Starting CUT&RUN for a new target? Don’t forget these key optimization steps
CUT&RUN is an ultra-sensitive chromatin mapping technology that is gaining widespread popularity in the epigenomics field. This streamlined approach is compatible with reduced cell inputs and lower sequencing depths compared to ChIP-seq, while simultaneously generating data with remarkable signal over background.
EpiCypher scientists have optimized CUT&RUN protocols for various targets, including histone post-translational modifications (PTMs) and transiently interacting chromatin-associated proteins such as transcription factors, chromatin reader proteins, and chromatin remodeling enzymes. We have adopted this approach for a variety of inputs, including varying numbers of native (non-fixed), crosslinked (fixed), and frozen cells/nuclei. Our most up-to-date CUT&RUN protocol is compatible with targets from each of the above categories and can generate robust data using as few as 5,000 cells/nuclei and 3 million sequencing reads.
As a result of these efforts, our researchers have gained valuable knowledge, and defined a set of best practices for developing CUT&RUN assays to new histone PTMs or chromatin associated proteins. Surprisingly, we have found that moderate crosslinking of cells/nuclei is advantageous for certain targets. Read on to learn more!
When to consider crosslinking for CUT&RUN? Depends on your target
It is important to note that the majority of targets work well under native conditions, so this method is preferred whenever possible to avoid potential negative effects from fixation. However, there are situations where crosslinking is necessary to improve CUT&RUN signal. Below we outline a few examples of targets that may particularly benefit from moderate crosslinking (Table 1):
CELL/NUCLEI PREPARATION
METHODS
WORKS WELL FOR
Native
No crosslinking
Majority of targets in CUT&RUN
Moderate crosslinking
0.1 - 1% formaldehyde, 1 min
Select targets in CUT&RUN, such as:
1) Labile histone PTMs, e.g. lysine acetylation
2) Readers of labile PTMs, e.g. bromodomain-containing reader proteins & complexes
3) Transiently interacting proteins, e.g. chromatin remodelers
Heavy crosslinking
1% formaldehyde, 10 min
Standard ChIP conditions; NOT recommended for CUT&RUN
However, it is nearly impossible to predict which targets will require crosslinking. Thus, for new or understudied targets, we recommend testing different strategies.
Developing CUT&RUN for a new target? Our advice
For researchers investigating a new target in CUT&RUN assays, EpiCypher scientists recommend a few key steps:
Source three to five antibodies to the protein / histone PTM of interest. NOTE: Claims of “ChIP-grade” do not guarantee success in CUT&RUN!
Test all of the antibodies in CUT&RUN, comparing cells (or nuclei) prepared under both native and moderately crosslinked conditions. Although CUT&RUN was originally developed using native cells, we (and others1,2) have developed alternate methods for crosslinked conditions as well.
EpiCypher scientists compare the results from this preliminary test by heatmap analysis of enrichment at annotated gene features and by comparing signal at individual loci (e.g. known target genes).
Always check for CUTANA Compatible Antibodies to the target of interest, since these have been directly confirmed to work in CUT&RUN, including optimized native or crosslinking protocol recommendations!
CUT&RUN to histone PTMs
As detailed in Table 1, moderate crosslinking can improve recovery for histone lysine acetylation PTMs. In Figure 1, we tested our highly specific H3K27ac antibody (EpiCypher 13-0045) in CUT&RUN assays, using native cells and cells treated with varying concentrations of formaldehyde for 1 to 10 minutes. Genome-wide heatmap analysis reveals improved signal over background with 0.1% and 1% formaldehyde treatment for 1 minute compared to no crosslinking. This is partly explained by the high lability of many acetyl marks, as acetylation is known to be dynamically regulated and display increased turnover relative to other classes of PTMs (e.g. methylation)3.
Interestingly, these results also show that over-fixation in CUT&RUN can reduce signal and recovery. Treatment with 1% formaldehyde for 10 minutes reduced yield and diminished signal over background compared to all other conditions tested, including non-fixed cells. These results are notable, as most ChIP-seq protocols recommend 1% formaldehyde crosslinking for 10 minutes. Thus, for acetylation PTMs, we highly recommend testing samples with moderate crosslinking protocols to investigate potential benefits for your assay.
Figure 1: CUT&RUN for H3K27ac is improved by moderate fixation. We used native cells (left column), and cells treated with 0.1-1% formaldehyde for 1 min-10 min (three right columns). 500K K562 cells were used for each CUT&RUN reaction with H3K27ac antibody (EpiCypher 13-0045). Heatmaps display CUT&RUN signal aligned to the transcription start site (TSS, +/- 2kb) for 18,793 genes. High and low signal are ranked by intensity (top to bottom) and reflected by red and blue colors, respectively. Gene rows in each heatmap are sorted from high to low signal relative to native cells.
Notably, while some PTMs are altered by fixation, others are unaffected. We used our robust CUTANA-compatible H3K4me3 antibody (EpiCypher 13-0041) to perform CUT&RUN using K562 cells subjected to diverse preparation methods. The results were identical across all samples (Figure 2), illustrating the stability of histone methylation3 and the reliability of our CUT&RUN protocol. In fact, the consistency of these data are the reason we suggest running H3K4me3 CUT&RUN as a positive control reaction in user experiments, and an aliquot of H3K4me3 antibody is included in our CUTANA ChIC/CUT&RUN Kit.
Figure 2: CUT&RUN for H3K4me3 is consistent across varying assay conditions. We used 500K K562 cells per CUT&RUN reaction. Cells / Nuclei were prepared as indicated (native, cryopreserved, crosslinked using the indicated concentrations of formaldehyde). Heatmaps display CUT&RUN signal aligned to the transcription start site (TSS, +/- 2kb) for 18,793 genes. High and low signal are ranked by intensity (top to bottom) and reflected by red and blue colors, respectively. Gene rows in each heatmap are sorted from high to low signal relative to Fresh Cells.
CUT&RUN to map chromatin-associated proteins
Importantly, CUT&RUN for non-histone PTM targets can also benefit from moderate fixation. EpiCypher is currently in the process of rigorously testing and releasing CUTANA Compatible Antibodies to a variety of chromatin-associated proteins, including chromatin remodelers and transcription factors. Our recently released antibodies to chromatin remodeling proteins BRM (SMARCA2) and SNF2L (SMARCA1) both display enhanced recovery and peak amplitude when using crosslinked vs. native samples.
Figure 3 shows representative CUT&RUN data for our BRM antibody, including heatmaps and loci-specific peak analysis at transcription start sites (TSSs). H3K4me3 data are shown for comparison, as chromatin remodelers typically function to rearrange nucleosomes at TSSs and are often correlated with H3K4me3 levels4. Fixation with 0.1% formaldehyde for 1 minute results in vastly improved signal : noise at TSSs, more closely replicating H3K4me3 peak patterns (Figure 3A) vs. non-fixed cells. These results are even more apparent when looking at BRM target loci, such as the transcription factor POU2F1 and microtubule-associated protein MAP4. In Figure 3B, genomic mapping data for native vs. crosslinked CUT&RUN reactions at these loci are outlined using red boxes. The native reactions display peaks barely discernable over background. In contrast, CUT&RUN data from moderately crosslinked cells exhibit increased peak amplitude, while maintaining extremely low backgrounds characteristic of CUT&RUN.
Figure 3: CUT&RUN for BRM/SMARCA2 is improved by moderate fixation. For native IgG, H3K4me3, and BRM CUT&RUN reactions, we used 500K K562 cells. For crosslinked BRM CUT&RUN, we used 500K K562 cells treated with 0.1% formaldehyde for 1 min. (A) Heatmaps display CUT&RUN signal aligned to the transcription start site (TSS, +/- 2kb) for 18,793 genes. High and low signal are ranked by intensity (top to bottom) and reflected by red and blue colors, respectively. Gene rows in each heatmap are sorted from high to low signal relative to BRM profiling in crosslinked cells. (B) Loci-specific analysis of CUT&RUN data shown in (A). Images were generated using the Integrative Genomics Viewer (IGV, Broad Institute).
It is not clear how or why fixation augments recovery for some of these chromatin-associated proteins. In the case of BRM, it may be due to the presence of a C-terminal bromodomain in the BRM protein. Bromodomains interact with acetylated histones; thus, fixation may stabilize lysine acetylation – BRM interactions, enabling improved detection by CUT&RUN.
The overarching summary of EpiCypher’s extensive research efforts is that, before embarking on a new CUT&RUN project, you should test a variety of conditions (antibodies, preparation methods) to find the optimal methods for your target of interest. In past blog posts we have focused on antibody specificity, which is paramount; but defining the best methods that generate the highest quality data for your target is also essential. This will accelerate your future studies by developing a robust CUT&RUN workflow, thereby reducing frustrating troubleshooting steps and increasing confidence in your data.
