Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Regulation of RNA polymerase II activation by histone acetylation in single living cells

Abstract

In eukaryotic cells, post-translational histone modifications have an important role in gene regulation. Starting with early work on histone acetylation1, a variety of residue-specific modifications have now been linked to RNA polymerase II (RNAP2) activity2,3, but it remains unclear if these markers are active regulators of transcription or just passive byproducts4,5. This is because studies have traditionally relied on fixed cell populations, meaning temporal resolution is limited to minutes at best, and correlated factors may not actually be present in the same cell at the same time. Complementary approaches are therefore needed to probe the dynamic interplay of histone modifications and RNAP2 with higher temporal resolution in single living cells2,5,6. Here we address this problem by developing a system to track residue-specific histone modifications and RNAP2 phosphorylation in living cells by fluorescence microscopy. This increases temporal resolution to the tens-of-seconds range. Our single-cell analysis reveals histone H3 lysine-27 acetylation at a gene locus can alter downstream transcription kinetics by as much as 50%, affecting two temporally separate events. First acetylation enhances the search kinetics of transcriptional activators, and later the acetylation accelerates the transition of RNAP2 from initiation to elongation. Signatures of the latter can be found genome-wide using chromatin immunoprecipitation followed by sequencing. We argue that this regulation leads to a robust and potentially tunable transcriptional response.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Figure 1: Covalent modifications to histones and RNAP2 can be imaged in living cells using FabLEM.
Figure 2: Fitting the RNAP2 transcription cycle.
Figure 3: Effect of perturbing array histone acetylation on RNAP2 transcription activation.
Figure 4: Sequencing to examine genome-wide acetylation and transcription.

Accession codes

Data deposits

Sequencing data was submitted to the DDBJ sequence read archive under accession number DRA000936.

References

  1. 1

    Struhl, K. Histone acetylation and transcriptional regulatory mechanisms. Genes Dev. 12, 599–606 (1998)

    CAS  PubMed  PubMed Central  Google Scholar 

  2. 2

    Zhou, V. W., Goren, A. & Bernstein, B. E. Charting histone modifications and the functional organization of mammalian genomes. Nature Rev. Genet. 12, 7–18 (2011)

    PubMed  Google Scholar 

  3. 3

    Bannister, A. J. & Kouzarides, T. Regulation of chromatin by histone modifications. Cell Res. 21, 381–395 (2011)

    CAS  PubMed  PubMed Central  Google Scholar 

  4. 4

    Berger, S. L. The complex language of chromatin regulation during transcription. Nature 447, 407–412 (2007)

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  5. 5

    Henikoff, S. & Shilatifard, A. Histone modification: cause or cog? Trends Genet. 27, 389–396 (2011)

    CAS  PubMed  Google Scholar 

  6. 6

    Ecker, J. R. et al. Genomics: ENCODE explained. Nature 489, 52–55 (2012)

    ADS  CAS  PubMed  Google Scholar 

  7. 7

    McNally, J. G., Müller, W. G., Walker, D., Wolford, R. & Hager, G. L. The glucocorticoid receptor: rapid exchange with regulatory sites in living cells. Science 287, 1262–1265 (2000)

    ADS  CAS  PubMed  Google Scholar 

  8. 8

    Egloff, S., Dienstbier, M. & Murphy, S. Updating the RNA polymerase CTD code: adding gene-specific layers. Trends Genet. 28, 333–341 (2012)

    CAS  PubMed  Google Scholar 

  9. 9

    Crump, N. T. et al. Dynamic acetylation of all lysine-4 trimethylated histone H3 is evolutionarily conserved and mediated by p300/CBP. Proc. Natl Acad. Sci. USA 108, 7814–7819 (2011)

    ADS  CAS  PubMed  Google Scholar 

  10. 10

    Clayton, A. L., Hazzalin, C. A. & Mahadevan, L. C. Enhanced histone acetylation and transcription: a dynamic perspective. Mol. Cell 23, 289–296 (2006)

    CAS  PubMed  Google Scholar 

  11. 11

    Hayashi-Takanaka, Y. et al. Tracking epigenetic histone modifications in single cells using Fab-based live endogenous modification labeling. Nucleic Acids Res. 39, 6475–6488 (2011)

    CAS  PubMed  PubMed Central  Google Scholar 

  12. 12

    Hayashi-Takanaka, Y., Yamagata, K., Nozaki, N. & Kimura, H. Visualizing histone modifications in living cells: spatiotemporal dynamics of H3 phosphorylation during interphase. J. Cell Biol. 187, 781–790 (2009)

    CAS  PubMed  PubMed Central  Google Scholar 

  13. 13

    Smith, C. L. A shifting paradigm: histone deacetylases and transcriptional activation. Bioessays 30, 15–24 (2008)

    CAS  PubMed  Google Scholar 

  14. 14

    Chao, S.-H. & Price, D. H. Flavopiridol inactivates P-TEFb and blocks most RNA polymerase II transcription in vivo. J. Biol. Chem. 276, 31793–31799 (2001)

    CAS  PubMed  Google Scholar 

  15. 15

    Darzacq, X. et al. In vivo dynamics of RNA polymerase II transcription. Nature Struct. Mol. Biol. 14, 796–806 (2007)

    CAS  Google Scholar 

  16. 16

    Stavreva, D. A. et al. Ultradian hormone stimulation induces glucocorticoid receptor-mediated pulses of gene transcription. Nature Cell Biol. 11, 1093–1102 (2009)

    CAS  PubMed  Google Scholar 

  17. 17

    Yamane, K. et al. PLU-1 is an H3K4 demethylase involved in transcriptional repression and breast cancer cell proliferation. Mol. Cell 25, 801–812 (2007)

    CAS  PubMed  Google Scholar 

  18. 18

    Roesch, A. et al. A temporarily distinct subpopulation of slow-cycling melanoma cells is required for continuous tumor growth. Cell 141, 583–594 (2010)

    CAS  PubMed  PubMed Central  Google Scholar 

  19. 19

    Bowers, E. M. et al. Virtual ligand screening of the p300/CBP histone acetyltransferase: identification of a selective small molecule inhibitor. Chem. Biol. 17, 471–482 (2010)

    CAS  PubMed  PubMed Central  Google Scholar 

  20. 20

    Jin, Q. et al. Distinct roles of GCN5/PCAF-mediated H3K9ac and CBP/p300-mediated H3K18/27ac in nuclear receptor transactivation. EMBO J. 30, 249–262 (2011)

    CAS  PubMed  Google Scholar 

  21. 21

    Yang, X.-J., Grégoire, S. & Class, I. I. Histone deacetylases: from sequence to function, regulation, and clinical implication. Mol. Cell. Biol. 25, 2873–2884 (2005)

    CAS  PubMed  PubMed Central  Google Scholar 

  22. 22

    Lin, C. et al. AFF4, a component of the ELL/P-TEFb elongation complex and a shared subunit of MLL chimeras, can link transcription elongation to leukemia. Mol. Cell 37, 429–437 (2010)

    CAS  PubMed  PubMed Central  Google Scholar 

  23. 23

    Byun, J. S. et al. ELL facilitates RNA polymerase II pause site entry and release. Nature Commun. 3, 633 (2012)

    ADS  Google Scholar 

  24. 24

    Rafalska-Metcalf, I. U., Powers, S. L., Joo, L. M., LeRoy, G. & Janicki, S. M. Single cell analysis of transcriptional activation dynamics. PLoS ONE 5, e10272 (2010)

    ADS  PubMed  PubMed Central  Google Scholar 

  25. 25

    Zhao, R., Nakamura, T., Fu, Y., Lazar, Z. & Spector, D. L. Gene bookmarking accelerates the kinetics of post-mitotic transcriptional re-activation. Nature Cell Biol. 13, 1295–1304 (2011)

    CAS  PubMed  Google Scholar 

  26. 26

    Kimura, H., Hayashi-Takanaka, Y. & Yamagata, K. Visualization of DNA methylation and histone modifications in living cells. Curr. Opin. Cell Biol. 22, 412–418 (2010)

    CAS  PubMed  Google Scholar 

  27. 27

    Manders, E. M., Kimura, H. & Cook, P. R. Direct imaging of DNA in living cells reveals the dynamics of chromosome formation. J. Cell Biol. 144, 813–822 (1999)

    CAS  PubMed  PubMed Central  Google Scholar 

  28. 28

    McNeil, P. L. & Warder, E. Glass beads load macromolecules into living cells. J. Cell Sci. 88, 669–678 (1987)

    PubMed  Google Scholar 

  29. 29

    Sugaya, K., Vigneron, M. & Cook, P. R. Mammalian cell lines expressing functional RNA polymerase II tagged with the green fluorescent protein. J. Cell Sci. 113, 2679–2683 (2000)

    CAS  PubMed  Google Scholar 

  30. 30

    Gronostajski, R. M. Roles of the NFI/CTF gene family in transcription and development. Gene 249, 31–45 (2000)

    CAS  PubMed  Google Scholar 

  31. 31

    Mueller, F., Morisaki, T., Mazza, D. & McNally, J. G. Minimizing the impact of photoswitching of fluorescent proteins on FRAP analysis. Biophys. J. 102, 1656–1665 (2012)

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  32. 32

    Kanda, T., Sullivan, K. F. & Wahl, G. M. Histone-GFP fusion protein enables sensitive analysis of chromosome dynamics in living mammalian cells. Curr. Biol. 8, 377–385 (1998)

    CAS  PubMed  Google Scholar 

  33. 33

    Mueller, F., Wach, P. & McNally, J. G. Evidence for a common mode of transcription factor interaction with chromatin as revealed by improved quantitative fluorescence recovery after photobleaching. Biophys. J. 94, 3323–3339 (2008)

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  34. 34

    Stasevich, T. J., Mueller, F., Brown, D. T. & McNally, J. G. Dissecting the binding mechanism of the linker histone in live cells: an integrated FRAP analysis. EMBO J. 29, 1225–1234 (2010)

    CAS  PubMed  PubMed Central  Google Scholar 

  35. 35

    Kimura, H., Tao, Y., Roeder, R. G. & Cook, P. R. Quantitation of RNA polymerase II and its transcription factors in an HeLa cell: little soluble holoenzyme but significant amounts of polymerases attached to the nuclear substructure. Mol. Cell. Biol. 19, 5383–5392 (1999)

    CAS  PubMed  PubMed Central  Google Scholar 

  36. 36

    Morris, D. P., Michelotti, G. A. & Schwinn, D. A. Evidence that phosphorylation of the RNA polymerase II carboxyl-terminal repeats is similar in yeast and humans. J. Biol. Chem. 280, 31368–31377 (2005)

    CAS  PubMed  PubMed Central  Google Scholar 

  37. 37

    Li, M. et al. Solution structure of the Set2–Rpb1 interacting domain of human Set2 and its interaction with the hyperphosphorylated C-terminal domain of Rpb1. Proc. Natl Acad. Sci. USA 102, 17636–17641 (2005)

    ADS  CAS  PubMed  Google Scholar 

  38. 38

    Phatnani, H. P. & Greenleaf, A. L. Phosphorylation and functions of the RNA polymerase II CTD. Genes Dev. 20, 2922–2936 (2006)

    CAS  PubMed  Google Scholar 

  39. 39

    Odawara, J. et al. The classification of mRNA expression levels by the phosphorylation state of RNAPII CTD based on a combined genome-wide approach. BMC Genomics 12, 516 (2011)

    CAS  PubMed  PubMed Central  Google Scholar 

  40. 40

    Trapnell, C., Pachter, L. & Salzberg, S. L. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105–1111 (2009)

    CAS  PubMed  PubMed Central  Google Scholar 

  41. 41

    Trapnell, C. et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nature Biotechnol. 28, 511–515 (2010)

    CAS  Google Scholar 

  42. 42

    Kimura, H., Hayashi-Takanaka, Y., Goto, Y., Takizawa, N. & Nozaki, N. The organization of histone H3 modifications as revealed by a panel of specific monoclonal antibodies. Cell Struct. Funct. 33, 61–73 (2008)

    CAS  PubMed  Google Scholar 

  43. 43

    Jee, J. et al. ACT: aggregation and correlation toolbox for analyses of genome tracks. Bioinformatics 27, 1152–1154 (2011)

    CAS  PubMed  PubMed Central  Google Scholar 

  44. 44

    Core, L. J., Waterfall, J. J. & Lis, J. T. Nascent RNA sequencing reveals widespread pausing and divergent initiation at human promoters. Science 322, 1845–1848 (2008)

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

This work was supported by grants-in-aid from the Japan Society for the Promotion of Science (JSPS) and the Ministry of Education, Culture, Sports, Science and Technology of Japan. T.J.S. and Y.H.-T. were supported by JSPS fellowships. We thank T. Kanda, A. Kitamura and T. Morisaki for the mCh-H2B construct, D. Stavreva and G. Hager for the mCherry-NF1A1.1 construct, and F. Mueller and D. Mazza for comments on the manuscript.

Author information

Affiliations

Authors

Contributions

T.J.S. performed most experiments and data analysis. Y.H.-T. performed ChIP and immunofluorescence, Y.H.-T., Y.S. and H.K. assisted with antibody preparation and molecular biology work. K.M. and Y.O. performed sequencing. M.T. and K.S.-S. created the mCherry-RPB1 construct. T.N. created HaloTag-tag constructs. J.G.M. provided materials and assisted with data interpretation. N.N. made hybridoma cell lines. H.K. and T.J.S. conceived the study and wrote the manuscript.

Corresponding authors

Correspondence to Timothy J. Stasevich or Hiroshi Kimura.

Ethics declarations

Competing interests

N.N. is a founder of Mab Institute Inc.

Extended data figures and tables

Extended Data Figure 1 Immunostaining the array with antibodies against RNAP2 and histones.

a, Monoclonal antibodies against the RPB1 subunit of RNAP2 (CTD) and its Ser 5 and Ser 2 phosphorylated forms (Ser 5ph and Ser 2ph) were evaluated by ELISA using the indicated peptides with specific phosphorylation patterns. Microtitre plates coated with the peptides were incubated with each antibody. After incubation with peroxidase-conjugated secondary antibody and washing, the colorimetric signal of tetramethylbenzidine was detected by measuring the absorbance at 405 nm using a plate reader. b, Immunofluorescence with monoclonal antibodies against RNAP2 (tested in a), H3K27ac and H3K4me2 in cells (arrays marked by GR, yellow arrows) fixed pre- and post-transcriptional activation (times indicated). c, Although H3K4me2 levels at the array are consistently high, H3K27ac levels are sometimes relatively low (pink arrows), as quantified in the histogram. d, Summary of a screen of histone modifications and variants at the MMTV array by immunostaining. Scale bars, 5 μm.

Extended Data Figure 2 FRAP determines the timescale of FabLEM experiments.

a, FRAP experiments on cells loaded with fluorescent Fabs (red) against RNAP2 (CTD, Ser 5ph and Ser 2ph) can be used to estimate an upper bound on how long it takes Fabs to track their targets. The FRAP recovery time is limited by the dissociation of photobleached Fab (grey) from target protein modifications (black dots) plus the association time of an unbleached Fab to the open modification. It is the latter time that corresponds to the tracking time of Fab and the temporal resolution of FabLEM. b, Fab FRAP recovery curves (coloured curves on right; ± s.e.m.) are complete in about 10 s. A control Fab with no target in the nucleus is shown in green to see how fast recoveries are when Fab do not bind (fits to n = 10 curves indicate this recovery is purely diffusive, yielding a Fab diffusion coefficient ± s.e.m. of DFab = 20 ± 8 μm2 sec−1). c, Select frames from single-cell FRAP experiments. Yellow lightening indicates the position of the bleach. On the right, reaction-diffusion fits of the bleach spot profiles (radial distance from centre of bleach spot) at times shown on the left, with earlier times having deeper profiles. Fits (± s.e.m.) to n = 43 (CTD), n = 30 (Ser 5ph), and n = 32 (Ser 2ph) FRAP experiments on three independent days yield the average effective diffusion coefficient Deff, binding time toff, and bound fraction BF for each Fab. The total bound fraction (BFtot) is computed: 1 − (Deff/DFab)(1 − BF). All Fab have high total bound fractions (>80%), indicating good signal to noise. Scale bars, 10 μm; a.u., arbitrary units.

Extended Data Figure 3 Testing fidelity of FabLEM for detecting RNAP2 and histone modifications.

ae, Representative FabLEM experiments in single living cells that were either untreated (a, b) or treated with inhibitors (ce). Scale bars 5 μm. c, Cells were treated with 1 μM flavopiridol for 1 h and then activated with hormone (time roughly corresponds to post-activation). Flavopiridol inhibits P-TEFb, which phosphorylates the RNAP2 CTD at Ser 2, preventing elongation at the gene array (marked by yellow arrow). FabLEM experiments confirm this, showing no array decondensation and no accumulation of Ser 2ph Fab (red) at the array upon hormone treatment even though Ser 5ph (purple) does accumulate, indicating RNAP2 initiation (although to a lesser extent than in untreated cells). d, The same experiment as above, but now examining histone acetylation levels at the array (blue, H3K27ac), which no longer go down post-activation, as in untreated cells. e, Cells were treated with 100 nM of the histone deacetylase inhibitor trichostatin A (TSA) for 1 h. As with flavopiridol treatment, the array no longer decondenses, H3K27ac levels remain high (and in fact the total nuclear intensity is higher, indicating global increases in H3K27ac levels) and levels of RNAP2 initiation are low, indicating little or no RNAP2.

Extended Data Figure 4 Testing the confidence of FabLEM correlations.

a, GR and Ser 2ph data from Fig. 1c were randomly split into two groups in 5,000 unique ways and the area between the average curves from each group was computed. A histogram of the obtained areas is shown. This reveals that the H3K27ac-based split of data into high and low groups in Fig. 1c scores in the top 5% of all GR splits (green) and in the top 10% of all Ser 2ph splits (red). This provides an estimate for the confidence of the measured correlation between H3K27ac and GR or Ser 2ph. b, Scatter plots of single-cell data from Fig. 1c, d with the initial H3K27ac array/nuclear intensity (2 ± 2 min) on the x-axis and the maximal GR (12 ± 2 min, green), Ser 2ph (23 ± 2 min, red) and Ser 5ph (21 ± 2 min, purple) array/nuclear intensities on the y-axis. Each point represents data from a single cell averaged over a four-minute time window (for example, each Ser 2ph point represents the mean of data from a single cell between 21 and 25 min). A positive correlation (quantified by the Pearson correlation coefficient and its corresponding P value calculated using the built-in Mathematica function PearsonCorrelationTest; Wolfram Research) is seen between H3K27ac and GR, and H3K27ac and Ser 2ph, but not between H3K27ac and Ser 5ph. c, To test if the nuclear concentration of loaded Fab has no deleterious effect on transcription at the array and is not responsible for the H3K27ac-dependence of GR recruitment and RNAP2 elongation (Ser 2ph) shown in Fig. 1c, immunostaining against Ser 2ph (red) was performed on a population of induced cells (30 min) expressing GFP–GR (green) in which a subset were bead loaded with the H3K27ac-specific Fab (blue). d, The intensity of arrays in control unloaded cells (n = 24) was the same within error (± s.e.m.) as in bead-loaded cells (n = 24), indicating the H3K27ac Fab do not alter Ser 2ph levels at the array. e, The Fab nuclear intensities of high/low sorted cells based on array intensities from Fig. 1c are statistically indistinguishable (the smallest P value from the Student’s t-test, the z-test and the Mann–Whitney median test is reported using the built-in Mathematica function LocationTest; Wolfram Research). This demonstrates that differing concentrations of Fab in the nucleus are not responsible for the measured correlation between H3K27ac and Ser 2ph.

Extended Data Figure 5 FabLEM quantification.

a, Frames from sample single-cell experiments show how the gene array (yellow arrow) is first bound by GFP–GR (green) after hormone is added to cells, followed by Fabs marking RNAP2 (orange, CTD, recruitment), Ser 5 phosphorylated RNAP2 (blue, Ser 5ph, initiation), and Ser 2 phosphorylated RNAP2 (red, Ser 2ph, elongation). Scale bars, 5 μm. Quantification of the array/nuclear intensity over time is shown below after adjusting time scale so GR curves are aligned (see Extended Data Fig. 9a for alignment details). b, Average single-cell recruitment curves (n = 12, ± s.e.m.). Insets show a normalized rolling average to illustrate the temporal ordering: gene activation, RNAP2 recruitment, initiation, and elongation. The arrows indicate, from left to right, when levels of GR, RNAP2 and RNAP2 Ser 5ph and Ser 2ph go up/down at the array. c–f, Summary of workflow for quantifying FabLEM data. Image stacks were aligned in time (step 1), background subtracted (step 2), and intensities in the red, green, and blue channels were measured in regions of interest covering both the array (yellow polygon in upper screen shot) and a representative portion of the nucleus (yellow polygon in lower screen shot). From this raw intensity data, the mean array/nuclear intensity was calculated for the cell. This was repeated for other cells and data averaged by aligning green GFP–GR curves (step 4). The maximum array relative to nuclear Ser 2ph signal is SSer2ph ≈ 1.1. d, The average volume of the array Varr and nucleus Vnuc were calculated from image stacks of cells expressing GFP–GR twenty minutes after transcription activation by dexamethasone. Image stacks were smoothed with a median filter (to remove single voxel speckle noise) and binarized by making all voxels with intensities above a threshold value black and all voxels equal to or below the threshold value white. The volume of the nucleus and the array could then be estimated by counting the number of black voxels. e, FRAP experiments on cells loaded with fluorescent Ser 2ph Fab were performed and fit with a reaction-diffusion model to estimate the total bound fraction of Ser 2ph Fab (BFtot) from which the free fraction could be calculated FFSer2ph = 1 – BFtot ≈ 0.04. f, Quantitative immunoblotting was used to estimate the total number of RNAP2 per cell, nCTD, as well as the number phosphorylated at Ser 2, nSer2ph. Together the estimates of SSer2ph, Varr, Vnuc, FFSer2ph and nSer2ph from cf were used to calculate the Ser 2ph renormalization factor and generate the final renormalized FabLEM Ser 2ph curve (step 5). The renormalized curves in Fig. 2a were generated in an analogous manner.

Extended Data Figure 6 Testing the quality of FabLEM fits.

a, When FRAP and FabLEM data are simultaneously fit to the same model, the top fits (within 5% error of the best fit) found after searching the full parameter space are better constrained compared to fits of only FabLEM data in b or fits of only FRAP data in c, so all parameters can be estimated within an order of magnitude. Note that FRAP experiments are performed in steady-state, so fits do not depend on or Δ. d, Ser 2ph data from experiments using Fab against H3K27ac (Fig. 1c) are consistent with data from experiments using Fab against Ser 5ph (Fig. 2a and Extended Data Figs 5a, b). This indicates the sorted cell Ser 2ph data from Fig. 1c can be used in place of the Ser 2ph data in Fig. 2a for fitting purposes, as shown in the lower panel. e, Fits to data taken from sorted cells with arrays having low (upper panel) or high (lower panel) H3K27ac levels before transcriptional activation. The three parameters (tinit = 1/kini, telong = 1/kt, and tesc = 1/kesc) that change the most significantly between these fits are plotted for comparison in f. The mean, the 10/90% quantiles, and data bounds are shown. If the 10/90% quantiles do not overlap, then the fitted parameters are statistically different with >99% confidence (that is, the top 10% of the top 10%). Of these parameters, only tesc did not have overlapping 10/90% quantiles. The 90% mean difference confidence interval Δ(90%) for the high/low fitted parameters is reported (calculated with the built-in Mathematica function MeanDifferenceCI; Wolfram Research). g, To confirm the statistical significance of the high/low H3K27ac fits, one or two random cells were dropped from each high (N = 9) and low (N = 10) group in all possible ways (Nhigh and Nlow denoted). The average curves of these subgroups were fit and results for tesc and telong are shown. In all cases, the 10/90% quantiles for tesc do not overlap between high/low groups. In contrast, the 10/90% quantiles for telong do overlap, meaning there is less statistical difference between telong in high/low H3K27ac cells. h, To further cross-validate, select random splits of the data from a histogram like the one in Extended Data Fig. 4a were fit in the same way as f and g. The black area of the histogram (inset) shows which bin splits were taken from for fitting (N splits from each bin). A comparison of f and h reveals that even though the H3K27ac sorted high/low split only ranks in the top 10% of random splits (ranked by area between split average curves, see Extended Data Fig. 4a), fitted tesc values in f and g between high/low groups are statistically as distinct as those in h from the top 5% of random splits. In contrast, fitted telong values are not statistically distinct in f and g, but they are in h. This demonstrates that sorting data by the initial levels of H3K27ac (like in f and g) is better at distinguishing fast tesc from slow telong, whereas random data sorted solely by the area between split curves (like in h) cannot distinguish these two effects. Thus, the difference between fitted tesc is statistically significant and supports a link between H3K27ac and the RNAP2 promoter escape rate rather than between H3K27ac and the RNAP2 elongation rate.

Extended Data Figure 7 Screening the MMTV array for histone-modifying enzymes.

a, The histone deacetylases HDAC4 (HaloTag, top, red) and HDAC7 (HaloTag, bottom, red) colocalize with the array (yellow arrows) both before (pre) and after (10 min) transcription activation by GR (green). b, The lysine acetyltransferases (KATs) p300 (top, red) and CBP (middle, red) also colocalize with the array pre-activation (as marked by H3K27ac, blue), as does the steroid receptor cofactor 1 (SRC1, bottom, red), an adaptor that bridges p300/CBP to GR after activation. c, HDAC7 (and HDAC4, data not shown) is distributed both in the cytoplasm and nucleus and the cytoplasmic/nuclear intensity ratio varies from cell to cell. When in the nucleus, they colocalize with the array (bottom two rows). d, AFF4 (red) does not colocalize with H3K27ac (blue) at arrays pre-activation. However, AFF4 can be seen 30 min after gene activation, along with GFP–GR (green) and H3K27ac (blue). e, A variety of histone-modifying enzymes were tested to see which localized at the array. All HDACs, KDMs, Sirts and also PHF8 were HaloTag-tagged and screened by transient transfection, while the remaining enzymes were screened by immunostaining. Scale bars 5 μm.

Extended Data Figure 8 The ratio of elongating to promoter-bound RNAP2 increases with H3K27ac levels independent of RNA expression level.

a, Histograms of H3K27ac levels above input at genes (transcription start site TSS ± 2,000 base pairs) with varying levels of activity according to RNA sequencing. Each plot corresponds to 1,000 genes, with the top 1,000 most active genes in the very top plot followed successively underneath by plots showing the 1,000–2,000, 2,000–3,000…7,000–8,000 most active genes. A trend can be seen, with the more active genes having on average slightly higher acetylation than less active genes, although there is huge variability in all bins suggesting some very active genes have little H3K27ac, while other inactive genes have lots of H3K27ac. b, When RNAP2 occupancy is mapped to these binned genes (with colours corresponding to the histograms), those with higher H3K27ac levels tend to have more elongating RNAP2 relative to the amount bound at the promoter (± 350 base pairs, easiest to see in the renormalized plots on the far right), regardless of RNA expression level.

Extended Data Figure 9 FabLEM data alignment, FRAP checks and immunoblot quantification.

a, The GFP–GR curve is used to align FabLEM curves by fitting to a phenomenological model describing the basic structure of the average GFP–GR curve rescaled from 0 to 1. The fitting curve has the general form a exp(−k2[t-t0])/(1−exp[−k1(t−t0)]) with and left as fitting parameters whose starting estimated values were determined by fitting the average of all GFP–GR recruitment curves. After fitting an individual curve to this function, we numerically determine when the fitted function has a value of 0.16. This time is used as the aligning time for the individual curve, which we define as 4.2 min post-activation (corresponding to frame 10 post-activation). b, To ensure reversible photobleaching was minimal in our FRAP experiments on the mCherry-labelled RPB1 subunit of RNAP2, mCherry–RPB1, we duplicated experiments on histone H2B (mCherry–H2B, n = 10). This showed very little fluorescence recovery 50 s post-bleach at the array (yellow arrow) where GFP–GR was colocalized (± s.e.m.), demonstrating negligible reversible photobleaching of mCherry. c, To check for the role of diffusion in mCherry–RPB1 recoveries after photobleaching at the gene array (see Fig. 2c for an example), the profile of the photobleach spot was measured with time (defined in cartoon to right). Top row: after normalizing from 0 to 1, the curves all fall on top of each other, indicating no diffusive recovery (which drives spatial distortions in the bleach spot shape). Only the first post-bleach time point (labelled t = 0.0 s here) shows a difference in shape (red arrows) from the others (0.7 s, 1.4 s,…), indicating diffusion plays a role in the recovery up until about 0.7 s. Bottom row: this is further confirmed by doing a rolling average of data (averaging three frames at a time), which shows no distortion of shape after normalization all the way up to 54 s post-bleach. d, FRAP was performed in cells transiently expressing mCherry–RPB1 and treated (n = 27) or untreated (n = 27) with the elongation inhibiting drug flavorpiridol (1 μM, 1 h). In treated cells, the recovery was 21% more complete at 50 s post-bleach (relative to the baseline recovery from mCherry–H2B FRAP in b) than in untreated cells, indicating this fraction of RNAP2 is elongating in the untreated cells. e, Sample immunoblots of whole-cell extract from 3617 and HeLa cells using monoclonal antibodies against the CTD (α-CTD) of RPB1, as well as its Ser-5 (α-Ser 5ph) and Ser-2 (α-Ser 2ph) phosphorylated forms. f, Sample immunoblot quantification. Images were digitized and the intensity of blots quantified (blue/green shapes). Here the intensity of HeLa blots was fitted to a line to determine the relative intensity of blots from 3617 cells (within the linear region of the fit) and thereby determine the ratios of the numbers of RNAP2 in each cell type.

Extended Data Table 1 Summary of estimated model parameters

Supplementary information

Supplementary Table 1

This file contains a table of the GR Response genes. (XLS 147 kb)

Supplementary Data

This zipped file details the Mathematica code for quantifying regions of interest in image sequences. Sample data and usage are included. (ZIP 8175 kb)

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Stasevich, T., Hayashi-Takanaka, Y., Sato, Y. et al. Regulation of RNA polymerase II activation by histone acetylation in single living cells. Nature 516, 272–275 (2014). https://doi.org/10.1038/nature13714

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing