A 3'-flanking NF-kappa B site mediates developmental silencing of the human zeta -globin gene
1999; Springer Nature; Volume: 18; Issue: 8 Linguagem: Inglês
10.1093/emboj/18.8.2218
ISSN1460-2075
Autores Tópico(s)RNA regulation and disease
ResumoArticle15 April 1999free access A 3′-flanking NF-κB site mediates developmental silencing of the human ζ-globin gene Zhibin Wang Zhibin Wang Howard Hughes Medical Institute and Departments of Genetics and Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104 USA Search for more papers by this author Stephen A. Liebhaber Corresponding Author Stephen A. Liebhaber Howard Hughes Medical Institute and Departments of Genetics and Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104 USA Search for more papers by this author Zhibin Wang Zhibin Wang Howard Hughes Medical Institute and Departments of Genetics and Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104 USA Search for more papers by this author Stephen A. Liebhaber Corresponding Author Stephen A. Liebhaber Howard Hughes Medical Institute and Departments of Genetics and Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104 USA Search for more papers by this author Author Information Zhibin Wang1 and Stephen A. Liebhaber 1 1Howard Hughes Medical Institute and Departments of Genetics and Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104 USA *Corresponding author. E-mail: [email protected] The EMBO Journal (1999)18:2218-2228https://doi.org/10.1093/emboj/18.8.2218 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The central developmental event in the human (h)α-globin gene cluster is selective silencing of the ζ-globin gene as erythropoiesis shifts from primitive erythroblasts in the embryonic yolk sac to definitive erythroblasts in the fetal liver. Previous studies have demonstrated that full developmental silencing of the hζ-globin gene in transgenic mice requires the proximal 2.1 kb of its 3′-flanking region. In the current report, we localize this silencing activity to a 108 bp segment located 1.2 kb 3′ to the ζ-globin gene. Protein(s) in nuclear extracts from cell lines representing the fetal/adult erythroid stage bind specifically to an NF-κB motif located at this site. In contrast, this binding activity is lacking in the nuclear extract of an embryonic-stage erythroid line expressing ζ-globin. This complex is quantitatively recognized by antisera to the NF-κB p50 and to a lesser extent to p65 subunits. A two-base substitution that disrupts NF-κB site protein binding in vitro also results in the loss of the developmental silencing activity in vivo. The data suggest that NF-κB complex formation is a crucial component of hζ-globin gene silencing. This finding expands the roles of this widely distributed transcriptional complex to include negative regulation in mammalian development. Introduction The human (h)α-globin gene cluster, located on chromosome 16p, contains three expressed genes arranged in the order of their developmental expression; 5′ ζ-α2-α1 3′ (Figure 1A; also see Higgs et al., 1989). hζ-globin is selectively expressed in the primitive erythroid cells of the embryonic yolk sac where it assembles into the embryonic hemoglobin tetramers, ζ2ϵ2 and ζ2γ2. Concurrent with the shift to definitive erythropoiesis in the fetal liver, occurring at gestational weeks 6–7 in humans and days 9–12 in mice, expression of the ζ-globin gene is suppressed to trace levels (Peschle et al., 1985; Albitar et al., 1989). In contrast, the two adjacent α-globin genes, α1 and α2, are co-expressed throughout development; they are initially activated in the primitive erythroid cells and their expression is subsequently increased several fold as the ζ-globin genes are silenced (Albitar et al., 1992). The mechanism(s) responsible for switching within the α-globin gene cluster, and specifically for selective silencing of the hζ-globin gene at the embryonic-to-fetal transition of erythropoiesis, remains undefined. Figure 1.Developmental profiles of the hζ-globin transgenes linked to HS-40. (A) The human α-globin gene cluster. The three functional genes (ζ, α2 and α1) are shown as filled boxes, the three pseudogenes as open boxes and the θ gene, the expression status of which is unclear, as a open box. The developmental specificity of each of the expressed genes is noted above the cluster. The position of the hα-globin locus major regulatory element, HS-40, located 40 kb upstream from hζ-globin gene, is noted. (B) Structures of the HS-40 ζEco and HS-40 ζ transgenes. The full-length hζ-globin gene on a 4.4 kb EcoRI fragment contains three exons (filled boxes). HS-40 is located on a 1.4 kb EcoRI–HindIII fragment. Nomenclature denotes the extent of 3′-flanking region. The angled arrow marks the site of transcription initiation. (C) RNase protection assay of embryonic and fetal erythroid RNA samples from representative transgenic mouse lines. The gestational age of each embryo and fetus is noted; day 9.5 samples are from individual yolk sacs, and day 14.5 and day 16.5 samples are from individual fetal livers. The origins of samples from transgenic (+) or non-transgenic (−) litter mates are indicated. The autoradiograph panels show analyses with probes corresponding to mα-globin mRNA (mα), mζ-globin mRNA (mζ) and hζ-globin mRNA (hζ). The 150 nucleotide protected fragment corresponds to expression of the hζ transgene; the 179 and 193 nucleotide protected fragments correspond to expression of the endogenous mα- and mζ-globin mRNAs, respectively (Liebhaber et al., 1996). (D) Expression of HS-40 ζ and HS-40 ζEco transgenes. The levels of transgene expression from five HS-40 ζ lines and two HS-40 ζEco lines are shown. Each value represents the average of at least two determinations carried out on tissues from three or more mice of each line at each time-point. Expression levels are calculated as a percentage of the combined expression of endogenous mζ- and mα-globin genes corrected for specific activities of the respective probes (see Materials and methods). The developmental age of each sample is as days p.c. The transgene copy number for each line is shown in the box to the right of the respective name. The vertical arrow in each frame indicates the mean fold-decrease for each transgenic construct between day 9.5 and day 16.5. Download figure Download PowerPoint Transgenic mice constitute an informative model system for the study of switching in the hα-globin gene cluster. The murine (m)α- and hα-globin gene clusters are conserved in their structures and developmental profiles; both clusters contain a single embryonic ζ- and two co-expressed α-globin genes (Pressley et al., 1980; Liebhaber et al., 1981). Therefore human α- and ζ-globin transgenes can be directly compared with the corresponding set of endogenous mouse genes during development. Previous studies have revealed normal developmental control of the human α-globin gene cluster in the transgenic mice (Gourdon et al., 1994). The developmental control of the hζ-globin expression is gene-autonomous; all the necessary information for appropriate developmental silencing is present within and/or directly flanking the ζ–globin transgene and this control does not require cis interactions with the fetal/adult α-globin gene (Albitar et al., 1991; Sabath et al., 1993). Linkage of the ζ-globin transgene to either the major α-globin cluster regulatory element (HS-40) or the β-globin cluster locus control region (microβ-LCR or HSII) establishes a permissive chromatin domain and overcomes site-of-integration effects on the transgene (Albitar et al., 1991; Sabath et al., 1993; Sharpe et al., 1993). A normal developmental profile of the hζ-globin transgene is maintained irrespective of which of these two auxiliary elements is used to enhance transgene expression (Albitar et al., 1991; Starck et al., 1994; Hug et al., 1996; Martin et al., 1996). The parallel organization of the murine and human α-globin gene clusters and the gene-autonomous nature of hζ-globin developmental control suggest that transgenic studies aimed at defining the cis determinants of hζ-globin gene silencing may be highly informative. The transcriptional activities of many structural genes are controlled by DNA sequences within the promoter and/or in the contiguous 5′-flanking region. Consistent with this generality, the hζ-globin gene with 550 bp of the proximal promoter was found to be appropriately silenced in transgenic mice (Pondel et al., 1992; Sabath et al., 1993, 1995). Subsequent 5′-deletional mapping of this region revealed that a hζ-globin transgene with as little as 128 bp of 5′-flanking sequence is also normally regulated (Sabath et al., 1993). However, extensive subsequent analysis of this promoter region failed to reveal elements that mediate silencing independently of global effects on transcription (Sabath et al., 1993). These studies suggest that cis-acting determinants required for stage-specific expression of the hζ-globin gene may reside within the transcribed or 3′-flanking sequences. Recent studies suggest that silencing of the hζ-globin gene is a complex process reflecting both transcriptional and post-transcriptional mechanisms. Using transgenic models, silencing determinants have been identified in both the transcribed region and the 3′-flanking region of the hζ-globin gene (Liebhaber et al., 1996). Replacement of the coding region of the hα-globin transgene with that of the hζ-globin gene results in partial gene silencing despite the presence of the intact α-globin promoter. This silencing activity of the ζ-globin transcribed region has been linked to the inability of its 3′UTR sequences to stabilize ζ-globin mRNA in the prolonged post-transcription stage of fetal/adult erythroid maturation (Russell et al., 1997, 1998) A second determinant of ζ-globin gene silencing has been localized to its 3′-flanking region. Previous studies using transgenic mouse models demonstrate that the hζ-globin transgene containing 2.1 kb of contiguous 3′-flanking sequences exhibits proper developmental regulation with a 50-fold decrease in ζ–globin mRNA levels from day 9.5 yolk sac to day 16.5 fetal liver (Liebhaber et al., 1996). In contrast, truncation of the 3′-flanking region to 106 bp 3′ to the poly(A) addition site significantly blunted silencing to a 10-fold decrease over the same developmental window (Liebhaber et al., 1996). The transcriptional silencing mediated by the 3′ flanking region can be detected as early as day 8.5–9.5, suggesting that the ζ-globin gene undergoes active silencing soon after it is activated. Thus, the combined effects of transcriptional silencing by the 3′-flanking region and accelerated clearance of residually expressed ζ-globin mRNA can account for most if not all of the developmental control observed for the intact ζ-globin transgene (Liebhaber et al., 1996). While the cis and trans determinants of ζ-globin mRNA instability have been studied in detail (Russell and Liebhaber 1998), the basis for silencing activity of the 3′-flanking region remains undefined. In the present report, we map, characterize and functionally test a cis determinant positioned within the 3′-flanking region of the hζ-globin gene, which mediates developmental silencing. Results A developmental silencing determinant is located 3′ to the human ζ-globin gene We have previously demonstrated that the hζ-globin transgene containing 0.55 kb of 5′-flanking sequences and 2.1 kb of 3′-flanking sequences is silenced 50-fold during the transition from primitive erythropoiesis (day 9.5 yolk sac) to definitive erythropoiesis (day 16.5 fetal liver) (Liebhaber et al., 1996). This 50-fold decrease in the steady-state concentration of the transgene mRNA defines ‘full developmental silencing’ (see Materials and methods; Liebhaber et al., 1996). This developmental control is markedly blunted by truncation of the 3′ flanking region suggesting that an element(s) critical to full silencing is located in this region (see Introduction). The hζ-globin transgenes in these prior studies were linked to the β-LCR to overcome site-of-integration effects (Talbot et al., 1989). Prior to further mapping of the putative silencing determinant, we confirmed its function in the context of the homologous αHS-40 major regulatory element (Figure 1A). This was accomplished by comparing the expression profiles of the HS-40 ζEco transgene containing the entire 2.1 kb of contiguous 3′-flanking sequence, with the HS-40 ζ transgene that is truncated 106 bp 3′ to the poly(A) addition site (Figure 1B). Multiple lines were established with each transgene and timed embryos were generated by crossing positive males (F1 or greater) from each line with wild-type females. Total RNA was isolated from day 9.5 yolk sacs and day 16.5 fetal livers of the transgenic embryos, and globin mRNAs were quantified by an RNase protection assay (Liebhaber et al., 1996). Representative studies are shown in Figure 1C. As expected, the mα- and mζ-globin genes were co-expressed in day 9.5 yolk sac while in day 16.5 fetal liver the level of mα-globin mRNA increased and mζ-globin gene was silenced. Analysis of transgene mRNA revealed equivalent expression of the HS-40 ζEco and HS-40 ζ transgenes at day 9.5, while at day 16.5 the HS-40 ζEco transgene was fully silenced and the HS-40 ζ transgene continued to be expressed. Quantitative analysis of mRNA expression profiles in five HS-40 ζ and two HS-40 ζEco mouse lines confirmed this pattern. The two HS-40 ζEco lines were silenced 60- and 118-fold, respectively (Figure 1D, right), while developmental silencing of the HS-40 ζ transgenes was markedly blunted (mean 15-fold decrease) (Figure 1D, left). These data substantiated the previous assignment of a silencing determinant to the 3′-flanking region and confirmed equivalent developmental control of the hζ-globin transgene in the presence of the major regulatory elements derived from either the α- or the β-globin gene clusters. The silencing element maps to a 108 bp sequence located 1.2 kb 3′ to the hζ-globin gene Full developmental silencing of the hζ-globin gene requires one or more elements within the proximal 2.1 kb of the 3′-flanking region (as described above). To facilitate subsequent studies, the sequence of this 3′-flanking region was determined (DDBJ/EMBL/GenBank accession No. AF078904). This sequence revealed the presence of two complete Alu elements in a head-to-head transcriptional orientation separated by 400 bp (Figure 2A). Native restriction sites were utilized to generate a set of 3′-truncation constructs: ζNco, ζBst and ζDra. These genes extend 0.93, 1.33 and 1.80 kb into the 3′-flanking sequences, respectively (Figure 2A). Each gene was linked to the microβ-LCR (see above), because it resulted in a more reproducible and robust enhancement of overall globin transgene expression levels than the HS-40 (10-fold higher transgene transcript levels; see below). Transgenic founders were generated from each construct and Southern analysis of genomic DNA from respective F1 progeny revealed the copy number and integrity of the transgenes. The developmental profile of the transgene in each line was determined using at least three individual embyros for each time-point. A study of a represenative ζNco line demonstrated that transgene expression was present in day 9.5 yolk sac and persisted in day 16.5 fetal liver (Figure 2B, left). Quantitative analysis of all six ζNco lines revealed that silencing of the transgene expression was blunted in five of six lines (6- to 14-fold decrease with a mean drop of 11-fold; Figure 2C, left). This incomplete silencing paralleled that of the ζ-globin transgene with the fully truncated 3′-flanking region (Figure 1; Liebhaber et al., 1996). These data suggested that the 3′-flanking region silencing element(s) was located more than 0.93 kb 3′ to the hζ-globin gene. Figure 2.Localization of a developmental-silencing element 3′ of the hζ-globin gene. (A) Structure of three hζ transgenes with truncations of the 3′-flanking region. A 4.4 kb EcoRI genomic fragment is shown. This fragment contains the full-length hζ-globin transgene gene as well as two head-to-head Alu repeats are represented as open boxes. The predicted transcriptional orientations of the Alu elements are indicated by arrows. A 6.5 kb DNA cassette containing all hypersensitive sites (HS) (vertical arrows) of the β-LCR is shown to the left of the cluster. The β-LCR cassette was ligated to each gene prior to microinjection. Nomenclature of the genes corresponds to the restriction site at the 3′ terminus (DraI, Bst1107I or NcoI site). (B) RNase protection assay of embryonic and fetal RNA samples. Results from three transgenic littermates of the indicated lines are shown for each construct. (C) Expression of ζNco, ζBst and ζDra transgenes. The levels of transgene expression from six ζNco lines, four ζBst and two ζDra lines are shown. Each value represents the average of at least two determinations carried out on tissues from three or more mice at each time-point (as described for Figure 1). Download figure Download PowerPoint The expression profiles of the ζBst and ζDra transgenes were analyzed next. RNase protection assay of a representative ζBst line (Figure 2B, middle) revealed full silencing of mRNA expression, paralleling the endogenous mζ-globin mRNA. Quantitative analysis of all four ζBst lines showed a remarkably consistent drop in transgene mRNA levels from day 9.5 yolk sac to day 16.5 fetal liver with a mean decrease of 45-fold (Figure 2C, middle). A similar result was found for the ζDra transgene with a more extensive 3′-flanking region; the mean silencing of transgene expression was 77-fold (Figure 2B right and C right). The expression profiles of the ζNco, ζBst and ζDra transgenes confirmed the presence of a silencing element in the 3′ flanking region and mapped it within a 408 bp sequence bracketed by the NcoI and Bst1107I sites 3′ to the hζ-globin gene. Inclusion of this region was necessary for full (∼50-fold) developmental silencing. A second set of 3′-deletion transgenes was constructed to more precisely map the silencing element within the 408 bp Nco–Bst fragment. These transgenes contained 1.1, 1.2 and 1.3 kb of contiguous 3′-flanking sequences, respectively (ζ1.1, ζ1.2 and ζ1.3; Figure 3A). RNA analysis of day 9.5 yolk sacs and day 16.5 fetal livers indicated incomplete developmental silencing of the ζ1.1 transgene with a mean 10-fold decrease from day 9.5 yolk sacs to day 16.5 fetal livers (Figure 3B and C). In marked contrast, the ζ1.2 transgene, containing an additional 108 bp, was fully silenced as was the ζ1.3 transgene (mean decreases of 90-fold in both cases; Figure 3B and C). These results sublocalized the 3′-silencing element to an 108 bp fragment beginning 1.1 kb 3′ to the poly(A) addition site; developmental silencing of transgenes lacking this 108 bp sequence was significantly blunted. Figure 3.Sublocalization of the hζ-globin 3′-silencing element to a 108 bp sequence. (A) Diagram of the 408 bp NcoI–Bst1107I sequence 3′ to the hζ-globin gene. This region, which contains the silencer activity (Figure 2), encompasses an Alu repeat that is situated in the opposite transcriptional orientation to the upstream hζ-globin gene. The Alu promoter elements (A and B boxes), the terminal direct repeats (DR) and the Alu termination signal (TTT) are each indicated. Shown below are three hζ transgenes which terminate at 1.1, 1.2 or 1.3 kb 3′ to the NcoI site. The 108 bp region between the 3′ termini of the ζ1.1 and ζ1.2 transgenes, containing the putative silencing activity, is boxed (Silencer). (B) RNase protection assay of embryonic and fetal RNA samples from littermates of the indicated representative mouse lines (as described in legend to Figure 1). (C) Expression of ζ1.1, ζ1.2 and ζ1.3 transgenes at two developmental time-points (day 9.5 and day 16.5) (as described in Figure 1). Results from four or five independent transgenic lines are shown for each construct. Download figure Download PowerPoint Trans-acting factors present in the nuclei of fetal/adult but not embryonic erythroid cell lines specifically recognized a 24 bp subsegment (SE24) of the putative silencing region To identify nuclear proteins that might interact with the putative silencing element, the 108 bp DNA fragment (SE108) defined by the transgenic studies was divided into four overlapping segments for electrophoretic mobility-shift assay (EMSA) (Figure 4A, top). Nuclear extracts were harvested from erythroid cell lines with an embryonic (K562; Lozzio and Lozzio, 1975), fetal/adult (HEL; Martin and Papayannopoulou, 1982) or adult (MEL; Friend et al., 1971) globin expression phenotypes. The first three segments (A–C) failed to form a complex with any of these three nuclear extracts (data not shown). In contrast, the 3′-most fragment (SE24) assembled a DNA–protein complex when incubated with equal quantities of nuclear extracts from MEL and HEL, but not embryonic K562 cells (Figure 4A, left panel). Competition was observed with the SE24 oligo as well as full-length SE108, but not by a non-specific (NS) probe (Figure 4A, right panels). To evaluate the quality of the nuclear extracts, all three nuclear extracts were shown to form equally strong complexes on a probe containing the binding site for a ubiquitous DNA binding protein (SP1; data not shown). The selective formation of complexes on the SE24 fragment with nuclear extracts isolated from cells with fetal and adult globin gene expression phenotypes suggested that such complexes may be involved in developmental silencing of the embryonic ζ-globin gene. Figure 4.A subsegment of the developmental silencer region bound a p50 NF-κB complex. (A) A DNA fragment within the putative silencing region specifically interacted with nuclear proteins present in erythroid cells with a fetal/adult globin expression profile but not in embryonic cells stably expressing the hζ-globin gene. The 108 bp silencing segment (SE108), divided into four overlapping subsegments [A, B, C and D (SE24)], is shown at the top. Below are the electrophoretic mobility gel shift analyses of SE24. The first three segments were inactive in the EMSA analysis. The 32P-labeled SE24 probe was incubated with nuclear extracts from the indicated cell lines and the resultant products were analyzed by EMSA. A complex formed between the SE24 probe and the MEL and HEL extracts, but not with the K562 extract, is indicated by the arrow. Competition studies were carried out using 200-fold excess of unlabeled SE24 probe, the encompassing SE108 segment [see (A)], or an unrelated oligonucleotide (see Materials and methods) further demonstrating the sequence specificity of the complex formation. (B) The SE24 complex assembled at an NF–κB motif. The sequences of SE24, SE24m and the consensus (human immunoglobulin κ chain enhancer) NF-κB binding motif are shown. SE24m is identical to SE24 with the exception of a GG→CC substitution within the NF-κB binding motif (underlined). 32P-labeled SE24 was incubated with MEL nuclear extract. The complex that was formed (Complex; first lane) was competed by the addition of unlabeled SE24 and NF-κB but not the mutant SE24m. The SE24m was also unable to directly assemble a complex (right lane). (C) The SE24 complex was recognized by an antibody to NF-κB p50 and p65 subunits. 32P-labeled SE24 was incubated with HEL cell nuclear extracts in the presence of antibodies specific to each of the noted NF-κB subunits: p50, p65 (RelA), p52, p68 (RelB) and p75 (c-Rel). The left panel shows gel mobility shift analysis without or with NF–κB antibodies. The lower arrow indicates the position of the native SE24 complex and the upper arrows indicate the supershifted complexes. In the right panel 32P-labeled SE24 was incubated with the indicated nuclear extracts or with human recombinant protein p50. The complexes formed with the recombinant p50 (p50 homodimer) were distinct from the native complex. Download figure Download PowerPoint SE24 contains an NF-κB motif and binds NF–κB in vitro Sequence analysis revealed that the SE24 contained a canonical NF-κB binding site (Figure 4B, bottom). NF–κB has been implicated in the control of a wide variety of gene expression pathways and has been specifically associated with selective inhibition of body patterning pathways in Drosophila. To evaluate NF-κB binding to this site, a two-base substitution of GG for CC at the highly conserved second and third positions in the NF–κB consensus sequence was introduced into SE24 sequence. This substitution would be expected to destroy NF-κB binding (SE24m mutation; Figure 4B, bottom). When analyzed by EMSA, the complex formed between SE24 and the MEL nuclear extract could be competed by SE24 as well as by the immunoglobulin κ chain NF-κB enhancer element, but not by SE24m (Figure 4B, left panel). Moreover, SE24m was unable to directly assemble the SE24 complex in MEL nuclear extracts (Figure 4B, right panel). These results indicated that one or more factors bound specifically to the NF-κB motif in the SE24 subsegment of the SE108 silencer domain. The SE24 complex is recognized by antibodies to NF-κB p50 and p65 There are five structurally related members of the mammalian NF-κB/Rel protein family: p50, p52, RelA (p65), RelB (p68) and c-Rel (p75) (Baeuerle and Baltimore, 1996; Baldwin, 1996) These proteins share an N-terminal Rel-homology domain that mediates binding to their cognate DNA sites as hetero- and/or homodimers. EMSA supershift assays were carried out to determine whether any of these NF-κB subunits were involved in formation of the SE24 complex (Figure 4C, left panel). Antibodies to p50 and p65 were able to recognize and supershift this complex: the supershift by the antibody to p50 was quantitative while that by the antibody to p65 was less pronounced. This complex was not supershifted with antibodies to any of the other three NF-κB/Rel family members (although convincing positive controls for these antibodies could not be generated). To determine whether p50 binds to SE24 as a homodimer or as a component of a heterodimer (Fujita et al., 1992), SE24 was incubated with recombinant human p50 to assemble a homodimer complex. The recombinant p50 contains no additional sequences and should be identical to the unmodified native protein (Ghosh et al., 1990; Kieran et al., 1990; Schmid et al., 1991). The result showed that the SE24 complexes formed with the native nuclear extract and with the recombinant p50 migrated at distinct positions (Figure 4C, right panel). Mutation of the NF-κB consensus sequence 3′ to the hζ-globin gene disrupted developmental silencing The deletion mapping, developmental studies and EMSA analyses detailed above suggest that the NF-κB motif 3′ to the hζ-globin gene is directly involved in its developmental silencing. To test this model, we introduced the same two-base substitution (GG for CC) at the NF–κB consensus sequence of ζ1.2 transgene that was shown to destroy complex assembly in vitro (see Figure 4B, bottom). The mutant transgene (ζ1.2m; Figure 5A) was introduced into the mouse genome. In contrast to the full silencing of the wild-type ζ1.2 transgene (Figure 3), introduction of the two-base substitution at the NF-κB element substantially blunted the developmental control in vivo (Figure 5). An RNase protection assay from a representative mouse line demonstrated continued expression of the transgene in day 16.5 fetal livers (Figure 5B). Quantitative determinations from four ζ1.2m lines demonstrated that developmental silencing was blunted by the NF-κB mutation; the ζ1.2m transgene exhibited a mean 16-fold decrease in mRNA concentration from day 9.5 to day 16.5 (Figure 5C). This loss of silencing activity is indistinguishable from that seen when the entire 108 bp silencer element is deleted (ζ, ζNco and ζ1.1) and was clearly blunted when compared with that of the wild-type ζ1.2 transgene (90-fold decrease; Figure 3). These data confirmed a role for the 3′-flanking NF-κB binding site in the developmental silencing of the hζ-globin gene. Figure 5.Interruption of the SE24 NF-κB motif blunts developmental silencing of the hζ transgene in vivo. (A) Diagram of the ζ1.2m transgene. The CC→GG substitution that destroyed the NF-κB site in the ζ1.2 is indicated. (B) RNase protection assay of day 9.5 embryonic yolk sac and day 16.5 fetal liver RNA samples from littermates of a representative line. (C) Expression of four individual ζ1.2m transgenic lines at two developmental time-points demonstrating blunted silencing activity secondary to the GG→CC mutation at the NF-κB site. Download figure Download PowerPoint Discussion Developmental switching within the globin gene cluster is a highly conserved program (Goodman et al., 1987). In mammalian species the embryonic globin genes are silenced as erythropoiesis switches to the definitive lineage. This definitive erythroblast series, which produces smaller, non-nucleated erythrocytes, is initially established in the fetal liver and subsequently shifts to the bone marrow around the time of birth. Thus, the initial event in globin switching is silencing of the embryonic genes. Whereas a variety of cis-acting elements and trans-factors, both general and erythroid specific, have been studied in conjunction with activation and enhancement of globin gene expression (Crossley and Orkin, 1993; Miller and Bieker, 1993; Raich and Romeo, 1993), the determinants of globin gene silencing are less well defined. Study of developmental silencing of the embryonic ϵ-globin in the hβ-globin gene cluster revealed a silencer 5′ to
Referência(s)