Vostok: A Looping Factor for the Organization of the Regulatory Genome in the Drosophila Brain

SUMMARY

Drosophila tethering elements mediate long-range enhancer-promoter interactions and connect promoters of distant paralogous genes. Micro-C maps identified 645 such loops in the Drosophila larval brain, spanning distances of 25 kb to 250 kb. Here, we demonstrate that the MADF-containing Vostok protein acts as a looping factor. It binds to GCAACA motifs that are overrepresented in brain tethering elements. There is a loss of 47 (7%) of the loops in Vostok mutants, resulting in diminished expression of associated genes. Vostok is largely independent of another looping factor, GAF (GAGA-associated factor). Only 9 loops are disrupted in both Vostok and GAF mutants, raising the possibility of a combinatorial code for tether-tether interactions. This is supported by the reliance of two previously identified meta-loops spanning 6 Mb on both GAF and Vostok. We discuss the prospects of using different combinations of looping factors to engineer 3D associations in animal genomes.

ETOC BLURB

In Drosophila, tethering elements are DNA regulatory regions facilitating long-range interactions between enhancers and promoters, as well as between promoters of distant paralogous genes. Hu et al. demonstrate that the Vostok protein binds to these elements in the larval brain and is essential for the correct expression of associated genes.

Graphical Abstract

INTRODUCTION

The 3D organization of the regulatory genome contributes to the precision of gene activity during development.^1–3 This organization depends on at least two classes of regulatory elements, boundaries (aka insulators) and tethering elements.^4–6 Boundaries delineate topological associating domains (TADs), constraining enhancers and silencers from interacting with inappropriate target genes. The Drosophila embryo also contains ~200 intra-TAD regulatory loops that foster timely activation of gene expression and coordinate regulation of duplicated paralogs by shared enhancers.⁷ For example, disruption of a 40 kb regulatory loop contained within the 90 kb TAD encompassing the Sex combs reduced (Scr) Hox locus leads to delayed activation of Scr by a distal enhancer located ~35 kb upstream of the promoter.⁷

In this study we sought to identify “looping factors”, which mediate interactions between pairs of tethering elements separated by distances of 25–250 kb within individual TADs. We have focused our efforts on the organization of the 3D regulatory genome of the larval brain. Our previous analysis of GAF (GAGA-associated factor) in the early Drosophila embryo was complicated by strong maternal expression, which could be diminished by expressing an F-box protein fused to a nanobody recognizing GFP to ubiquitinate endogenously GFP-tagged GAF.^8,9 We also detected losses of tether-tether loops in wing imaginal discs derived from wandering larvae of GAF mutants. Here, we have switched our emphasis from discs to brains since a survey of different tissues revealed more loops in the brain (645) than any other tissue.

High-resolution Micro-C assays reveal a loss of 64 of 645 loops in hand-dissected brains derived from homozygous mutant wandering larvae lacking the GAF-POZ domain. A slightly smaller fraction of loops (14 of 209) is lost in wing imaginal discs of mutant larvae.⁸ However, GAF is also known to function as a pioneer factor to stabilize paused RNA Pol II when bound to promoter-proximal regions. ^10–13 Moreover, GAGA and GAF have also been implicated in the activities of PREs (Polycomb-response elements), thereby complicating emerging evidence that a subset of tethering elements in the early embryo might also function as PREs at later stages of development.^14–16 We therefore sought to identify additional looping factors that are not encumbered by associations with PREs and Polycomb-mediated gene silencing.

A computational survey of embryonic tethering elements identified a number of over-represented sequence motifs, including GCAACA. This motif is also over-represented in tethering elements active in the larval brain. Previous studies identified CG11504 as a putative transcription factor that binds this motif.^17–20 Using a combination of genetic analysis, ChIP-Seq assays, Micro-C maps, and RNA-Seq analysis, we present evidence that CG11504, hereafter called Vostok, functions as a looping factor that is essential for 3D genome organization and gene regulation in the Drosophila brain.

RESULTS

We aimed to identify additional looping factors based on their ability to bind tethering elements. To achieve this, we performed a motif enrichment analysis using MEME-ChIP Suite 4.11.2.²¹ To avoid a preferential search for GAGA motifs, the GAF-bound tethering elements were separated from others into an independent group using the available GAF embryo ChIP-seq data. As expected, tethering elements enriched for GAF binding revealed GAGA repeats (Fig. S1A). Interestingly, those lacking GAGA/GAF exhibited significant enrichment for a different motif, GCAACA, which closely matches the recognition sequence for the MADF protein encoded by CG11504 (Fig. S1B).¹⁹

To validate CG11504/Vostok binding, we used ENCODE ChIP-seq data obtained from 0–18h embryos expressing CG11504-GFP. Analysis of the ChIP-seq data confirmed that Vostok binds the GCAACA motif in vivo (Fig. S1C) and thereby interacts with a subset of tethering elements (Fig. S1D).

To study the role of Vostok in 3D genome organization, we generated and analyzed Vostok mutants via genome editing using CRISPR-Cas9. A frameshift mutation was introduced at codon 19, predicted to produce a truncated peptide lacking the critical recognition helix in the MADF DNA-binding domain (Fig. 1A, S2A). Vostok is expressed throughout the larval body, with slightly higher expression in the brain (Fig. S2B). Homozygous mutants survived until the wandering larval stage and proceeded to puparium formation but died during pupation. This suggests that maternal contributions of Vostok may be sufficient for early survival, while the loss of Vostok, primarily in brain and other tissues, leads to severe developmental defects during pupation and lethality. We examined mutant wandering larvae in the remainder of our experiments.

Figure 1. Vostok is enriched at specific loop anchors in the Drosophila larval brain.

(A) Structure of Vostok protein, 410 aa. MADF - myb/SANT-like domain in Adf-1 domain, CC - coiled-coil region, and LC - low-complexity regions.

(B) Vostok DNA-binding motif defined by ChIP-seq in the Drosophila larvae brains. The top250 Vostok binding peaks −+250 bp from peak summit were analyzed by MEME suite 5.5.6.

(C) Distribution of Vostok and GAF on Vostok Differential Bound ChIPseq peaks. Vostok(+)TE an Vostok(−)TE – Vostok peaks −+1000 bp from peak summit that overlap and don’t overlap tethering elements (TE), respectively. Vostok binding is lost in ΔVostok mutants, while GAF remains mainly unaffected.

Analysis of Vostok mutants

To validate the efficiency of the frameshift mutation we generated a Vostok-specific antibody and performed immunofluorescence (IF) assays on larval brains. We observed a substantial loss of Vostok staining in mutants (Fig. S2C) as compared with yw line used as wild-type control (Fig. S2C′). The IF results were further validated by Western blot (Fig. S2D). However, at this stage, brain morphology remained largely unchanged, as indicated by the DAPI signal (Fig. S2C, S2C′) and the expression of early brain developmental genes, including dpn, ase, and wor (Fig. S2E). ChIP-Seq assays confirmed that Vostok binds to GCAACA motifs in brains (Fig. 1B, Fig. S3A-B) with 31% of peaks having a dimeric motif, GCAACATGTTGC (Fig. S3C). According to differential binding analysis, there are 369 Vostok binding peaks showing a significant reduction in mutants (Vostok DiffBind peaks) out of a total of 1440, and 76 of them overlap with putative tethering elements (Fig. 1C). By contrast, ΔVostok homozygous mutant brains do not display significant changes in the binding of GAF (Fig. 1C), suggesting that it does not depend on Vostok.

We utilized high-resolution Micro-C contact maps to determine whether Vostok mutants affect chromatin loops. We identified 645 loops in the Micro-C maps of brains derived from yw (control) wandering larvae (Table S1). These loops mostly span distances of ~25–250 kb and are contained within individual TADs. We refer to these as local loops to distinguish them from meta-loops described previously.²² Micro-C contact maps were then generated for hand-dissected brains from Vostok mutant wandering larvae. A comparison of loop strengths between yw and Vostok mutants revealed losses or significant reductions of 47 of the 645 identified loops (Table S2, Fig. 2A and 2B). The reduction of these 47 loops was confirmed using an aggregation plot (Fig. S4). Additionally, there were 45 increased loops, though the increase was not dramatic (Fig. 2A, Fig. S4). Notably, there were no significant levels of Vostok at tethering elements exhibiting slight increasing in looping frequencies (Fig. S5A, B). Therefore, we focused on the reduced loops in the remainder of this study and considered the increased loops as part of the unaffected loop set.

Figure 2. Some of chromatin loops are reduced in ΔVostok mutants.

(A) Differential loop strength analysis of all larval brain loops yw vs ΔVostok mutants, highlighting significantly altered loops (cyan).

(B) A total of 47 out of 645 chromatin loops show reduced strength in ΔVostok mutants.

(C) Of these, 19 reduced loops are associated with decreased Vostok protein binding.

(D, F) Micro-C maps illustrating reduced loops at the kek1 locus (D) and the NetA-NetB locus (F) in ΔVostok mutants.

(E, G) Corresponding reductions in Vostok binding at the kek1 (E) and NetA-NetB (G) loci in ΔVostok mutants. The positions of tethering elements and genes are shown at the bottom of the maps.

The top 10 reductions (Fig. 2A) include genetic loci encoding neural adhesion proteins, such as Netrin-A(Net-A), Fasciclin 2 (Fas2), and neuropeptide receptor CCKLR-17D1. ~40% of 47 reduced loops display a significant reduction of Vostok binding in mutants based on ChIP-seq assays (Fig. 2C). Consistently, the Vostok binding signals are more centralized and stronger at these loops that are diminished and display reduced binding in mutants as compared with unaffected loops (Fig. S5B). Notable examples (Fig. 2D-G) include a ~30 kb loop at the kek1 locus (inhibitor of EGF signaling implicated in axonogenesis,²³ Fig. 2D and 2E) and a ~100 kb loop that connects the promoter regions of Netrin paralogs (Net-A and Net-B, Fig. 2F and 2G), which encode classical axonal guidance cues.²⁴ These observations suggest that Vostok functions as a looping factor that fosters interactions among a subset of tethering elements in the larval brain containing GCAACA motifs.

3D Loops and Gene Regulation

We next performed RNA-Seq assays on larval brains from yw line and Vostok mutants to determine whether losses of Vostok-dependent regulatory loops correlate with changes in gene expression. We have previously shown that deletions of specific tethering elements result in reduced or delayed patterns of gene expression in the early Drosophila embryo due impaired long-range enhancer-promoter interactions.²² GAF mutants also cause losses of specific regulatory loops in wing imaginal discs, along with corresponding changes in gene expression.²⁵

RNA-Seq assays suggest that a number of genes exhibiting reductions or loss of tether-tether regulatory loops also display diminished levels of expression (Fig. 3A). This set includes several neurogenic genes, such as Pvf3 (a growth factor for receptor tyrosine kinases),^26,27 CCKLR-17D1, Fas3, kek1, and dpr5 (Fig. 3B). These genes exhibit a coordinated reduction in Vostok binding at their associated tethering elements. Pvf3 shows a particularly marked reduction in expression in Vostok mutants (Fig. 3F). It contains a ~30 kb regulatory loop that connects alternate promoters for the Pvf3 transcription unit (Fig. 3C-F). Each of the promoter-proximal tethering elements contain several GCAACA motifs and bind Vostok, but not GAF (Fig. 3D and 3E). The Pvf3 loop is unaffected in GAF mutants (Fig. 3C).

Figure 3. Pvf3 transcription is reduced in ΔVostok mutants, with associated loop and Vostok protein binding reductions.

(A) Differentially expressed genes in ΔVostok mutants, with genes associated with reduced loops highlighted in red (41 genes with log2 padj > 50 were cut off to zoom on the genes of interest).

(B) Differentially expressed genes in ΔVostok mutants, highlighting genes associated with both reduced loops and diminished Vostok protein binding in red.

(C) Reduced chromatin looping at the Pvf3 locus in ΔVostok mutants, but not in ΔPOZ GAF mutants.

(D-F) ΔVostok mutants show reduced Vostok binding (D), absence of GAF binding (E), and decreased RNA-seq signal at the Pvf3 locus (F). P1 and P2 – two alternate promoters for the Pvf3 transcription unit.

(G, H) HCR FISH analysis reveals reduced Pvf3 transcription in the larval brain of ΔVostok mutants (H) compared to yw control (G).

To further validate the RNA-seq results (Fig. 3F), we performed hybridization chain reaction fluorescence in situ hybridization (HCR FISH) using Pvf3 probes in brains isolated from wandering larvae. In control brains, Pvf3 was widely expressed (Fig. 3G), whereas in Vostok mutants, the signal was significantly diminished (Fig. 3H). Altogether, these results establish a close correspondence between Vostok binding motifs, the formation of a tether-tether loop, and reduced Pvf3 expression in mutant larvae.

Combinatorial activities

To explore potential combinatorial activities of Vostok and GAF, we compared brain Micro-C maps from control, GAF, and Vostok mutants. From this analysis, 64 out of the 645 local loops were found to be reduced in GAF mutants (Fig. 4A and Fig. S6A). As indicated earlier, 47 loops are lost or reduced in Vostok mutants. Altogether, there are 102 loops altered in GAF or Vostok mutants, but just nine of these appear to depend on both (Fig. 4A). The majority depend on either Vostok (e.g., the kek1 locus; Fig. S6B) or GAF (e.g., the elB/noc locus; Fig. S6C). To rule out the possibility that the lack of overlap is due to insufficient statistical power for detecting differences—or that weaker reductions in looping intensity may be shared by the factors but fall below the threshold for statistical significance—we performed an aggregation plot analysis. The aggregated signal for the GAF mutant sample, focusing on the reduced loops in the Vostok mutant, is not noticeably different from that of the yw sample (Fig. S4), and vice versa (Fig. S4). These findings suggest that GAF and Vostok mainly function independently in regulating these two sets of loops.

Figure 4. Nine local loops and two meta-loops are reduced in both ΔVostok and ΔPOZ GAF mutants.

(A) Classification of local loops reduced in either ΔVostok or ΔPOZ GAF mutants, or both.

(B) Example Micro-C map of a local mamo locus loop reduced in both mutants.

(C) Reduced Vostok binding at mamo locus loop anchors in ΔVostok mutants.

(D) GAF binding is unaffected at mamo locus loop anchors in ΔVostok mutants.

(E) Two meta-loops are reduced in both mutants.

(F, G) Vostok (F) and GAF (G) binding remain largely unchanged within the hbs-containing TAD in ΔVostok mutants. However, the comparison of the Vostok ChIP-seq signal normalized to the Input in yw and mutant flies by subtraction (yw-ΔVostok*) demonstrates higher level of Vostok protein enrichment in the yw line. The possible reasons for Vostok stability at hbs TE are discussed in the “Limitations of the study” section.

(H, I) At the sns locus, Vostok binding is low in both control and ΔVostok mutants (H), and GAF binding remains unaltered in ΔVostok mutants (I).

Among the nine overlapping loops, one notable example is mamo, which encodes a POZ-containing protein (Fig. 4B).^28,29 The mamo locus contains an ~90 kb loop spanning the promoter region and 5’ flanking sequence. Vostok binds to the two tethering elements, but is lost in brains derived from Vostok mutants (Fig. 4C). GAF binding is not altered in Vostok mutants (Fig. 4D), suggesting that the two looping factors bind independently of one another. There is a reasonable correlation between the levels of Vostok and GAF binding to the mamo tethering elements, and the extent to which focal contacts are diminished in Micro-C maps of the corresponding mutants (Fig. 4B, C and D).

It was previously reported that two meta-loops associated with the meta-domain containing sticks and stones (sns) and hibris (hbs) were reduced in GAF mutants in the larval brain.²² The two genes are paralogs that encode IgSF (immunoglobulin superfamily) neural adhesion proteins implicated in the formation of synapses in the optic lobe and other regions of the nervous system.³⁰ Both of these meta-loops are also diminished in Vostok mutants, although GAF mutants display somewhat stronger losses (Fig. 4E and Fig. S7A-B). As seen for mamo, Vostok mutants do not alter binding of GAF to the tethering elements, suggesting independent binding. Curiously, the sns promoter-proximal tethering element contains GAF (Fig. 4I) but not Vostok (Fig. 4H), while the hbs tethering elements bind to both GAF (Fig. 4G) and Vostok (Fig. 4F). It is possible that these differences in binding influence how sns and hbs respond to shared enhancers contained within the meta-domain. For example, GAF mutants show reduced expression of sns but not hbs in RNA-Seq assays²², while Vostok mutants alter hbs but not sns (Fig. S7C).

DISCUSSION

We have presented evidence that CG11504/Vostok encodes a sequence-specific looping factor in Drosophila. It binds to a subset of tethering elements in the larval brain to form 47 3D regulatory loops fostering enhancer-promoter interactions. For example, the Pvf3 locus contains a 30 kb tether-tether regulatory loop that is lost in Vostok mutants. This loss is associated with a marked reduction of Pvf3 expression in different neurons of the larval CNS based on both RNA-seq and HCR in situ hybridization assays (e.g., Fig. 3).

We also presented evidence that a previously identified looping factor, GAF (GAGA-associated factor), functions in the larval brain to produce 64 regulatory loops. As seen for Vostok, many of the regulatory loops are associated with genes, including Net-A, Net-B, CCKLR-17D1 and dpr5, involved in neurogenesis, neuronal function, or the formation of specific synapses.^31,32 The brain contains more loops than other larval tissues, suggesting an added level of complexity in the control of gene expression.²² In this regard we note that many of the genes containing regulatory loops encode cell surface adhesion proteins, particularly members of the immunoglobulin superfamily (IgSF genes). Many of these genes (e.g., Dprs, Dips, Sides and Beats) have been shown to be expressed in dynamic patterns across different regions of the CNS as post-mitotic neurons form specific synaptic interconnections.^32–35

Altogether, Vostok and GAF account for 16% of the regulatory loops seen in the larval brain. This suggests that there are additional looping factors yet to be identified. We expect these factors to share some of the properties seen for Vostok and GAF, namely, recognition of simple repeated sequence motifs, phase separating properties, and zygotic lethality during pupal development when the adult CNS is established. It is possible that additional members of the MADF and POZ families will be found to function as looping factors. Prime candidates include Adf1 (MADF) and Lola (POZ).^36,37

For the most part, Vostok and GAF appear to function independently of one another (Fig. S6). Mutants alter different subsets of regulatory loops, and mutations in one gene do not significantly alter the binding activities of the proteins encoded by the other. Nonetheless, there are a few loops (9/102) that are disrupted in both Vostok and GAF mutants. It is of particular interest that both GAF and Vostok are associated with meta-loops connecting sns and hbs, IgSF paralogs separated by ~6 Mb along the right arm of chromosome 2. These meta-loops bypass at least 50 tethering elements forming local loops in some of the genes located between the sns and hbs loci. Perhaps a combinatorial code of looping factors underlies this specificity, but it will be necessary to identify additional looping factors to validate or disprove this possibility.

Vostok contains a MADF domain, which is unique to animal genomes although lacking in many mammals including humans. This might offer an opportunity for editing our 3D genomes. We have found that the strongest Vostok-specific tethering elements contain clusters of at least two motifs with the GCAACA core hexanucleotides. It might be possible to produce ectopic 3D loops within the human genome using a combination of dimeric motifs (synthetic tethering elements) and conditional Vostok expression.

Limitations of the study

We used CRISPR-Cas9 to mutate Vostok that caused a frameshift and premature termination of the Vostok coding sequence. Although we observed a significant decrease in protein levels through immunofluorescence, a small amount of fluorescence signal could still be detected in mutants (Fig. S2C). This suggests that there might be some perdurance of maternal proteins in the larval brain. Alternatively, it is possible that the Vostok antibody cross-reacts with additional proteins, including additional members of the MADF family.

We observed the changes in the positions of boundaries in our Micro-C maps of Vostok mutants (e.g., Fig. 2D) and investigated whether these changes were caused by mutations or other structural variations in the affected regions. Using cooltools, we identified a list of “changed boundary regions” (Table S3). However, we did not observe large fragment rearrangement around the “changed boundary regions” from the Micro-C map. We suspect that the loss of Vostok may lead to the mislocalization of certain boundary-associated proteins, which presents an interesting direction for future research.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to the lead contact, Darya Chetverina (daria.chetverina@gmail.com).

Materials availability

Unique materials generated by this study, including transgenic fly lines, are available upon request from the lead contact. This study did not generate new unique reagents.

Data and code availability

• Data for yw and GAF POZ mutant larval brain Micro-C was previously published under accession number GSE228095.²² The Micro-C and RNA-seq data for Vostok mutants and controls (RNA-seq only) generated in this study are available at GSE285744 and are publicly available as of the date of publication. The ChIP-seq data generated for this work are accessible through the GEO Series accession number GSE285842 and are publicly available as of the date of publication. All other source data have been deposited at Mendeley, including Unprocessed Original Images for Western blot (DOI:10.17632/8nd5hdc3md.1) and microscopy (DOI: 10.17632/km724c2kpt.1) and are publicly available as of the date of publication.

• Codes for differential loop strength analysis, differential gene expression, aggregation plot, “changed boundary regions” analyses, and other figure plotting are available at https://github.com/xl5525/Vostok_CG11504 (DOI: 10.5281/zenodo.15358274).

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Drosophila fly strains

The detailed fly strain and origin of the lines are described in KEY RESOURCE TABLE. All the flies were raised at 25 °C with 12h-12h day-night cycle on standard food. The halo-GAF-ΔPOZ/TM6B line was described previously ⁴⁰. The details of generation of Vostok mutant line, Vostok-mutants/TM6B, by CRISPR-Cas9 is described in Method Details section.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal Vostok antibody	This paper	N/A
Rabbit polyclonal GAF antibody	Erokhin et al., 2015 38	N/A
Rabbit polyclonal CP190 antibody	Mazina et al., 2020 39	N/A
Chemicals, peptides, and recombinant proteins
DSG (disuccinimidyl glutarate)	Thermo Fisher Scientific	20593
EGS (ethylene glycol bis(succinimidyl succinate))	Thermo Fisher Scientific	21565
E. coli Spike-in DNA	EpiCypher	18–1401
Ampure XP beads	Beckman Coulter	A63880
Hoechst 33342	Sigma-Aldrich	23491–52-3
UltraPure™ SSC, 20X	Thermo Fisher Scientific	15557044
ProLong Gold Antifade Mountant	Thermo Fisher Scientific	P36934
Critical commercial assays
HCR buffer set	Molecular Instruments	www.molecularinstruments.com
HCR B3 amplifier	Molecular Instruments	www.molecularinstruments.com
HCR B4 amplifier	Molecular Instruments	www.molecularinstruments.com
Deposited data
Micro-C for Vostok mutants in the larvae brains	This study, GEO	GSE285744
Micro-C for yw control line and GAF-ΔPOZ in the larvae brains	Xiao et al., 2023 8	GSE228095
RNA-seq for Vostok mutants and yw control line in the larvae brains	This study, GEO	GSE285744
Vostok and GAF ChIP-seq for Vostok mutants and yw  control line in the larvae brains	This study, GEO	GSE285842
Unprocessed Original Images for Western blot	This study, Mendeley data	DOI:10.17632/8nd5hdc3md.1
Microscopy images	This study, Mendeley data	DOI:10.17632/km724c2kpt.1
Github code deposition: Codes for differential loop strength analysis, differential gene expression, aggregation plot, “changed boundary regions” analyses, and other figure plotting	This study, Zenodo data	DOI:10.5281/zenodo.15358274
Experimental models: Organisms/strains
Drosophila melanogaster/ halo-GAF-ΔPOZ/TM6B	Tang et al., 2022 40	N/A
Drosophila melanogaster/ Vostok-mutants/TM6B	This study	N/A
Oligonucleotides
Pvf3-L probe	This study, table S4	N/A
Pvf3-S probe	This study, table S5	N/A
Software and algorithms
BWA	Li and Durbin, 2009 41	https://bio-bwa.sourceforge.net/
pairtools	Goloborodko et al., 2019 42	https://github.com/open2c/pairtools
Cooler	Abdennur and Mirny, 2020 43	https://github.com/open2c/cooler
Mustache	Ardakany et al., 2020 44	https://github.com/ay-lab/mustache
BEDtools	Quinlan and Hall, 2010 45	https://bedtools.readthedocs.io/en/latest/
DESeq2	Love et al., 2017 46	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
ggplot2	Villanueva and Chen, 2019 47	https://ggplot2.tidyverse.org/
cooltools	Abdennur et al., 2024 48	https://github.com/open2c/cooltools
deepTools	Ramírez et al., 2016 49	https://deeptools.readthedocs.io/en/develop/
MEME	Timothy et al., 2015 50	https://meme-suite.org/meme/tools/meme
MEME-ChIP	Machanick et al., 2011 21	https://meme-suite.org/meme/tools/meme-chip
FIMO	Grant et al., 2011 51	https://meme-suite.org/meme/tools/fimo
Tomtom Motif Comparison Tool	Gupta et al., 2007 52	https://meme-suite.org/meme/tools/tomtom
Salmon	Patro et al., 2017 53	https://combine-lab.github.io/salmon/
nf-core/RNA-seq	Ewels et al., 2020 54	https://nf-co.re/rnaseq/3.18.0/
Bowtie2	Langmead and Salzberg, 2012 55	https://bowtie-bio.sourceforge.net/bowtie2/index.shtml
HISAT2	Kim et al., 2015 56	http://daehwankimlab.github.io/hisat2/
SAMtools	Danecek et al., 2021 57	http://www.htslib.org/
MACS2	Zhang et al., 2008 58	https://github.com/macs3-project/MACS
UCSC-tools	Kuhn et al., 2013 59	https://genome.ucsc.edu/util.html
RNASTAR	Widmann et al., 2012 60	https://github.com/alexdobin/STAR
ImageJ	Collins, 2007 61	https://imagej.nih.gov/ij/
Easy_HCR	Elagoz et al., 2022 62,63	https://github.com/SeuntjensLab/Easy_HCR?tab=readme-ov-file

Ethic approval

Animal handling for the antibody production in rabbits was carried out strictly according to the procedures outlined in the NIH (USA) Guide for the Care and Use of Laboratory Animals. The protocols used were approved by the Committee on Bioethics of the Institute of Gene Biology, Russian Academy of Sciences. All procedures were performed under the supervision of a licensed veterinarian, under conditions that minimize pain and distress.

Rabbits were purchased from a licensed specialized nursery, Manihino. Soviet chinchilla rabbits used in the study are not endangered or protected. Only healthy rabbits, certified by a licensed veterinarian were used. The rabbits were individually housed in standard size, stainless steel rabbit cages, and provided an ad libitum access to alfalfa hay, commercial rabbit food pellets, and water. The appetite and behavior of each rabbit was monitored daily by a licensed veterinarian. Body weight and temperature of each rabbit were evaluated prior to and daily following the immunization. No animals became ill or died at any time prior to the experimental endpoint. At the end of the study period all rabbits were euthanized by intravenous injection of barbiturate anesthetics.

METHOD DETAILS

Generation of Vostok mutation by CRISPR-Cas9 method

The fly CRISPR target finder online program (https://flycrispr.org/ )⁶⁴ was used to select optimal guide RNA target sites (gRNA, CRISPRs) for the 5′ end of the VostokORF. The following Vostok-target CRISPR was selected: 5′- CCGCAGCCGACGCGGGCTGT −3′. The Vostok CRISPR was cloned by a PCR-based method into the prAc_yiless_U6:1:3_attB vector downstream U6:1 promoter. The prAc_yiless_U6:1:3_attB vector⁶⁵ is based on pCFD4-U6:1_U6:3tandemgRNAs plasmid (Addgene # 49411 ⁶⁶) with additional insertion of the yellow-intron-less gene under control of Actin5C promoter.

Vostok-target CRISPR plasmid was injected into embryos of ZH-attP-86Fb line. ⁶⁷ Injected embryos were grown to adulthood and crossed with y-w1118. The flies with the insertion of the Vostok target CRISPR were identified by phenotypic analysis using the yellow gene expression. To create a mutation in the Vostok gene, the CRISPR-expressing flies were crossed with flies from line y(1) M{Act5C-Cas9.P.RFP-}ZH-2A w(1118) DNAlig4(169) line (Bloomington Drosophila Stock Center #58492) that expressed Cas9 under control of the Actin5C promoter. Flies with a mutation in Vostok gene have been identified using the PCR and DNA sequencing methods.

Antibodies

Antibodies against the Vostok protein (CG11504, full length, UniProt ID Q7K4R2) were raised in rabbits. Antigen for antibody production was expressed as 6×His-tagged fusion protein in Escherichia coli, affinity purified on Ni Sepharose 6 Fast Flow (GE Healthcare), according to the manufacturer’s protocol, and injected into rabbits following standard immunization procedures. Antibodies were affinity-purified from serum on the same antigen as was used for the immunization and tested by immunoprecipitation/Western blotting (IP/WB) to confirm their specificity. Antibodies against the GAF ³⁸ and CP190 ³⁹ proteins were created previously.

ChIP-seq

All flies were maintained at 25°C on the standard yeast medium. For the dissection, homozygous wandering 3rd instar larvae were collected approximately 4–6 hours before the puparium formation. Dissected brains aimed for ChIP-Seqs were collected in tubes containing PBS/0.1% NP-40 and placed on ice. Dissected tissues were stored on ice for no more than 1 hour. After the dissection, the tissues were warmed up to room temperature and fixated by adding of formaldehyde up to 1% for 10 min. Fixation was stopped by adding of 1/20V 2.5M Gly to the reaction for 5 min. After the fixation, the Sonication buffer (ChIP lysis buffer) containing 0.1% SDS and PIC (Roche) was added to the tissues and they were stored at −70°C. The sonication step was performed after the thawing. For each ChIP-Seq sample (single biological replicate) we used 30 brains of wandering larvae. The chromatin immunoprecipitation (ChIP) was performed and analyzed exactly as previously described.^68–70 ChIP-Seq libraries were obtained using the NEBNext UltraTM II DNA library preparation kit (New England Biolabs). Next generation sequencing was performed with the Illumina NovaSeq6000 sequencer. The CG11504, GAF and Input paired-end ChIP-Seq data were obtained in 2 biological replicates for each condition (yw and CG11504 mutation).

Micro-C library preparation

Micro-C library related experiments were performed as described previously.⁸ Homozygous wandering 3rd instar larvae were collected and dissected in PBST. 25 brains were collected with cuticles for each biological replicate and two replicates for each genotype were prepared. Brains were incubated in 500 μL 1% formaldehyde/PBST for 15 min, passively rotating at room temperature. The fixation was terminated by addition of 370 μL 2 M Tris-HCl pH 7.5. Secondary fixations with 3 mM DSG (Thermo Fisher Scientific) and EGS (Thermo Fisher Scientific) were performed as previously described after furtherly dissected out brains.⁸ Fixed tissues were permeabilized with MB1 (50 mM NaCl, 10 mM Tris, 5 mM MgCl2, 1 mM CaCl2, 0.2% NP-40) on ice for 20 min. Chromatin from permeabilized tissues was digested with 30 U MNase for 10 min at 37C with 850rpm rotation in MB1. After two washes in MB2 (50 mM NaCl, 10 mM Tris-HCl pH 7.5, 10 mM MgCl2), the tissues were treated with 25 U T4 PNK at 37 °C for 15 min and 25 U Klenow Fragment at 37°C for 15 min, for chewing 3’ ends for biotin. Biotin labeling was achieved by adding biotin-dATP, biotin-dCTP, dTTP, and dGTP and incubating at 25°C for 45 min and terminated with 30 mM EDTA incubating at 65 °C for 20 min. Finally, fragmented and labeled DNA ends were ligated with 5000 U T4 ligase in 250 μL ligation mix (50 mM Tris, 10 mM MgCl2, 1 mM ATP, 10 mM DTT, 100 μg/mL BSA) at room temperature for 180 min with rotation after rinsing in 200 μL MB3 (50 mM Tris-HCl pH 7.5, 10 mM MgCl2) 3 times. Unligated ends were removed by exonuclease III for 10 min at 37C. After reverse-crosslinking in 1 mg/mL Proteinase K and 1% SDS at 65 °C overnight, DNA was purified using PCI and ethanol precipitation, treated with 2 μl RNase A at 37 °C for 30 min and purified on a ZymoClean spin column. Informative fragments were immobilized on MyONE Strptavidin C1 Dynabeads. Sequencing libraries were prepared using NEBNext Ultra II DNA Library Prep Kit with NEBNext Multiplex Oligos and amplified with KAPA HiFi Hot Start Mix for 8 cycles. The library was finally purified with 0.9x Ampure XP beads.

The libraries were sequenced by the Princeton University Genomics Core Facility with NovaSeq S1 100nt Flowcell v1.5.

RNA-sequencing

10 brains were dissected from homozygous wandering 3rd instar larvae in PBS and transferred to TRIzol (Thermo Fisher Scientific) for each biological triplicate. The total RNA was extracted based on standard protocols. The libraries were prepared using NEBNext^® Ultra^™ II Directional RNA Library Prep and sequencing were performed by the Princeton University Genomics Core Facility.

In situ HCR

The split probes were designed using Easy_HCR ^62,63 with default parameters. The HCR buffers and HCR amplifiers were purchased from Molecular Instruments. The Pvf3 probe was associated with B3 amplifier sequences.

Homozygous wandering 3rd instar larvae were dissected in PBST and fixed in 4% formaldehyde/PBST for 15min at room temperature. The samples were rinsed in PBST 3 times at room temperature for 5min. Then the samples were processed as described in https://files.molecularinstruments.com/MI-Protocol-RNAFISH-GenericSolution-Rev7.pdf with small modifications. Volumes for all the steps were reduced to 100 μL and concentrations for the probes were adjusted to 16 pmol. The hybridization incubation time was reduced to 90min. DAPI was diluted in 5xSSCT at 1 μg/ml final concentration to label nuclear and applied at the 30min final washing step. The sample was mounted in VECTORSHEILD Mountant.

The brains were imaged by a Zeiss LSM 880 confocal microscope (Zen software 2.3 SP1) with a Plan-Apochromat ×40/1.3 N.A. oil-immersion objective. Three laser lines at 405 nm for DAPI, 633 nm for B3–647 amplifier were used. The z-stack was taken 2 μm apart for 55–60 steps and all steps were projected for the final figures. The pictures were adjusted and the total intensity of each channel was measured in ImageJ⁶¹.

QUANTIFICATION AND STATISTICAL ANALYSIS

ChIP-seq data analysis

The following publicly available single-end ChIP-Seq data for the Drosophila embryos were used: CG11504-GFP 0–18h embryo stage - GSE256994 (ENCFF895KIP) ⁷¹, GAF embryo stage 5 - GSE152770 GAF (GSM4626049 for the GAF and GSM4626048 for the Input).⁹ We have not used any figures or text from the previously published manuscripts—only analyzed the data deposited at the free public access databases. The raw reads in FastQ format were trimmed with Trim Sequences (Galaxy Version 1.0.2), then mapped to the dm6 Drosophila genome assembly using Bowtie2 (Galaxy Version 2.3.4.3) ⁵⁵ and filtered (with a minimum MAPQ quality score = 5) ⁷². BigWig files were generated using bamCoverage (Galaxy Version 3.5.1.0.0) with scores representing number of reads normalized by the size of the library (the protein-binding levels were normalized to the genome content, calculated as RPGC: number of reads per bin/(total number of mapped reads * fragment length/effective genome size) ⁴⁹. The final BigWig files (representing the protein-binding profiles) for both, CG11504 and GAF, were obtained using bigwigCompare tool (Galaxy Version 3.5.1.0.0)⁴⁹ as a ratio of ChIP signal to GSM4626048 Input. The peaks of CG11504 and GAF protein binding were defined by MACS2 ⁵⁸ callpeak (Galaxy Version 2.1.1.20160309.6) with the following parameters: --qvalue 0.01 --mfold 5 50 --bw 350 ⁷³. The GSM4626048 input DNA was used as a control for peak calling for both CG11504 and GAF.

Here and after, the required sequences were extracted from dm6 coordinates by bedtools getfasta tool (Galaxy Version 2.31.1)⁴⁵ ). The definition of DNA motifs for CG11504 binding was performed with MEME-ChIP suite 4.11.2²¹ (using (−+)250bp intervals from peaks summits of the Top500 peaks with the highest MACS2 callpeak scores.

Tethering elements⁷ were divided into two groups: overlapping and non-overlapping GAF peaks using bedtools Intersect intervals tool ⁴⁵. Motifs were defined by MEME-ChIP suite 4.11.2 ²¹. Candidate proteins were searched using Tomtom Motif Comparison Tool Version 5.5.5 ⁵².

In case of the larvae brain paired-end CHIPseq, reads in FastQ format were mapped to the Drosophila genome assembly dm6 using HISAT2⁵⁶. The 2 biological replicates for each protein or Input were merged using Merge BAM Files tool (Galaxy Version 1.2.0) and filtered with a minimum MAPQ quality score = 10 ⁷². BigWig files were generated using bamCoverage (Galaxy Version 3.5.4) with scores representing number of reads normalized by the size of the library (the protein-binding levels were normalized to the genome content, calculated as RPGC: number of reads per bin/(total number of mapped reads * fragment length/effective genome size) ⁴⁹. The final bigwig files (representing the protein binding profiles) were obtained using the bigwigCompare tool (Galaxy Version 3.5.4) as a subtract of the Input ChIP signal (all inputs were preliminary smoothed over a 1 kb window)⁴⁹. The peaks of CG11504 and GAF protein binding were defined by MACS2 callpeak relative control Input (Galaxy Version 2.2.9.1) with the following parameters: --qvalue 0.01 --mfold 5 50 --bw 300⁷³.

The definition of DNA motifs for CG11504 or GAF binding was performed using MEME 5.5.6⁷⁴ and MEME-ChIP suite 4.11.2 ²¹ using (−+)250bp intervals from the peaks summits of the top250 peaks with the highest intensity MACS2 callpeak scores.

Differential binding analysis of the CG11504 ChIP-Seq peak data was performed using the DiffBind package and the DESeq2 method (Galaxy Version 2.10.0) with the following conditions: FDR Threshold 0.05, FC<=−1, p<=0.05⁷⁵ comparing yw control and CG11504 mutant conditions. For the DiffBind analysis the filtered BAM Files and MACS2 callpeak relative control Input (narrow peaks) were recounted independently for each of the biological replicates using the same conditions as described above.

The CG11504 DiffBind peaks were aligned to the corresponding Vostok peak summits and the required peak intervals were generated. The sequences of the 369 Vostok differential binding (DiffBind) peaks −+250 bp from the peak summit were analyzed by FIMO suite Version 5.5.7 ⁵¹.

The plotHeatmaps were generated using the Deeptool2 package computeMatrix and plotHeatmap tools⁴⁹. Analysis of ChIP-Seq data was performed on the Galaxy-P platform⁷⁶. All of the processed files (both bigwig and bed) were obtained with coordinates of the Drosophila genome assembly dm6.

Micro-C data processing

The raw data were mapped onto the dm6 reference genome as described by bwa⁴¹. Validated paired reads were filtered and finalized using pairtool⁴², and further processed into multi-resolution matrix by Cooler⁴³. pyGenomeTracks⁷⁷ was used to generate the figures for example contacts. The detailed processing information is described as previously.⁸

Systematic loop calling and differential looping analysis

Focal contacts in brains of yw were detected by Mustache⁴⁴. The parameters were set as pt_0.0001_st_0.88_sz_2.0. The loops were annotated by nearest transcription start sites by BEDtools. We used previous lists of loops in the early embryo and wing discs ⁸ as back ground loop list and combined these two loop list with the brain loop list as standard loop list for our analysis.

The standard loop lists were used to obtain paired reads from individual samples. The dumped matrix with 100bp bin size from the cooler files were compared with the standard loop list with anchors normalized to 3.2kb, using the pairToPair function from BEDtools⁴⁵. The reads were summed for each standard loop and summarized into a count table. The count tables were imported for differential looping analysis with DESeq2⁴⁶. P-value < 0.05. The differential looping results were volcano plotted by ggplot2⁴⁷ and ggrepel⁷⁸.

RNA-seq data analysis

Processing of raw fastq files was performed using the nextflow pipeline^54,79. The differential gene analysis was performed with DESeq2⁴⁶. Significantly differentially expressed genes were defined by adjusted p-value < 0.05. The results were visualized by ggplot2⁴⁷ and ggrepel⁷⁸.

Analyses of “changed boundary regions” in Vostok mutants

The strong boundaries were called by cooltools for both, which were identified by cooltools as strong boundaries at 1600, 3200 and 4800bp windows and at 1600bp resolution. ⁴⁸

Supplementary Material

1

Spreadsheet. Table S1. The list of 645 loops identified in the Micro-C maps of brains derived from yw control wandering larvae. Related to Figure 2.

Supplementary File 1. Supplementary Figures S1-S7 and Tables S2-S5.

Spreadsheet. Table S1.

HIGHLIGHTS.

The MADF transcription factor, Vostok, establishes brain-specific regulatory loops

Vostok was identified by a computational survey of tethering elements.

Meta-loops depend on a combination of two looping factors, Vostok and GAF.