Chromodomain helicase DNA-binding 4 (CHD4) regulates early B cell identity and V(D)J recombination

Abstract

B lymphocytes develop from uncommitted precursors into immunoglobulin (antibody)-producing B cells, a major arm of adaptive immunity. Progression of early progenitors to antibody-expressing cells in the bone marrow is orchestrated by the temporal regulation of different gene programs at discrete developmental stages. A major question concerns how B cells control the accessibility of these genes to transcription factors. Research has implicated nucleosome remodeling ATPases as mediators of chromatin accessibility. Here, we describe studies of chromodomain helicase DNA-binding 4 (CHD4; also known as Mi-2β) in early B cell development. CHD4 comprises multiple domains that function in nucleosome mobilization and histone binding. CHD4 is a key component of Nucleosome Remodeling and Deacetylase, or NuRD (Mi-2) complexes, which assemble with other proteins that mediate transcriptional repression. We review data demonstrating that CHD4 is necessary for B lineage identity: early B lineage progression, proliferation in response to interleukin-7, responses to DNA damage, and cell survival in vivo. CHD4-NuRD is also required for the Ig heavy-chain repertoire by promoting utilization of distal variable (V_H ) gene segments in V(D)J recombination. In conclusion, the regulation of chromatin accessibility by CHD4 is essential for production of antibodies by B cells, which in turn mediate humoral immune responses to pathogens and disease.

Keywords: B cell identity; CHD4; V(D)J recombination; chromodomain helicase DNA-binding 4.

FIG. 1-. CHD4 and NuRD complex composition.

A. Domain structure of CHD4 (murine isoform X6). Isoform X6 comprises 1912 residues (CHD4-www.ncbi.nlm.nih.gov/protein/XP_017176784.1). A 558 amino acid ATPase/helicase domain functions as a molecular motor to reposition nucleosomes. It includes a DEAH-box that is essential for chromatin remodeling function. Other functional domains include the 54 amino acid N-terminal (NT) domain, an HMG box-like structure that binds poly(ADP-ribose). The NT domain may have a role in DNA repair. Two tandem PHD fingers, PHD1 and PHD2, bind unmodified (highest affinity) or methylated (e.g., H3K9me3) N-terminal tails of histone H3. Tandem chromodomains CD1 and CD2 interact with DNA and are required for chromatin remodeling by the ATPase/helicase domain. The last third of CHD4 includes two domains of unknown function (DUF) and the C-terminal (CT) domain, a feature that is conserved between in other CHD-like helicases. B. Schematic diagram of CHD4-NuRD complexes (adapted from 27). Labels indicate different protein family members (e.g., MTA1/MTA2) that may each be alternately incorporated into NuRD complexes. These subunits add to NuRD’s heterogeneity and functional diversity. Stoichiometry (26) was estimated for purified NuRD complexes using mass spectrometry and biochemical assembly of complexes to be approximately 2:2:4:1:1:1 for MTA:HDAC:RBBP:MBD:GATAD2:CHD4. Heterogeneity of NuRD complexes was observed with variable numbers of associated subunits. Contacts between GATAD2 and CHD4 are important for assembly of NuRD complexes. MTA3 can be substituted for MTA1 or MTA2 in CHD4-NuRD but is not expressed in BM B cell progenitors.

FIG. 1-. CHD4 and NuRD complex composition.

A. Domain structure of CHD4 (murine isoform X6). Isoform X6 comprises 1912 residues (CHD4-www.ncbi.nlm.nih.gov/protein/XP_017176784.1). A 558 amino acid ATPase/helicase domain functions as a molecular motor to reposition nucleosomes. It includes a DEAH-box that is essential for chromatin remodeling function. Other functional domains include the 54 amino acid N-terminal (NT) domain, an HMG box-like structure that binds poly(ADP-ribose). The NT domain may have a role in DNA repair. Two tandem PHD fingers, PHD1 and PHD2, bind unmodified (highest affinity) or methylated (e.g., H3K9me3) N-terminal tails of histone H3. Tandem chromodomains CD1 and CD2 interact with DNA and are required for chromatin remodeling by the ATPase/helicase domain. The last third of CHD4 includes two domains of unknown function (DUF) and the C-terminal (CT) domain, a feature that is conserved between in other CHD-like helicases. B. Schematic diagram of CHD4-NuRD complexes (adapted from 27). Labels indicate different protein family members (e.g., MTA1/MTA2) that may each be alternately incorporated into NuRD complexes. These subunits add to NuRD’s heterogeneity and functional diversity. Stoichiometry (26) was estimated for purified NuRD complexes using mass spectrometry and biochemical assembly of complexes to be approximately 2:2:4:1:1:1 for MTA:HDAC:RBBP:MBD:GATAD2:CHD4. Heterogeneity of NuRD complexes was observed with variable numbers of associated subunits. Contacts between GATAD2 and CHD4 are important for assembly of NuRD complexes. MTA3 can be substituted for MTA1 or MTA2 in CHD4-NuRD but is not expressed in BM B cell progenitors.

FIG. 2-. B cell development in mouse bone marrow and conditional deletion of Chd4 genes in vivo.

A. Progressive stages necessary to generate conventional B cells from pluripotent long-term hematopoietic stem cells (LT-HSC). Intermediate stages (double arrow) between LT-HSC and common lymphoid progenitors (CLP) are not shown. Uncommitted pre-pro-B cells express transcription factors including EBF1 and PAX5, which activate genes necessary for lineage commitment. Expression of the lineage marker CD19 in early pro-B cells is associated with B cell commitment. Status of immunoglobulin heavy chain (Igh) genes at each stage is indicated as germline (GL, or unrearranged), having completed rearrangements of diversity and joining segments (D_HJ_H; B cells representing subsequent stages of development beyond pro-B cells may retain one D_HJ_H allele) and variable (V_H) to D_HJ_H loci (V_HD_HJ_H). Similarly, status of light chain genes encoding κ or λ light chains is indicated as GL or as having completed rearrangements of variable to joining (V_LJ_L) segments. Expression of key marker genes at each stage is indicated below. ‘+’ indicates high expression, ‘lo’ indicates significantly lower expression, and ‘−’ represents very low or absence of expression (ImmGen; reference and our data, not shown). λ5 and Vpreb1 surrogate light chains are encoded by the coordinately expressed, tandem Igll1:Vpreb1 locus. BP-1 is encoded by Enpep. CD2 is expressed by B cells that express u heavy chains (initiated by display of the pre-BCR). The black arrow reflects earliest expression of the Cd79a-cre gene, which directs deletion of floxed Chd4 loci (FIG. 3) in early pro-B cells (28, 30) . Following differentiation to become surface IgM+ immature B cells, these cells exit the bone marrow to the periphery (right of dashed line). Multiple stages of antigen-dependent differentiation in the periphery can ultimately result in production of antibody-secreting plasma cells, which may return to the bone marrow. B. Conditional deletion of floxed Chd4 genes in B cell progenitors of mice. Insertion of loxP sites to generate floxed Chd4 alleles was reported previously (59). C57BL6/N mice with floxed Chd4 alleles were bred with mice with Cd79a-cre genes (28, 30) or Cd2-Cre (29) to generate conditionally deleted (cKO) mice with early B cells lacking functional CHD4. Deletion of exons between 12 and 23 results in out-of-frame joining and loss of ATPase/helicase domains.

FIG. 2-. B cell development in mouse bone marrow and conditional deletion of Chd4 genes in vivo.

A. Progressive stages necessary to generate conventional B cells from pluripotent long-term hematopoietic stem cells (LT-HSC). Intermediate stages (double arrow) between LT-HSC and common lymphoid progenitors (CLP) are not shown. Uncommitted pre-pro-B cells express transcription factors including EBF1 and PAX5, which activate genes necessary for lineage commitment. Expression of the lineage marker CD19 in early pro-B cells is associated with B cell commitment. Status of immunoglobulin heavy chain (Igh) genes at each stage is indicated as germline (GL, or unrearranged), having completed rearrangements of diversity and joining segments (D_HJ_H; B cells representing subsequent stages of development beyond pro-B cells may retain one D_HJ_H allele) and variable (V_H) to D_HJ_H loci (V_HD_HJ_H). Similarly, status of light chain genes encoding κ or λ light chains is indicated as GL or as having completed rearrangements of variable to joining (V_LJ_L) segments. Expression of key marker genes at each stage is indicated below. ‘+’ indicates high expression, ‘lo’ indicates significantly lower expression, and ‘−’ represents very low or absence of expression (ImmGen; reference and our data, not shown). λ5 and Vpreb1 surrogate light chains are encoded by the coordinately expressed, tandem Igll1:Vpreb1 locus. BP-1 is encoded by Enpep. CD2 is expressed by B cells that express u heavy chains (initiated by display of the pre-BCR). The black arrow reflects earliest expression of the Cd79a-cre gene, which directs deletion of floxed Chd4 loci (FIG. 3) in early pro-B cells (28, 30) . Following differentiation to become surface IgM+ immature B cells, these cells exit the bone marrow to the periphery (right of dashed line). Multiple stages of antigen-dependent differentiation in the periphery can ultimately result in production of antibody-secreting plasma cells, which may return to the bone marrow. B. Conditional deletion of floxed Chd4 genes in B cell progenitors of mice. Insertion of loxP sites to generate floxed Chd4 alleles was reported previously (59). C57BL6/N mice with floxed Chd4 alleles were bred with mice with Cd79a-cre genes (28, 30) or Cd2-Cre (29) to generate conditionally deleted (cKO) mice with early B cells lacking functional CHD4. Deletion of exons between 12 and 23 results in out-of-frame joining and loss of ATPase/helicase domains.

FIG. 3-. Loss of CHD4 results in proximal bias of V_H to D_HJ_H rearrangements in early pro-B cells.

Linear map of 106 V_H segments and frequencies of their rearrangements detected using the Ion AmpliSeq Mouse BCR IGH assay (Thermo Fisher Scientific; Waltham, MA, USA). The assay identifies unique B-cell clones through targeting of the highly diverse complementarity-determining region 3 (CDR3) of IgH chains from genomic DNA templates. Nucleic acids were isolated from CD19+cKit+ early pro-B cells were sorted from BM of two littermates each (~5 weeks old) using standard kit methodology (Qiagen; Hilden, Germany). Library construction proceeded from isolation of nucleic acid, followed by multiplex PCR with primers targeting the V_HD_HJ_H rearrangements to generate amplicons for sequencing using the Thermo Fisher Ion suite of sequencing systems. The nucleotide sequence of the IGH CDR3 region serves as a natural barcode to enable clone tracking and measurements of B-cell clonal expansion and diversity. V_H gene clusters are arranged from distal (left) to proximal (right) segments upstream of D_H and J_H segments (not shown) in wild type (WT) and Chd4 cKO mice (all C57BL6/N). Each bar represents all rearrangements including each V_H. Numbers of reads and clones, respectively, identified for each sample in BCR-seq were: WT1, 2,734,991 and 2717; WT2, 850,465 and 8450; Chd4 cKO1, 4,527,579 and 2372; Chd4 cKO2, 1,357,677 and 4436. Raw data are tabulated in Supplemental Table S1. Mice were bred and housed in the Biological Resources Center at National Jewish Health. All experiments were performed under a protocol approved by the Institutional Animal Care and Use Committee at National Jewish Health.