ATP-dependent chromatin remodeling enzymes (remodelers) alter the structure and organization of nucleosomes, the basic building blocks of eukaryotic chromatin (Becker and Workman, 2013; Bracken et al., 2019; Clapier et al., 2017; Jungblut et al., 2020; Sundaramoorthy and Owen-Hughes, 2020). Core nucleosomes comprise 147 bp of DNA wrapped around a histone protein octamer formed by a H3-H4 tetramer that is flanked on either side by a H2A-H2B dimer. Remodelers modulate the accessibility of regulatory DNA elements by unwrapping or sliding nucleosomes, thereby presenting a fundamental level of control of transcription and other processes that require access to DNA (Brahma and Henikoff, 2020). Remodelers are large, multi-subunit complexes that harbor a Snf2-class ATPase motor (Clapier et al., 2017; Jungblut et al., 2020; Sundaramoorthy and Owen-Hughes, 2020). Nucleosome remodeling is the result of ATP-dependent DNA translocation by the ATPase, which is locked onto the nucleosome through binding additional DNA- and histone sites. ATPase activity is directed by accessory subunits that determine targeting to specific genomic loci, and the outcome of the remodeling reaction. Different classes of remodelers, defined by their ATPase and unique sets of accessory subunits, are dedicated to distinct chromatin regulatory functions.

Members of the SWI/SNF family of remodelers have mainly been implicated in generating open chromatin at gene regulatory DNA elements (Becker and Workman, 2013; Bracken et al., 2019; Brahma and Henikoff, 2020; Cakiroglu et al., 2019; Clapier et al., 2017; Pillidge and Bray, 2019). Recent structural studies revealed that SWI/SNF-class remodelers engage the nucleosome in a highly modular manner (Han et al., 2020; He et al., 2020; Jungblut et al., 2020; Mashtalir et al., 2020; Patel et al., 2019; Sundaramoorthy and Owen-Hughes, 2020; Wagner et al., 2020; Ye et al., 2019). The motor domain of the ATPase binds the nucleosomal DNA at superhelical position +2, while other modules contact the opposite DNA gyre, both faces of the histone octamer, and possibly the exiting DNA, and the histone H4 tail. SWI/SNF remodelers do not closely encompass the nucleosome, and the nucleosome remains mostly accessible. ATPase activity is only required for binding to a subset of genomic loci, but loss of ATPase activity can have dominant effects on chromatin binding, enhancer accessibility, and gene expression (Gelbart et al., 2005; Hodges et al., 2018; Pan et al., 2019). Mutations in key functional domains of SWI/SNF are associated with human cancer and developmental disorders (Bracken et al., 2019; Cenik and Shilatifard, 2021; Mashtalir et al., 2020; Sundaramoorthy and Owen-Hughes, 2020).

There are two main subclasses of SWI/SNF complexes that are conserved from yeast to man. The first includes yeast SWI/SNF, Drosophila BAP and mammalian BAF, while the second class includes yeast RSC, fly PBAP, and mammalian PBAF (Bracken et al., 2019; Clapier et al., 2017; Mohrmann and Verrijzer, 2005). Complexes of both classes share a common core, comprising related or identical subunits, associated with a set of signature subunits that define either SWI/SNF-BAF or RSC-PBAF. In mammalian cells, a third type of SWI/SNF complex has been described, named GBAF or non-canonical BAF (ncBAF; Alpsoy and Dykhuizen, 2018; Gatchalian et al., 2018; Mashtalir et al., 2018). Drosophila BAP and PBAP share a single ATPase, BRM, which is the ortholog of yeast Swi2/Snf2 and Sth1, and human SMARCA2 and SMARCA4. BRM associates with seven additional subunits to form the core of both BAP and PBAP. OSA (human ARID1A/B), D4/TTH (DPF1-3) and SS18 associate with the core subunits to form BAP, whereas POLYBROMO (PBRM1), BRD7, BAP170 (ARID2), and SAYP (PHF10) are specific for PBAF. The signature subunits play a major role in genomic targeting and functional selectivity of BAP and PBAP (Bracken et al., 2019; Chalkley et al., 2008; Mohrmann et al., 2004; Moshkin et al., 2007; Moshkin et al., 2012). Thus, SWI/SNF remodelers can be considered holoenzymes, in which different modules provide different functionalities that direct the remodeling activity of the ATPase. Although BAP and PBAP have shared activities, for example, both antagonize Polycomb repression, they also function in unique gene expression programs that control development, cell proliferation and differentiation (Chalkley et al., 2008; Mohrmann et al., 2004; Moshkin et al., 2007; Moshkin et al., 2012). Here, we use (P)BAP when making general statements that apply to both BAP and PBAP.

The natural amplification of Drosophila larval salivary gland polytene chromosomes allows the visualization of fluorescent-tagged transcription factors on interphase chromatin at native genetic loci in living cells. This has yielded fundamental insights in RNA polymerase II (RNAPII) recruitment and transcriptional dynamics (Yao et al., 2007; Zobeck et al., 2010). Salivary gland polytene chromosomes are the result of typically 10 rounds of DNA replication without segregation, resulting in cable-like super chromosomes comprising up to 1024 closely aligned chromatids. Interphase polytene chromosomes are visible by light microscopy and provided the first view of a physical genetic map (Bridges, 1935). The characteristic banding pattern of polytene chromosomes reflects the degree of interphase condensation, i.e., the ratio between the length of a stretched DNA molecule and the length of the corresponding chromosomal domain. The compacted polytene bands have been linked to transcriptionally repressed topologically associated domains (TADs), which are conserved in diploid cells (Eagen et al., 2015). Interbands contain gene regulatory regions, origins of replication, and are characterized by marks of active, open chromatin (Zykova et al., 2018). Polytene bands can be divided into moderately condensed gray chromatin, harboring the coding regions of active genes, and the highly compacted black chromatin that contains inactive developmental genes, and under-replicated, gene-poor intercalary heterochromatin (Eagen et al., 2015; Zykova et al., 2018). Very high levels of transcriptional activity results in puffing, the local uncoiling of individual chromosome strands due to the accumulation of the gene expression machinery and RNA. Pertinently, there is a close correspondence between chromatin structure of polytene chromosomes and that in diploid cells (Eagen et al., 2015; Zykova et al., 2018).

The steroid hormone 20-hydroxyecdysone (Ec) is the key regulator of Drosophila development and controls major transitions such as molting and metamorphosis (Hill et al., 2013). Exposure of larval salivary glands to Ec leads to the formation of early chromosomal puffs, starting a cascade of developmental gene expression and puffing. The Ec hormone mediates gene activation via binding to the Ec receptor (EcR), which associates with specific DNA elements. EcR belongs to the nuclear receptor family of transcription factors that mediate hormone-driven gene expression. Upon hormone binding, nuclear receptors cooperate with a slew of coactivators, including the SWI/SNF remodelers, to stimulate transcription (Hoffman et al., 2018; Paakinaho et al., 2017).

Previous live-cell microscopy of fluorescent-tagged remodelers in diploid cells suggested a fast exchange dynamic, similar to many transcription factors (Erdel et al., 2010; Johnson et al., 2008; Mehta et al., 2018; Paakinaho et al., 2017; Swinstead et al., 2016; Voss and Hager, 2014). However, the in vivo kinetics of remodeler interactions with specific natural genomic loci has not been studied. In particular, the role of ATP hydrolysis in remodeler kinetics remains unclear. Here, we used vital imaging to determine the spatial and temporal dynamics of (P)BAP interactions with endogenous target loci and the effect of chromatin condensation. Our results show that ATP-hydrolysis fuels a continuous probing of the genome by (P)BAP. In the absence of ATP-hydrolysis, there is increased chromatin retention and reduced turnover of (P)BAP. We discuss the implications of our in vivo findings for understanding remodeler function and chromatin dynamics.

We studied the role of ATP-hydrolysis during exploration of the genome by (P)BAP chromatin remodelers. We took advantage of the parallel amplification of the Drosophila genome in salivary gland polytene chromosomes to monitor (P)BAP interactions with endogenous targets in real time. This allowed us to distinguish between nucleoplasm and natural loci in interphase chromosomes by live-cell imaging, which is not possible in diploid cells. We found that a surprisingly small portion of (P)BAP (~5%) is associated with chromatin at any given time. This is caused by the coupling of ATP-hydrolysis during nucleosome remodeling to (P)BAP release. Consequently, (P)BAP acts through a continuous transient probing of the genome (Figure 9A). (P)BAP mainly interacts with chromatin in non-condensed interbands or puffs. Conversely, (P)BAP levels within highly condensed polytene bands are much lower than in the nucleoplasm. We note that the loss of (P)BAP does not affect chromatin condensation and the polytene banding pattern (Figure 1—figure supplement 1A). Based on these observations, we postulate that chromatin condensation excludes (P)BAP, thereby reducing its genomic search space. At the height of E74 and E75 transcription, a substantial portion of RNAPII is retained locally in what Lis and colleagues named a transcription compartment (Yao et al., 2007). On the same chromatin puffs, however, PBAP exchanges rapidly (within ~2–4 s) without a substantial immobile fraction. Thus, although essential for expression of E74 and E75, PBAP is not retained within a transcription compartment. The rapid exchange of (P)BAP indicates continuous cycles of nucleosome remodeling. Transient binding to chromatin might explain the relatively poor capture of remodelers by ChIP, as the crosslinking reaction has temporal constraints (Gelbart et al., 2005; Schmiedeberg et al., 2009). Indeed, a mutation in the ATP-binding pocket increased the residence time on chromatin and concomitantly enhanced ChIP signals of BRM-K804R. Consistent with our earlier developmental genetic analysis (Chalkley et al., 2008), we found that PBAP, but not BAP, is required for Ec-regulated gene expression. This finding expands the functional diversification between BAP and PBAP remodeling complexes, which is to a large extent determined by their signature subunits (Mohrmann et al., 2004; Moshkin et al., 2007; Chalkley et al., 2008; Moshkin et al., 2012). We detected the fly homolog of GLTSCR1 in GFP-BRM IPs from S2 cells (Figure 4B). Indeed, a recent study identified and characterized the fly BRM complex that is orthologous to mammalian GBAF/ncBAF (Barish et al., 2020). Thus, although GBAP appears to be mainly expressed in the nervous system, it might contribute to our analysis of BRM dynamics. We note that the dynamic characteristics of SNR1, D4 and BRD7, which are all absent from GBAP, is highly similar to that of BRM.

Figure 9

Download asset Open asset

Histones displayed only limited exchange, which was more than 2 orders of magnitude slower than that of (P)BAP. FRAP analysis of GFP-H2B in polytene chromosomes showed that it takes about 40 min to obtain a ~18% recovery (Figure 5H). These results agree with the notion that histone eviction is restricted to small regulatory regions of the genome and is associated with the placement of variant histones (Brahma and Henikoff, 2020; Cakiroglu et al., 2019; Deal et al., 2010; Pillidge and Bray, 2019). Indeed, it is difficult to envision how the limited amount of free canonical histones available in the nucleoplasm could support substantial turnover of chromosomal histones. A comparison of DNA accessibility and histone occupancy during transcriptional induction revealed that changes in the former predominate the latter (Mueller et al., 2017). Likewise, several recent studies indicated that promoter chromatin harbors so called fragile nucleosomes with accessible DNA, which depend on chromatin remodeler activity (Brahma and Henikoff, 2020). These observations suggest that the prevalent outcome of chromatin remodeling in vivo is restructuring or sliding of nucleosomes rather than eviction.

Vital imaging of fluorophore-tagged histones recapitulated the polytene banding pattern and local chromatin condensation. This enabled us to match the GFP-H2B density distribution in living cells with chromatin states derived from histone marks and DNA accessibility (Eagen et al., 2015; Filion et al., 2010; Kharchenko et al., 2011; Zykova et al., 2018). Our analysis indicated that DNA in gray bands is on average ~6-fold more condensed than that in interbands, whereas heterochromatic DNA can be up to ~150-fold more condensed. The level of gray band condensation measured by live-cell imaging agrees well with the value derived from HiC analysis of diploid Kc167 cells (Eagen et al., 2015), whereas our imaging-based estimate for black chromatin condensation is ~5-fold higher. We suspect that this discrepancy is caused by the very high degree of chromatin condensation at the chromocenter and telomers of polytene chromosomes. Indeed, measurements of black chromatin bands flanking the E74 and E75 loci, suggested a ~25-fold condensation compared to open chromatin, which falls in the range derived from HiC in diploid cells (Figure 3E; Eagen et al., 2015). Our real time measurements of histone density in polytene chromosomes in living cells complements other methods to estimate condensation that all depend on chromatin fixation (Eagen et al., 2015; Ou et al., 2017; Zykova et al., 2018). Collectively, these studies suggest that euchromatin and heterochromatin are formed by the same basic nucleosomal array, but at different degrees of condensation.

Although central to the function of Snf2 remodelers, our understanding of the role of ATP hydrolysis in the nucleosome remodeling cycle remains incomplete. A mutation in the ATP-binding pocket of BRM caused up to 5-fold stronger chromatin binding and a ~2-times slower turnover of BRM-K804R in S2 cells. In polytene cells, BRM-K804R clamped down onto chromatin and displayed no measurable turnover. Likewise, in permeabilized salivary gland cells that lack ATP, wt BRM accumulated onto polytene chromosome interbands. Chromatin binding persisted for over 1 hr, but following the addition of ATP, BRM released immediately. Most likely, the high local density of target loci in polytene chromosomes amplifies the effect of prolonged chromatin binding by BRM. We speculate that in the absence of ATP hydrolysis, slow release combined with rapid re-binding to a paired chromatid, might results in the local trapping of (P)BAP on polytene chromosomes. We note that under normal physiological conditions, ATP concentrations in cells would not become depleted to a level (< 1 mM) that would impair remodeler function. Finally, our results argue against a role for liquid-liquid phase separation in the clustering of (P)BAP in the absence of ATP-hydrolysis. Local (P)BAP accumulation is insensitive to the disruption of weak hydrophobic interactions by 1,6-hexanediol. Moreover, FRAP analysis showed that clustered BRM has a strongly reduced mobility, indicating it does not form a condensate. We consider prolonged chromatin-binding the most plausible cause of local accumulation of BRM-K804R, or wt BRM in the absence of ATP.

Studies in diploid cells often equate the formation of low-mobility aggregates of a fluorophore-tagged transcription factor with persistent chromatin binding and functional activity. However, given that chromatin typically only occupies about 3% of the nuclear volume, this interpretation remains speculative. E.g., in a study on human ISWI remodelers, an increase in the immobile fraction upon DNA damage was rationalized as an increase in activity on chromatin (Erdel et al., 2010). Moreover, it was concluded that loss of ATP-binding reduced chromatin association and increased mobility. In contrast, our results with BRM showed that chromatin binding is independent of ATP, but that its release is stimulated by ATP-hydrolysis. At the peak of PBAP- and EcR-dependent transcription of E74 and E75 genes, both retained a rapid exchange, indicating ceaseless cycles of nucleosome remodeling. Consistent with this notion, the use of small-molecule inhibitors recently showed that chromatin accessibility required continuous remodeler activity (Iurlaro et al., 2021). Transient remodeler dynamics enables competition at regulatory elements between epigenetic regulators with opposing effects on chromatin structure (Bracken et al., 2019; Cenik and Shilatifard, 2021; Kubik et al., 2019; Mohd-Sarip et al., 2017).

There are both parallels and differences in the roles of ATP in either chromatin remodeling or in RNA-duplex unwinding by DEAD-box helicases. ATP-hydrolysis is dispensable for RNA-binding and duplex unwinding by DEAD-box helicases, but is required for fast enzyme release from RNA (Liu et al., 2014). However, the formation of long-lived RNA-DEAD-box helicase complexes requires ATP-binding but not its hydrolysis. In contrast, chromatin binding by BRM is independent of ATP. The ATP hydrolysis-fueled transient probing of the genome by BRM we described here, dovetails well with structural observations on Swi2/Snf2 remodelers (Xia et al., 2016). When Snf2 is free in solution, the two domains that harbor the ATP-binding motifs are oriented away from one another and Snf2 cannot bind ATP. Binding to naked DNA or nucleosomal DNA induces a reorientation of the two ATPase lobes so that they now form a functional ATP-binding pocket. Combined with our results, these structural observations suggest a model in which nucleosome binding occurs independent of ATP but induces the formation of an ATP-binding pocket (Figure 9B). ATP hydrolysis can now fuel nucleosome remodeling coupled to an increased probability of remodeler release, resulting in rapid cycles of nucleosome remodeling and remodeler recycling. The absence of ATP-hydrolysis leads to prolonged chromatin binding of the remodeler and impedes nucleosome remodeling.

Our observation that ATP hydrolysis promotes remodeler release from chromatin raises the question of how many translocation steps a single remodeler takes before dissociation from a nucleosome in vivo. Previous single molecule studies on the RSC remodeler determined that an ATP concentration above 1 mM is required for optimal translocation speed and processivity (Zhang et al., 2006; Sirinakis et al., 2011). This value derived from in vitro experiments is remarkably similar to our in vivo estimates. Note that many biophysical studies on remodelers use much lower ATP concentrations to slow down the reaction speed and determine individual kinetic steps (see e.g. Harada et al., 2016). In the presence of 1 mM ATP in vitro, RSC translocated at a speed of 25 bp/s on naked DNA and ~13 bp/s on nucleosomes (Zhang et al., 2006; Sirinakis et al., 2011). Enzyme processivity in these experiments was ~35 bp on naked DNA and (albeit it with a widespread) on average ~100 bp per round of translocation on nucleosomes. Stopped-flow pre-steady state kinetic analysis of RSC translocation along naked DNA suggested a series of uniform steps yielding an average translocation distance of 20 bp before dissociation (Fischer et al., 2007). The (P)BAP turnover time of ~2–4 s in our in vivo FRAP experiments agrees well with the in vitro single molecule analyses and is compatible with multiple kinetic steps and substantial nucleosome remodeling prior to remodeler disengagement. Remodeler release might be considered as a probability during each ATPase cycle. Full remodeling, sliding or even eviction may require the cooperation of a cascade of remodelers acting subsequently.

BRM-K804R clamps to chromatin and impedes histone turnover, emphasizing that remodeling and remodeler recycling are closely integrated. Surprisingly, BRM-K804R also traps wt BRM onto polytene chromosomes. Thus, reflecting early genetic studies (Elfring et al., 1998), BRM-K804R behaves as a molecular dominant mutant. The retention of wt BRM onto chromatin by BRM-K804R suggests that BRM complexes might interact on chromatin. Consequently, remodeling might involve a nucleosome hand-off mechanism between different BRM complexes. Additionally, translocation by one BRM complex might help to push another one off the chromatin. This way, the persistent chromatin binding of BRM-K804R might create a ‘traffic jam’ that blocks the release of wt BRM.

In summary, our visualization of BRM remodelers engaging their endogenous genomic targets revealed highly transient and dynamic interactions. (P)BAP binds chromatin independent of ATP. However, ATP-hydrolysis couples nucleosome remodeling to (P)BAP release, resulting in rapid cycles of remodeling and remodeler turnover. Remodelers might interact on chromatin, as suggested by the trapping of wt BRM by the ATP-binding mutant BRM-K804R. Importantly, an accompanying report by Wu and colleagues, studying budding yeast remodelers through live-cell single molecule tracking, also revealed the key role of the ATPase in promoting dissociation from chromatin and fast remodeler kinetics (Kim et al., 2021). Collectively, these in vivo results provide a framework for understanding remodeler function in genome regulation.

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://www.proteomexchange.org/) via the PRIDE partner repository with the dataset identifier PXD025474. All data generated or analyzed during this study are included in the manuscript and supporting files. Source data files have been provided for Figures 1, 2 and Figure 2—figure supplement 1.

1. 1. Verrijzer CP

   (2021) ProteomeXchange

   ID PXD025474. In vivo analysis reveals that ATP-hydrolysis couples remodeling to SWI/SNF release from chromatin.

   http://proteomecentral.proteomexchange.org/cgi/GetDataset?ID=PXD025474

1. 1. Alpsoy A
   2. Dykhuizen EC

   (2018) Glioma tumor suppressor candidate region gene 1 (GLTSCR1) and its paralog GLTSCR1-like form SWI/SNF chromatin remodeling subcomplexes

   Journal of Biological Chemistry 293:3892–3903.

   https://doi.org/10.1074/jbc.RA117.001065

   • Google Scholar
2. 1. Armstrong JA
   2. Papoulas O
   3. Daubresse G
   4. Sperling AS
   5. Lis JT
   6. Scott MP
   7. Tamkun JW

   (2002) The Drosophila BRM complex facilitates global transcription by RNA polymerase II

   The EMBO Journal 21:5245–5254.

   https://doi.org/10.1093/emboj/cdf517

   • PubMed
   • Google Scholar
3. 1. Barish S
   2. Barakat TS
   3. Michel BC
   4. Mashtalir N
   5. Phillips JB
   6. Valencia AM
   7. Ugur B
   8. Wegner J
   9. Scott TM
   10. Bostwick B
   11. Murdock DR
   12. Dai H
   13. Perenthaler E
   14. Nikoncuk A
   15. van Slegtenhorst M
   16. Brooks AS
   17. Keren B
   18. Nava C
   19. Mignot C
   20. Douglas J
   21. Rodan L
   22. Nowak C
   23. Ellard S
   24. Stals K
   25. Lynch SA
   26. Faoucher M
   27. Lesca G
   28. Edery P
   29. Engleman KL
   30. Zhou D
   31. Thiffault I
   32. Herriges J
   33. Gass J
   34. Louie RJ
   35. Stolerman E
   36. Washington C
   37. Vetrini F
   38. Otsubo A
   39. Pratt VM
   40. Conboy E
   41. Treat K
   42. Shannon N
   43. Camacho J
   44. Wakeling E
   45. Yuan B
   46. Chen CA
   47. Rosenfeld JA
   48. Westerfield M
   49. Wangler M
   50. Yamamoto S
   51. Kadoch C
   52. Scott DA
   53. Bellen HJ
   54. Undiagnosed Diseases Network

   (2020) BICRA, a SWI/SNF complex member, is associated with BAF-Disorder related phenotypes in humans and model organisms

   The American Journal of Human Genetics 107:1096–1112.

   https://doi.org/10.1016/j.ajhg.2020.11.003

   • PubMed
   • Google Scholar
4. 1. Becker PB
   2. Workman JL

   (2013) Nucleosome remodeling and epigenetics

   Cold Spring Harbor Perspectives in Biology 5:a017905.

   https://doi.org/10.1101/cshperspect.a017905

   • PubMed
   • Google Scholar
5. 1. Bernardo TJ
   2. Dubrovskaya VA
   3. Xie X
   4. Dubrovsky EB

   (2014) A view through a chromatin loop: insights into the ecdysone activation of early genes in Drosophila

   Nucleic Acids Research 42:10409–10424.

   https://doi.org/10.1093/nar/gku754

   • PubMed
   • Google Scholar
6. 1. Bracken AP
   2. Brien GL
   3. Verrijzer CP

   (2019) Dangerous liaisons: interplay between SWI/SNF, NuRD, and polycomb in chromatin regulation and Cancer

   Genes & Development 33:936–959.

   https://doi.org/10.1101/gad.326066.119

   • PubMed
   • Google Scholar
7. 1. Brahma S
   2. Henikoff S

   (2020) Epigenome regulation by dynamic nucleosome unwrapping

   Trends in Biochemical Sciences 45:13–26.

   https://doi.org/10.1016/j.tibs.2019.09.003

   • PubMed
   • Google Scholar
8. 1. Bridges CA

   (1935)

   Salivary chromosome maps: with a key to the banding of the chromosomes of Drosophila Melanogaster

   The Journal of Heredity 26:60–64.

   • Google Scholar
9. 1. Cakiroglu A
   2. Clapier CR
   3. Ehrensberger AH
   4. Darbo E
   5. Cairns BR
   6. Luscombe NM
   7. Svejstrup JQ

   (2019) Genome-wide reconstitution of chromatin transactions reveals that RSC preferentially disrupts H2AZ-containing nucleosomes

   Genome Research 29:988–998.

   https://doi.org/10.1101/gr.243139.118

   • PubMed
   • Google Scholar
10. 1. Cenik BK
    2. Shilatifard A

    (2021) COMPASS and SWI/SNF complexes in development and disease

    Nature Reviews Genetics 22:38–58.

    https://doi.org/10.1038/s41576-020-0278-0

    • PubMed
    • Google Scholar
11. 1. Chalkley GE
    2. Moshkin YM
    3. Langenberg K
    4. Bezstarosti K
    5. Blastyak A
    6. Gyurkovics H
    7. Demmers JA
    8. Verrijzer CP

    (2008) The transcriptional coactivator SAYP is a trithorax group signature subunit of the PBAP chromatin remodeling complex

    Molecular and Cellular Biology 28:2920–2929.

    https://doi.org/10.1128/MCB.02217-07

    • PubMed
    • Google Scholar
12. 1. Chalkley GE
    2. Verrijzer CP

    (2004) Immuno-depletion and purification strategies to study chromatin-remodeling factors in vitro

    Methods in Enzymology 377:421–442.

    https://doi.org/10.1016/S0076-6879(03)77028-1

    • PubMed
    • Google Scholar
13. 1. Clapier CR
    2. Iwasa J
    3. Cairns BR
    4. Peterson CL

    (2017) Mechanisms of action and regulation of ATP-dependent chromatin-remodelling complexes

    Nature Reviews Molecular Cell Biology 18:407–422.

    https://doi.org/10.1038/nrm.2017.26

    • PubMed
    • Google Scholar
14. 1. Cox J
    2. Matic I
    3. Hilger M
    4. Nagaraj N
    5. Selbach M
    6. Olsen JV
    7. Mann M

    (2009) A practical guide to the MaxQuant computational platform for SILAC-based quantitative proteomics

    Nature Protocols 4:698–705.

    https://doi.org/10.1038/nprot.2009.36

    • PubMed
    • Google Scholar
15. 1. Deal RB
    2. Henikoff JG
    3. Henikoff S

    (2010) Genome-wide kinetics of nucleosome turnover determined by metabolic labeling of histones

    Science 328:1161–1164.

    https://doi.org/10.1126/science.1186777

    • PubMed
    • Google Scholar
16. 1. Dietzel S
    2. Niemann H
    3. Brückner B
    4. Maurange C
    5. Paro R

    (1999) The nuclear distribution of polycomb during Drosophila Melanogaster development shown with a GFP fusion protein

    Chromosoma 108:83–94.

    https://doi.org/10.1007/s004120050355

    • PubMed
    • Google Scholar
17. 1. Eagen KP
    2. Hartl TA
    3. Kornberg RD

    (2015) Stable chromosome condensation revealed by chromosome conformation capture

    Cell 163:934–946.

    https://doi.org/10.1016/j.cell.2015.10.026

    • PubMed
    • Google Scholar
18. 1. Elfring LK
    2. Daniel C
    3. Papoulas O
    4. Deuring R
    5. Sarte M
    6. Moseley S
    7. Beek SJ
    8. Waldrip WR
    9. Daubresse G
    10. DePace A
    11. Kennison JA
    12. Tamkun JW

    (1998) Genetic analysis of brahma: the Drosophila homolog of the yeast chromatin remodeling factor SWI2/SNF2

    Genetics 148:251–265.

    https://doi.org/10.1093/genetics/148.1.251

    • PubMed
    • Google Scholar
19. 1. Erdel F
    2. Schubert T
    3. Marth C
    4. Längst G
    5. Rippe K

    (2010) Human ISWI chromatin-remodeling complexes sample nucleosomes via transient binding reactions and become immobilized at active sites

    PNAS 107:19873–19878.

    https://doi.org/10.1073/pnas.1003438107

    • PubMed
    • Google Scholar
20. 1. Filion GJ
    2. van Bemmel JG
    3. Braunschweig U
    4. Talhout W
    5. Kind J
    6. Ward LD
    7. Brugman W
    8. de Castro IJ
    9. Kerkhoven RM
    10. Bussemaker HJ
    11. van Steensel B

    (2010) Systematic protein location mapping reveals five principal chromatin types in Drosophila cells

    Cell 143:212–224.

    https://doi.org/10.1016/j.cell.2010.09.009

    • PubMed
    • Google Scholar
21. 1. Fischer CJ
    2. Saha A
    3. Cairns BR

    (2007) Kinetic model for the ATP-dependent translocation of Saccharomyces cerevisiae RSC along double-stranded DNA

    Biochemistry 46:12416–12426.

    https://doi.org/10.1021/bi700930n

    • PubMed
    • Google Scholar
22. 1. Gatchalian J
    2. Malik S
    3. Ho J
    4. Lee DS
    5. Kelso TWR
    6. Shokhirev MN
    7. Dixon JR
    8. Hargreaves DC

    (2018) A non-canonical BRD9-containing BAF chromatin remodeling complex regulates naive pluripotency in mouse embryonic stem cells

    Nature Communications 9:5139.

    https://doi.org/10.1038/s41467-018-07528-9

    • PubMed
    • Google Scholar
23. 1. Gelbart ME
    2. Bachman N
    3. Delrow J
    4. Boeke JD
    5. Tsukiyama T

    (2005) Genome-wide identification of Isw2 chromatin-remodeling targets by localization of a catalytically inactive mutant

    Genes & Development 19:942–954.

    https://doi.org/10.1101/gad.1298905

    • PubMed
    • Google Scholar
24. 1. Han Y
    2. Reyes AA
    3. Malik S
    4. He Y

    (2020) Cryo-EM structure of SWI/SNF complex bound to a nucleosome

    Nature 579:452–455.

    https://doi.org/10.1038/s41586-020-2087-1

    • PubMed
    • Google Scholar
25. 1. Harada BT
    2. Hwang WL
    3. Deindl S
    4. Chatterjee N
    5. Bartholomew B
    6. Zhuang X

    (2016) Stepwise nucleosome translocation by RSC remodeling complexes

    eLife 5:e10051.

    https://doi.org/10.7554/eLife.10051

    • PubMed
    • Google Scholar
26. Book

    1. Harlow E
    2. Lane D

    (1998)

    Antibodies: A Laboratory Manual

    NY: Cold Spring Harbor Laboratory Press.

    • Google Scholar
27. 1. He S
    2. Wu Z
    3. Tian Y
    4. Yu Z
    5. Yu J
    6. Wang X
    7. Li J
    8. Liu B
    9. Xu Y

    (2020) Structure of nucleosome-bound human BAF complex

    Science 367:875–881.

    https://doi.org/10.1126/science.aaz9761

    • PubMed
    • Google Scholar
28. 1. Hill RJ
    2. Billas IM
    3. Bonneton F
    4. Graham LD
    5. Lawrence MC

    (2013) Ecdysone receptors: from the Ashburner model to structural biology

    Annual Review of Entomology 58:251–271.

    https://doi.org/10.1146/annurev-ento-120811-153610

    • PubMed
    • Google Scholar
29. 1. Hodges HC
    2. Stanton BZ
    3. Cermakova K
    4. Chang CY
    5. Miller EL
    6. Kirkland JG
    7. Ku WL
    8. Veverka V
    9. Zhao K
    10. Crabtree GR

    (2018) Dominant-negative SMARCA4 mutants alter the accessibility landscape of tissue-unrestricted enhancers

    Nature Structural & Molecular Biology 25:61–72.

    https://doi.org/10.1038/s41594-017-0007-3

    • PubMed
    • Google Scholar
30. 1. Hoffman JA
    2. Trotter KW
    3. Ward JM
    4. Archer TK

    (2018) BRG1 governs glucocorticoid receptor interactions with chromatin and pioneer factors across the genome

    eLife 7:e35073.

    https://doi.org/10.7554/eLife.35073

    • PubMed
    • Google Scholar
31. 1. Iurlaro M
    2. Stadler MB
    3. Masoni F
    4. Jagani Z
    5. Galli GG
    6. Schübeler D

    (2021)

    Mammalian SWI/SNF continuously restores local accessibility to chromatin

    Nature Genetics 8:279–287.

    • Google Scholar
32. 1. Johnson TA
    2. Elbi C
    3. Parekh BS
    4. Hager GL
    5. John S

    (2008) Chromatin remodeling complexes interact dynamically with a glucocorticoid receptor-regulated promoter

    Molecular Biology of the Cell 19:3308–3322.

    https://doi.org/10.1091/mbc.e08-02-0123

    • PubMed
    • Google Scholar
33. 1. Jungblut A
    2. Hopfner KP
    3. Eustermann S

    (2020) Megadalton chromatin remodelers: common principles for versatile functions

    Current Opinion in Structural Biology 64:134–144.

    https://doi.org/10.1016/j.sbi.2020.06.024

    • PubMed
    • Google Scholar
34. 1. Kharchenko PV
    2. Alekseyenko AA
    3. Schwartz YB
    4. Minoda A
    5. Riddle NC
    6. Ernst J
    7. Sabo PJ
    8. Larschan E
    9. Gorchakov AA
    10. Gu T
    11. Linder-Basso D
    12. Plachetka A
    13. Shanower G
    14. Tolstorukov MY
    15. Luquette LJ
    16. Xi R
    17. Jung YL
    18. Park RW
    19. Bishop EP
    20. Canfield TK
    21. Sandstrom R
    22. Thurman RE
    23. MacAlpine DM
    24. Stamatoyannopoulos JA
    25. Kellis M
    26. Elgin SC
    27. Kuroda MI
    28. Pirrotta V
    29. Karpen GH
    30. Park PJ

    (2011) Comprehensive analysis of the chromatin landscape in Drosophila melanogaster

    Nature 471:480–485.

    https://doi.org/10.1038/nature09725

    • PubMed
    • Google Scholar
35. 1. Kim JM
    2. Visanpattanasin P
    3. Jou V
    4. Liu S
    5. Tang X
    6. Zheng Q
    7. Li KY
    8. Snedeker J
    9. Lavis LD
    10. Lionnet T
    11. Wu C

    (2021) Single-molecule imaging of chromatin remodelers reveals role of ATPase in promoting fast kinetics of target search and dissociation from chromatin

    eLife 10:e69387.

    https://doi.org/10.7554/eLife.69387

    • Google Scholar
36. 1. Koulouras G
    2. Panagopoulos A
    3. Rapsomaniki MA
    4. Giakoumakis NN
    5. Taraviras S
    6. Lygerou Z

    (2018) EasyFRAP-web: a web-based tool for the analysis of fluorescence recovery after photobleaching data

    Nucleic Acids Research 46:W467–W472.

    https://doi.org/10.1093/nar/gky508

    • PubMed
    • Google Scholar
37. 1. Kubik S
    2. Bruzzone MJ
    3. Challal D
    4. Dreos R
    5. Mattarocci S
    6. Bucher P
    7. Libri D
    8. Shore D

    (2019) Opposing chromatin remodelers control transcription initiation frequency and start site selection

    Nature Structural & Molecular Biology 26:744–754.

    https://doi.org/10.1038/s41594-019-0273-3

    • PubMed
    • Google Scholar
38. 1. Liu F
    2. Putnam AA
    3. Jankowsky E

    (2014) DEAD-box helicases form nucleotide-dependent, long-lived complexes with RNA

    Biochemistry 53:423–433.

    https://doi.org/10.1021/bi401540q

    • PubMed
    • Google Scholar
39. 1. Livak KJ
    2. Schmittgen TD

    (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta delta C(T)) Method

    Methods 25:402–408.

    https://doi.org/10.1006/meth.2001.1262

    • PubMed
    • Google Scholar
40. 1. Mashtalir N
    2. D'Avino AR
    3. Michel BC
    4. Luo J
    5. Pan J
    6. Otto JE
    7. Zullow HJ
    8. McKenzie ZM
    9. Kubiak RL
    10. St Pierre R
    11. Valencia AM
    12. Poynter SJ
    13. Cassel SH
    14. Ranish JA
    15. Kadoch C

    (2018) Modular organization and assembly of SWI/SNF family chromatin remodeling complexes

    Cell 175:1272–1288.

    https://doi.org/10.1016/j.cell.2018.09.032

    • PubMed
    • Google Scholar
41. 1. Mashtalir N
    2. Suzuki H
    3. Farrell DP
    4. Sankar A
    5. Luo J
    6. Filipovski M
    7. D'Avino AR
    8. St Pierre R
    9. Valencia AM
    10. Onikubo T
    11. Roeder RG
    12. Han Y
    13. He Y
    14. Ranish JA
    15. DiMaio F
    16. Walz T
    17. Kadoch C

    (2020) A structural model of the endogenous human BAF complex informs disease mechanisms

    Cell 183:802–817.

    https://doi.org/10.1016/j.cell.2020.09.051

    • PubMed
    • Google Scholar
42. 1. McSwiggen DT
    2. Mir M
    3. Darzacq X
    4. Tjian R

    (2019) Evaluating phase separation in live cells: diagnosis, caveats, and functional consequences

    Genes & Development 33:1619–1634.

    https://doi.org/10.1101/gad.331520.119

    • PubMed
    • Google Scholar
43. 1. Mehta GD
    2. Ball DA
    3. Eriksson PR
    4. Chereji RV
    5. Clark DJ
    6. McNally JG
    7. Karpova TS

    (2018) Single-Molecule analysis reveals linked cycles of RSC chromatin remodeling and Ace1p transcription factor binding in yeast

    Molecular Cell 72:875–887.

    https://doi.org/10.1016/j.molcel.2018.09.009

    • PubMed
    • Google Scholar
44. 1. Mohd-Sarip A
    2. Teeuwssen M
    3. Bot AG
    4. De Herdt MJ
    5. Willems SM
    6. Baatenburg de Jong RJ
    7. Looijenga LHJ
    8. Zatreanu D
    9. Bezstarosti K
    10. van Riet J
    11. Oole E
    12. van Ijcken WFJ
    13. van de Werken HJG
    14. Demmers JA
    15. Fodde R
    16. Verrijzer CP

    (2017) DOC1-Dependent recruitment of NURD reveals antagonism with SWI/SNF during Epithelial-Mesenchymal transition in oral Cancer cells

    Cell Reports 20:61–75.

    https://doi.org/10.1016/j.celrep.2017.06.020

    • PubMed
    • Google Scholar
45. 1. Mohrmann L
    2. Langenberg K
    3. Krijgsveld J
    4. Kal AJ
    5. Heck AJ
    6. Verrijzer CP

    (2004) Differential targeting of two distinct SWI/SNF-related Drosophila chromatin-remodeling complexes

    Molecular and Cellular Biology 24:3077–3088.

    https://doi.org/10.1128/MCB.24.8.3077-3088.2004

    • PubMed
    • Google Scholar
46. 1. Mohrmann L
    2. Verrijzer CP

    (2005) Composition and functional specificity of SWI2/SNF2 class chromatin remodeling complexes

    Biochimica Et Biophysica Acta (BBA) - Gene Structure and Expression 1681:59–73.

    https://doi.org/10.1016/j.bbaexp.2004.10.005

    • Google Scholar
47. 1. Moshkin YM
    2. Mohrmann L
    3. van Ijcken WF
    4. Verrijzer CP

    (2007) Functional differentiation of SWI/SNF remodelers in transcription and cell cycle control

    Molecular and Cellular Biology 27:651–661.

    https://doi.org/10.1128/MCB.01257-06

    • PubMed
    • Google Scholar
48. 1. Moshkin YM
    2. Chalkley GE
    3. Kan TW
    4. Reddy BA
    5. Ozgur Z
    6. van Ijcken WF
    7. Dekkers DH
    8. Demmers JA
    9. Travers AA
    10. Verrijzer CP

    (2012) Remodelers organize cellular chromatin by counteracting intrinsic histone-DNA sequence preferences in a class-specific manner

    Molecular and Cellular Biology 32:675–688.

    https://doi.org/10.1128/MCB.06365-11

    • PubMed
    • Google Scholar
49. 1. Mueller B
    2. Mieczkowski J
    3. Kundu S
    4. Wang P
    5. Sadreyev R
    6. Tolstorukov MY
    7. Kingston RE

    (2017) Widespread changes in nucleosome accessibility without changes in nucleosome occupancy during a rapid transcriptional induction

    Genes & Development 31:451–462.

    https://doi.org/10.1101/gad.293118.116

    • PubMed
    • Google Scholar
50. 1. Mutskov V
    2. Felsenfeld G

    (2004) Silencing of transgene transcription precedes methylation of promoter DNA and histone H3 lysine 9

    The EMBO Journal 23:138–149.

    https://doi.org/10.1038/sj.emboj.7600013

    • PubMed
    • Google Scholar
51. 1. Myohara M
    2. Okada M

    (1987) Induction of ecdysterone-stimulated chromosomal puffs in permeabilized Drosophila salivary glands: a new method for assaying the gene-regulating activity of cytoplasm

    Developmental Biology 122:396–406.

    https://doi.org/10.1016/0012-1606(87)90304-6

    • Google Scholar
52. 1. Ou HD
    2. Phan S
    3. Deerinck TJ
    4. Thor A
    5. Ellisman MH
    6. O'Shea CC

    (2017) ChromEMT: visualizing 3D chromatin structure and compaction in interphase and mitotic cells

    Science 357:eaag0025.

    https://doi.org/10.1126/science.aag0025

    • PubMed
    • Google Scholar
53. 1. Paakinaho V
    2. Presman DM
    3. Ball DA
    4. Johnson TA
    5. Schiltz RL
    6. Levitt P
    7. Mazza D
    8. Morisaki T
    9. Karpova TS
    10. Hager GL

    (2017) Single-molecule analysis of steroid receptor and cofactor action in living cells

    Nature Communications 8:15896.

    https://doi.org/10.1038/ncomms15896

    • PubMed
    • Google Scholar
54. 1. Pan J
    2. McKenzie ZM
    3. D'Avino AR
    4. Mashtalir N
    5. Lareau CA
    6. St Pierre R
    7. Wang L
    8. Shilatifard A
    9. Kadoch C

    (2019) The ATPase module of mammalian SWI/SNF family complexes mediates subcomplex identity and catalytic activity-independent genomic targeting

    Nature Genetics 51:618–626.

    https://doi.org/10.1038/s41588-019-0363-5

    • PubMed
    • Google Scholar
55. 1. Patel AB
    2. Moore CM
    3. Greber BJ
    4. Luo J
    5. Zukin SA
    6. Ranish J
    7. Nogales E

    (2019) Architecture of the chromatin remodeler RSC and insights into its nucleosome engagement

    eLife 8:e54449.

    https://doi.org/10.7554/eLife.54449

    • PubMed
    • Google Scholar
56. 1. Pillidge Z
    2. Bray SJ

    (2019) SWI/SNF chromatin remodeling controls Notch-responsive enhancer accessibility

    EMBO Reports 20:e46944.

    https://doi.org/10.15252/embr.201846944

    • PubMed
    • Google Scholar
57. 1. Plys AJ
    2. Kingston RE

    (2018) Dynamic condensates activate transcription

    Science 361:329–330.

    https://doi.org/10.1126/science.aau4795

    • PubMed
    • Google Scholar
58. 1. Schmiedeberg L
    2. Skene P
    3. Deaton A
    4. Bird A

    (2009) A temporal threshold for formaldehyde crosslinking and fixation

    PLOS ONE 4:e4636.

    https://doi.org/10.1371/journal.pone.0004636

    • PubMed
    • Google Scholar
59. 1. Sirinakis G
    2. Clapier CR
    3. Gao Y
    4. Viswanathan R
    5. Cairns BR
    6. Zhang Y

    (2011) The RSC chromatin remodelling ATPase translocates DNA with high force and small step size

    The EMBO Journal 30:2364–2372.

    https://doi.org/10.1038/emboj.2011.141

    • PubMed
    • Google Scholar
60. 1. Strom AR
    2. Emelyanov AV
    3. Mir M
    4. Fyodorov DV
    5. Darzacq X
    6. Karpen GH

    (2017) Phase separation drives heterochromatin domain formation

    Nature 547:241–245.

    https://doi.org/10.1038/nature22989

    • PubMed
    • Google Scholar
61. 1. Sundaramoorthy R
    2. Owen-Hughes T

    (2020) Chromatin remodelling comes into focus

    F1000Research 9:1011.

    https://doi.org/10.12688/f1000research.21933.1

    • Google Scholar
62. 1. Swinstead EE
    2. Paakinaho V
    3. Presman DM
    4. Hager GL

    (2016) Pioneer factors and ATP-dependent chromatin remodeling factors interact dynamically: a new perspective: multiple transcription factors can effect chromatin pioneer functions through dynamic interactions with ATP-dependent chromatin remodeling factors

    BioEssays : News and Reviews in Molecular, Cellular and Developmental Biology 38:1150–1157.

    https://doi.org/10.1002/bies.201600137

    • PubMed
    • Google Scholar
63. 1. Szabo Q
    2. Bantignies F
    3. Cavalli G

    (2019) Principles of genome folding into topologically associating domains

    Science Advances 5:eaaw1668.

    https://doi.org/10.1126/sciadv.aaw1668

    • PubMed
    • Google Scholar
64. 1. Voss TC
    2. Hager GL

    (2014) Dynamic regulation of transcriptional states by chromatin and transcription factors

    Nature Reviews Genetics 15:69–81.

    https://doi.org/10.1038/nrg3623

    • PubMed
    • Google Scholar
65. 1. Wagner FR
    2. Dienemann C
    3. Wang H
    4. Stützer A
    5. Tegunov D
    6. Urlaub H
    7. Cramer P

    (2020) Structure of SWI/SNF chromatin remodeller RSC bound to a nucleosome

    Nature 579:448–451.

    https://doi.org/10.1038/s41586-020-2088-0

    • PubMed
    • Google Scholar
66. 1. Xia X
    2. Liu X
    3. Li T
    4. Fang X
    5. Chen Z

    (2016) Structure of chromatin remodeler Swi2/Snf2 in the resting state

    Nature Structural & Molecular Biology 23:722–729.

    https://doi.org/10.1038/nsmb.3259

    • PubMed
    • Google Scholar
67. 1. Yao J
    2. Ardehali MB
    3. Fecko CJ
    4. Webb WW
    5. Lis JT

    (2007) Intranuclear distribution and local dynamics of RNA polymerase II during transcription activation

    Molecular Cell 28:978–990.

    https://doi.org/10.1016/j.molcel.2007.10.017

    • PubMed
    • Google Scholar
68. 1. Yao J
    2. Zobeck KL
    3. Lis JT
    4. Webb WW

    (2008) Imaging transcription dynamics at endogenous genes in living Drosophila tissues

    Methods 45:233–241.

    https://doi.org/10.1016/j.ymeth.2008.06.004

    • PubMed
    • Google Scholar
69. 1. Ye Y
    2. Wu H
    3. Chen K
    4. Clapier CR
    5. Verma N
    6. Zhang W
    7. Deng H
    8. Cairns BR
    9. Gao N
    10. Chen Z

    (2019) Structure of the RSC complex bound to the nucleosome

    Science 366:838–843.

    https://doi.org/10.1126/science.aay0033

    • PubMed
    • Google Scholar
70. 1. Zhang Y
    2. Smith CL
    3. Saha A
    4. Grill SW
    5. Mihardja S
    6. Smith SB
    7. Cairns BR
    8. Peterson CL
    9. Bustamante C

    (2006) DNA translocation and loop formation mechanism of chromatin remodeling by SWI/SNF and RSC

    Molecular Cell 24:559–568.

    https://doi.org/10.1016/j.molcel.2006.10.025

    • PubMed
    • Google Scholar
71. 1. Zobeck KL
    2. Buckley MS
    3. Zipfel WR
    4. Lis JT

    (2010) Recruitment timing and dynamics of transcription factors at the Hsp70 loci in living cells

    Molecular Cell 40:965–975.

    https://doi.org/10.1016/j.molcel.2010.11.022

    • PubMed
    • Google Scholar
72. 1. Zykova TY
    2. Levitsky VG
    3. Belyaeva ES
    4. Zhimulev IF

    (2018) Polytene chromosomes - A portrait of functional organization of the Drosophila Genome

    Current Genomics 19:179–191.

    https://doi.org/10.2174/1389202918666171016123830

    • PubMed
    • Google Scholar

Acceptance summary:

ATP-dependent chromatin remodelers scan the genome and can open up chromatin (to enable access of transcription or DNA repair factors) via nucleosome movement or eviction, and then release from that location to conduct new work. Currently, it is challenging to quantify remodeler-nucleosome dynamics in vivo and the role of ATP binding and hydrolysis in the retention and recycling of remodelers on chromatin/nucleosomes. Here, Tilly and co-workers address these questions using elegant live-cell imaging, applied to monitor the SWI/SNF-family Brahma (BRM) remodeler in Drosophila polytene nuclei, and show that fast kinetics and likely cooperation between remodelers enables chromatin opening. This manuscript will be of considerable interest to the remodeler and chromatin communities.

Decision letter after peer review:

Thank you for submitting your article "in vivo Analysis Reveals that ATP-hydrolysis Couples Remodeling to SWI/SNF Release from Chromatin" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by Sebastian Deindl as the Reviewing Editor and Kevin Struhl as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Bradley R Cairns (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) The study presents a wealth of data in a rather compact format. As a result, some arguments appear a bit abbreviated, some methods could be explained more thoroughly, and some controls are lacking for a more stringent argumentation. By addressing the corresponding reviewers' comments below, this highly interesting manuscript can be significantly improved.

Although additional data could further strengthen the manuscript, we feel that the points raised by the reviewers can be adequately addressed with textual revision and by consolidating and more rigorously quantifying the data, and more explicitly stating controls, etc. Note that no new experimentation is called for.

2) One worry is that the tagged proteins are not functional, but there may be good arguments as to why this is less of an issue (e.g., the effect of ATP depletion and KR mutation). The authors should make these arguments in the paper.

3) All three reviewers felt that the clarity of the presentation could be improved by bringing together the sections related to ATP binding and hydrolysis. We strongly recommend such a change in the order of presentation.

Reviewer #1:

Strength: Chromatin remodelers play critical roles in mobilization of nucleosomes to modulate chromatin-templated functions, but the molecular mechanisms underlying the dynamics of remodelers interacting with chromatin remains incompletely understood. By using cellular imaging, this study reports that chromatin remodeler mutant, which is defective in ATP-binding, reside longer at chromatin than wild-type one, and further indicates that ATP-hydrolysis stimulates remodeler release from chromatin. By leveraging the characteristic of polytene chromatin being visible under fluorescence microscopy, the authors study that the dynamics of chromatin remodelers at native loci, providing further evidence that chromatin remodelers dynamically interact with chromatin with turn-over time less than 5 sec. Overall, these studies indicate that chromatin remodelers rapidly probe the genome and ATP-binding and hydrolysis play important roles in release of remodelers from chromatin.

Weakness: There are a few studies that have reported that remodelers dynamically exchange with chromatin. The novelty of this study is the discovery of the roles of ATP-binding and hydrolysis on the dissociation of remodelers from chromatin; however, to this reviewer, there is lack of functional significance of the destabilization or mechanistic insights into the destabilization.

1) The authors study the dynamics of BRM-K804R within polytene nuclei (Figure 7). The authors also investigate the dynamics of wild-type BRM at the E74 and E75 loci (Figure 5). It appears to this reviewer that one nice experiment would be to investigate the exchange dynamics of BRM-K804R at these two loci and to study the effects of BRM-K804R on their transcriptional levels. Thus, it would be possible to correlate the binding dynamics and transcription and to provide potential functional significances of rapid exchange of remodelers at chromatin.

2) To this review, it is not easily to follow the logic of the manuscript. It would be helpful to readers if the authors give sufficient reasonings for some experiments. For instance, why do the authors study PBAP and BAP for Ec-induced gene expression (Page 12). To me, the authors should also reorganize some figures and change the order of the presentation.

3) I would suggest that the authors quantify FRAP data to extract kinetic parameters based kinetic modelling. In the current version, when comparing t_1/2 between different cellular systems, it is not informative.

4) 1,6-hexanediol has been widely used in phase-separation studies. It primarily disrupts hydrophobic interactions. Given that LLPS involves multiple non-covalent interactions, the authors should re-phase some of their conclusions.

5) A schematic that describes PBAP and BAP would help non-remodeler readers in Figure 1.

6) Figure 1e shows difference in FRAP curves between wild-type BRM and BRM-K804R. BRM-K804R forms clusters while BRM not. It would be helpful if FRAP of BRM-K804R is also done in cluster and no-cluster regions.

7) Page 6, the last conclusion sentence seems over-stated since BRM-K804R impacts ATP-binding rather than hydrolysis.

8) Page 10, the second to last sentence, there is no data to support the author's claim.

9) Page 12, the conclusion remark of the first paragraph, there is no data to support hit-and-run' model.

10) To this reviewer, it appears that the retention of BRM-K804R at chromatin at polytene chromatin cannot be simply explained by a slow dissociation rate and a fast on-rate. This is because BRM-K804R can trap wild-type BRM at polytene chromatin and BRM-K804R can be recovery rapidly in S2 cells. Alternative interpretation should be given.

Reviewer #2:

ATP-dependent chromatin remodelers must scan the genome for locations to open chromatin (to help provide access to transcription and DNA repair factors) via nucleosome movement or eviction, and then release from that location to conduct new work. Currently, it is challenging to quantify remodeler-nucleosome dynamics in vivo, and the role of ATP binding and hydrolysis on the retention and recycling of remodelers at nucleosomes – aside from the known role of ATP in the process of DNA translocation. These questions are of general interest as chromatin remodelers are intimately involved in the regulation of DNA accessibility for processes such as transcription and DNA repair, and as remodelers must both associate with regions productively yet release properly to search for and conduct the scale of work needed.

Here, Tilly et al., (Verrijzer Lab) and colleagues address these questions using elegant live-cell imaging, applied to monitor the SWI/SNF-family Brahma (BRM) remodeler in Drosophila polytene nuclei. This method combined with cleverly chosen reagents, mutants, or physiological contexts, allowed the authors to measure chromatin condensation along with the dynamics governing the interaction of BRM with chromatin. Several conclusions of interest could be reached through the work:

– With the vast majority of BRM is unbound at any given time, BRM remodeler interacts very transiently with chromatin, browsing nucleosomes largely located in decondensed regions.

– ATP binding reduces BRM retention at chromatin, as revealed by characterizing the in vivo dynamics of a Walker A motif mutant.

– ATP hydrolysis increases BRM release from chromatin, as revealed by characterizing the in vivo dynamics in the presence of ADP, or a slow hydrolysable ATP analogue, or a non-hydrolysable ATP ground state mimic.

– When applied to the specific physiological context of highly transcribed ecdysone-induced loci, only the (BRM-containing) PBAP remodeler appears essential; and, remarkably, PBAP fulfills this function without a local retention as shown for RNAPII, but by a rapid exchange.

– Consistent with – and expanding on – the previous conclusions, the presence of the ATP-binding deficient BRM mutant leads to the clustered binding of not only mutated BRM but also wild-type BRM, without the involvement of a phase separation (as 1,6-hexanediol did not affect the punctate/clustered pattern) but likely by accumulation due to excessive retention. Notably, as the authors rightly comment, these interesting results lead to the possibility of a direct cooperation or a hand-off mechanism between different tandem SWI/SNF remodelers, and will deserve further future investigation.

The conclusions listed above are relatively convincingly as they are well supported by the data. The discussion is clearly and thoughtfully written.

Altogether, this work rigorously addresses how ATP binding and hydrolysis cycle of the Brm remodeler, affects the in vivo dynamics of nucleosome browsing, binding and recycling. I think this will be of considerable interest to the remodeler and chromatin community, and recommend publication after revision in eLife.

1. To improve the manuscript, I will suggest that the authors consider slightly reorganizing the manuscript in order to bring in close proximity all the sections related to ATP binding and to ATP hydrolysis, which are currently separated by the characterization of the BRM dynamics at the ecdysone-induced loci.

2. I will also ask that the authors note that the reorientation of the two ATPase lobes is not just induced by nucleosomes, but also efficiently elicited by naked DNA, so nucleosomes do not gate that conformational change.

3. I appreciate how the kinetics of the in vivo work was linked to former in vitro studies in the Discussion – and here the authors might consider (in addition to the works cited) how prior pre-steady state stopped-flow kinetic analysis of SWI/SNF-family remodelers and resulting kinetic and probabilistic models (Fischer et al., Biochemistry 2007) also generally align with their work – and how thinking about release as a probability during each ATPase cycle might be of interest.

4. Finally – the identification of a GLTSCR1 ortholog (Figure 1) is really interesting. he authors mention this but do not explore more deeply. This affects the work in two ways – the possible existence of a GBAF/ncBAF related complex in flies, and that this complex may be participating in the processes observed. I understand that a full characterization of a possible separate fly GBAF/ncBAF complex (in addition to BAF and PBAF) is beyond the scope of the paper, but I will request that the authors revisit and discuss this issue in the Discussion to note its likely existence and possible participation in the data and processes.

Reviewer #3:

The polytenic chromosomes of Drosophila larval salivary glands allow the visualization of interphase chromosomes in a linear, extended form. The authors applied confocal life-imaging confocal microscopy and FRAP analysis to polytene chromosomes to study the dynamic interactions of the (P)BAP remodeling complexes with chromatin. They observe through a series of intriguing experiments that the BRM ATPase interacts with open chromatin only transiently in the presence of hydrolysable ATP. Rapid redistribution of the complex in the nucleus requires continuous ATP hydrolysis. Polytene squashes remove ATP and lead to artefactual trapping of the complex at the sites of otherwise transient interaction. The salivary glands can be cultivated for some time and remain active so that treatment with the ecdysone triggers the characteristic site-specific 'puffing' of the chromosome as a measure of transcription. The authors manage to subject individual 'puffs' to FRAP analyses, a nice demonstration of their technical proficiency.

Although the very transient interactions of other nucleosome remodeling factors have been reported by Erdel and Rippe, the case of BRM is of interest to the chromatin community. The imaging of polytene chromosomes in their squashed and 'life' state is particularly illuminating. The authors push the system and attempt to correlate histone turnover with the presence of the BAP complex.

The study presents a wealth of data in a very compact format. As a result, some arguments appear a bit abbreviated, some methods poorly explained and some controls are lacking for stringent argumentation. This interesting manuscript that can be significantly improved by addressing the following comments.

1. Many findings are illustrated by single individual pictures (e.g. most panels in figure 2). While these illustrate the observations well, they represent individual cells and lack statistical robustness. It will be reassuring if the authors mention at strategic places that all images are representative of the large majority of cells and that similar ones have been obtained in biological replicates (i.e. same experiment on different days). Should this not be the case the readers must be warned accordingly, and the type and source of heterogeneity be detailed.

2. The authors summarize published profiles to say that 70% of chromatin can be considered 'black'. Because they find only 1% of chromatin stained with GFP-H2B at 'black' intensity, they calculate a 25- to 150-fold level of chromatin compaction. Can we assume that the GFP-H2B intensities scale linearly with compaction? Since the GFP-histones are provided by transgenes, their expression is presumably not cell-cycle regulated. What are the levels of GFP-histones relative to the endogenous ones? Can we assume that the GFP-H2B is fully functional, i.e. a substrate to remodelers?

3. Figure 2 Supplement 1C. It is stated that GFP-BRM is efficiently incorporated into remodeler complexes. Here the level of GFP-BRM appears at least 10-fold higher that the level of endogenous BRM. This suggests that GFP-BRM is overexpressed to a worrying extent, where it may compete for the endogenous ATPase. How come that the excessive amount of BRM is not seen in the input? The question here is whether the properties of the tagged ATPase reflect those of the endogenous protein. Devil's advocate says that the endogenous protein constitutes the 'immobile fraction' that is, of course, not detected.

4. "Similar FRAP kinetics were observed for a low-expressing GFP-BRM Drosophila line (Figure 3—figure supplement 1A-C)." How is low expression defined? Given the overexpression of BRM (point 5 above), why was the analysis not done with a line with more physiological levels (since it was apparently available)?

5. It is said that "Depletion of BRM affects neither the polytene chromosome condensation pattern nor the binding of RNAPII (Figure 2—figure supplement 1A)". However, the shown image suggests otherwise as the level of pol II staining appears clearly diminished. If this statement was to be maintained some kind of quantification needs to be done.

6. The authors classify polytene banding as 'white', 'grey' and 'black' using defined fluorescence intensity thresholds. How does the live signal relate to the corresponding squashed chromosomes? How well do the histone-GFP values correlate to the DNA staining?

7. Figure 2 supplement 1D: Moira antibody appears to stain everything. How do we know the staining is specific?

8. Figure 2 supplement 1E lacks all controls (Western blots).

9. It is concluded that GFP-BRM is incorporated in (P)BAP and targeted correctly. What are the criteria for correct targeting and how has it been measured?

10. It is stated: "Pertinently, BRM and SNR1 levels within the polytene chromosome bands are much lower than in the nucleoplasm. This indicates that chromosome condensation leads to (P)BAP exclusion." This latter statement about causation cannot be derived from the correlative data.

11. What is the basis of the statement: "Note that loss of BRM does not affect polytene chromosome condensation (Figure 2-S1A)"? Is this just by looking at the image or has any other analysis been done? What is the sensitivity of the analysis – limitation of the statement?

12. It is stated "Pertinently, there was no appreciable difference in recovery kinetics between BRM in the nucleoplasm or associated with chromatin." What is the basis of this conclusion? Just looking at Figure 3A one would not be able to tell.

13. From Figure 3- supplement 1A-C one may conclude that the lower-expressing GFP-BRM binds better to the chromosome. Is this conclusion justified (and if not, why not?).

14. Figure 3C: which band was frapped for RPB3? Do the recovery curves also summarize the results of 17 band-FRAPs?

15. Related to the RPB3 FRAP profile it is stated "The rapid recovery at the start probably reflects dynamic binding to the promoter, whereas the linear phase corresponds to elongating RNAPII." This is pure speculation, should be removed from the 'results'.

16. The conclusion "We conclude that (P)BAP acts through a continuous and rapid "hit-and-run" probing of the genome" depends on a more stringent argumentation that the protein is functional (see above).

17. Figure 4f: Please indicate the number of biological replicates (assuming each biological replicate consists of three technical replicates).

18. Figure 4a shows the level of BRM recruitment to Ec-induced puffs. This is presumably fromn a wt larva. How does the same picture look like if the GFB-BRM overexpression line is used and stained for either GFP or BRM?

19. Figure 5C: the puffing is not visible, unlike stated in the text.

20. Figure 5H. "The absence of a full recovery of RNAPII at the height of E74 and E75 expression and puffing is similar to its behavior on fully induced heat shock loci 87A and 87C, and is indicative of local RNAPII recycling (Yao et al., 2007)." Might it also be possible that a certain fraction of pol II is stalled under culture conditions and thus 'trapped' in the topologically associated elongating form?

21. Figure 5J: Concluding about the absence of the D4 subunit requires some kind of a positive control and statistics.

22. Figure 6b: what is red-blue?

23. Figure 6c, 7c: How many bands were frapped?

24. "IF of polytene chromosome spreads revealed co-localization of GFP-BRM-K804R with endogenous MOR (SMARCC1/2), indicating largely normal targeting of this mutant". What are the criteria for 'Largely normal targeting' – what would be your criteria for 'not normal'?

25. Figure 7k. BRM binds interbands where histones are barely detectable. It is unclear how both proteins could be frapped at the same site. For H2B FRAP- was the neighbouring band frapped? What is the resolution of the assay?

26. The authors conclude that chromatin remodeling and remodeler relased are coupled. This conclusion is premature given the possibility of indirect effects (e.g. through elongating pol II -mediated by different transcription levels). Furthermore, the discussion states: "These observations suggest that the prevalent outcome of chromatin remodeling in vivo is restructuring or sliding of nucleosomes rather than eviction." The authors should acknowledge that they have very little spatial resolution: the histone dynamics could be very local. The data do not allow concluding about nucleosome remodeling mechanism.

27. The discussion states: "However, given that chromatin typically only occupies about 1.5% of the nuclear volume, …" Where does that number come from? That seems to be very odd.

Essential revisions:

1) The study presents a wealth of data in a rather compact format. As a result, some arguments appear a bit abbreviated, some methods could be explained more thoroughly, and some controls are lacking for a more stringent argumentation. By addressing the corresponding reviewers' comments below, this highly interesting manuscript can be significantly improved.

Although additional data could further strengthen the manuscript, we feel that the points raised by the reviewers can be adequately addressed with textual revision and by consolidating and more rigorously quantifying the data, and more explicitly stating controls, etc. Note that no new experimentation is called for.

We thank the reviewers for their constructive comments and suggestions, which have helped us to improve our manuscript. We now provide a more extensive and explicit explanation of the quantification of the polytene nuclei images in the legends and methods, which were actually extremely consistent across different animals and different nuclei in the salivary glands. For all other experiments quantitative information was already provided. In addition, we have reorganized the manuscript according to the reviewers suggestions, and changed the text to increase clarity. Finally, we have included additional data (detailed below) to clarify some of the concerns of reviewer #3.

2) One worry is that the tagged proteins are not functional, but there may be good arguments as to why this is less of an issue (e.g., the effect of ATP depletion and KR mutation). The authors should make these arguments in the paper.

Tagging may indeed affect a protein’s function. However, all major conclusions derived from tagged factors (such as remodeler distribution, the effects of ATP depletion, ATP add-back and the BRM-K804R mutation) were confirmed by IF of endogenous (P)BAP subunits. These effects were entirely consistent across multiple (P)BAP subunits. Moreover, other tagged chromosomal regulators (RNAPII, Polycomb, HP1, EcR) that were analyzed in parallel to (P)BAP behaved differently. We have expanded these arguments in the text.

3) All three reviewers felt that the clarity of the presentation could be improved by bringing together the sections related to ATP binding and hydrolysis. We strongly recommend such a change in the order of presentation.

We thank the reviewers for this valuable suggestion. We have completely reorganized the manuscript. We now first present our results on chromosomal organization and (P)BAP intranuclear distribution. Next, we present the data on the role of ATP in remodeler dynamics. The following changes were made (new<old): Figure 1<Figure 2, Figure 2<Figure 4, Figure 3<Figure 5A-D, Figure 4<Figure 1, Figure 5<Figure 3, Figure 6<Figure 5F-K, Figure 7<Figure 6, Figure 8<Figure 7, Figure 9<Figure 8. To clarify some issues raised by reviewer #3, we have provided new data in Figure 2C, Figure 3C, Figure 6A-B, Figure 5-suppl.1G-H, and Figure 6-suppl.1C-D.

Reviewer #1:

Strength: Chromatin remodelers play critical roles in mobilization of nucleosomes to modulate chromatin-templated functions, but the molecular mechanisms underlying the dynamics of remodelers interacting with chromatin remains incompletely understood. By using cellular imaging, this study reports that chromatin remodeler mutant, which is defective in ATP-binding, reside longer at chromatin than wild-type one, and further indicates that ATP-hydrolysis stimulates remodeler release from chromatin. By leveraging the characteristic of polytene chromatin being visible under fluorescence microscopy, the authors study that the dynamics of chromatin remodelers at native loci, providing further evidence that chromatin remodelers dynamically interact with chromatin with turn-over time less than 5 sec. Overall, these studies indicate that chromatin remodelers rapidly probe the genome and ATP-binding and hydrolysis play important roles in release of remodelers from chromatin.

Weakness: There are a few studies that have reported that remodelers dynamically exchange with chromatin. The novelty of this study is the discovery of the roles of ATP-binding and hydrolysis on the dissociation of remodelers from chromatin; however, to this reviewer, there is lack of functional significance of the destabilization or mechanistic insights into the destabilization.

Our work shows that remodeling and release are mechanistically coupled. This is significant because it shows that the remodeling process is intrinsically dynamic and does not induce a stable static state. Moreover, we show that during remodeling, remodelers turn over faster than histones. An open chromatin state requires continuous remodeler activity and allows for quick regulation. The use of polytene nuclei enabled us to visualize these processes on endogenous, natural gene loci in living cells. We show that induction of E74/75 depends on PBAP. Thus, we are studying PBAP and chromatin organization in living tissue on physiological target genes that depend on PBAP for its regulation. This has not been done previously.

1) The authors study the dynamics of BRM-K804R within polytene nuclei (Figure 7). The authors also investigate the dynamics of wild-type BRM at the E74 and E75 loci (Figure 5). It appears to this reviewer that one nice experiment would be to investigate the exchange dynamics of BRM-K804R at these two loci and to study the effects of BRM-K804R on their transcriptional levels. Thus, it would be possible to correlate the binding dynamics and transcription and to provide potential functional significances of rapid exchange of remodelers at chromatin.

BRM-K804R clamps to chromosomes and is also completely immobile on E74/75. As first shown by Tamkun and colleagues (Armstrong et al., 2002), BRM-K804R is a dosagedependent dominant mutant that interferes with all RNAPII transcription, including from the E74/75 locus. We have now included results showing that BRM kinetics on RNAPII loci is indistinguishable from that on non-transcribed loci (Figure 5-suppl.1G-H)

2) To this review, it is not easily to follow the logic of the manuscript. It would be helpful to readers if the authors give sufficient reasonings for some experiments. For instance, why do the authors study PBAP and BAP for Ec-induced gene expression (Page 12). To me, the authors should also reorganize some figures and change the order of the presentation.

We have reorganized the manuscript according to the recommendations of the reviewers. The importance of the experiments on (P)BAP and Ec-induced transcription is that these provide the foundation to study PBAP dynamics by life imaging on genes that dependent on that remodeler in vivo.

3) I would suggest that the authors quantify FRAP data to extract kinetic parameters based kinetic modelling. In the current version, when comparing t_1/2 between different cellular systems, it is not informative.

We have used the half time of recovery and the (im-)mobile fraction to compare the kinetics of the interactions of individual (P)BAP subunits as well as with other known chromatin binding proteins. In addition, we used these parameters to describe the differences in chromatin interaction between WT and mutant BRM. To extract additional kinetic parameters by modelling, a (currently unavailable) model containing detailed information about the interaction(s) between (P)BAP and chromatin is required. Moreover, we refrain from kinetic modelling on our FRAP data because these types of analyses can easily lead to over- or misinterpretations (discussed by e.g., Mueller et al., Curr. Opin. Cell Biol. 23, 403-411, 2010). Importantly, our FRAP data provides an upper limit to (P)BAP residence time and establishes that (P)BAP has a very fast turnover, whereas histones are comparatively stable. Additional kinetic data on remodeler/chromatin interactions are reported in the accompanying study by Kim et al. using single molecule analysis, a system more suitable for detailed kinetic analysis. The findings of Kim et al. (https://doi.org/10.1101/2021.04.21.440742) agree well with our results. We have included a reference to this manuscript.

4) 1,6-hexanediol has been widely used in phase-separation studies. It primarily disrupts hydrophobic interactions. Given that LLPS involves multiple non-covalent interactions, the authors should re-phase some of their conclusions.

We have included the 1,6-HD experiments because every time we presented these data in a seminar, someone asked about LLPS and 1,6-HD. So, we used it simply as an additional test that provides further evidence that LLPS appears to play no substantial role in the effects we describe in this paper.

5) A schematic that describes PBAP and BAP would help non-remodeler readers in Figure 1.

We have added cartoons of BAP and PBAP in Figure 2C.

(6) Figure 1e shows difference in FRAP curves between wild-type BRM and BRM-K804R. BRM-K804R forms clusters while BRM not. It would be helpful if FRAP of BRM-K804R is also done in cluster and no-cluster regions.

This would indeed be a very interesting experiment. In this figure, the kinetics of GFP-BRM and GFP-BRM-K804R were studied by strip FRAP on Drosophila S2 cells. Because of the size of the individual nuclei (Ø < 4 µm), a 16-pixel wide strip was bleached in these experiments. Due to limitations in resolution, discriminating between clustered and non-clustered BRM-K804R was not possible.

7) Page 6, the last conclusion sentence seems over-stated since BRM-K804R impacts ATP-binding rather than hydrolysis.

We have re-phrased this sentence (p10, line 4).

8) Page 10, the second to last sentence, there is no data to support the author's claim.

The bleached area in the FLIP-FRAP experiments presented in figure 5A consists of both nucleoplasmic and chromatin-associated GFP-BRM. As can be concluded from these images, recovery of both compartments occurs at a similar rate. We have re-phrased this sentence (p10, line 23).

9) Page 12, the conclusion remark of the first paragraph, there is no data to support hit-and-run' model.

We removed “hit-and-run”.

10) To this reviewer, it appears that the retention of BRM-K804R at chromatin at polytene chromatin cannot be simply explained by a slow dissociation rate and a fast on-rate. This is because BRM-K804R can trap wild-type BRM at polytene chromatin and BRM-K804R can be recovery rapidly in S2 cells. Alternative interpretation should be given.

We have modified our discussion on the difference between polytene nuclei and S2 cells for greater clarity (p16, line 14). Although this remains speculative, we propose that the high local density of (P)BAP loci in polytene chromosomes amplifies the difference in retention time between wtBRM and BRM-K804R.

Reviewer #2:

ATP-dependent chromatin remodelers must scan the genome for locations to open chromatin (to help provide access to transcription and DNA repair factors) via nucleosome movement or eviction, and then release from that location to conduct new work. Currently, it is challenging to quantify remodeler-nucleosome dynamics in vivo, and the role of ATP binding and hydrolysis on the retention and recycling of remodelers at nucleosomes – aside from the known role of ATP in the process of DNA translocation. These questions are of general interest as chromatin remodelers are intimately involved in the regulation of DNA accessibility for processes such as transcription and DNA repair, and as remodelers must both associate with regions productively yet release properly to search for and conduct the scale of work needed.

Here, Tilly et al., (Verrijzer Lab) and colleagues address these questions using elegant live-cell imaging, applied to monitor the SWI/SNF-family Brahma (BRM) remodeler in Drosophila polytene nuclei. This method combined with cleverly chosen reagents, mutants, or physiological contexts, allowed the authors to measure chromatin condensation along with the dynamics governing the interaction of BRM with chromatin. Several conclusions of interest could be reached through the work:

– With the vast majority of BRM is unbound at any given time, BRM remodeler interacts very transiently with chromatin, browsing nucleosomes largely located in decondensed regions.

– ATP binding reduces BRM retention at chromatin, as revealed by characterizing the in vivo dynamics of a Walker A motif mutant.

– ATP hydrolysis increases BRM release from chromatin, as revealed by characterizing the in vivo dynamics in the presence of ADP, or a slow hydrolysable ATP analogue, or a non-hydrolysable ATP ground state mimic.

– When applied to the specific physiological context of highly transcribed ecdysone-induced loci, only the (BRM-containing) PBAP remodeler appears essential; and, remarkably, PBAP fulfills this function without a local retention as shown for RNAPII, but by a rapid exchange.

– Consistent with – and expanding on – the previous conclusions, the presence of the ATP-binding deficient BRM mutant leads to the clustered binding of not only mutated BRM but also wild-type BRM, without the involvement of a phase separation (as 1,6-hexanediol did not affect the punctate/clustered pattern) but likely by accumulation due to excessive retention. Notably, as the authors rightly comment, these interesting results lead to the possibility of a direct cooperation or a hand-off mechanism between different tandem SWI/SNF remodelers, and will deserve further future investigation.

The conclusions listed above are relatively convincingly as they are well supported by the data. The discussion is clearly and thoughtfully written.

Altogether, this work rigorously addresses how ATP binding and hydrolysis cycle of the Brm remodeler, affects the in vivo dynamics of nucleosome browsing, binding and recycling. I think this will be of considerable interest to the remodeler and chromatin community, and recommend publication after revision in eLife.

1. To improve the manuscript, I will suggest that the authors consider slightly reorganizing the manuscript in order to bring in close proximity all the sections related to ATP binding and to ATP hydrolysis, which are currently separated by the characterization of the BRM dynamics at the ecdysone-induced loci.

As detailed above, we have reorganized the manuscript according to the recommendations of the reviewers.

2. I will also ask that the authors note that the reorientation of the two ATPase lobes is not just induced by nucleosomes, but also efficiently elicited by naked DNA, so nucleosomes do not gate that conformational change.

We have corrected the text (p17, line 20).

3. I appreciate how the kinetics of the in vivo work was linked to former in vitro studies in the Discussion – and here the authors might consider (in addition to the works cited) how prior pre-steady state stopped-flow kinetic analysis of SWI/SNF-family remodelers and resulting kinetic and probabilistic models (Fischer et al., Biochemistry 2007) also generally align with their work – and how thinking about release as a probability during each ATPase cycle might be of interest.

We have included these prescient findings in our discussion and included the insightful notion of release as a probability during each ATPase cycle (p18, line 7).

4. Finally – the identification of a GLTSCR1 ortholog (Figure 1) is really interesting. he authors mention this but do not explore more deeply. This affects the work in two ways – the possible existence of a GBAF/ncBAF related complex in flies, and that this complex may be participating in the processes observed. I understand that a full characterization of a possible separate fly GBAF/ncBAF complex (in addition to BAF and PBAF) is beyond the scope of the paper, but I will request that the authors revisit and discuss this issue in the Discussion to note its likely existence and possible participation in the data and processes.

We have corrected the discussion and included a reference to a recent publication from the Bellen lab (Barish et al., 2020) that includes a further characterization of GBAF/ncBAF in Drosophila (page 14, line 9).

Reviewer #3:

The polytenic chromosomes of Drosophila larval salivary glands allow the visualization of interphase chromosomes in a linear, extended form. The authors applied confocal life-imaging confocal microscopy and FRAP analysis to polytene chromosomes to study the dynamic interactions of the (P)BAP remodeling complexes with chromatin. They observe through a series of intriguing experiments that the BRM ATPase interacts with open chromatin only transiently in the presence of hydrolysable ATP. Rapid redistribution of the complex in the nucleus requires continuous ATP hydrolysis. Polytene squashes remove ATP and lead to artefactual trapping of the complex at the sites of otherwise transient interaction. The salivary glands can be cultivated for some time and remain active so that treatment with the ecdysone triggers the characteristic site-specific 'puffing' of the chromosome as a measure of transcription. The authors manage to subject individual 'puffs' to FRAP analyses, a nice demonstration of their technical proficiency.

Although the very transient interactions of other nucleosome remodeling factors have been reported by Erdel and Rippe, the case of BRM is of interest to the chromatin community. The imaging of polytene chromosomes in their squashed and 'life' state is particularly illuminating. The authors push the system and attempt to correlate histone turnover with the presence of the BAP complex.

The study presents a wealth of data in a very compact format. As a result, some arguments appear a bit abbreviated, some methods poorly explained and some controls are lacking for stringent argumentation. This interesting manuscript that can be significantly improved by addressing the following comments.

1. Many findings are illustrated by single individual pictures (e.g. most panels in figure 2). While these illustrate the observations well, they represent individual cells and lack statistical robustness. It will be reassuring if the authors mention at strategic places that all images are representative of the large majority of cells and that similar ones have been obtained in biological replicates (i.e. same experiment on different days). Should this not be the case the readers must be warned accordingly, and the type and source of heterogeneity be detailed.

IF images of polytene spreads and whole mount images of salivary glands (or imaginal disks or embryos) are normally highly reproducible, therefore we did not state this explicitly. However, we have now included statements on the reproducibility of images of individual nuclei in the legends to Figure 1, 7 and 8. More importantly, we note that the IF images of different PBAF subunits in different figures are very similar. Moreover, IF and live cell imaging of polytene nuclei yield very similar images. Collectively, these comparisons emphasize the reproducibility of the images presented. In response to some later criticisms from this reviewer, we have included additional data (Figure 3C, Figure 6A-B, Figure 5-suppl.1G-H, and Figure 6-suppl.1C-D), which further illustrates the reproducibility of our imaging.

2. The authors summarize published profiles to say that 70% of chromatin can be considered 'black'. Because they find only 1% of chromatin stained with GFP-H2B at 'black' intensity, they calculate a 25- to 150-fold level of chromatin compaction. Can we assume that the GFP-H2B intensities scale linearly with compaction? Since the GFP-histones are provided by transgenes, their expression is presumably not cell-cycle regulated. What are the levels of GFP-histones relative to the endogenous ones? Can we assume that the GFP-H2B is fully functional, i.e. a substrate to remodelers?

1) All data were collected using conditions that preserved linearity during data acquisition as much as possible. However, we cannot exclude small deviations at very low and very high fluorescence intensities. Moreover, there might be shielding and other effects that interfere with strict linearity. The GFP-H2B fluorescence was quantitated from 10 individual nuclei, 9 slices/nucleus. From this data set we concluded that 5% (not 1%) of the chromatin volume is “black” and 43% is “grey”. As explained in the methods, the chromocenter and telomers are recognizable structures that were used to calibrate the signal from black chromatin. Likewise, interbands and puffs were used to calibrate white chromatin. Chromatin bands that fall between “white” and “black” were assigned grey. This classification does not depend on linear scalability of GFP-H2B. The goal of these experiments was not to provide a very precise estimate of the amount of white, grey, and black chromatin, but to obtain an impression of their distribution in living cell polytene nuclei. What we find amazing is how well our estimates derived from vital imaging agree with those of Eagen et al. (2015), who used a completely different method (Hi-C) relying on fixed cells.

2) GFP-H2A and mCh-H2B fully co-localize (Figure 1F) and recapitulate the canonical chromosomal banding pattern. We cannot detect appreciable amounts of free histones in the nucleoplasm or cytoplasm, indicating there is no substantive over-expression. Moreover, the Drosophila expressing the tagged histones develop, pupate, and propagate normally. Thus, we have no indications that these tagged histones are not behaving normally.

3. Figure 2 Supplement 1C. It is stated that GFP-BRM is efficiently incorporated into remodeler complexes. Here the level of GFP-BRM appears at least 10-fold higher that the level of endogenous BRM. This suggests that GFP-BRM is overexpressed to a worrying extent, where it may compete for the endogenous ATPase. How come that the excessive amount of BRM is not seen in the input? The question here is whether the properties of the tagged ATPase reflect those of the endogenous protein. Devil's advocate says that the endogenous protein constitutes the 'immobile fraction' that is, of course, not detected.

We cannot detect BRM or GFP-BRM in the input extract from dissected salivary glands by western blotting. This is probably due to interfering proteins in that size range. BRM and GFPBRM are only detected in the MOR co-IPs, thus by definition as part of a remodeler complex. Note that we have no problems detecting GFP-SNR1, SNR1 and MOR. Loss of the architectural subunit MOR, leads to disintegration of the (P)BAP complexes and loss of GFPBRM (see Figure 1.suppl.1E, and Moshkin et al., 2007). This observation argues strongly that GFP-BRM only exists as part of a complex and not as a free protein. This interpretation is also in line with the published literature (see e.g., Mashtalir et al., 2018 Cell 175, p1272). Most likely, free endogenous BRM is degraded because its place in the (P)BAP complex is taken by GFP-BRM (see also Figure 4A). Importantly, all our major conclusions are supported by analysis of endogenous proteins and do not rely solely on the tagged remodeler subunits. Finally, the FRAP kinetics of GFP-SNR1, GFP-D4 and GFP-BRD7 are all very similar to those of GRPBRM (including the low expressing line).

4. "Similar FRAP kinetics were observed for a low-expressing GFP-BRM Drosophila line (Figure 3—figure supplement 1A-C)." How is low expression defined? Given the overexpression of BRM (point 5 above), why was the analysis not done with a line with more physiological levels (since it was apparently available)?

The primary reason is that we generated the Sgs3>GFP-BRM line before the #59784 fly line became available. Expression of GFP-BRM in the salivary gland nuclei of the 59784 line is very low compared to Sgs3>GFP-BRM, as judged by fluorescence. Importantly, nuclear distribution and kinetics of GFP-BRM in #59784 nuclei is indistinguishable from that in GFPBRM in Sgs3>GFP-BRM. Thus, our measurements are not influenced by the expression level of GFP-BRM.

5. It is said that "Depletion of BRM affects neither the polytene chromosome condensation pattern nor the binding of RNAPII (Figure 2—figure supplement 1A)". However, the shown image suggests otherwise as the level of pol II staining appears clearly diminished. If this statement was to be maintained some kind of quantification needs to be done.

We respectfully disagree. The RNAPII and DAPI stains of polytenes from the wt and BRM knockdown line both fall within the normal range of variability we observe for polytene spreads (whereas BRM is no longer detectable). Thus, we do not detect a gross change in the polytene banding pattern or RNAPII binding. These assays are qualitative rather than quantitative. We have changed the text for greater clarity (p5, line 19).

6. The authors classify polytene banding as 'white', 'grey' and 'black' using defined fluorescence intensity thresholds. How does the live signal relate to the corresponding squashed chromosomes? How well do the histone-GFP values correlate to the DNA staining?

The correspondence is so good that we can use the polytene maps to navigate the chromosomes in a living nucleus and identify the chromocenter, telomers, and specific loci like E74/75 in living gland nuclei.

7. Figure 2 supplement 1D: Moira antibody appears to stain everything. How do we know the staining is specific?

MOR/(P)BAP binds many interband loci. The anti-MOR antibody is specific and validated previously by depletion and IP-mass spectrometry. See e.g., Moshkin et al., 2007; Chalkley et al., 2008 and Mohrmann et al., 2004.

8. Figure 2 supplement 1E lacks all controls (Western blots).

The absence of GFP-BRM in the salivary gland polytene nuclei from MOR^KD larvae is clear. As mentioned above, after blotting of extracts from dissected salivary glands (which is not trivial), BRM is not detectable, possibly due to other proteins in this range. The observation that MOR is a key architectural subunit (with all control blots) has been published previously (Moshkin et al., 2007).

9. It is concluded that GFP-BRM is incorporated in (P)BAP and targeted correctly. What are the criteria for correct targeting and how has it been measured?

As described in the results, we concluded that GFP-BRM is incorporated into (P)BAP and targeted correctly based on:

1) Mass spectrometric analyses of anti-GFP IPs (Figure 4B).

2) Western blots of MOR-IPs, using anti-GFP, anti-BRM, anti-SNR1 and anti-MOR antibodies (Figure 1suppl.1C).

3) GFP-BRM expression in MOR^KD larvae (Figure 1suppl.1E).

4) Comparison of GFP-BRM distribution on polytene chromosomes with MOR

(Figure 1suppl.1D).

5) Comparison of the nuclear distribution of GFP-BRM with endogenous PBAP (compare Figure 1B and I).

6) GFP-BRM binds canonical BRM target loci (Figure 3B, Figure 4F)

10. It is stated: "Pertinently, BRM and SNR1 levels within the polytene chromosome bands are much lower than in the nucleoplasm. This indicates that chromosome condensation leads to (P)BAP exclusion." This latter statement about causation cannot be derived from the correlative data.

Although we do not formally prove causation, (P)BAP is excluded from condensed polytene chromatin bands, whereas (P)BAP is not required to form interbands (Fig1suppl.1A). Thus, a parsimonious explanation would be that a high level of chromatin condensation causes the exclusion of (P)BAP. Nevertheless, we have re-phrased this statement to reflect this notion is a possibility, not a conclusion (p7, line 4; p14, line 23).

11. What is the basis of the statement: "Note that loss of BRM does not affect polytene chromosome condensation (Figure 2-S1A)"? Is this just by looking at the image or has any other analysis been done? What is the sensitivity of the analysis – limitation of the statement?

Yes, this is based on inspection of images, which showed that in the absence (P)BAP the classical banding pattern of polytene chromosomes is not changed appreciably. Analysis of polytene chromosomes allows for an easy “whole genome” analysis of global chromosome structure. It is not a high-resolution technique. However, the work of Eagen and colleagues (2015), established a close correspondence between the banding pattern of polytene chromosomes and TADs.

12. It is stated "Pertinently, there was no appreciable difference in recovery kinetics between BRM in the nucleoplasm or associated with chromatin." What is the basis of this conclusion? Just looking at Figure 3A one would not be able to tell.

The bleached area in the FLIP-FRAP experiments presented in new Figure 5A consists of both nucleoplasmic and chromatin-associated GFP-BRM. As can be concluded from these images, recovery of both compartments occurs at a similar rate. We have re-phrased this sentence (p10, line 22).

13. From Figure 3- supplement 1A-C one may conclude that the lower-expressing GFP-BRM binds better to the chromosome. Is this conclusion justified (and if not, why not?).

In our opinion, sg>GFP-BRM and line GFP-BRM_low show comparable distributions.

Importantly, quantitative FRAP experiments confirmed this observation. The results from the FRAP experiments are very comparable with respect to t_½ ( GFP-BRM_low , t_½ = 4.2 ± 1.6 s, n=15 and GFP-BRM, t_½ = 3.1 ± 1.1, n=17) and mobile fraction: 0.9 ± 0.1 for both GFP-BRM_low and GFP-BRM.

14. Figure 3C: which band was frapped for RPB3? Do the recovery curves also summarize the results of 17 band-FRAPs?

New Figure 5G shows a representative single FRAP curve for RBP3 (we have performed multiple experiments in different lines). The kinetics we observe for both RBP3 and RBP2 (Figure 6E) is very similar to those obtained earlier by the Lis lab, using different lines, and analyzing different loci. We refer to the extensive analysis of RNAPII by Lis et al., in our results. As we basically confirm their earlier results, we see no need to expand the RNAPII analysis, which functions as a reference for (P)BAP.

15. Related to the RPB3 FRAP profile it is stated "The rapid recovery at the start probably reflects dynamic binding to the promoter, whereas the linear phase corresponds to elongating RNAPII." This is pure speculation, should be removed from the 'results'.

Again, we refer to the rigorously tested conclusions published earlier by Lis and colleagues, and indicate that our RNAPII FRAP curves are highly similar.

16. The conclusion "We conclude that (P)BAP acts through a continuous and rapid "hit-and-run" probing of the genome" depends on a more stringent argumentation that the protein is functional (see above).

As discussed above, our conclusions are fully supported by results obtained with the endogenous (P)BAP. We removed “hit-and-run”.

17. Figure 4f: Please indicate the number of biological replicates (assuming each biological replicate consists of three technical replicates).

We used 3 independent biological replicates. This information was added to the legends.

18. Figure 4a shows the level of BRM recruitment to Ec-induced puffs. This is presumably fromn a wt larva. How does the same picture look like if the GFB-BRM overexpression line is used and stained for either GFP or BRM?

Unless stated otherwise, Drosophila images and data shown are derived from wild type animals. We use the GFP-BRM for life cell imaging and it binds the same loci as the endogenous protein (see Figures 2, 3 and 5).

19. Figure 5C: the puffing is not visible, unlike stated in the text.

We have included a lower magnification of the same nucleus to show the zoom area. In our opinion this is a clear puff with a diameter much larger than that of a regular polytene region (Figure 3C). This is confirmed by direct measurements of chromosomal areas (Figure 3E).

20. Figure 5H. "The absence of a full recovery of RNAPII at the height of E74 and E75 expression and puffing is similar to its behavior on fully induced heat shock loci 87A and 87C, and is indicative of local RNAPII recycling (Yao et al., 2007)." Might it also be possible that a certain fraction of pol II is stalled under culture conditions and thus 'trapped' in the topologically associated elongating form?

We cannot exclude this possibility. Therefore, we referred to the results by Yao et al. (2007), who made similar observations on the HS loci and provided additional evidence that RNAPII was recycled, not stalled. In our study, the kinetics of GFP-RPB2 was studied 30 min after ecdysone addition, at the peak of RNAPII accumulation and induction of E74/75, making it unlikely this is stalled RNAPII. However, we have modified the text to indicate this possibility (p11, line 21).

21. Figure 5J: Concluding about the absence of the D4 subunit requires some kind of a positive control and statistics.

We have included a GFP-D4 image showing a full nucleus (Figure 6.suppl.1D).

22. Figure 6b: what is red-blue?

The false color scale bar represents full scale grey values (8 bit) from 0 (black) to 255 (white). This information was added to the legend.

23. Figure 6c, 7c: How many bands were frapped?

Figure 6C, at least 6; Figure 7C, at least 8. These FRAP curves are highly invariable across different animals.

24. "IF of polytene chromosome spreads revealed co-localization of GFP-BRM-K804R with endogenous MOR (SMARCC1/2), indicating largely normal targeting of this mutant". What are the criteria for 'Largely normal targeting' – what would be your criteria for 'not normal'?

This figure shows the strong co-localization of GFP-BRM-K804R and endogenous MOR, indicated by the predominantly yellow merge, and by zooming, which shows a very similar pattern for GFP and MOR. Clearly separated red and green in the merge would indicate not normal targeting of GFP-BRM-K804R.

25. Figure 7k. BRM binds interbands where histones are barely detectable. It is unclear how both proteins could be frapped at the same site. For H2B FRAP- was the neighbouring band frapped? What is the resolution of the assay?

Bleached area is 4 x 2 µm (there is a scale bar in this panel). Thus, we do not reach band/interband resolution. Nevertheless, we observe a clear and significant reduction in histone turnover. We also note that although BRM or GFP-BRM levels are low in bands they are not completely absent (see e.g., Figure 1B,C and Figure 1.suppl.1F).

26. The authors conclude that chromatin remodeling and remodeler relased are coupled. This conclusion is premature given the possibility of indirect effects (e.g. through elongating pol II -mediated by different transcription levels). Furthermore, the discussion states: "These observations suggest that the prevalent outcome of chromatin remodeling in vivo is restructuring or sliding of nucleosomes rather than eviction." The authors should acknowledge that they have very little spatial resolution: the histone dynamics could be very local. The data do not allow concluding about nucleosome remodeling mechanism.

1) Remodeling and release are coupled because both depend on ATP-hydrolysis. The possibility that there might be indirect effects does not change that.

2) There are many (P)BAP-binding loci (actually the majority) that do not coincide with sites of active transcription. We did not observe a difference in BRM dynamics between transcribed versus non-transcribed loci. We have now added these results (Figure 5.suppl.1G,H).

3) In our discussion, we first review and cite results from other labs, before we suggest that histone eviction is not the prevalent outcome of in vivo remodeling. Clearly we do not solely rely on our own results.

4) We explicitly state that “ … histone eviction is restricted to small regulatory regions of the genome and is associated with the placement of variant histones (Brahma and Henikoff, 2020; Cakiroglu et al., 2019; Deal et al., 2010; Pillidge and Bray, 2019).”

5) Nowhere in the manuscript do we state that there is no histone turnover. In fact, we quantify histone turnover in Figure 5H and Figure 8K.

27. The discussion states: "However, given that chromatin typically only occupies about 1.5% of the nuclear volume, …" Where does that number come from? That seems to be very odd.

The diameter of a typical human nucleus is 10 μm (r=5 μm). V=4/3πr^3, yielding a volume of ~420 μm^3.

Human genome: diploid genome ~ 6.10^9 bp. A fully packaged genome, assuming 39 bp linker DNA between nucleosomes would be roughly 80% nucleosomal and 20% linker DNA. Volume bp ~ 1nm^3. Thus volume linker DNA: 0.2 x ~ 6.10^9 x bp ~ 1nm^3 = ~1.2 μm^3. The shell volume of a nucleosome is 3.28e+5 Å^3. nucleosomal genome: 0.8 x 6.10^9 / 147 = 3.3.10^7 nucleosomes, resulting in an approximate volume of: 0.33e+8 x 3.28e+2 nm^3 = 1.1e+10 nm^3 = 11 μm^3

Total volume chromatin: 1.2+11 = 12.2 μm^3. This corresponds to ~2.9% of the nuclear volume.

If the human genome was not packaged into chromatin.

Human genome: diploid genome ~ 6.10^9 bp, Volume bp ~ 1nm^3, thus human genome occupies ~6 μm^3. This corresponds to ~1.4% of the nuclear volume.

Author details

1. Ben C Tilly

   Department of Biochemistry, Rotterdam, Netherlands

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing

   For correspondence

   b.tilly@erasmusmc.nl

   Competing interests

   No competing interests declared

2. Gillian E Chalkley

   Department of Biochemistry, Rotterdam, Netherlands

   Contribution

   Data curation, Validation, Investigation, Visualization, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

3. Jan A van der Knaap

   Department of Biochemistry, Rotterdam, Netherlands

   Contribution

   Conceptualization, Data curation, Validation, Investigation, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

4. Yuri M Moshkin

   Department of Biochemistry, Rotterdam, Netherlands

   Contribution

   Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology

   Competing interests

   No competing interests declared

5. Tsung Wai Kan

   Department of Biochemistry, Rotterdam, Netherlands

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

6. Dick HW Dekkers

   1. Department of Biochemistry, Rotterdam, Netherlands
   2. Proteomics Center, Erasmus University Medical Center, Rotterdam, Netherlands

   Contribution

   Data curation, Formal analysis, Investigation

   Competing interests

   No competing interests declared

7. Jeroen AA Demmers

   1. Department of Biochemistry, Rotterdam, Netherlands
   2. Proteomics Center, Erasmus University Medical Center, Rotterdam, Netherlands

   Contribution

   Data curation, Formal analysis, Supervision, Investigation, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

8. C Peter Verrijzer

   Department of Biochemistry, Rotterdam, Netherlands

   Contribution

   Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing - original draft, Project administration, Writing - review and editing

   For correspondence

   c.verrijzer@erasmusmc.nl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6476-3264

• 1,336

  Page views

• 233

  Downloads

• 5

  Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.