Article | Published:

Dynamic interplay between enhancer–promoter topology and gene activity

Nature Genetics (2018) | Download Citation


A long-standing question in gene regulation is how remote enhancers communicate with their target promoters, and specifically how chromatin topology dynamically relates to gene activation. Here, we combine genome editing and multi-color live imaging to simultaneously visualize physical enhancer–promoter interaction and transcription at the single-cell level in Drosophila embryos. By examining transcriptional activation of a reporter by the endogenous even-skipped enhancers, which are located 150 kb away, we identify three distinct topological conformation states and measure their transition kinetics. We show that sustained proximity of the enhancer to its target is required for activation. Transcription in turn affects the three-dimensional topology as it enhances the temporal stability of the proximal conformation and is associated with further spatial compaction. Furthermore, the facilitated long-range activation results in transcriptional competition at the locus, causing corresponding developmental defects. Our approach offers quantitative insight into the spatial and temporal determinants of long-range gene regulation and their implications for cellular fates.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Benoist, C. & Chambon, P. In vivo sequence requirements of the SV40 early promotor region. Nature 290, 304–310 (1981).

  2. 2.

    Levine, M. Transcriptional enhancers in animal development and evolution. Curr. Biol. 20, R754–R763 (2010).

  3. 3.

    Long, H. K., Prescott, S. L. & Wysocka, J. Ever-changing landscapes: transcriptional enhancers in development and evolution. Cell 167, 1170–1187 (2016).

  4. 4.

    Buecker, C. & Wysocka, J. Enhancers as information integration hubs in development: lessons from genomics. Trends Genet. 28, 276–284 (2012).

  5. 5.

    Kim, T. K. & Shiekhattar, R. Architectural and functional commonalities between enhancers and promoters. Cell 162, 948–959 (2015).

  6. 6.

    Vernimmen, D. & Bickmore, W. A. The hierarchy of transcriptional activation: from enhancer to promoter. Trends Genet. 31, 696–708 (2015).

  7. 7.

    Consortium, E. P. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012).

  8. 8.

    Tolhuis, B., Palstra, R. J., Splinter, E., Grosveld, F. & de Laat, W. Looping and interaction between hypersensitive sites in the active beta-globin locus. Mol. Cell. 10, 1453–1465 (2002).

  9. 9.

    Uslu, V. V. et al. Long-range enhancers regulating Myc expression are required for normal facial morphogenesis. Nat. Genet. 46, 753–758 (2014).

  10. 10.

    Zhang, Y. et al. Chromatin connectivity maps reveal dynamic promoter-enhancer long-range associations. Nature 504, 306–310 (2013).

  11. 11.

    Arnold, C. D. et al. Genome-wide quantitative enhancer activity maps identified by STARR-seq. Science 339, 1074–1077 (2013).

  12. 12.

    Ghavi-Helm, Y. et al. Enhancer loops appear stable during development and are associated with paused polymerase. Nature 512, 96–100 (2014).

  13. 13.

    Kvon, E. Z. et al. Genome-scale functional characterization of Drosophila developmental enhancers in vivo. Nature 512, 91–95 (2014).

  14. 14.

    Levine, M., Cattoglio, C. & Tjian, R. Looping back to leap forward: transcription enters a new era. Cell 157, 13–25 (2014).

  15. 15.

    Kagey, M. H. et al. Mediator and cohesin connect gene expression and chromatin architecture. Nature 467, 430–435 (2010).

  16. 16.

    Mifsud, B. et al. Mapping long-range promoter contacts in human cells with high-resolution capture Hi-C. Nat. Genet. 47, 598–606 (2015).

  17. 17.

    Andrey, G. et al. A switch between topological domains underlies HoxD genes collinearity in mouse limbs. Science 340, 1234167 (2013).

  18. 18.

    Spitz, F. Gene regulation at a distance: from remote enhancers to 3D regulatory ensembles. Semin. Cell Dev. Biol. 57, 57–67 (2016).

  19. 19.

    Carter, D., Chakalova, L., Osborne, C. S., Dai, Y. F. & Fraser, P. Long-range chromatin regulatory interactions in vivo. Nat. Genet. 32, 623–626 (2002).

  20. 20.

    Sanyal, A., Lajoie, B. R., Jain, G. & Dekker, J. The long-range interaction landscape of gene promoters. Nature 489, 109–113 (2012).

  21. 21.

    Fujioka, M., Wu, X. & Jaynes, J. B. A chromatin insulator mediates transgene homing and very long-range enhancer-promoter communication. Development 136, 3077–3087 (2009).

  22. 22.

    Fujioka, M., Sun, G. & Jaynes, J. B. The Drosophila eve insulator Homie promotes eve expression and protects the adjacent gene from repression by polycomb spreading. PLoS Genet. 9, e1003883 (2013).

  23. 23.

    Fujioka, M., Mistry, H., Schedl, P. & Jaynes, J. B. determinants of chromosome architecture: insulator pairing in cis and in trans. PLoS Genet. 12, e1005889 (2016).

  24. 24.

    Larson, D. R., Zenklusen, D., Wu, B., Chao, J. A. & Singer, R. H. Real-time observation of transcription initiation and elongation on an endogenous yeast gene. Science 332, 475–478 (2011).

  25. 25.

    Hocine, S., Raymond, P., Zenklusen, D., Chao, J. A. & Singer, R. H. Single-molecule analysis of gene expression using two-color RNA labeling in live yeast. Nat. Methods 10, 119–121 (2013).

  26. 26.

    Fukaya, T., Lim, B. & Levine, M. Enhancer control of transcriptional bursting. Cell 166, 358–368 (2016).

  27. 27.

    Dubarry, N., Pasta, F. & Lane, D. ParABS systems of the four replicons of Burkholderia cenocepacia: new chromosome centromeres confer partition specificity. J. Bacteriol. 188, 1489–1496 (2006).

  28. 28.

    Saad, H. et al. DNA dynamics during early double-strand break processing revealed by non-intrusive imaging of living cells. PLoS Genet. 10, e1004187 (2014).

  29. 29.

    Gasser, S. M. Visualizing chromatin dynamics in interphase nuclei. Science 296, 1412–1416 (2002).

  30. 30.

    Sinclair, P., Bian, Q., Plutz, M., Heard, E. & Belmont, A. S. Dynamic plasticity of large-scale chromatin structure revealed by self-assembly of engineered chromosome regions. J. Cell Biol. 190, 761–776 (2010).

  31. 31.

    Bystricky, K. Chromosome dynamics and folding in eukaryotes: insights from live cell microscopy. FEBS Lett. 589, 3014–3022 (2015).

  32. 32.

    Garcia, H. G., Tikhonov, M., Lin, A. & Gregor, T. Quantitative imaging of transcription in living Drosophila embryos links polymerase activity to patterning. Curr. Biol. 23, 2140–2145 (2013).

  33. 33.

    Fukaya, T., Lim, B. & Levine, M. Rapid rates of Pol II elongation in the Drosophila embryo. Curr. Biol. 27, 1387–1391 (2017).

  34. 34.

    Lucas, J. S., Zhang, Y., Dudko, O. K. & Murre, C. 3D trajectories adopted by coding and regulatory DNA elements: first-passage times for genomic interactions. Cell 158, 339–352 (2014).

  35. 35.

    Guo, Y. et al. CRISPR inversion of CTCF sites alters genome topology and enhancer/promoter function. Cell 162, 900–910 (2015).

  36. 36.

    Rao, S. S. et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 159, 1665–1680 (2014).

  37. 37.

    Deng, W. et al. Reactivation of developmentally silenced globin genes by forced chromatin looping. Cell 158, 849–860 (2014).

  38. 38.

    Franke, M. et al. Formation of new chromatin domains determines pathogenicity of genomic duplications. Nature 538, 265–269 (2016).

  39. 39.

    Dixon, J. R. et al. Chromatin architecture reorganization during stem cell differentiation. Nature 518, 331–336 (2015).

  40. 40.

    Krijger, P. H. et al. Cell-of-origin-specific 3D genome structure acquired during somatic cell reprogramming. Cell Stem Cell 18, 597–610 (2016).

  41. 41.

    Raser, J. M. & O’Shea, E. K. Control of stochasticity in eukaryotic gene expression. Science 304, 1811–1814 (2004).

  42. 42.

    Voss, T. C. & Hager, G. L. Dynamic regulation of transcriptional states by chromatin and transcription factors. Nat. Rev. Genet. 15, 69–81 (2014).

  43. 43.

    Sanchez, A., Garcia, H. G., Jones, D., Phillips, R. & Kondev, J. Effect of promoter architecture on the cell-to-cell variability in gene expression. PLoS Comput. Biol. 7, e1001100 (2011).

  44. 44.

    Hug, C. B., Grimaldi, A. G., Kruse, K. & Vaquerizas, J. M. Chromatin architecture emerges during zygotic genome activation independent of transcription. Cell 169, 216–228 e19 (2017).

  45. 45.

    Rubin, A. J. et al. Lineage-specific dynamic and pre-established enhancer-promoter contacts cooperate in terminal differentiation. Nat. Genet. 49, 1522–1528 (2017).

  46. 46.

    Hnisz, D., Shrinivas, K., Young, R. A., Chakraborty, A. K. & Sharp, P. A. A phase separation model for transcriptional control. Cell 169, 13–23 (2017).

  47. 47.

    Sexton, T., Umlauf, D., Kurukuti, S. & Fraser, P. The role of transcription factories in large-scale structure and dynamics of interphase chromatin. Semin. Cell Dev. Biol. 18, 691–697 (2007).

  48. 48.

    Cho, W. K. et al. RNA polymerase II cluster dynamics predict mRNA output in living cells. eLife 5, e13617 (2016).

  49. 49.

    Williamson, I., Lettice, L. A., Hill, R. E. & Bickmore, W. A. Shh and ZRS enhancer colocalisation is specific to the zone of polarising activity. Development 143, 2994–3001 (2016).

  50. 50.

    Stadler, M. R., Haines, J. E. & Eisen, M. B. Convergence of topological domain boundaries, insulators, and polytene interbands revealed by high-resolution mapping of chromatin contacts in the early Drosophila melanogaster embryo. Elife 6(2017).

  51. 51.

    Negre, N. et al. A comprehensive map of insulator elements for the Drosophila genome. PLoS Genet. 6, e1000814 (2010).

  52. 52.

    Kyrchanova, O. & Georgiev, P. Chromatin insulators and long-distance interactions in Drosophila. FEBS Lett. 588, 8–14 (2014).

  53. 53.

    Chetverina, D. et al. Boundaries of loop domains (insulators): determinants of chromosome form and function in multicellular eukaryotes. Bioessays 39(2017).

  54. 54.

    Lupianez, D. G. et al. Disruptions of topological chromatin domains cause pathogenic rewiring of gene-enhancer interactions. Cell 161, 1012–1025 (2015).

  55. 55.

    Bartman, C. R., Hsu, S. C., Hsiung, C. C., Raj, A. & Blobel, G. A. Enhancer regulation of transcriptional bursting parameters revealed by forced chromatin looping. Mol. Cell. 62, 237–247 (2016).

  56. 56.

    Wu, B., Chen, J. & Singer, R. H. Background free imaging of single mRNAs in live cells using split fluorescent proteins. Sci. Rep. 4, 3615 (2014).

  57. 57.

    Sladitschek, H. L. & Neveu, P. A. MXS-chaining: a highly efficient cloning platform for imaging and flow cytometry approaches in mammalian systems. PLoS One 10, e0124958 (2015).

  58. 58.

    Vodala, S., Abruzzi, K. C. & Rosbash, M. The nuclear exosome and adenylation regulate posttranscriptional tethering of yeast GAL genes to the nuclear periphery. Mol. Cell 31, 104–113 (2008).

  59. 59.

    Dubarry, M., Loiodice, I., Chen, C. L., Thermes, C. & Taddei, A. Tight protein-DNA interactions favor gene silencing. Genes Dev. 25, 1365–1370 (2011).

  60. 60.

    Bateman, J. R., Lee, A. M. & Wu, C. T. Site-specific transformation of Drosophila via phiC31 integrase-mediated cassette exchange. Genetics 173, 769–777 (2006).

  61. 61.

    Small, S. In vivo analysis of lacZ fusion genes in transgenic Drosophila melanogaster. Methods Enzymol. 326, 146–159 (2000).

  62. 62.

    Little, S. C., Tkacik, G., Kneeland, T. B., Wieschaus, E. F. & Gregor, T. The formation of the Bicoid morphogen gradient requires protein movement from anteriorly localized mRNA. PLoS Biol. 9, e1000596 (2011).

  63. 63.

    Little, S. C., Tikhonov, M. & Gregor, T. Precise developmental gene expression arises from globally stochastic transcriptional activity. Cell 154, 789–800 (2013).

  64. 64.

    Dubuis, J. O., Samanta, R. & Gregor, T. Accurate measurements of dynamics and reproducibility in small genetic networks. Mol. Syst. Biol. 9, 639 (2013).

  65. 65.

    Gao, Y. & Kilfoil, M. L. Accurate detection and complete tracking of large populations of features in three dimensions. Opt. Express 17, 4685–4704 (2009).

  66. 66.

    Dijkstra, E. W. A note on two problems in connexion with graphs. Numer. Math. 1, 269–271 (1959).

Download references


We thank K. Bystricky for introducing us to the ParB/parS system, and F. Payre and P. Valenti for sharing a ParB-eGFP plasmid and the parS sequence. We also thank S. Blythe, H. Garcia, H. Grabmayr, T. Fukaya, M. Levine, S. Little, P. Ratchasanmuang, S. Ryabichko, P. Schedl, E.F. Wieschaus, B. Zoller, and the Bloomington Drosophila Stock Center. This study was funded by grants from the National Institutes of Health (U01 EB021239, U01 DA047730, R01 GM097275, R01 GM117458) and from the National Science Foundation (PHY-1734030). H.C. was supported by the Charles H. Revson Biomedical Science Fellowship. M.L. was supported by the Rothschild, EMBO and HFSP fellowships.

Author information


  1. Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ, USA

    • Hongtao Chen
    • , Michal Levo
    •  & Thomas Gregor
  2. Department of Molecular Biology, Princeton University, Princeton, NJ, USA

    • Lev Barinov
  3. Department of Biochemistry and Molecular Biology, and the Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA, USA

    • Miki Fujioka
    •  & James B. Jaynes
  4. Joseph Henry Laboratories of Physics, Princeton University, Princeton, NJ, USA

    • Thomas Gregor
  5. Department of Developmental and Stem Cell Biology, Institut Pasteur, Paris, France

    • Thomas Gregor


  1. Search for Hongtao Chen in:

  2. Search for Michal Levo in:

  3. Search for Lev Barinov in:

  4. Search for Miki Fujioka in:

  5. Search for James B. Jaynes in:

  6. Search for Thomas Gregor in:


H.C. and T.G. conceived the main ideas regarding live-cell image generation, processing and analysis. M.F. and J.B.J. developed the homie-eve system to create quantifiably distinct architectural and transcriptional states. H.C. designed the study to overlap these technologies. H.C., M.F. and J.B.J. designed and generated the transgenic flies. H.C. and L.B. performed the imaging experiments. H.C., L.B., M.L. and T.G. analyzed the data and wrote the manuscript.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Thomas Gregor.

Integrated supplementary information

  1. Supplementary Figure 1 FISH reveals eve enhancer-dependent expression of a reporter gene located 142 kb upstream of the endogenous eve locus.

    a, Genomic design of a synthetic long-range enhancer–promoter interaction. An ectopic homie insulator sequence with an eve promoter driving lacZ is integrated at ~142 kb upstream of the eve locus. Embryos homozygous for this construct are hybridized with single-molecule FISH probes to label endogenous eve (red) and lacZ (green) mRNA. b, Top, surface view of a 2.5-h-old Drosophila embryo hybridized with eve-atto633 probes. Anterior is to the left. Bottom, z-stack projection of the marked region in the top panel. LacZ activity (labeled with lacZ-atto565 probes) only occurs sporadically within the limits of the eve pattern (red). This lacZ pattern appears in all 13 embryos imaged (2–3 h old) and a representative sample is shown here.

  2. Supplementary Figure 2 The eve-MS2 allele recapitulates the expression pattern and transcriptional activity of the endogenous eve gene.

    a, Editing the endogenous eve locus (top) to obtain the eve-MS2 allele (bottom). Arrowheads indicate primers for PCR genotyping. Green and red lines mark sequences targeted by smFISH probes (lacZ-atto565 and eve-atto633, respectively). Loci are not drawn to scale. b, Genotyping the eve-MS2 allele. The PCR result from a single fly carrying the eve-MS2 allele is shown with DNA ladder. The 466-bo band was verified by sequencing. Primers are shown in a. cg, smFISH quantification of the transcriptional activity of the eve-MS2 allele from a representative embryo at ~45 min into nc14. Maximum z projections are shown for the lacZ-atto565 channel (c) and the eve-atto633 channel (d) of an eve-MS2/eve+ embryo. eve stripes 5 to 7 (from left to right) are shown. e, Magnified view of the square in c and d. Nuclear regions are marked with yellow dashed lines. Arrows indicate examples of eve-MS2 transcription loci that are labeled by both probes. f, Cytoplasmic spots and active transcription spots were identified by image analysis routines (Methods). A cytoplasmic unit (CU) that corresponds to the fluorescent intensity of a single cytoplasmic mRNA is extracted. The panel shows the number of RNA polymerase II (PolII) on the eve-MS2 loci from 93 nuclei in which a transcription spot in the eve-atto633 channel was observed at the eve-MS2 locus. The PolII numbers are inferred from either the CU derived from lacZ-atto565 (x axis) or eve-atto633 (y axis) measurements. The inset shows the calculation of cytoplasmic unit for eve. Specifically, a sliding window of 220 × 220 × 23 pixels (16.5 × 16.5 × 7.4 μm3) is applied to the raw image stack (c and d) and the total pixel values in the window are plotted against the number of cytoplasmic spots found in the window. A linear fit in the range of 0–100 cytoplasmic spots is applied to extract CU for each probe set. g, Comparison of the PolII number on the eve-MS2 locus and on the endogenous eve locus (mean ± s.d.). Note that the numbers reported in f and g are for two sister chromatids. The number of nuclei analyzed for stripes 1 to 7 was 25, 28, 24, 23, 25, 27 and 54, respectively. Analysis performed on other embryos (n= 3) imaged at different stages in nc14 also showed no significant difference in PolII numbers on the eve-MS2 locus and the endogenous eve locus.

  3. Supplementary Figure 3 Spot localization precision and measurement error.

    a, Genetic design of a transgene that colocalizes all three reporter systems. MS2 and PP7 stem loops are alternated and repeated 24 times. A knirps (kni) reporter gene, which includes the kni CDS (with the start codon removed) and 3′ UTR, is driven by a hunchback P2 (hbP2) promoter, resulting in expression in all nuclei located in the anterior 10%–45% of the embryo. b, Analysis of chromatic aberration and localization error (Methods). Panels show the linear distance (along the x coordinate only) for each blue-green spot pair as a function of the pair’s x position for eve-MS2 embryos carrying the parS-homie-eve-PP7 transgene (left, n= 34 embryos), embryos carrying the three-color colocalization transgene from a (middle, n= 9 embryos), and TetraSpec beads (right, n= 5 independent data sets), respectively. Blue data points are for all spot pairs at all time frames for all embryos analyzed. Yellow data points are from one of the embryos (or one set of experiment for the beads). Linear fits in each panel report on the chromatic aberrations between blue and green spots in the x direction. As slopes and intercepts for the different samples show no significant differences, chromatic aberrations can be corrected for each individual embryo dataset internally. c, Summary of the distributions of spot pair distances (after chromatic aberration correction) for the three configurations in b. Each direction (x, y and z) is shown for each color combination. For example, for the blue-green (MS2-parS) distances in the x direction, the s.d. of the parS-homie-eve-PP7 transgene (labeled –142 kb) corresponds to the solid black bar shown in the left panel of b. Spot localization errors are estimated from the s.d. measured with the three-color colocalization control embryos (labeled 0 kb). Center values, means; solid lines, s.d.; bars, 25%–75% quantiles. d, Dependence of localization precision on signal intensities. Since localization precision scales directly with the square root of the number of photons, we can assess localization error of the three-color colocalization control embryos from the localization error measured with immobilized beads of similar fluorescent intensity values (photon counts). Thus, differences in y-axis offset are not due to differences in photon counts but are due to ‘motion blurring’ of the moving spot during acquisition, which amounts to about two-thirds of the total localization. The remaining one-third (corresponding to the error obtained from immobile beads) stems from optical measurement noise and our analysis pipeline. eh, Optical characterization of nascent transcription sites and parS foci. For each fluorescent channel, all identified fluorescent spots are classified into eight groups according to their raw intensities. A ‘super-spot’ for each group is obtained by aligning all spots of a group with the brightest pixel at the center of a 25 × 25 × 13 voxel region of interest and by taking the average intensity per voxel in that region over all spots. The intensity profiles along the x (e, f) and z (g, h) cross-sections for the blue MS2 super-spot (e, g) and green parS super-spot (f, h) are plotted (darker curves represent brighter spots). Dashed lines are from equivalent measurements of TetraSpec beads. Images of the super-spots for the brightest blue (MS2; e, g) or green (parS; f, h) spots (top) and for the beads (bottom) are shown as panel insets.

  4. Supplementary Figure 4 Different genomic labeling approaches report on similar chromatin dynamics and transcription kinetics.

    a, Three methods of labeling genomic loci. b, The measured blue-green (MS2-parS) distances are not sensitive to labeling approach. The box plot shows the distributions of the instantaneous distance between spot pairs in the same nuclei. Distances shown are after chromatic aberration corrections. For all three genomic settings, the MS2-parS (blue-green) distances showed no significant differences (one-way Kruskal–Wallis test on individual embryo mean distances, n= 34, 9 and 6 embryos for sets A, B and C, respectively). This was observed regardless of the absence (Red-OFF, P = 0.17, χ2 = 4.3, d.f.=50) or presence (Red-ON, P = 0.60, χ2 = 1.04, d.f. = 49) of PP7 activity. Center values, medians; boxes, interquantile ranges (25–75% quantiles); whiskers, 1.5 times the interquantile range. The 0 kb control is the hbP2-MS2PP7-kni embryo described in Supplementary Fig. 3. c, The distances between spot pairs reflect their genomic arrangement. Distributions of the instantaneous distance between spot pairs are plotted. Box-and-whisker plots are as described in b. Distances shown are after chromatic aberration corrections. Note that the parS-PP7 (green-red) distance is significantly shorter when the parS tag is located at the 3′ side of the PP7 reporter (P = 4.5 × 10–6, two-tailed Wilcoxon rank-sum test). d, Mean square displacement (MSD) plots for set A and set B. Each MSD trace is a result of the population ensemble of all nuclei in a single embryo (embryo-averaged MSD; Methods). Results from the two genomic settings display subdiffusive characteristics with a scaling power of ~0.24, and their anomalous diffusion coefficients show no significant difference (two-tailed Student’s t test, P = 0.87, t = –0.1534, d.f. = 90; linear fits with mean ± s.d. across embryos). eg, Transcriptional activation of eve-PP7 is not affected by labeling approach. The fraction of eve-MS2-expressing nuclei that also contain active PP7 (mean ± SE) is plotted as a function of time for three genomic settings. It seems that neither the presence nor the location of the parS tag interferes with either enhancer action or transcriptional activation. This is consistent with the hypothesis that the ParB-DNA complex is formed from specific ParB-parS nucleation sites followed by stochastic binding and trapping.

  5. Supplementary Figure 5 Sustained physical proximity is required for transcription initiation and maintenance: individual traces.

    a, Transcriptional activity (red spot (PP7) intensity) and instantaneous enhancer–promoter distance (blue-green distance) as a function of time for 286 nuclei transitioning from the Red-OFF to the Red-ON state. Time series for individual nuclei are aligned such that PP7 activity starts at 0 min (red dashed lines). Individual traces are sorted according to the mean enhancer–promoter distance in the 5 min before PP7 activity is observed. b, Distribution of the instantaneous enhancer–promoter distance as a function of time for the Red-OFF to Red-ON transition. Calculated from a. c, Transcription and instantaneous enhancer–promoter distance as a function of time for 203 nuclei transitioning from the Red-ON to the Red-OFF state. Time series for individual nuclei are aligned such that PP7 activity ends at 0 min (red dashed lines). Individual traces are sorted according to the mean enhancer–promoter distance in the 5 min before PP7 activity disappears. d, Distribution of the instantaneous enhancer–promoter distance as a function of time for the Red-ON to Red-OFF transition. Calculated from c.

  6. Supplementary Figure 6 Analysis of enhancer–promoter distance for individual embryos and individual nuclei.

    ac, The time-averaged r.m.s. distance between MS2 (blue) and parS (green) spots (enhancer–promoter distance) is depicted as a scatterplot for each nucleus from 84 embryos carrying the parS-homie-evePr-PP7 construct (a), 29 embryos carrying the parS-homie-noPr-PP7 construct (b), and 15 embryos carrying the parS-lambda-evePr-PP7 construct (c) located at –142 kb with respect to the eve-MS2 locus. d, Time-averaged r.m.s. enhancer–promoter distance for ten embryos carrying the parS-homie-evePr-PP7 construct at –589 kb with respect to the eve-MS2 locus. Data points marked in red are calculated from the Red-ON part of enhancer–promoter trajectories in nuclei displaying PP7 activity. Data points marked with blue are calculated from full enhancer–promoter trajectories in nuclei that never show PP7 during the imaging time window (25–55 min in nc14). Notice that the number of time points (for example, length of enhancer–promoter trajectories) used for calculating r.m.s. distance varies among nuclei depending on the nuclear anterior–posterior position and the view of the image. e, r.m.s. enhancer–promoter distance as a function of the length of the trajectories used for r.m.s. distance calculation. All r.m.s. enhancer–promoter distance samples from the 84 embryos carrying the parS-homie-evePr-PP7 construct at –142 kb are shown.

  7. Supplementary Figure 7 Tuning stability of the homie element.

    ad, The enhancer–promoter distance (r.m.s. distance) distribution for four experimental constructs: parS-lambda-evePr-PP7 (a), parS-homie½-evePr-PP7 (b), parS-homie¾-evePr-PP7 (c) and parS-homie-evePr-PP7 (d). A 5-min sliding window along each time trace is used to calculate r.m.s. enhancer–promoter distances. homie½ (chr2R:9,988,934–9,988,750, dm6) and homie¾ (chr2R:9,989,025–9,988,750, dm6) are two truncated homie elements. Red bars in c and d show the probability density of r.m.s. distance samples accompanied by continuous PP7 transcription. eh, Quantile–quantile plots against the standard normal distribution for the r.m.s. enhancer–promoter distances shown in ad, respectively. Short enhancer–promoter distances resulting from the paired (Poff and Pon) states are progressively enriched as the stability of the homie element increases. Insets show the complete quantile–quantile plots. il, Fraction of each topological state for the constructs shown in ad, respectively. See Supplementary Fig. 8 and the Methods for details about topological state classification.

  8. Supplementary Figure 8 Training of a Bayesian classifier and characterization of the three topological states.

    a, An enhancer–promoter distance vector (dx,y,z) is calculated at each time point, corrected for chromatic aberration. b, The relative velocity (vx,y,z) between the enhancer (blue MS2 spot) and promoter (green parS spot) is calculated from the two consecutive distance vectors. The instantaneous distance vector and the two velocity vectors that connect the two adjacent time points are used for training a binary classifier using a naive Bayes method (Methods). Two training samples are used. For the open state (O state), enhancer–promoter trajectories from the parS-lambda-evePr-PP7 control are used. For the paired state (P state), enhancer–promoter trajectories from the Red-ON part of nuclei displaying PP7 activity are used. The last 4 min of these Red-ON trajectories are removed from the training sample because of PolII elongation. cf, Joint distribution of the selected dimensions of the distance vectors for the O state (c, e) and P state (d, f) training samples. gj, Joint distribution of the selected dimensions of the velocity vectors for the O state (g, i) and P state (h, j) training samples. From cj, z projections are raw data and the probability density functions of 2D Gaussian fits are shown. k, l, r.m.s. distance (k) and fraction (l) for each topological state calculated for individual embryos (n is the number of embryos).

  9. Supplementary Figure 9 A kinetic model captures transition rates among the three topological states.

    a, A series of first-order reactions are used to model the transition kinetics between the Ooff, Poff and Pon states. Based on the finding that physical proximity is required for transcriptional activation, we assume in this model that Pon occurs only after Poff is established. The coupled ODEs describe evolution of the system given the initial conditions. For parS-homie-noPr-PP7, only the Ooff and Poff states are present and we assume the same f1 and b1 values as for parS-homie-evePr-PP7. b, Fraction of the Poff state for homie-noPromoter-PP7 as a function of developmental time. 0 on the x axis corresponds to 25 min in nc14. The mean ± SE is shown (n= 29 embryos). This curve, together with time series curves obtained from the parS-homie-evePr-PP7 construct (dashed lines; same as in Fig. 3e), is used to infer the kinetic parameters with Markov chain Monte Carlo (MCMC) simulations (Methods). cg, Marginal posterior distributions of the five kinetic parameters in a constructed from 90,000 stationary MCMC samples. Medians are labeled. Error bars span from the 25th-percentile to the 75th-percentile quantile (also shown in square brackets). Insets in f and g show the joint distribution of (b1, b2) and (b1, b3), respectively. Darker color represents higher density. h, The inferred parameters for the disappearance of Pon recapitulate the distribution of lifespans of PP7 activity. To calculate PP7 lifespan distribution, PP7 traces are grouped into cohorts according to the maximum measurable lifespan for each trace (Methods). For each PP7 cohort, a cumulative distribution function (CDF) for the PP7 lifespan is calculated (gray curves). Because the lifespan distribution is truncated at the maximum measurable time, the tails of the CDFs (corresponding to CDF = 1) are removed. The solid red line shows the median of these truncated CDFs, which is the CDF of the lifespans of PP7 activity. The dashed red curve comes from the CDF of an exponential distribution with mean= (b2 + b3)–1 = (0.014 + 0.011)–1 min. This exponential CDF is shifted horizontally to account for a deterministic elongation time of 4 min, which coincides with the lifespan of the shortest PP7 trace.

  10. Supplementary Figure 10 Scoring mutant phenotypes resulting from promoter competition.

    a, Cross schemes to test the phenotypic effects of competition between the endogenous eve promoter and the ectopic eve promoter that is activated on the formation of new topological states. Single males, either homie-evePr-lacZ or lambda-evePr-lacZ, are used for the crosses (note removal of PP7 sequences). For each single cross, the patterning phenotypes in adult abdominal segments A4, A6 and/or A8 are scored. No conspicuous phenotype in other abdominal segments is noticed. bf, The adult abdominal phenotypes most likely result from haploinsufficiency of eve, as shown in the cuticles derived from a cross between homozygous homie-evePr-lacZ males and Df(2R)eve-/CyO females. A wild-type cuticle is shown in b. Strong eve phenotypes, that is, loss or perturbation of denticle bands in the even-numbered abdominal segments, are observed (cf).

Supplementary information

  1. Supplementary Text and Figures

    Supplementary Figures 1–10 and Sequences

  2. Reporting Summary

  3. Supplementary Video 1

    Live imaging of the eve-MS2 expression pattern. The embryo shown in the video was generated by crossing HisRFP;MCP-eGFP females with males carrying the eve-MS2 CRISPR construct. Anterior is to the left. The video starts at late nuclear cycle 13 (nc13) and ends at germband elongation. Time 0 indicates the onset of nc14, and gastrulation occurs at ~62 min in nc14. Nuclei (red) are marked with His-RFP and the green spots are nascent mRNAs transcribed at the endogenous eve locus. For three-color data acquisition, we focus on the posterior regions (yellow rectangle) during a 30-min time window (~25–55 min into nc14). The video shown here is representative of 102 embryos imaged

  4. Supplementary Video 2

    Three-color live imaging of enhancer–promoter interactions. The embryo in the video was generated by the crosses shown in Fig. 1a. Anterior is to the left. Blue spots mark eve-MS2 transcription as well as the physical position of the eve locus and eve enhancers. Red spots mark transcriptional activity from the PP7 reporter. Green spots are from the parS sequence, which marks the location of the reporter construct. The cartoon below the video depicts the genomic construct and the labeling scheme for the three-color imaging. Time 0 is the onset of nc14. Four eve-MS2 stripes (stripes 3 to 6, left to right) are visible. This video is representative of 84 embryos imaged. Results of quantitative analysis on these videos are described in the main text

  5. Supplementary Video 3

    Enhancer–promoter physical proximity is necessary for transcriptional activity. Three representative nuclei are labeled with numbers. PP7 transcription (red signal) is not detected in nucleus 1 or 2 (RedOFF). In these two nuclei, a clear physical separation between the blue (eve-MS2 marking the locations of eve enhancers) and the green spot (parS marking the locations of the PP7 reporter) is observed. In nucleus 3 where the PP7 reporter is active (Red-ON), the blue and green spots appear to overlap. Pixel size is 50 nm × 50 nm. Quantitative analysis of all Red-OFF (n= 7,163) and Red-ON (n= 720) trajectories is described in the main text

  6. Supplementary Video 4

    Enhancer–promoter physical proximity is necessary for continuous initiation of transcription. a–c, Representative nuclei in which the PP7 reporter switches from Red-OFF to Red-ON (n= 286). d–f, Representative nuclei in which the PP7 reporter switches from Red-ON to Red-OFF (n= 203). Time 0 marks the frame at which the OFF-to-ON or ON-to-OFF switch occurs

  7. Supplementary Data 1

    Raw spot localization dataset 1

  8. Supplementary Data 2

    Raw spot localization dataset 2

About this article

Publication history