Preparation of “stress-free” concanavalin A-conjugated Dynabeads® magnetic beads for CUT&Tag

Epigenome research has employed various methods to identify genomic location of proteins of interest, such as transcription factors and histone modifications. A recently established method called CUT&Tag uses a Protein-A Tn5 transposase fusion protein, which cuts the genome and inserts adapter sequences nearby the target protein. Throughout most of the CUT&Tag procedure, cells are held on concanavalin A (con A)-conjugated magnetic beads. Proper holding of cells would be decisive for the accessibility of Tn5 to the chromatin, and efficacy of the procedure of washing cells. However, BioMag®Plus ConA magnetic beads, used in the original CUT&Tag protocol, often exhibit poor suspendability and severe aggregation. Here, we compared the BioMag beads and Dynabeads® magnetic particles of which conjugation of con A was done by our hands, and examined the performance of these magnetic beads in CUT&Tag. Among tested, one of the Dynabeads, MyOne-T1, kept excessive suspendability in a buffer even after overnight incubation. Furthermore, the MyOne-T1 beads notably improved the sensitivity in CUT&Tag assay for H3K4me3. In conclusion, the arrangement and the selection of MyOne-T1 refine the suspendability of beads, which improves the association of chromatin with Tn5, which enhances the sensitivity in CUT&Tag assay.

remarkably reduced number of cells, while the sequence profiling resolution is remarkably increased. In both methods, cells are bound to concanavalin A (con A)-coated magnetic beads and are handled throughout the process until the DNA extraction step. To date, options for commercially available con A-coated magnetic beads were limited at this point.
Further, the con A-coated magnetic beads, BioMag®Plus ConA magnetic beads (here after BioMag), used in the original protocols are difficult to handle because of its poor suspendability and severe aggregation. One might argue that efficiency of enzyme reaction in either protocol would be suffered.
To solve this issue, we evaluated different kinds of Dynabeads® magnetic beads (hereafter Dynabeads) that vary in size, cell binding capacity, or water binding capacity in order to verify whether the Dynabeads can be an alternative choice for the conventional con A-coated magnetic beads.

Preparation of Con A-coated beads
Four different streptavidin-conjugated Dynabeads, M-270, M-280, MyOne C1, and MyOne T1, that are capable of binding to biotin-conjugated concanavalin A (con A) were purchased from Thermo Fisher (Table 1). To conjugate con A, 100 μL of each beads is washed with 1× PBS (pH 6.8) for three times and resuspended in 100 μL of 1× PBS (pH 6.8) containing 0.01% Tween-20. Fifty μL of biotin-conjugated con A solution (2.3 mg/mL, Sigma Aldrich, C2272) in 1× PBS (pH 6.8) containing 0.01% Tween-20 is added to the beads, and rotated for 30 min at room temperature (RT). The beads are briefly spun down, and the supernatant is removed and used for measuring unbound con A in a "con A binding assay".
The remaining beads are resuspended in 100 μL of 1x PBS (pH 6.8) containing 0.01% Tween-20, and then used for "cell-binding assay" or the CUT&Tag procedure. BioMag from

Con A binding assay
Nanodrop One, a micro-UV/Vis spectrophotometer (Thermo Fisher) is used to measure the unbound con A in the supernatant. An absorbance based on the Protein A280 is measured with three technical replicates ( Figure 1A).

Cell-binding assay
To obtain cultured cells, human lymphoma cell line Daudi (JCRB9071) and Large T- is added and incubated at 4 °C overnight. The fusion protein is eluted with total 17.5 mL of elution buffer (HEGX buffer containing 100 mM DTT), and dialyzed with 500 mL of Dialysis buffer (100 mM HEPES-KOH (pH 7.2), 100 mM NaCl, 0.2 mM EDTA, 2 mM 2-ME, 0.2% Triton X-100, 20% Glycerol). the fusion protein is further purified in IEX-A buffer by using HiTrap SP1 and ACTA start (GE Healthcare Life Sciences). The Protein A-Tn5 fusion protein is mixed with pre-annealed ME-A and ME-B oligonucleotides. Ten μL of either con A-

Con A binding assay
To quantify the binding capacities of Dynabeads to con A, we first measured the concentration of unbound con A in the supernatant after con A-Dynabeads coupling reaction ( Figure 1A). Amount of absorbed (i.e. beads-bound) con A was calculated by subtracting the concentration of unbound con A from the original concentration (2.3 µg/µL). Among the four tested Dynabeads, M-280 was less capable of binding to con A, while other 3 types of Dynabeads (M-270, MyOne C1, and My One T1) exhibited comparable binding capacity in a range of 0.2-0.3 µg biotin-conjugated recombinant con A/µL beads solution (Figure 1B).

Cell-binding assay
We next compared cell binding ability of the Dynabeads to three different cell types (Daudi, HEK293T, and mouse testicular cells). After removing con A-bound cells, the number of con A-unbound cells in the supernatant was counted ( Figure 1C). We found that M-280 and M-270, which were bigger in diameter compared with those of MyOne C1 and T1 (2.8 µm vs. 1.05 µm), demonstrated relatively low capability to capture cells, especially for Daudi, although ~90% of cells were captured in all cell types ( Figure 1D). Notably, cell-binding ability of MyOne T1 and C1 beads appeared to be equivalent or even better than that of BioMag beads (1.05 µm) used in the original protocol ( Figure 1D). Particularly, MyOne T1 exhibited the best performance in all cell types, indicating the usefulness of these smallsized Dynabeads for efficient and stable cell capture.

CUT&Tag and data analyses
To further test the practicality of MyOne T1 in CUT&Tag procedure, we next performed CUT&Tag in Daudi cells and compared the usability of BioMag and MyOne T1 beads. Previously, we have experienced the inconvenience of using BioMag beads during the CUT&Tag procedure in the following two points; one is the difficulty of uniform suspension of BioMag in the solution as it is easily aggregated after attaching cells to the beads (Figure 2A). Another is serious attachment to the wall of an Eppendorf tube. These issues could cause an inefficient cell capture. The result clearly demonstrated that MyOne T1 exhibited more efficient and uniform suspendability compared with BioMag, and there were little residual beads attached to the wall of a tube (Figure 2B). The beads-bound cells were then mixed with antibody against H3K4me3 or control IgG, and incubated for overnight at 4 °C. After incubation, cell-bound MyOne T1 were uniformly dissolved in the buffer, while cell-bound BioMag were severely aggregated ( Figure 2C). Subsequently, size distribution and concentration of the library were analyzed by capillary electrophoresis. The original CUT&Tag using BioMag reported nucleosomal ladders when anti-H3K4me3 antibody is used (Kaya-Okur et al., 2019). In agreement with the original study, nucleosomal ladders were evidently observed in BioMag, Notably, MyOne T1 showed reduced relative amount of DNA in higher (> 1000 bp) molecular weight, indicating the better performance of Tn-5 in the MyOne T1 samples (Figure 2E).
We next performed paired-end 150 bp-sequencing of the libraries that were made with either BioMag or MyOne T1 beads. Genome-wide distribution of the peaks were identified by comparing with the IgG control samples. Overall, the peak distribution pattern was similar between BioMag and MyOne T1 beads in a chromosome-wide level (Figure 3A).
In a 1-2 Mb window, H3K4me3 was observed as either narrow peaks (~100 bp, Figure 3B) or broad peaks (~400 bp, Figure 3C) depending on the genome loci, and in both cases, the peak distribution and pattern were consistent between BioMag and MyOne T1 beads.
Further bioinformatic analyses demonstrated that 98.46 % of narrow peaks and 99.26 % of broad peaks from BioMag were overlapped with those from MyOne T1 beads ( Figure 4A). In addition, peak distribution around genes was also similar between BioMag and MyOne T1 as the majority of H3K4me3 peaks were mainly distributed at upstream and genic regions rather than intergenic regions in both samples with a significant enrichment at transcription start sites (TSS) (Figure 4B, 4C). Furthermore, MyOne T1 showed notable numbers of MyOne T1-specific narrow and broad peaks (2723 and 1727 peaks, respectively) (Figure 4A), and these peaks contained relatively less upstream and more intergenic peaks as compared with the total MyOne T1 peaks (Figure 4B). Heatmaps and line plots of H3K4me3 peaks (both narrow and broad) from BioMag or MyOne-T1 showed similar pattern (Figure 4D). MA plot analyses further indicated that 4.4-10.9 % of the peaks were uniquely detected in MyOne T1 although their read counts exhibited relatively lower values (Figure 4E).

DISCUSSION
The present study showed that the use of Dynabeads, especially MyOne T1, can be better alternatives to BioMag used in an original CUT&Tag protocol in terms of easy handling, and equivalent data quality (Kaya-Okur et al., 2020; Kaya-Okur et al., 2019). Unpreferable aggregation and absorption to the wall of tubes, that were frequently observed when BioMag was used, can be minimized by using MyOne T1, potentially preventing the loss of cells and increasing the efficiency of washing process. This advantage is probably because of the hydrophobic surface of MyOne T1. Interestingly, we also found that MyOne T1 or C1 magnetic beads, that were smaller in size compared with Dynabeads M-280 or M-270, showed slightly higher ability to capture three kinds of tested cells, suggesting that smaller in size is more capable beads due to the increased surface area of beads when the same volume of beads were used.
In addition, our subsequent CUT&Tag using the MyOne T1 followed by sequencing observed remarkable similarity in H3K4me3 peaks between MyOne T1 and BioMag (Figure   3 and 4). Importantly, 4.4-10.9% of the H3K4me3 peaks observed using MyOne T1 showed higher enrichment at detected peak regions compared with those of BioMag, while almost no peaks showed lower enrichment (Figure 4E). Also considering the increased number of H3K4me3 peaks in MyOne T1 samples (Figure 4A), using MyOne T1 may increase the sensitivity of the CUT&Tag. This may be because the improved suspendability by MyOne T1 enhanced accessibility of IgG to the nucleus. On the other hand, MyOne T1 slightly increases the H3K4me3 peaks in intergenic regions (Figure 4B) implying the possibility of non-specific binding as well. Therefore to verify the specificity of MyOne T1 beads, further examination is needed.
Nevertheless, we propose that the home-made con A-conjugated Dynabeads magnetic beads (i. e. MyOne T1) can be a "stress-free" alternative choice for CUT&RUN and CUT&Tag methods. It may need to test other types of magnetic beads to find the best ones for the cell types used in your experiments.

Conflict of Interest
This study was supported in part by VERITAS Corporation (Tokyo, Japan). Y.T. is an employee of VERITAS.