Concentration- and co-factor-dependent Runx DNA binding site choices during early T-development

Runx1 and Runx3 function redundantly in early T-development, and together drive T-lineage developmental progression by regulating distinct sets of genes in different stages. Context-specific Runx target genes are particularly enriched near Runx binding sites that dynamically shift from pre-T-commitment stages (Phase 1) to post-T-commitment stages (Phase 2). As total Runx activities (Runx1+Runx3) are maintained stably throughout early T-development stages, yet Runx factors physically interact with multiple collaborators, we hypothesized that different Runx-collaborating transcription factors compete to recruit a limited pool of Runx transcription factors. To test whether increasing Runx availability can alter Runx DNA-binding site choices, Runx1 overexpression vectors were introduced to two different systems. First, we increased Runx1 expression in a DN3-like cell line (representing a post-commitment, Phase 2 stage) in the presence or absence of exogenous Phase 1 co-factor, PU.1. Second, we increased Runx1 expression in Phase 1 primary cells which naturally express PU.1 and other Phase 1 collaborators. Then, Runx1 binding behaviors were measured using ChIP-seq or CUT&RUN (C&R). A modest increase of Runx1 levels (about 2-3 fold increase) substantially increased the number of Runx binding sites seen and the intensity of occupancy in both systems. When Runx expression was at physiological levels, PU.1 dominated Runx1 site choice before T commitment by recruiting Runx1 to PU.1 sites. The Phase 2 cell line system showed that it did this while depleting Runx1 from alternative high quality Runx motif sites. However, when Runx1 was overexpressed, Runx1 was still recruited to PU.1 sites, but this recruitment did not evacuate Runx1 occupancy from default preferred sites. Notably, Runx1 overexpression in primary Phase 1 cells caused precocious occupancy of post-commitment, Phase 2-specific sites. We found that these are often co-occupied with TCF1, E2A, and HEB, but have minimal co-binding with PU.1. In addition, Runx1 overexpression resulted in new binding sites that were not normally observed in pro-T cells, which are mostly sequestered by closed chromatin normally although they harbor more numerous Runx motifs. Thus, these data suggest that Runx DNA binding site choices are sensitive to Runx concentration and co-factors during early T-development. CUT&RUN (C&R), ChIP-seq, ATAC-seq To measure Runx binding sites in unperturbed DN1 cells, freshly obtained bone-marrow-derived progenitor cells from 8-10 week old C57BL/6J wild-type (WT) animals were cultured on OP9-Dll1 cells with cytokines (IL-7 and Flt3-ligand, 10 ng/ml each). On day 5, live lineage (TCRb, TCRgd, CD19, NK1.1, CD11b, CD11c, Ly6G/C) negative, CD45+, cKithi, and CD25- cells were FACS sorted. Approximately 250K -500K cells were utilized to conduct CUT&RUN (C&R). For unperturbed thymic DN3 C&R, thymus was obtained from 5-6 weeks old C57BL/6J WT mice and live Lineage marker-negative DN3 cells (cKitlow, CD44low, and CD25-) were sorted. 500K cells were subjected to C&R analysis. To detect Runx1 occupancies with Runx1 overexpression in Phase 1, freshly obtained bone-marrow progenitor cells from B6.Bcl2 transgenic mice were co-cultured with OP9-Dll1 cells in medium supplemented with 10ng/ml of IL-7 and Flt3-ligand. On day 2, MSCV-Runx1-mCherry (Runx1 OE) or MSCV-empty control-mCherry vectors were delivered using spinfection. Two days after delivering retrovirus (total 4 days of culture), live Lineage-, mCherry+, CD45+, cKithi Phase 1 cells were FACS sorted. 180-250K cells were utilized to measure Runx1 binding using C&R. ChIP-seq DN3-like Scid.adh.2c2 (P2C2) cells were retrovirally infected with pMXs-PU.1-human NGFR (hNGFR) and MSCV-Runx1-HA-mCherry vectors. For control conditions, empty vectors for pMXs-hNGFR and MSCV-mCherry vectors were introduced. On day 2 post-infection, P2C2 cells were stained with anti-NGFR-PE antibody and hNGFR-expressing P2C2 cells were selected using anti-PE microbeads (Miltenyi Biotec). Approximately 10 million cells were fixed at RT in 1 mg/ml disuccinimidyl glutarate (DSG) and 1% formaldehyde and subjected to chromatin immunoprecipitation (ChIP) using anti-HA antibody. ATAC-seq To assess chromatin accessibility changes upon Runx1 overexpression, control- or Runx1 OE introduced-pro-T cells (the same as the C&R samples) were FACS sorted and washed with 1X HBSS with 10mM HEPES twice. Then, chromatin accessibility profiles were measured using ATAC-seq.