Dissecting the roles of MBD2 isoforms and domains in regulating NuRD complex function during cellular differentiation

Abstract

The Nucleosome Remodeling and Deacetylation (NuRD) complex is a crucial regulator of cellular differentiation. Two members of the Methyl-CpG-binding domain (MBD) protein family, MBD2 and MBD3, are known to be integral, but mutually exclusive subunits of the NuRD complex. Several MBD2 and MBD3 isoforms are present in mammalian cells, resulting in distinct MBD-NuRD complexes. Whether these different complexes serve distinct functional activities during differentiation is not fully explored. Based on the essential role of MBD3 in lineage commitment, we systematically investigated a diverse set of MBD2 and MBD3 variants for their potential to rescue the differentiation block observed for mouse embryonic stem cells (ESCs) lacking MBD3. While MBD3 is indeed crucial for ESC differentiation to neuronal cells, it functions independently of its MBD domain. We further identify that MBD2 isoforms can replace MBD3 during lineage commitment, however with different potential. Full-length MBD2a only partially rescues the differentiation block, while MBD2b, an isoform lacking an N-terminal GR-rich repeat, fully rescues the Mbd3 KO phenotype. In case of MBD2a, we further show that removing the methylated DNA binding capacity or the GR-rich repeat enables full redundancy to MBD3, highlighting the synergistic requirements for these domains in diversifying NuRD complex function.

Fig. 1. Neuronal differentiation in absence of MBD2 and MBD3.

a Microscopy images of in vitro derived neurons from WT, Mbd3 KO, and Mbd2 KO ESCs shown at 20× magnification. Same results were obtained from 3-5 independent replicates. b Microscopy images of in vitro derived neurons from Mbd3 KO ESCs stably expressing MBD3a (high) or MBD2b under a strong promoter (CAG) or MBD3a (low) under a weaker promoter (CMV). Shown at 20× magnification. Same results were obtained from 3 independent replicates. c Representative flow cytometry analysis of FLK1 surface expression and flow cytometry analysis. d, e Number of live cells (d) and percentage of FLK1+ cells (e) in EB cultures of WT, Mbd3 KO, and Mbd3 KO cell lines expressing MBD3a or MBD2b. f–h Flow cytometry analysis showing number of live cells (WT to Mbd3 KO p = 0.0006, Mbd3 KO to Mbd2 KO p = 0.0147) (f), CD24+ cells (WT to Mbd3 KO p = 0.0002, Mbd3 KO to Mbd2 KO p = 0.0451) (g) and CD24 + CD56+ cells (WT to Mbd3 KO p = 0.0002, Mbd3 KO to Mbd2 KO p = 0.0186) (h). Each data point (in d–h) represents individually generated cell lines. For each cell line 3 independent clones were analyzed, n = 2 biological replicates. Error bars represent mean +/− SD. p-values were calculated using an unpaired, two-tailed t test (Mann-Whitney). Source data for (d–h) are provided as a Source Data file.

Fig. 2. The MBD domain of MBD3 is dispensable for ESC neuronal lineage differentiation.

a Schematic of MBD3 and MBD2 isoforms present in ESCs. MBD: MBD domain, cc: coil-coiled domain, GR: glycine- arginine rich repeat, TRD: transcriptional repressor domain. b Flow cytometry analysis indicating (from left to right) the number of live cells (WT to Mbd3 KO p = 0.0020, Mbd3 KO to +MBD3a p = 0.0001), CD24+ (WT to Mbd3 KO p = 0.0001, Mbd3 KO to +MBD3a p = 0.0002), or CD24 + CD56 + NPC cells (WT to Mbd3 KO p = 0.0020, Mbd3 KO to +MBD3a p = 0.0002) at day 8 of WT, Mbd3 KO, and Mbd3 KO stably expressing MBD3a or MBD3ΔMBD variants. Each data point represents an individual cell line. For each cell line 3 independent clones were analyzed, n = 2 biological replicates. Error bars represent mean +/− SD. p values were calculated using an unpaired, two-tailed t test (Mann-Whitney). Source data is provided as a Source Data file. c Microscopy images of in vitro derived neurons from Mbd3 KO stably expressing MBD3ΔMBD at 20× magnification. Identical results were obtained from 3 independent replicates. d–f MA-plots showing differential gene expression between Mbd3 KO and WT NPCs (d), Mbd3 KO + MBD3a vs WT NPCs (e), and Mbd3 KO + MBD3ΔMBD vs WT NPCs (f). Red and blue dots indicate genes with significant changes in gene expression (edgeR, log2FC > 1 | < −1 and adjusted p-value < 0.05).

Fig. 3. Full-length MBD2a partially rescues the differentiation block in Mbd3 KO ESCs.

a Flow cytometry analysis indicating the number of live cells (left, WT to Mbd3 KO p = 0.0003, Mbd3 KO to +MBD2b p = 0.0091), CD24+ (middle, WT to Mbd3 KO p = 0.0002, Mbd3 KO to +MBD2b p = 0.0091), and CD24 + CD56+ (right, WT to Mbd3 KO p = 0.0001, Mbd3 KO to +MBD2b p = 0.0070) NPCs at day8 of neuronal differentiation of WT, Mbd3 KO, or Mbd3 KO stably expressing MBD2b, MBD2a or MBD2t. Each data point represents individual cell lines. For each cell line 3 independent clones were analyzed (biological replicates), except for MBD2t, where three technical replicates of one clone were analyzed. Error bars represent mean +/− SD. p values were calculated using an unpaired, two-tailed t test (Mann-Whitney). Source data is provided as a Source Data file. Microscopy images of in vitro derived neurons from Mbd3 KO stably expressing MBD2a (b) or MBD2t (c) at 20× magnification. Similar results were obtained from 3-5 independent replicates. d Flow cytometry analysis indicating the number of live cells (left, WT to +MBD2a p = 0.0016, +MBD2a to +MBDaΔGR p = 0.0012), CD24+ (middle, WT to +MBD2a p = 0.0011, +MBD2a to +MBDaΔGR p = 0.0040), and CD24 + CD56+ (right, WT to +MBD2a p = 0.0004, +MBD2a to +MBDaΔGR p = 0.0002) NPCs at day8 of neuronal differentiation of WT or Mbd3 KO stably expressing MBD2a, MBD2aΔGR or MBD2aR191C. Each data point represents individual cell lines. For each cell line 3 independent clones were analyzed, n = 2 biological replicates. Error bars represent mean +/− SD. p values were calculated using an unpaired, two-tailed t test (Mann-Whitney). Source data is provided as a Source Data file. e Microscopy images of in vitro derived neurons from Mbd3 KO stably expressing MBD2aΔGR or MBD2aR191C at 20× magnification. Similar results were obtained from 3-5 independent replicates.

Fig. 4. The MBD domain and GR repeat of MBD2 serve non-redundant functions during neuronal differentiation.

a–d MA-plots showing differential gene expression between Mbd3 KO+MBD2aΔGR vs. WT NPC (a), Mbd3 KO + MBD2aR191C vs WT NPCs (b), Mbd3 KO + MBD2a vs WT NPCs (c), Mbd3 KO + MBD2t vs WT NPCs (d). Red and blue dots indicate genes with significant changes in gene expression (edgeR, log2FC > 1 | < −1 and adjusted p-value < 0.05). e Multi-dimensional scaling (MDS) plot indicates the degree of similarity for all RNA-seq datasets obtained from NPCs at day8. Each point represents an individually derived clone except for Mbd3 KO + MBD2t, where triplicates of one clone were analyzed. f Flow cytometry analysis indicating (from left to right) the number of live cells, CD24+, and CD24 + CD56+ NPCs at day 8 of neuronal differentiation of WT or Mbd3 KO stably expressing MBD3a, MBD3_MBD^MBD2, or MBD3_Nterm^MBD2. Each data point represents individual cell lines. For each cell line 3 independent clones were analyzed, n = 1 biological replicate. Source data is provided as a Source Data file. g Microscopy images of in vitro derived neurons from Mbd3 KO stably expressing MBD3_MBD^MBD2, or MBD3_Nterm^MBD2at 20x magnification. Similar results were obtained from 3 independent replicates.

Fig. 5. Impact of GR and MBD mutations on MBD2a localization.

a Stoichiometry determination of NuRD complex members of MS biotin-Co IPs in Mbd3 KO ESCs stably expressing MBD2a (gray) or MBD3a (orange). The iBAQ value of each protein group is divided by the iBAQ value of MBD2 or MBD3, respectively, then graphed with MBD2/3 set to 1. Proteins with shared peptides are collapsed. n = 3 independent pull-downs. Source data is provided as a Source Data file. b MBD2 variant localization to an example chromosomal region. Shown is the enrichment for all MBD2a, MBD2aΔGR, and MBD2R191C proteins at the Peg10 promoter, which harbors a methylated CpG island (marked as a green bar below the gene profile). MBD protein enrichment is calculated as library-normalized number of tags per 100 bp and two replicates are shown. c Heatmaps showing library-normalized ChIP-seq counts in 40 bp windows covering 2 kb upstream and downstream of methylated CpG islands. Average density profiles are indicated on top.

Fig. 6. Differential nuclear localization of MBD2a- and MBD3a-NuRD complexes influences neuronal differentiation.

a, b Representative immunofluorescence images indicating MBD2a (a) and MBD3a (b) localization in Mbd2,Mbd3-DKO cells. c CHD4 localization in Mbd2,Mbd3-DKO cells alone and expressing either MBD2a or MBD3a. d MBD2 localization in Mbd2,Mbd3-DKO cells stably expressing MBD2a, MBD2aΔGR or MBD2aR191C. e CHD4 localization in Mbd2,Mbd3-DKO cells stably expressing MBD2a, MBD2aΔGR or MBD2aR191C. DAPI staining (magenta) reveals chromocenters. Similar results were obtained from 2 independent replicates. Scale bars, 5 μm.