VSports - Building multifunctionality into a complex containing master regulators of hematopoiesis

Discriminatory Functions of a Non-DNA Binding Component of a Complex Containing Master Regulators of Hematopoiesis.

The contribution of non-DNA binding components of the GATA-1-Scl/TAL1 complex to activity of the complex is not well defined. We analyzed the role of one of these components, ETO2, in GATA-1 activity using GATA-1–null G1E cells (37) stably expressing GATA-1 fused to the estrogen receptor ligand-binding domain (G1E-ER-GATA-1 cells) (38, 39). We tested whether reducing ETO2 levels influences the capacity of ER-GATA-1 to regulate target genes. After transfection of G1E-ER-GATA-1 cells with anti-Cbfa2t3 [Cbfa2t3 encodes ETO2 (40)] siRNA, ER-GATA-1 was activated with β-estradiol (Fig. 1A). Cbfa2t3 mRNA (Fig. S1A) and protein (Fig. S1B, Upper) expression strongly decreased, and ER-GATA-1 levels were unaltered (Fig. S1B, Lower).

We compared expression profiles in β-estradiol-induced G1E-ER-GATA-1 cells transfected with control and anti-Cbfa2t3 siRNA. This analysis revealed altered expression of established GATA-1 targets and those not known to be GATA-1-regulated (Fig. 1B). The ETO2-regulated genes included those in which ER-GATA-1–mediated activation (Fig. 1C) and repression (Fig. 1D) were significantly enhanced by the knockdown (P < 0.05). An additional cohort of GATA-1 targets was unaffected (Fig. 1E). The reduction in ETO2 levels therefore derepressed ER-GATA-1 activity, increasing its activation and repression of targets. ETO2 occupied GATA switch sites bound by GATA-2 (Fig. S2E) and GATA-1 (Fig. S2A) in uninduced and induced G1E-ER-GATA-1 cells, respectively, at representatives of this group (Fig. S2C), indicating a direct function at these loci. As certain GATA-1 targets were unaffected, ETO2 mediates GATA-1 function in a context-dependent manner.

The gene exhibiting the greatest sensitivity to ETO2 knockdown was Slc4a1 (Fig. 1C, Left), a GATA-1 target gene encoding an anion exchanger with pivotal roles in red cells (41). GATA-1 (Fig. S3B) and ETO2 (Fig. S3C) occupies the Slc4a1 promoter and intron 1 (+1.4 kb) (Fig. S3A) and regulates Slc4a1 transcription in a FOG-1–dependent manner (36, 42), but other aspects of this transcriptional mechanism are unknown.

To establish how knocking-down ETO2 renders Slc4a1 hypersensitive to ER-GATA-1, we analyzed the epigenetic landscape at Slc4a1 in untreated and β-estradiol–treated G1E-ER-GATA-1 cells. Histone H3 di-acetylated at K9/K14 was considerably higher at sites of ER-GATA-1 occupancy, whereas tetra-acetylated histone H4 was distributed throughout the locus. ER-GATA-1–mediated activation of Slc4a1 transcription was accompanied by increased H3-diacK9/K14 and tetraacetylated H4 (Fig. S3 D and E). Analogous to H3-diacK9/K14, H3-dimeK4 was highest at ER-GATA-1 occupancy sites and increased upon activation, albeit not at the promoter (Fig. S3F). Two marks associated with repression, H3-trimeK9 and H3-trimeK27 (43), were unusually high (comparable or higher than the inactive Necdin promoter) at the inactive and active Slc4a1 locus (Fig. S3 G and H). The presence of active and repressive marks resembles bivalent chromatin domains, which may reflect a state in which the gene is poised to be activated (44), but such domains have not been described in erythroid cells.

We asked whether the ETO2 knockdown altered all or specific components of this nucleoprotein architecture in untreated and β-estradiol–treated G1E-ER-GATA-1 cells. In untreated cells, knocking-down ETO2 significantly, but modestly, increased ER-GATA-1 (Fig. 2A) and Scl/TAL1 (Fig. 2B) occupancy at the promoter and the +1.4-kb site. ER-GATA-1 (Fig. 2A) and Scl/TAL1 (Fig. 2B) occupancy was unaffected in β-estradiol–treated cells. The knockdown significantly increased H3-diacK9/K14 at the promoter and intron 1, without influencing this mark at the active and inactive RPII215 and Necdin promoters, respectively (Fig. 2C). Whereas the knockdown did not affect H3-dimeK4 (Fig. 2D) and H3-trimeK9 (Fig. 2E) at Slc4a1 and other targets, in the presence of active ER-GATA-1, the knockdown strongly reduced H3-trimeK27 at the promoter and intron 1, without affecting this mark at control promoters (Fig. 2F). Although ETO2 regulation of histone modifications had not been described, given its binding to HDACs (45), increased H3-diacK9/K14 is consistent with this mechanism. In contrast, its regulation of H3-trimeK27 was surprising. It is attractive to consider that high-level H3-trimeK27 might uniquely characterize ETO2-regulated GATA-1 targets. Although ETO2-regulated targets (BC056923, Tmod1, and Rragd), but not the nonresponsive target (Abcb10), have high-level H3-trimeK27 (Fig. S4), ETO2 did not regulate H3-trimeK27 at all targets (Fig. 2F, Right). ETO2 suppressed H3-diacK9/K14 at all responsive targets, but not the nonresponsive targets (Fig. 2C, Right). ETO2 also suppressed ER-GATA-1–mediated repression at select loci, and this action at Plekhg3 involved suppressing H3-trimeK27 and maximizing H3-dimeK4 (Fig. S5). Thus, ETO2-dependent transcriptional repression involves a widely used epigenetic mechanism to suppress H3-diacK9/K14 and a locus-selective action to regulate H3-trimeK27 (Fig. 2G).

GATA-1-Scl/TAL1 Complex Structure/Function: Qualitatively Distinct Mechanistic and Biological Functions of Individual Complex Components.

Although ETO2, LMO2, and LDB1 are components of GATA-1-Scl/TAL1 complexes (22, 29, 30, 46), whether these components function as a unit or independently to regulate GATA-1 target genes is unknown. We knocked-down LMO2 mRNA (Fig. S1C) and protein (Fig. S1D, Upper), which did not alter ER-GATA-1 levels (Fig. S1D, Lower), and conducted expression profiling to establish the LMO2-regulated gene ensemble (Fig. 3A). Surprisingly, the knockdown significantly up-regulated all but five of the LMO2-regulated genes, indicating an unexpected predominant repressive function of LMO2. As LMO2 overexpression inhibited erythropoiesis, our loss-of-function analysis provides a conceptual framework for understanding this result (47). Genes encoding prototypical erythroid proteins (e.g., heme biosynthetic enzymes and red cell cytoskeletal proteins) were conspicuously absent from this gene ensemble. LMO2 targets included genes shown previously to be FOG-1–independent, GATA-1–repressed genes (21), including Lyl1, Rgs18, Clec4d, Vim, Sept9, Lgals9, and Klf10 (Fig. 3A). Because the mechanism underlying FOG-1–independent GATA-1–mediated repression has been elusive, and these targets include the important regulator of HSCs LYL1 (48), we focused on this set of genes.

Quantitative RT-PCR analysis confirmed that the LMO2 knockdown up-regulated FOG-1–independent, GATA-1–repressed genes (Rgs18, Vim, Clec4d, Lyl1) (Fig. 3B). Even in uninduced G1E-ER-GATA-1 cells before GATA-1 occupancy, the knockdown up-regulated Rgs18, Vim, Clec4d, and Lyl1 (Fig. 3B). In contrast, the ETO2 knockdown decreased Rgs18 expression (Fig. 1D). Although knocking down ETO2 yielded ER-GATA-1–mediated hyperinduction of Slc4a1 (Fig. 1C), knocking-down LMO2 significantly inhibited GATA-1–mediated induction of Slc4a1 (Fig. 3C, Left). Although ETO2 and LMO2 exert qualitatively distinct responses, both occupy Lyl1, Rgs18, and Slc4a1 (Fig. S2 C and D). Because LMO2 facilitates ER-GATA-1–mediated repression of FOG-1–independent genes, we tested whether it also affects FOG-1–independent activation of genes. The LMO2 knockdown did not affect ER-GATA-1–mediated induction of the prototypical FOG-1–independent gene Fog1 nor repression of the prototypical FOG-1–dependent gene Gata2 (Fig. 3C, Right).

Because ETO2 and LMO2 knockdowns yielded qualitatively distinct outputs, and both factors bind Scl/TAL1, we asked whether the factor-specific activities are Scl/TAL1-dependent or -independent. If Scl/TAL1 is critical for ETO2 or LMO2 to regulate their gene ensembles, knocking-down Scl/TAL1 should recapitulate the gene dysregulation resulting from ETO2 or LMO2 knockdowns. Knocking-down Scl/TAL1 (Fig. S1 E and F), which did not alter ER-GATA-1 levels (Fig. S1F), significantly reduced ER-GATA-1–mediated activation of Slc4a1, resembling the LMO2 knockdown (Fig. 3C, Left), although the reduction was greater with the Scl/TAL1 knockdown (Fig. 3D, Left). Also resembling the LMO2 knockdown, knocking-down Scl/TAL1 significantly reduced ER-GATA-1–mediated repression of the FOG-1–independent gene Lyl1, but not the FOG-1–dependent gene Gata2 (Fig. 3D, Right). As LMO2 and Scl/TAL1 exert qualitatively similar influences on GATA-1 targets, this finding supports the notion that LMO2 and Scl/TAL1 function collectively in the FOG-1–independent GATA-1 repression mechanism. In contrast, knocking-down Scl/TAL1 did not yield hyperactivation of Slc4a1 in response to ER-GATA-1 activation. The distinct influences of ETO2 and Scl/TAL1 on the FOG-1–dependent, GATA-1–activated gene Slc4a1 are inconsistent with ETO2 contributing to a solitary function of the GATA-1-Scl/TAL1 complex.

The different ETO2 and LMO2 mechanisms (Fig. 4A) raises the question as to whether the respective gene ensembles mediate similar or distinct biological pathways. Gene Ontology analysis revealed that the majority of ETO2- and LMO2-regulated genes control distinct pathways (Fig. 4B). Most ETO2- and LMO2-regulated genes (61 and 138, respectively) differed, but both factors coregulated only 16 genes (Fig. 4C). The coregulated genes included four ER-GATA-1–activated (Hba-a1, Hbb-b1, Cotl1, Gng2), four ER-GATA-1–repressed (Mical1, Hdc, Lilrb3, Plekhg3), and eight not regulated by ER-GATA-1 (Aldoa, Fau, Gsn, Fes, Fcgr3, Fcgr2b, Alox5ap, and Ptprz1). ETO2 and LMO2 coregulate a minority of GATA-1–regulated and –insensitive genes (Fig. 4C). Thus, a comparable reduction in the level of two components of the GATA-1-Scl/TAL1 complex dysregulated distinct gene ensembles, and the genes comprising these ensembles differed considerably vis-à-vis their cellular functions.

LMO2 had not been implicated in FOG-1–independent GATA factor pathways or any other mode of GATA-1 function. We tested whether LMO2 occupies FOG-1–independent, GATA-1–repressed genes, which would imply a direct function at these loci. ER-GATA-1 occupied the Lyl1 and Rgs18 promoters, but not the control Necdin promoter (Fig. S2A). Scl/TAL1 occupied the promoters similarly in uninduced and induced cells (Fig. S2B), consistent with its colocalization with GATA-2, before GATA-1 occupancy (34, 36). LMO2 occupancy resembled that of Scl/TAL1 (Fig. S2D), and coupled with dysregulated transcription upon LMO2 knockdown, these results indicate that LMO2 functions directly to control these FOG-1–independent, GATA-1–repressed genes.

Multimeric Complexes Containing Master Developmental Regulators: Mechanistic Principles and Biological Implications. (VSports注册入口)

Critical developmental regulators that establish and modulate genetic networks often reside in multimeric complexes. Many questions remain unanswered regarding how individual components contribute to complex function. Perhaps the simplest mechanism involves the coordination of the actions of all components of a complex to yield a sole molecular output, such as activation or repression. Individual components may have dedicated functions to mediate recruitment of the complex to factors occupying target genes, to stabilize intracomplex interactions, to catalyze chromatin modification/remodeling, to interact with the transcriptional machinery, or to integrate cell signals. Given these intertwined activities, perturbations of individual components would yield qualitatively similar defects. Alternatively, distinct components may function in unique pathways. Rather than collectively mediating complex recruitment and activity, individual components might selectively mediate recruitment to distinct cohorts of targets. Similarly, components may have intrinsic enzymatic activities essential for the control of select targets, and therefore building multiple components with specialized activities into the complex would diversify complex function.

Because the GATA-1-Scl/TAL1 complex has been defined biochemically (15, 22, 29, 46, 47, 49), its components colocalize at chromatin sites (36, 50, 51), and it has critical biological functions (1), this is a particularly attractive system for addressing mechanisms underlying the function of multimeric complexes. High-efficiency knockdowns of two components of this complex revealed major differences in their activities to control target genes. Gene Ontology analysis revealed extreme differences in the spectrum of biological activities attributed to ETO2 vs. LMO2 targets. ETO2 and LMO2 therefore exert qualitatively distinct functions, inconsistent with a model in which the tightly linked actions of these components confer a solitary activity. The qualitatively distinct activities indicate that these components can convey independent molecular and cellular functions, entirely unexpected based on the GATA-1-Scl/TAL1 complex paradigm.

Our results illustrate how trans-acting factors that bind DNA sequence specifically can serve as a scaffold to build a complex with multifunctionality. Although ETO2 and LMO2 clearly exert independent functions, ETO2 and LMO2 coregulate a small cohort of genes (Fig. 4C). Thus, the independent actions of these factors do not preclude their collective function, at least not at select loci. In principle, an alternative possibility is that ETO2 and LMO2 function both within and independent of the GATA-1-Scl/TAL1 complex. ETO2 interacts with the intracellular domain of Notch1 and the Notch-regulated transcription factor CBF1, thereby repressing Notch target genes (52).

Considering parameters dictating the composition of multiprotein complexes at endogenous chromatin sites, at the most rudimentary level, the requisite DNA and protein interaction modules are critically important. The precise configuration of cis-elements may influence the conformation of DNA-bound components, favoring or disfavoring nucleation of scaffold constituents. Regarding the GATA-1-Scl/TAL1 complex, a deeper understanding of the molecular structure of the complex in solution and when bound to chromatin is required to assess the relative importance of cis-element–instigated conformational transitions. Differences in the signaling milieu characteristic of distinct subnuclear compartments may also control scaffold composition in a spatially restricted manner. Although GATA-1 target genes regulated via distinct mechanisms can reside in different subnuclear compartments (53), whether subnuclear organization and spatially restricted signaling affect GATA-1-Scl/TAL1 complex structure and function is unknown.

Given our strong evidence that components of a multimeric complex containing master regulators of development exert qualitatively distinct functions, it is instructive to consider the pathophysiological implications of this work. As disruptions of different GATA-1-Scl/TAL1 complex constituents yielded qualitatively distinct phenotypes, this may explain why different complex components are linked to the genesis of distinct leukemias (26, 31, 54). To understand how perturbations of individual components of a complex containing multiple developmental regulators yield diverse pathophysiologies, it will be important to define the structure and function of the complex in distinct cellular contexts, and importantly in distinct neighborhoods within the 3D confines of the nucleus.

Cell Culture.

Cultures were grown as described in SI Materials and Methods.

Antibodies.

Antibodies are described in SI Materials and Methods.

Quantitative ChIP Assay.

Real-time PCR-based quantitative ChIP analysis was conducted as described (55) and in SI Materials and Methods.

Small Interfering RNA-Mediated Knockdown.

ETO2 knockdown in G1E-ER-GATA-1 cells was as described previously (36). The SMARTpool siRNA sequences are described in SI Materials and Methods. The siGENOME Non-Targeting siRNA pool (D-001206–13, Dharmacon/Thermo Fisher Scientific) was used as a control. Small interfering RNA was transfected into 3 × 10⁶ G1E-ER-GATA-1 cells by Amaxa Nucleofector kit R (Amaxa Inc.). The siRNA was transfected twice at 0 and 24 h, cells were induced with β-estradiol at 24 h, and harvested after 48 h (Fig. 1A).

VSports在线直播 - Quantitative RT-PCR Analysis and Transcription Profiling.

Total RNA was analyzed as described in SI Materials and Methods.

Western Blotting. (V体育平台登录)

Western blotting was conducted as described in SI Materials and Methods.