Survey of activation‐induced genome architecture reveals a novel enhancer of Myc

Abstract The transcription factor Myc is critically important in driving cell proliferation, a function that is frequently dysregulated in cancer. To avoid this dysregulation Myc is tightly controlled by numerous layers of regulation. One such layer is the use of distal regulatory enhancers to drive Myc expression. Here, using chromosome conformation capture to examine B cells of the immune system in the first hours after their activation, we reveal a previously unidentified enhancer of Myc. The interactivity of this enhancer coincides with a dramatic, but discrete, spike in Myc expression 3 h post‐activation. However, genetic deletion of this region, has little impact on Myc expression, Myc protein level or in vitro and in vivo cell proliferation. Examination of the enhancer deleted regulatory landscape suggests that enhancer redundancy likely sustains Myc expression. This work highlights not only the importance of temporally examining enhancers, but also the complexity and dynamics of the regulation of critical genes such as Myc.


INTRODUCTION
In order to mount an appropriate immune response, immune cells process activation signals from pathogens and immune accessory cells. The sum of these signals induces proliferation, differentiation and, in the case of lymphocytes, clonal expansion. The magnitude of this response is critical. An underestimation of the threat could be fatal, while an overestimation of the danger could lead to autoimmune damage. Thus, activation induces dramatic and rapid, but tightly controlled molecular changes in immune cells, including transcriptional, 1 epigenetic, 2,3 proteomic, 4 metabolomic 5 and three-dimensional genome organizational changes. 1 Among the most archetypal and well studied of these changes are those that occur to the gene of the transcription factor Myc and its protein product. 6 Upon lymphocyte activation, Myc is rapidly upregulated in a highly controlled manner in order to oversee further transcriptional changes central to immune cell activation. [7][8][9] Unsurprisingly, given its importance in both the immune context and in all healthy and diseased cell proliferation, 10 the body of research on the regulation of Myc is extremely rich. 11,12 Interestingly, many of the fundamental discoveries involving the regulation of Myc were made in immune cells. For example, shortly after the first reports of the existence and function of distal regulatory enhancers, 13 Myc was among the first eukaryotic genes in which regulation by these gene regulatory elements was explored. 14,15 While these early studies explored the role of aberrant enhancement of Myc expression driven largely by translocation, the normal genomic location of Myc, within a large gene-poor region in both mice and humans, made it an excellent model for the exploration of distal gene regulation. A large fraction of this regulation occurs in three-dimensions, with distal enhancers being physically drawn to the Myc promoter to drive expression. Thus, it is unsurprising that the invention of chromosome conformation capture, which can reveal the three-dimensional proximity of DNA to other DNA, 16,17 drove significant advancement of our understanding of the distal regulation of Myc. 11 These studies revealed an elaborate, cell-type-and cell-statespecific three-dimensional enhancer landscape controlled by umpteen transcription factors, histone modifications, DNA methylation, long non-coding RNAs, 18 among other mechanisms. 11 Here, via exploring changes in three-dimensional genome organization in the first hours post-B cell activation, we reveal a previously uncharacterized upstream enhancer of Myc, apparent for mere hours following activation, which accompanies the rapid and dramatic spike in post-activation Myc expression. However, genetic removal of this enhancer leads to minimal impact on either Myc expression or Myc protein level, which is likely to be due to differential enhancer usage or enhancer redundancy sustaining critical levels of Myc.

Activation induces genome reorganization upstream of Myc promoter
To begin exploring the three-dimensional genome architectural change that may regulate early activationinduced transcriptional change, we examined paired chromosome conformation capture (in situ HiC) and RNA-Seq data of activation-induced B cell differentiation 1 with a particular focus on the changes between na€ ıve and 3 h activated B cells (Figure 1a).
First, we examined all the differential interactions (DIs) between na€ ıve and 3 h activated B cells associated with the top 100 differentially expressed genes (DEs) between the same cells (Supplementary table 1). Differential interactions are statistically significant differences in the DNA-DNA interaction frequency between any two points on the linear genome between samples, determined using diffHiC, 19 and denote changes in enhancer-promoter DNA loops, or other genome architectural changes. The top 100 differentially expressed genes associate with just 19 DIs, four of which associate with the promoter of the Myc gene ( Figure 1b). This result suggests that Myc is one of the few genes regulated by early activation-induced genome architectural change in B cells.
To further define the nature and position of the Myc associated activation-induced genome architectural changes, we generated contact matrices of the genomic regions containing the DIs (Figure 1c) and performed virtual 4C of the same region using the Myc promoter (2 kbp upstream and 5 kbp downstream) as the viewpoint (Figure 1d, viewpoint shown in red). This analysis revealed a previously unknown DNA-DNA interaction between the Myc promoter and an upstream region (Figure 1c, d in green). This statistically significant change in interactivity (P = 0.001, unpaired two-tailed t-test,) appears to be highly 3 h specific ( Figure 1e) and is the strongest architectural change in the region (Supplementary figure 1) between na€ ıve and 3 h activated B cells. Interestingly, the interactivity of this region and the Myc promoter ( Figure 1e) across activation-induced B cell differentiation is reflective of the expression pattern of Myc in the same cells (Figure 1f). This may reflect a regulatory role for the region in the expression of Myc.

Deletion of one putative enhancer of Myc alters Myc expression
To explore potential gene regulatory roles for the putative activation-induced enhancers discovered upstream of Myc, we first sought to clarify the potential functions of the regions using available epigenetic data. Overlaying our in situ HiC data with publicly available ATAC-sequencing 20 and H3K27 acetylation and H3K4 mono-methylation chromatin immunoprecipitation data 21 (generally associated with DNA accessibility, active enhancers and active/primed enhancers, respectively), we revealed part of the epigenetic landscape of the region in na€ ıve B cells (Figure 2a). This analysis highlighted four genomic regions of particular interest (regions 1-4). Regions 1, 2 and 3 (R1-3) were both accessible and contained epigenetic marks consistent with enhancers. Region 4 was accessible and in contact with the Myc promoter, but contained no detectable H3K27ac or H3K4me1 modifications.
To definitively characterize the function of these regions in regulating Myc expression, we genetically removed each region using clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 technologies in the A20 B cell line (Supplementary figure 2a). First, we confirmed that activation induced Myc expression shows a similar pattern in the A20 cell line as in primary B cells (Figure 2b). We then quantified the impact of the genetic  the Myc transcriptional start site (promoter) with FIMO from the MEME suite. We filtered transcription factor motifs for only those differentially expressed between na€ ıve B cells and 3 h activated B cells 1 (calculated with fold changes significantly above 1.5 with TREAR and FDR < 0.05; Supplementary table 3) and normalized motif incidence to region size. We find no transcription factor motifs unique to R1; however,   We also tested the immune response generated in allergic lung inflammation using intraperitoneal ovalbumin sensitization followed by nebulized OVA challenge (Figure 4c). 22 Cell populations in the BAL were examined the day after the final OVA challenge. Enumeration of total cells, B cells (Figure 4d) and other immune cell types (Supplementary figure 4c) revealed no significant differences between Myce mice and littermate controls.
Together with the unchanged numbers or proportions of immune cells in steady state Myce mice (Supplementary figure 3a-k), these experiments suggest that the deletion of this enhancer of Myc has little to no impact on the maintenance or activation response of immune cells in vivo.

Myce mice cells display altered chromatin interactions with the Myc promoter
We next sought to determine why deletion of the Myc enhancer had little to no impact on either Myc expression, Myc protein levels or immune responses in cells from the Myce mice. To do so, we examined the consequence of the deletion on the genome architecture upstream of Myc in both the Myce mice and littermate controls. In brief, we isolated B cells from either Myce mice or wild-type littermate controls, activated them in vitro for 3 h then sorted the activated and na€ ıve B cells from the culture. A proximity ligation protocol followed by DNA precipitation was then performed on these cells, before qPCR was used to determine the frequency of DNA-DNA interaction between the Myc promoter and a number of upstream genomic regions.
The frequency of interaction determined by this qPCRbased method correlated well with in situ HiC data from the same region (Figure 5a-c) with regions of high or low interactivity detected similarly by both methods. Importantly, it also revealed that in 3 h activated B cells the interaction frequency at each examined region upstream of the deleted enhancer is greater in the Myce mice B cells than in littermate controls (Figure 5b, left of the green dotted line). This is in contrast to the regions downstream of the deletion in both Myce mice and littermate controls (Figure 5b, right of the green dotted line) and in na€ ıve B cells either up-or downstream (Figure 5c), suggesting the change in interactivity is deletion and activation induced. This change in interactivity can be more clearly observed in the frequency of interaction in the activated Myce mice B cells relative to the wild-type (Figure 5d). Furthermore, summing all relative threshold cycle (Ct) values upstream or downstream of the enhancer deleted region in both na€ ıve and activated B cells reveals a significant (P = 0.04, paired two-tailed t-test) increase in interaction frequency upstream of the deletion, specifically in activated B cells (Figure 5e).  These experiments suggest that the reason that Myc levels, or indeed immune responses, are unchanged in the Myce mice B cells under various activation conditions is that in the absence of the enhancer the Myc promoter simply continues searching the upstream DNA to find another enhancer to sustain its expression. This is reflected in the increased interaction frequencies upstream of the enhancer deletion in Myce mice activated B cells.

DISCUSSION
Here, through a comprehensive examination of genome organizational changes in the hours after B cell activation, we identified a previously undetected enhancer of Myc, in a highly scrutinized regulatory landscape. 11 However, genetic deletion of this enhancer has little to no impact on Myc expression, protein levels or in vitro and in vivo cell proliferation or survival. Further exploration revealed that Myc expression is maintained in the absence of this activation-induced enhancer likely via differential enhancer usage or enhancer redundancy.
Enhancer redundancy is the process by which gene expression is sustained upon the deletion of an enhancer of a gene by the action of other enhancers of the gene. 23 This process has been widely documented across the kingdoms of complex life [24][25][26][27] and appears to be enriched in genes important in development and health, supporting robust transcription buffered against environmental and genetic perturbation. For example, the total number of predicted enhancers and the redundancy of transcription factor binding within these enhancers is predictive of a genes potential pathogenicity. 28,29 Experimentally testing this potential is challenging as the deletion of enhancers in these redundant regulatory landscapes frequently results in little to no phenotypic impact. 23 However, this is not true in all cases. For example, in the Myc regulatory landscape, deletion of eight different enhancers within 2 megabases downstream of the Myc promoter, 30,31 or large sections of the region, 32 each lead to significant reductions in Myc expression.
Our detailed exploration of the three-dimensional organization in the enhancer deleted genome provides interesting insights into potential mechanisms of enhancer redundancy. As such, in 3 h activated B cells, deletion of the activation induced enhancer appears to increase DNA-DNA interactions between the Myc promoter and numerous upstream enhancers. Furthermore, the magnitude of these interaction changes suggests multiple new interactions per B cell. This suggests that the redundancy observed is not simply a case of the Myc promoter contacting and utilizing the next upstream enhancer to maintain appropriate regulation, but potentially the whole upstream regulatory landscape.
Why some enhancers exhibit redundancy and others do not is unknown. What is clear is that the combined function of enhancers is extremely context specific. For example, the enhancers of hunchback in Drosophila melanogaster will behave either additively (the induction of expression in the presence of both enhancers is the sum of both individually) or subadditively (the induction of expression in the presence of both enhancers is less than the sum of both individually) depending on the concentration of the transcription factor, Bicoid. 33 Similarly, two enhancers of pomc in mice have been demonstrated to function additively in adult mice neurons, but superadditively (the induction of expression in the presence of both enhancers is greater than the sum of both individually) in young mice neurons. 34 Thus, while it is clear that transcription factors can influence enhancer usage, epigenetics, the relative position of the enhancer to the promoter or other enhancers, among many other variables, likely impact enhancer redundancy. Transcription factors likely play a role in the enhancer redundancy observed in this study as the deleted enhancer was enriched for transcription factor motifs, including NFjb. NFjB has previously been shown to regulate Myc expression via binding near the promoter, 35 so under normal conditions our enhancer likely forms part of this NFjB -Myc regulatory system.
Here we have revealed a slight, but significant, change in steady state Myc expression levels upon deletion of a novel enhancer, revealed by a detailed examination of genome organization just hours after cell activation. However, enhancer redundancy has made it challenging to accurately dissect the normal function of this enhancer. Nonetheless, the correlation between the temporal specificity of the interaction between our novel enhancer and the Myc promoter and the dramatic and discrete spike in Myc expression at 3 h post-activation is compelling. Thus, we propose that under normal conditions this activation-induced enhancer assists in regulating this critical and dramatic increase in Myc expression. The development and application of new super-resolution live imaging technologies that allow concurrent visualization of multiple regions of DNA and associated factors at single molecule resolution, 36,37 will likely soon enable the visualization and dissection of not only this function, but the discussed genome architecture of the enhancer deleted Myc regulatory landscape and definition of the factors regulating redundancy.

Intracellular Myc staining
For intracellular staining of Myc, the cells were harvested, centrifuged at 5009 g at 4°C for 5 min before resuspension in fixation buffer (0.5% PFA (Sigma-Aldrich) + 0.2% Tween 20 (Sigma-Aldrich) in PBS) and incubation at 4°C for 24 h. The cells were washed with FACS buffer before pelleting at 5009 g at 4°C for 5 min and maintained in FACS buffer at 4°C until staining of all samples within each experiment could be performed simultaneously. For staining, the cells were incubated with either anti-Myc (clone D84C12, Cell Signaling, Danvers, USA) or a rabbit IgG isotype control antibody in PBS-0.1% BSA (Sigma-Aldrich) for 45 min at room temperature. The cells were washed with PBS-0.1% BSA before incubating with anti-rabbit IgG Alexa Fluor 647 (A21244, Life Technologies, Carlsbad, USA) in PBS-0.1% BSA, for 45 min at room temperature. The cells were washed and flow cytometry analysis performed.

Calculation of total cell and cohort number
The total cell numbers were calculated based on the ratio of live cells to counting beads detected by flow cytometry in each sample, after a known number of beads was added prior to analysis. Dead cell exclusion was performed using propidium iodide (0.2 lg mL À1 , Sigma-Aldrich). The cohort number and the mean division number over time were calculated based upon CellTrace Violet dilution, using the "precursor cohort method" as described previously. 40

Influenza infection and analysis
Myce À/À and wildtype mice were infected with influenza A/ H3N2/X31 virus at 1 9 10 4 pfu in 25 lL by intranasal insufflation on day 0. On day 8, the mice were killed by CO 2 inhalation after which BAL was performed (250 lL twice with sterile PBS).~200 lL of blood was collected into nonheparinized tubes by retro-orbital bleeding, and the mice were then killed by CO 2 inhalation after which BAL was performed (250 lL twice with sterile PBS). Flow cytometry was performed fresh on BAL cells using the following staining panel: CD19-BUV395 Clone#1D3, CD8a-PerCPeFluor710 Clone#52-6.7 (BD), CD11c-FITC Clone #N418, GR1-PECy7 Clone#RB6-8C5, TCRb-APCeFluor780 Clone #H57-597 (eBioscience), CD4-Alexa647 Clone#GK1.5, CD11b-Alexa700 Clone#M1/70 from WEHI Antibody Facility (Supplementary  table 4). MHC Class I tetramers PA224-BV421 and NP366-PE (a generous gift from Professor Katherine Kedzierska, The University of Melbourne) were used to detect influenzaspecific CD8 + T cells. Flow cytometry was performed on a BD LSRFortessa X-20 analyzer. SYTOX Blue Dead Cell Stain was used to exclude dead cells from analysis and SPHERO Rainbow Beads were used to calculate the absolute cell counts.

Enzyme-linked immunosorbent assay (ELISA)
Antibody binding to influenza hemagglutinin protein was measured by ELISA. The 96-well Maxisorp plates (Thermo Fisher) were coated overnight at 4°C with 2 lg mL À1 recombinant X31 HA ectodomain from A/Aichi/2/1968 (SinoBiological; 40 059-V08H). The plates were blocked with 1% FCS in PBS before duplicate wells of serially diluted mouse plasma were added and incubated at room temperature for 2 h. Bound antibody was detected using 1:10 000 dilution of HRP-conjugated anti-mouse IgG (KPL) and the plates developed using TMB substrate (Sigma-Aldrich), stopped using 0.12 M sulfuric acid and read at OD 450 nm. Endpoint titers were calculated using as the reciprocal serum dilution giving signal 29 background using a fitted curve (Graphpad Prism; 4 parameter log regression).

In vitro transcription of sgRNA
The sgRNA used in this study were generated via in vitro transcription, as described previously. 41 In brief, the transcription template was generated by PCR using Q5 high fidelity DNA polymerase (New England Biolabs, Ipswich, USA) with dNTP (Promega, Madison, USA) and an annealing temperature of 58°C. Universal primers as well as a specific primer bearing the sgRNA flanked by T7 promoter sequence and scaffold (as the format T7-sgRNA-scaffold) were used (Supplementary table 2

Ribonucleoprotein (RNP) assembly and delivery
For Cas9 RNP, 150 pmol of in vitro transcribed sgRNA was incubated with 100 pmol of recombinant Cas9 nuclease (Integrated DNA Technologies, Coralville, USA) at room temperature for 15 min. RNP with 100 pmol of electroporation enhancer (IDT) were subsequently transfected into cells via electroporation. Delivery into A20 cells was performed via 4D-Nucleofector (Lonza, Basel, Switzerland) with buffer SF and pulse code FF113.

Generating and confirming the deletion of putative Myc enhancers
Deletion of candidate regions in the A20 cell line genome was performed using electroporation delivered RNP containing paired sgRNAs (Supplementary table 2). At 48 h post-electroporation, single cell clones were sorted into U-bottom 96-well plates and expanded as outlined above. A fraction of the expanded population was lysed in Direct PCR lysis reagent (Viagen, Cedar Park, USA) with proteinase K (Roche, Basel, Switzerland). Genotyping was performed directly from the lysates with region targeting primers (Supplementary table 2

Quantitative reverse transcription PCR (qRT-PCR)
RNA was extracted using NucleoSpin RNA Plus (Macherey-Nagel, Duren, Germany) with gDNA removal, 1 lg of RNA was then reverse transcribed with random hexamer and anchored oligo dT primers using SensiFast cDNA synthesis kit (Bioline). Quantitative PCR was performed on the diluted cDNA using SensiFast probe (Bioline) as per the manufacturers' protocol with b-actin acting as an endogenous reference. Primer and probe sequences are detailed in Supplementary table 2. Gene expression was normalized to the endogenous control and relative expression was evaluated using DDCt method.

In situ HiC
The in situ HiC data are available as GEO series GSE147467. The HiC libraries were processed and DIs were assessed as described previously. 1

3C from HiC
Fixation, proximity ligation and DNA precipitation were performed as for in situ HiC, 42 except the 0.4 mM biotinylated dATP (Life Technologies) in the end blunting step was replaced with 10 mM unbiotinylated dATP (New England Biolabs). The protocol was ceased after DNA precipitation and before DNA shearing. The proximity ligated DNA was then quantified using a Nanodrop (Thermo Fisher).
The frequency of proximity ligated fragments containing the Myc promoter and other downstream regions was then quantified on the BioRad C1000 Thermocycler using the SensiMix SYBR NO-ROX kit (Bioline) with 60°C annealing temperature, water or 5 ng of proximity ligated DNA per 25 lL reaction in duplicate. Primers P1 and P4 (Supplementary  table 2) were used to capture the ligation events between the Myc promoter itself. The P3 primer to all other regions numbered sequentially according to their location across the 1 Mb upstream of the Myc promoter (Supplementary table 2) were used to quantify interactions between the Myc promoter and other regions. Primer #22 targeted the Myce deleted region.
Data were plotted as the average raw threshold cycle (Ct) in DNA samples per primer pair minus the average threshold cycle of the paired water controls. Relative to wild-type (WT) was calculated by subtracting the Myce mice Ct value from the associated WT Ct.

Virtual 4C analysis
The virtual 4C profiles were produced with the diffHic package v1.26.0 in R and then plotted using Graphpad PRISM. The Myc promoter was defined with the TxDb.Mmusculus.UCSC.mm10.knownGene package v3.10.0 and by applying the promoters function from the GenomicFeatures package v1.46.2 with upstream = 2 kbp and downstream = 5 kbp. Interactions between the Myc promoter and the entire genome were counted across all samples with the connectCounts function from diffHic using regions = Myc promoter with a filter set to 0 and the second.region = 20 kbp. The counts per million for each bin were calculated with the cpm function from the edgeR package v3.36.0.

Visualization of HiC
Normalized contact matrices at a 20 kbp resolution were produced with the HOMER HiC pipeline for visualization. With the summed biological-replicate tag directories, the analyzeHiC function was used with the -balance option. Contact matrices were plotted using the plotHic function from the Sushi R package v1.34.0. 43 The color palette was inferno from the viridisLite package v0.4.0. 44

Motif analysis
The web based version of FIMO v5.5.0 (MEME suite) with default settings was used to scan for motifs matches in the sequences of the regions 1-4 and the 2 kb region upstream of the Myc transcriptional start site (promoter). 45 We used the HOCOMOCOv11 core mouse mono motif database 46 and motifs were matched to the transcription factor using annotation from https://hocomoco11.autosome.org/ downloads_v11. Only motifs from transcription factors that are differential expressed between the 3 h activation B cells and resting B cells were used in the analysis. Heatmaps were plotted with pheatmap package in R.

RNA-sequencing analysis
The RNA-seq data are available as GEO series GSE147496. The RNA-Seq libraries were processed and differential expression was assessed as previously described. 1