beads should be resuspended in TE and saved at 4°C for potential library rescue.

18Do not use TE buffer here, as EDTA may interfere with the subsequent PCR.

19It is not uncommon that the library concentration after 5-7x cycles at step 3.20. is still below the sensitivity range of the bioanalyzer. Therefore, we typically do not bioanalyze the library at this stage, but we keep a small aliquot of the sample for troubleshooting.

20Each sample may have a different number of required cycles. In that case, set up the PCR program to the largest number of cycles, and take individual samples out of the PCR machine after the appropriate number of cycles. For example, for the three samples in Figure 3, set up an 11 cycles PCR and remove samples 1 and 2 at the end of the extension step of cycle 4 and 9, respectively.

21At both size selection steps 3.24 and 3.25, it is important to measure the exact volume of the sample or supernatant, so that the ratio of SPRI beads slurry to sample is accurate. The volumes may slightly vary between samples. The ratio of SPRI beads slurry to sample determines the molecular weight cutoff for DNA retention on the beads. Here, for the high and low molecular weight removal steps (steps 3.24 and 3.25), we use a bead slurry to sample ratio of rhigh=0.6 and rlow=0.9, respectively. These ratios should select molecules ranging from ~200bp to ~500bp. One may adjust these rations to adjust the fragment size distribution if necessary. For example, a ratio rhigh of 0.7 can be used to further trim out the population of molecules with high molecular weight. The volume of SPRI beads Vhigh to add at step 3.24 is equal to

where Vsample is the volume of the sample after PCR.

The volume of SPRI beads Vlow to add at step 3.25 can be calculated based on the supernatant volume Vsup with the following formula:

22If the bioanalyzer reveals the presence of primer dimers or if the fragment length distribution has a significant tail with molecules larger than ~500 bp, it is necessary to repeat the final two size selections (steps 3.23-3.24). Adjust the ratios rlow and rhigh according to note 21. When the library is sequenced with 2×150 bp reads, molecules (adapter + insert) larger than ~425 bp are undesired. Indeed, the bridge sequence may be located in the part of the insert in between the two sequenced ends. This will result in a lower proportion of useable reads, where the cDNA and DNA parts can be unambiguously identified. Alternatively, if the molecules (including adapters) are too short (80% of the reads.