ALBERT

Description: Understanding the molecular mechanisms that regulate γ-globin gene expression is essential for development of new therapeutic strategies for individuals with sickle cell disease and β-thalassemia. We have previously identified a tissue- and developmentally- specific multiprotein transacting factor complex, the human stage selector protein (SSP), which facilitates the interaction of the g-globin gene promoters with the upstream locus control region enhancer in fetal erythoid cells. This complex interacts with the stage selector element (SSE) in the proximal g-globin promoter, a regulatory motif phylogenetically conserved in primate species with a distinct fetal stage of β-globin like gene expression. Given these observations, we hypothesized that a similar complex modulates γ-globin in the rhesus macaque, a non-human primate model that has been utilized to study β-globin like gene expression. We focused our efforts on NF-E4, given that a human isoform of this factor confers erythroid and fetal specificity to the SSP complex. Fetal liver erythroblasts were obtained from rhesus embryos and analyzed by reverse transcriptase(RT)-PCR analysis for NF-E4 expression. NF-E4 like transcripts were identified in day 60, 80 and 120 embryonic erythroblasts, but not other rhesus tissues, demonstrating an erythroid-specific pattern of expression. Utilizing 5′ RACE, we cloned a full length NF-E4 transcript, identifying an open reading frame encoding a 131 amino acid polypeptide. This 20kD polypeptide shares a high degree of homology with human NF-E4, especially in its carboxy-terminal domain. Like human NF-E4, GST pulldown chromatography confirmed the ability of the rhesus factor to interact directly with CP2 and ALY, the other core components of the SSP. To evaluate rNF-E4 function in vivo, we utilized retrovirally mediated gene transfer to enforce expression of this factor in K562 cells, a model of human fetal erythropoiesis. Initial co-immunoprecipitation studies confirmed the in vivo interaction of rNF-E4 with other components of the SSP. Interestingly, we observed a specific 3-fold induction of γ-globin gene expression in rNF-E4 expressing cells when compared to controls. Moreover, we demonstrated that, like enforced expression of human NF-E4, rNF-E4 induced a significant increase in ε-globin gene expression. Taken together, our results suggest a conservation of NF-E4 expression and function in species with a fetal stage of globin gene expression. Moreover, the identification of rNF-E4 provides a platform for the pre-clinical development of therapeutic agents that induce high levels of NF-E4 in adult erythroblasts.

Description: The zinc finger-encoding transacting factor EKLF binds key regulatory elements of many erythroid-specific genes, and is essential for definitive erythropoiesis. Mice lacking this factor (EKLF−/−) die of anemia by E15.5 of gestation, failing to activate β-globin gene transcription, and demonstrating a block in the erythroid differentiation program at the primitive erythroblast stage. In contrast, megakaryocytic progenitors are amplified in EKLFnull embryos, with increased Fli-1 gene expression (a marker of early megakaryocytic differentiation), consistent with the idea that EKLF modulates the megakaryocyticerythroid (M-E) differentiation switch. We have demonstrated that an amino terminal mutant of EKLF (Δ221EKLF), is required to induce chromatin remodeling at the β-globin promoter in an EKLF-null erythroid cell line, but additional amino terminal sequences are required for initiation of β-globin gene transcription (Brown et al., 2002). To evaluate the role of this chromatin remodeling (CR) domain in erythroid and megakaryocytic differentiation in vivo, we have generated a knock-in allele of EKLFCR allele. Similar to EKLF-null embryos, mice homozygous for this mutant allele die of anemia by E15.5 of gestation. In contrast to erythroid cells lacking EKLF, EKLFCR/CR progenitors demonstrate appropriate binding of the CR encoding domain to all EKLF-regulatory sequences; a block in erythropoiesis at a more a mature stage in differentiation a chromatin architecture and histone modification pattern at erythroid-specific genes that recapitulates the events observed in EKLF+/+ erythroblasts at a similar stage of erythroid ontogeny; a failure of terminal erythroid gene transcription. Examining the role of EKLFCR in megakaryopoiesis, we observed inhibition of megakaryocytic progenitor amplification in EKLFCR/CR fetal hematopoietic cell populations when compared to EKLF-null embryos; loss of Fli-1 gene expression in EKLFCR expressing cells; binding of the EKLFCR mutant protein to the Fli-1 promoter with inhibition of gene transcription; a repressed chromatin architecture at megakaryocytic gene loci. In contrast to these results, mice homozygous for a knockin allele encoding the zinc finger DNA binding domain alone (Δ253EKLF), a region shown previously to be sufficient for chromatin remodeling in vitro, demonstrate erythroid and megakaryocytic phenotypes that resemble those observed in EKLF-null hematopoietic progenitors. Taken together, our results suggest strongly that the unique EKLFCR domain is necessary and sufficient to modulate the chromatin-specific roles of EKLF at erythroid- and megakaryocytic-specific loci in definitive hematopoietic cells in vivo.

Description: Abstract 2094 Krüppel-like factor 1 (KLF1) is a zinc finger-encoding transcription factor that recognizes CACC elements, and is essential for maximal erythroid-specific gene transcription. Several critical mechanisms dependent on KLF1 and required for gene activation have been elucidated, predominantly using the beta-globin locus. KLF1 has been associated with the ordered recruitment of SWI/SNF and RNA polymerase-II complexes, necessary for chromatin remodeling and gene transcription respectively. KLF1 has also been reported to influence erythroid-specific heme biosynthesis. Studies in KLF1-null fetal erythroblasts and a KLF-1 deficient cell line have demonstrated that mRNA levels of the first three enzymes of the biosynthetic pathway are underrepresented. However, although in vitro studies of the rate-limiting enzymes ALAS2 and PBGD suggested a potential regulatory role for KLF1, in vivo studies failed to validate these findings. ALAD is the second enzyme of the pathway. Complete loss of ALAD expression in erythroid cells results in catastrophic events during zebrafish ontogeny. Interestingly, no human erythropoietic defect has been reported as a consequence of aberrant ALAD expression. To extend the analysis of KLF1's regulation of heme biosynthesis, we evaluated KLF1 binding of enzyme regulatory sequences by EMSA and ChIP studies, identifying a KLF1 binding CACC element in the erythroid-specific ALAD promoter. This regulatory element was transactivated specifically by a KLF1 transgene in KLF1-deficient cells. Using a unique 4-OH-Tamoxifen (4-OHT) mediated KLF1-inducible erythroid cell line (K1-ERp), we identified KLF1 as an essential, and early (within 2 hours of induction) activator of transcription of the endogenous ALAD, but not ALAS2 or PBGD genes. Further studies in K1-ERp cells, including DNAseI hypersensitivity and ChIP assays revealed that KLF1 occupancy at the erythroid-specific ALAD promoter triggers a series of molecular events including histone modifications, and enhanced recruitment of the sequence-specific transcription factors, GATA-1, NF-E2 and the TAL-1/SCL multiprotein complex. Importantly, we identified differences in the kinetics of recruitment of the closely related histone acetyltransferases proteins CBP and p300 and the SWI/SNF ATPase Brg1. The latter complex was recruited subsequent to KLF1 binding, although the ALAD promoter was already DNAseI hypersensitive. These results suggest strongly that KLF1 plays a major role in the regulation of heme biosynthesis in erythroid cells. Furthermore, our data challenges a model in which an identical temporal cascade of molecular events are required for transcription at KLF1-dependent promoters. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 3189 Binding of the Krüppel-like Factor 1 (KLF1) to the β-globin gene promoter is required for developmental stage-specific chromatin remodeling and transcriptional initiation. Interaction with a long-range enhancer, the Locus Control Region (LCR) is also required for maximal gene expression. KLF1 binds the LCR, and has been described recently as a facilitator of the proximal clustering of the LCR with the β-globin gene promoter region. To elucidate the role(s) of KLF1 at the LCR and their relationship to β-globin gene activation, we evaluated KLF1-directed events across the β-globin locus using a 4-OH-Tamoxifen KLF-1 inducible erythroid cell line. KLF1 binding was maximal 1 hour post-induction at the LCR, whereas 2 hours elapsed prior to maximal occupancy at the β-globin promoter. The site-specific differential in factor occupancy is consistent with the notion that the LCR serves as a nucleating docking element for sequence-specific transcription factors as well as the RNA Pol-II complex. LCR occupancy by Pol-II, p45 NF-E2, GATA-1 and the SCL/LMO2/Ldb1 complex was KLF1-independent. In contrast, the occupancy of the β-globin promoter by Pol-II and erythroid-restricted factors was dependent on maximal KLF-1 binding. Interestingly, the SCL/LMO2/Ldb1 complex was recruited first, consistent with the idea that this complex is required for LCR/promoter interaction. Interestingly, we have identified a direct protein-protein interaction between the carboxy terminal (aa221–396) domain of KLF1 (Δ221KLF1) and Ldb1, the factor implicated as critical for distal regulatory loci interactions at the β-globin locus. To dissect these events in vivo, we took advantage of a novel KLF1 knock-in strain (Δ221KLF1). In these animals, the endogenous KLF1 gene is replaced with a carboxy-terminal domain (aa221–396) expression cassette, resulting in DNA binding of the mutant factor, with β-globin chromatin remodeling. However, Δ221KLF1 fetal liver erythroblasts fail to activate the KLF1-dependent gene network, and have the identical differentiation defect characteristic of KLF1-null erythroblasts. Chromatin remodeling at the β-globin failed to result in recruitment of GATA-1, NF-E2, or the SCL/LMO2/Ldb1 complex. In contrast, ChIP analyses of the LCR in Δ221KLF1 erythroblasts revealed rescue of RNA Pol-II, GATA-1 and the SCL/LMO2/Ldb1 complex occupancy to wild type levels. Structural studies, utilizing the chromosome conformation capture (3C) assay, revealed that an incomplete re-configuration of the locus in Δ221 KLF1 mice. In conclusion, the chromatin remodeling domain of KLF1 is sufficient to reconfigure the LCR. However, KLF1-dependent Ldb1 recruitment at the LCR is insufficient to promote adequate communication between the two regulatory elements of the β-globin gene. Preliminary data suggest that an adjacent domain (aa163–221) is sufficient to rescue β-globin transcription in vivo. Together, these domains (aa163–396) are sufficient to promote the appropriate loading of transcription factors at the β-globin promoter in conjunction with locus structural re-configuration. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 461 Krüppel-like factor 1 (KLF1) is essential for erythroid gene expression. Key molecular mechanisms modulated by this transacting factor have been elucidated at the b-globin locus. KLF1 has been associated with recruitment of SWI/SNF and RNA polymerase (PolII) complexes necessary for chromatin remodeling and gene transcription respectively, and for facilitating the apposition of the promoter with the far-upstream locus control region. More recently, KLF1 has been implicated in the regulation of an erythroid-specific gene program unlinked to the b-globin locus. Coordinated expression of these genes, including Alpha Hemoglobin-Stabilizing Protein (AHSP), a factor required for globin tetramer stability, and the red cell membrane protein Dematin, are critical for erythroid ontogeny. To compare the role(s) of KLF1 at these loci, we have used a unique 4-OH-Tamoxifen (4-OHT) inducible erythroid cell line, which facilitates the characterization of the temporal kinetics of KLF1-dependent erythroid gene activation. In preliminary experiments, we observed that KLF1 binding was maximal at the three loci within 60 minutes of 4-OHT induction. AHSP and dematin primary RNA transcripts followed similar kinetics, being maximal at 60-90 minutes post-induction. In contrast, b-globin gene transcription reached a plateau 4-6 hours post-induction. From these observations, we hypothesized that transcriptional activation at AHSP and dematin differs from that observed at the b-globin cluster. Consistent with this hypothesis, we observed significant differences in chromatin remodeling at the three loci. At the b-globin promoter, we observed a small but statistically significant increase in DNaseI sensitivity, a measure of chromatin remodeling, with KLF1 binding. In contrast, we observed a complete loss of DNaseI resistance after KLF1 binding at the AHSP and dematin promoters. Consistent with these findings, we observed a five-fold reduction in histone H3 occupancy at the AHSP and dematin promoters, contrasting with no significant change in occupancy at the b-promoter. Importantly, these differences were not observed in regions 1-5 kb upstream of the promoters. These observations, coupled with similar differences in DNaseI hypersensitivity and histone occupancy in fetal liver erythroblasts from wild type and KLF1-null mice, suggest a profound difference in the mechanisms of chromatin remodeling at KLF1-dependent erythroid gene loci. To explore the potential mechanisms underlying these differences in chromatin accessibility, we examined the kinetics of recruitment of other transacting factors and co-activators to the three loci. We observed similar increases in binding of serine-5 phosphorylated PolII, GATA-1, and p45NF-E2 at the promoters. In contrast, binding of BRG1, the core ATPase component of the SWI/SNF complex differed between the b–promoter and the other erythroid genes. Although BRG1 binding was co-incident with KLF1 binding to the b-gene, we observed significant albeit weak binding of this complex to the AHSP and Dematin promoters only after maximal gene transcription had occurred. Our results suggest that different KLF1 multiprotein complexes are recruited to remodel target gene promoters in vivo. Furthermore, we propose that KLF1's chromatin remodeling capabilities are not limited to the recruitment of the SWI/SNF complexes Disclosures: No relevant conflicts of interest to declare.

Description: Binding of EKLF to the proximal promoter CACC motif is essential for high-level tissue-specific β-globin gene expression. More recent studies have demonstrated that EKLF regulates expression of other erythroid-specific genes, suggesting a broad role for EKLF in co-ordinating gene transcription in differentiating erythroblasts. Given these observations, we hypothesized that EKLF may play a role in synchronizing α- and β-globin gene expression. Supporting this model, studies of fetal erythroblasts derived from EKLF-null embryos revealed a 3-fold reduction in murine α-globin gene expression in fetal erythroblasts when compared to wild type littermate controls. A similar reduction in primary α-globin RNA transcripts was observed in these studies. To further examine the molecular consequences of EKLF function at the α- and β-globin genes in vivo, we utilized an erythroid cell line derived from EKLF null fetal liver cells. We have demonstrated previously that introduction into these cells of the wildtype EKLF cDNA, fused in frame with a mutant estrogen response element results in tamoxifen-dependent rescue of β-globin gene expression. Consistent with our observations in primary erythroblasts, α-globin gene expression is present in the absence of functional EKLF. However, with tamoxifen induction, we observed a 3–5 fold increase in α-globin gene transcription. Interestingly, the kinetics of the changes in transcription of the α- and β-gene transcripts were similar. Enhancement in α-gene transcription was associated with EKLF binding at the α- and β-globin promoters as determined by a quantitative chromatin immunoprecipitation (ChIP) assay. Interestingly, maximal EKLF binding and α-gene transcription was observed within 2 hours of tamoxifen induction. We hypothesized that the role of EKLF may differ function at the promoters, given that a basal level of α-globin gene expression occurs in absence of EKLF binding. Supporting this hypothesis, we observed sequential recruitment of p45NF-E2, RNA polymerase II (Pol II) and the co-activator CBP to the β-promoter with tamoxifen induction. No change in GATA-1 binding was observed. In contrast, p45NF-E2 does not bind to the α-promoter and the kinetics of GATA-1 and PolII association is unchanged after tamoxifen induction. Taken together, our results demonstrate that EKLF regulates the co-ordinate high-level transcription of the α- and β-globin genes, binding in a kinetically identical manner to the gene promoters. However, the effects of EKLF on transacting factor recruitment (and chromatin modification) differ between the promoters, consistent with the idea that EKLF acts in a context-specific manner to modulate gene transcription.