Single-molecule imaging reveals discrete GATA2 kinetic modes in hematopoietic progenitor cells.

(A) Schematic of the live-cell single-molecule imaging workflow. Suspension cells were secured to the surface of an imaging dish using CD43-mediated immobilization and imaged by HiLo microscopy using 500-ms exposure times to preferentially detect chromatin-bound GATA2 molecules with multi-second residence times. (B) Representative time-lapse sequences illustrating two measurable GATA2 interaction modes. Short-lived interactions (<1 s) appear as brief sampling contacts in which GATA2 rapidly binds and unbinds chromatin. Long-lived interactions (>5 s) correspond to multi-second chromatin-bound events in which GATA2 remains localized at a single site. (C) Residence time distribution of GATA2 trajectories from a single representative cell fit with a double-exponential model, resolving a short-lived kinetic component (<1 s) and a long-lived component (>5 s) corresponding to chromatin engagement. (D) Schematic depiction of the interaction framework used throughout this study. Left: short-lived target-search behavior, in which GATA2 rapidly engages and disengages chromatin during chromatin scanning. Right: long-lived chromatin engagement, in which GATA2 remains localized at a defined regulatory element (RE) for several seconds. (E) Quantification of baseline kinetic parameters in Basal progenitor cells (0 h), including total binding events per cell, fraction of molecules in the long-lived binding population (>5 s), and per-cell dwell-time distributions. Each point represents a single cell (N = 47), revealing substantial cell-to-cell heterogeneity in GATA2–Halo chromatin binding. Created using BioRender.com.

GATA2 binding dynamics exhibit a transitory strengthening of long-lived chromatin engagement during Early erythroid differentiation in G1E-ER4 cells.

(A) Schematic of the G1E-ER4 differentiation system. Activation of GATA1–ER with 4-hydroxytamoxifen induces progression from the Basal progenitor (0 h) to Early erythroid (2 h) and Late erythroid (24 h) states. (B) Total number of detectable GATA2 chromatin binding events per cell across erythroid stages. Binding-event frequency is unchanged between Basal and Early states but is reduced in the Late state relative to both Basal and Early erythroid cells. (C) Fraction of GATA2 molecules engaging in long-lived chromatin interactions across erythroid stages. The proportion of long-lived binding events remains constant across Basal, Early, and Late states (approximately 20%). (D) Residence times of long-lived GATA2 chromatin interactions across erythroid stages. Residence time is significantly prolonged in the Early erythroid state relative to the Basal progenitor state and is significantly reduced in the Late erythroid state. Each point represents a single cell (Basal, N = 47; Early, N = 62; Late, N = 76). Bars indicate mean ± SEM. Statistical significance was assessed using the Brown–Forsythe and Welch ANOVA with Games–Howell post hoc correction. Significance levels are indicated as follows: * P < 0.05; ** P < 0.01; **** P < 0.0001. Created using BioRender.com.

Early erythroid entry in HPC7 progenitors is marked by a transitory increase in long-lived GATA2 chromatin engagement.

(A) Schematic of the HPC7 erythroid differentiation paradigm. Erythropoietin (EPO) stimulation induces progression from the Basal progenitor state to Early and Late erythroid stages. (B) Total number of detectable GATA2 chromatin binding events per cell across erythroid stages. Binding-event frequency increases significantly from the Basal to Early stage following EPO stimulation and remains elevated in the Late stage, with no significant difference between Early and Late populations. (C) Fraction of GATA2 molecules engaging in long-lived chromatin interactions across erythroid stages. The long-lived binding fraction is significantly increased in Early erythroid cells relative to Basal progenitors and declines in Late erythroid cells to levels comparable to Basal. (D) Residence times of long-lived GATA2 chromatin interactions across erythroid stages. Residence times are similar between Basal and Early stages and are modestly reduced in Late erythroid cells relative to the Early stage. Each point represents a single cell. Bars indicate mean ± SEM. Statistical significance was assessed using the Brown–Forsythe and Welch ANOVA with Games–Howell post hoc correction. Statistical significance was defined as P < 0.05. Significance levels are indicated as follows: * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001. Created using BioRender.com.

Endogenous GATA2 exhibits transitory enrichment of long-lived chromatin interactions during in vivo erythroid maturation.

(A) Schematic of the Gata2–SNAP knock-in mouse used for endogenous single-molecule imaging, together with the CD71/Ter119 flow-cytometry gating strategy used to define Basal, Early, and Late erythroid populations. (B) Total detectable GATA2 chromatin binding events per cell across Basal, Early, and Late erythroid populations. Binding-event frequency is highest in Early cells and reduced in Late cells. (C) Fraction of GATA2 molecules engaging in long-lived chromatin interactions across Basal, Early, and Late erythroid populations. The long-lived binding fraction is highest in Early cells and declines upon progression to the Late stage. (D) Residence times of long-lived GATA2 chromatin interactions across erythroid stages. Residence times are similar across all three populations, indicating that regulation primarily occurs through changes in the occupancy of the long-lived binding state rather than binding duration. Each point represents a single cell. Bars indicate mean ± SEM. Statistical significance was evaluated using the Brown–Forsythe and Welch ANOVA test with Games–Howell post hoc correction. Statistical significance was defined as P < 0.05. Significance levels are indicated as follows: * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001. Created using BioRender.com.

Early erythroid–restricted GATA2 peaks reveal temporally specific recruitment, two regulatory subclasses, and distinct motif and genomic features.

(A) Categorical classification of GATA2 CUT&Tag peaks across Basal progenitor (0 h), Early erythroid (2 h), and Late erythroid (24 h) G1E-ER4 cells. Peaks were grouped into Early-enriched, shared, Late-acquired, and monotonic categories, providing a global view of how GATA2 genomic occupancy evolves during erythroid differentiation. (B) Identification of Early erythroid–restricted GATA2 peaks. A stringent subset of peaks was defined as sites detected by MACS2 exclusively at the Early (2 h) erythroid state (q < 0.01) and absent at both Basal (0 h) and Late (24 h) stages using identical processing parameters. These peaks represent genomic regions uniquely recruited during Early erythroid entry. (C) CUT&Tag signal heatmap for all Early erythroid–restricted GATA2 peaks (n = 1,167) aligned to peak centers (±2 kb). Peaks show strong, centered GATA2 enrichment at the Early (2 h) stage with minimal signal at Basal (0 h) and Late (24 h) stages, indicating stage-specific recruitment. (D) Genome browser views of representative Early erythroid–restricted GATA2 peaks (highlighted by red boxes), including a promoter-proximal peak near Nono, an intronic site within Nr3c1, and an intergenic site near Gata3. Each locus exhibits GATA2 binding selectively in the Early erythroid state and loss in the Late erythroid state. Blue boxes indicate adjacent genomic regions lacking stage-dependent changes in GATA2 signal. (E) Classification of Early erythroid–restricted peaks based on overlap with GATA1 CUT&Tag data. Peaks were subdivided into GATA2-only peaks (n = 700) and peaks bound by both GATA2 and GATA1 (n = 467), indicating that Early GATA2 recruitment can occur either independently of, or in conjunction with, GATA1. (F) Motif enrichment analysis for GATA2-only peaks and peaks bound by both GATA2 and GATA1. GATA2-only peaks are enriched for canonical GATA motifs and RUNX-associated elements, whereas peaks bound by both factors display composite GATA/E-box motifs characteristic of TAL1-centered regulatory assemblies. (G) Genomic distribution of GATA2-only peaks versus peaks bound by both GATA2 and GATA1 among Early erythroid–restricted sites. GATA2-only peaks are preferentially promoter-proximal, whereas peaks bound by both factors are enriched at intronic and intergenic regions, supporting distinct promoter- and enhancer-associated regulatory roles during Early erythroid differentiation. Created using BioRender.com.

A kinetic transition state uncouples GATA2 chromatin engagement from protein abundance during erythroid differentiation.

(A) Expression Dynamics (Classical View). The canonical model of the GATA switch, where GATA2 protein levels (blue line) decline monotonically as GATA1 levels (red line) rise to drive differentiation. (B) Chromatin Engagement Dynamics (Kinetic View). In contrast to expression, single-molecule measurements reveal that GATA2 chromatin engagement (blue line) is transiently strengthened during the Early erythroid transition. This creates a kinetic peak that uncouples binding activity from protein abundance, defining a distinct biophysical phase of lineage commitment. Created using BioRender.com.

Short-lived GATA2 chromatin binding dynamics across differentiation in G1E-ER4, HPC7, and primary mouse progenitors.

(A–B) G1E-ER4 cells. The fraction of short-lived (<1 s) GATA2–Halo interactions and the corresponding short-lived residence times did not differ significantly across the Basal (0 h), Early (2 h), and Late (24 h) stages. (C–D) HPC7 progenitors. The short-lived fraction decreased modestly at the Early (2 h) stage and increased again at the Late (24 h) stage (C). Short-lived residence times declined significantly from the Early to the Late stage (D). (E–F) Primary Gata2–SNAP bone-marrow progenitors. The short-lived fraction differed across all three stages, with the largest decrease at the Early stage (E). Short-lived residence times increased from Basal to Early cells and decreased again after the Late stage commitment (F). Statistical comparisons used Brown–Forsythe and Welch ANOVA with Games–Howell post-hoc tests. Significance was defined as P < 0.05, with significance levels indicated as: * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001. Created using BioRender.com.

Validation of erythroid induction and GATA factor exchange in the GATA2–Halo reporter line.

These analyses confirm that the GATA2–Halo reporter line undergoes normal erythroid induction and GATA1 activation. (A) Heatmap of GATA1 CUT&Tag signal aligned across gene bodies at Basal (0 h), Early (2 h), and Late (24 h) stages following 4-hydroxytamoxifen treatment. Rows are ordered by increasing signal intensity. GATA1 occupancy increases progressively over the time course, with strongest enrichment at 24 h. (B) Representative genome-browser views comparing GATA1 (red) and GATA2–Halo (blue) CUT&Tag occupancy at the Gata2, Klf1, and Kit loci. Tracks illustrate progressive gain of GATA1 binding accompanied by a reduction or redistribution of GATA2–Halo occupancy during erythroid progression. Red boxes highlight sites of reciprocal GATA factor exchange (“GATA switch”). Y-axis scales are fixed across time points within each factor to enable direct visual comparison. (C) Western blot analysis of GATA1 protein abundance at 0, 2, 8, and 24 h after 4-hydroxytamoxifen treatment. β-actin serves as a loading control. GATA1 protein levels increase over time, with maximal expression at 24 h. Created using BioRender.com.

Global CUT&Tag signal patterns across GATA2 peak classes during erythroid differentiation.

CUT&Tag signal heatmaps showing GATA2 occupancy across peak categories defined in Figure 5A, including peaks shared across all stages, Basal progenitor–specific (0 h), Early erythroid–specific (2 h), and Late erythroid–specific (24 h) regions. Heatmaps confirm stage-restricted enrichment and validate the temporal classification of GATA2 binding sites used for downstream analyses. (A) All GATA2 peaks. (B) Basal progenitor–specific peaks (0 h). (C) Early erythroid–specific peaks (2 h). (D) Late erythroid–specific peaks (24 h). Created using BioRender.com.