henhaobio
Sign in

bioRxivGenomicsPreprint

KAS-CUT&Tag for direct mapping of transcription bubbles

Weifang WuiD, Jacob E. GreeneiD, Kami AhmadiD, Steven HenikoffiD

PDFDOICode

Abstract

Transcription by RNA Polymerase II (Pol II) generates dynamic transcription bubbles as it moves forward, but existing methods map transcriptional activity indirectly. Here, we introduce KAS-CUT&Tag, a method that combines N3-kethoxal labeling of exposed single-stranded DNA with CUT&Tag to directly map transcription bubbles. Bubble density correlates with nascent transcription levels and varies across Pol II-bound genes at promoters, gene bodies, and termination regions. Transcription bubbles are enriched at genes marked by active histone modifications H3K27ac, H3K4me3, and H3K36me3 and by chromatin-bound splicing factor U2AF2, but enrichment is highest at replication-coupled histone genes in K562 and GM12878 cells. KAS-CUT&Tag also detects co-occupancy of the histone gene transcription factor NPAT with Pol II at the Histone Locus Body, suggesting that in these cells NPAT is juxtaposed to transcription bubbles via interaction with Pol II. Together, KAS-CUT&Tag enables direct mapping of Pol II and regulatory interactions at transcription bubbles, providing a powerful tool for precise analysis of transcriptional activity.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (9 Oct 2026) · CC BY 4.0

N3-kethoxal labeling of single-stranded DNA within CUT&Tag maps Pol II transcription bubbles; bubble density tracks nascent transcription, shifts with transcription inhibitors and TFIIS loss, and peaks at histone genes.

RNA-based assays infer Pol II activity indirectly, KAS-seq single-stranded DNA signal can also arise from DNA damage, R-loops or replication, and KAS-ATAC cannot attribute bubbles to a specific protein. Tying kethoxal capture to antibody-targeted CUT&Tag links the bubble signal to Pol II or a co-bound factor such as NPAT, and the work reports a kethoxal-induced Tn5 blind spot near TSSs not previously recognized in KAS-ATAC.

Key results

  1. ShownBubble density (the ratio of IP to total) varies widely among active promoters that have similar Pol II occupancy, and it tracks nascent transcription measured by PRO-seq. It did not follow R-loop signal at active TSSs.

    Of 3,859 active TSSs in K562 (the top 20% of 19,294 by Pol II occupancy), 674 (17%) were bubble-rich (log2 > 1) and 194 (5%) bubble-poor (log2 < -1). PRO-seq medians: 6.70 rich, 5.49 moderate, 3.55 poor, 0.982 inactive (all pairwise BH-adjusted p ≤ 3.10 × 10^-11). R-loops from three HEK293T methods were more frequent at active TSSs but did not correlate with bubble density. · KAS-CUT&Tag maps transcription bubbles; Figure legends

  2. ShownThe bubbles come from engaged Pol II. Blocking initiation lowers bubble density at TSSs, while blocking pause release or elongation raises it. TFIIS knockout lowers it at TSSs, gene bodies and post-TES regions without matching losses of occupancy.

    K562, 1 h: triptolide reduced Pol II occupancy and bubble density at TSSs; flavopiridol and actinomycin D increased both (two-sided KS, Holm-adjusted P ≤ 1.039 × 10^-224 for bubble density). HEK293T TFIIS KO: TSS bubble density P = 1.35 × 10^-219 despite higher occupancy; gene-body Pol II-Ser2p density p < 9.99 × 10^-30 with occupancy p = 0.062; post-TES density P = 3.5 × 10^-51. · KAS-CUT&Tag maps transcription bubbles; Transcription bubble density varies across gene…; Human cell culture; Figure legends

  3. ShownReplication-coupled (RC) histone genes have the highest bubble density in K562 and GM12878. In K562, KAS-CUT&Tag and CUT&RUN.ChIP show NPAT on these genes together with Pol II. Colocalized NPAT and Pol II-Ser5p foci appear throughout interphase in K562 but mainly in S phase in GM12878.

    In K562, 51/57 active RC histone genes ranked in the top 10% of 3,859 active genes by bubble density (47/57 by PRO-seq). CUT&RUN.ChIP median ratios to Pol II-Ser5p input: NPAT 2.59 vs IgG 2.03 at histone TSSs, 1.15 vs 1.00 at other active TSSs (RPB3 6.48, 1.63). Cells with colocalized foci: GM12878 about 90% in S phase vs 21% in gap phase; K562 75% vs 70%. · Transcription factor NPAT is juxtaposed to…; Figure legends

  4. ShownGene bodies with dense Pol II-Ser2p bubbles carry more H3K36me3 and more chromatin-bound U2AF2. The bubble-rich, U2AF2-high genes include RNA-splicing genes.

    K562, 3,754 active genes: H3K36me3 medians 4.66 (rich, n = 342), 2.47 (moderate, n = 3,349), 0.20 (poor, n = 63), BH p ≤ 2.77 × 10^-17. High U2AF2 in 63% (217/342) of rich, 26% (872/3,349) of moderate and none of the poor gene bodies; GO linked 18% of the 217 to RNA splicing. In HEK293, U2AF2 was about 10-fold (rich) and 4-fold (moderate) above poor. · Transcription bubbles are enriched at genes…; Figure legends

  5. ShownN3-kethoxal blocks Tn5 insertion into labeled single-stranded DNA just downstream of TSSs. This depletes short fragments in kethoxal-treated CUT&Tag and in published KAS-ATAC data, while Pol II-Ser2p profiles over gene bodies are largely unaffected.

    Kethoxal-treated Pol II-Ser5p CUT&Tag lost signal from the TSS to about +500 bp, where K562 KAS-seq signal is enriched; V-plots over 12,397 protein-coding TSSs showed a loss of subnucleosomal fragments. Kethoxal-treated HEK293T ATAC-seq from two laboratories showed the same loss, plus depleted Tn5 insertion sites downstream of TSSs. IP signal did not track GC content. · KAS-CUT&Tag detects single-stranded DNA bound by…; Transcription bubble density varies across gene…; Data analysis and visualization; Figure legends

Methods, data and reproducibility

Cells on concanavalin A beads are treated with N3-kethoxal, tagmented by antibody-tethered pAG-Tn5, and the fragments biotinylated by click chemistry. Bubble density is the ratio of a streptavidin-enriched IP library to a total library (reported as log2(IP/total)). Pol II, Pol II-Ser2p and NPAT profiles were compared with public PRO-seq and probed with inhibitors, TFIIS knockout and FACS cell-cycle sorting.

  • K562 cells: KAS-CUT&Tag (Pol II, Pol II-Ser2p, Pol II-Ser5p, NPAT), 1 h inhibitor treatments, FACS-sorted G1/S/S-G2 fractions, CUT&RUN, CUT&RUN.ChIP and NPAT/Pol II-Ser5p/EdU immunofluorescence · human · n = 30,000 cells per KAS-CUT&Tag reaction; at least two biological replicates · GSE317877
  • HEK293T wild-type and TFIIS knockout cells: Pol II and Pol II-Ser2p KAS-CUT&Tag · human · GSE317877
  • GM12878 cells (Coriell): Pol II KAS-CUT&Tag and NPAT/Pol II-Ser5p/EdU immunofluorescence · human · n = at least 300 cells analyzed by imaging (figure covers GM12878 and K562) · GSE317877
  • Public nascent-transcription data: K562 PRO-seq, HEK293T PRO-seq, GM12878 GRO-seq · human · GSM1480327; GSM6735112-GSM6735115; GSM1480326
  • Public K562 chromatin profiles: H3K36me3, H3K9me3 and H3K27ac CUT&Tag, U2AF2 CUT&RUN, KAS-seq · human · GSM8343606; GSM8343607; GSM4835704; GSM4842201; GSM3536514; GSM8339598-GSM8339602; GSM6219605
  • Public HEK293T R-loop maps: MapR, R-loop CUT&RUN, R-loop CUT&Tag · human · GSM3406747; GSM3406749; GSM3406753; GSM3406755; GSM3406759; GSM4730182; GSM4730198
  • Public HEK293T ATAC-seq, standard and N3-kethoxal-treated (KAS-ATAC total libraries) · human · SRR10319911; SRR10319912; SRR28028139; SRR28028140; SRR28764289
  • K562 rRNA-depleted total RNA-seq from ENCODE · human · ENCFF782PCD
Methods & tools
KAS-CUT&Tag, CUT&Tag, CUT&RUN, CUT&RUN.ChIP, N3-kethoxal, pAG-Tn5, Cutadapt 4.4, Bowtie2 2.5.1, MACS2, deepTools 3.5.1, BEDTools 2.30.0, Imaris 11.0.1
Software & environment
Cutadapt 4.4; Bowtie2 2.5.1 against UCSC hg19; BEDTools 2.30.0; deepTools 3.5.1; MACS2 and SAMtools (versions not given); Imaris 11.0.1. No container or requirements file stated.
Reproducibility gaps
Code links are protocols.io protocols (wet-lab KAS-CUT&Tag; CUT&Tag read processing); analysis scripts are not shared. All primary sequence data, interpreted tracks and the CUT&RUN.ChIP step-by-step protocol are at GEO GSE317877, given as an accession, not a URL.
Limitations (4)
  • N3-kethoxal labeling stops Tn5 from inserting into labeled single-stranded DNA between the TSS and the +1 nucleosome. This technical blind spot removes short fragments in both KAS-CUT&Tag and KAS-ATAC. · stated by the authors
  • Two ideas are presented as inference or one possible explanation, not tested directly: that contacts between disordered regions hold NPAT next to Pol II, and that persistent Pol II-NPAT association at histone genes speeds cell-cycle progression in cancer cells. · stated by the authors
  • Promoter bubble density is IP over a total library, and the paper shows kethoxal depletes that total library from the TSS to about +500 bp. How this depletion affects promoter ratios is not separately quantified. · noted in this summary
  • In the non-histone NPAT analysis, the control peaks described as outside HLBs cluster on chromosome 6p themselves (17 of 33; 13 within 0.5 Mb of HIST1), so they are not a clean distal control. The Discussion also cites a Fig. 7e that the legends do not contain. · noted in this summary

Read this if you study Pol II pausing, elongation or transcription-factor contacts at promoters and want an antibody-targeted, low-input readout of transcription bubbles, or if you use KAS-ATAC and need to know about its TSS blind spot.

Citations
1

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Similarity Map
SelectedSame fieldCross-field
THISOhishi '22Henikoff '20Janssens '21Miura '20Wang '20Chen '19
56% semantic matchMolecular Biology
Hiroaki Ohishi, Seiru Shimada, Satoshi Uchino+9 more·Nature Communications·Dec 20, 2022· 40 citations