Partial purification of a nuclear protein that binds to the CCAAT box of the mouse alpha 1-globin gene.

We enriched a fraction from nuclear extracts of murine erythroleukemia cells which contains a protein able to form stable complexes with the promoter region of the alpha 1-globin gene. Binding activity, which is present in mouse brain and a variety of cultured mouse and human cell lines, is not erythroid cell specific. Binding studies with alpha-globin gene promoter deletion mutants as well as DNase I footprinting and dimethyl sulfate protection studies showed that the factor bound specifically to the CCAAT box of the alpha 1 promoter. Enriched factor preparations exhibited weak binding to the promoter region of the beta maj-globin gene. This suggests that this protein could bind differentially to these two promoters in vivo. The enriched factor may be a ubiquitous nuclear protein involved in the differential regulation of the alpha 1- and beta maj-globin genes.

Purification and characterization of an erythroid cell-specific factor that binds the murine alpha- and beta-globin genes.

An erythroid cell-specific nuclear factor that binds tightly to a sequence motif (5'-GATAAGGA-3') shared by many erythroid cell-specific promoters was purified to homogeneity by DNA sequence affinity chromatography. Visualization of the purified factor, which we term EF-1, showed a simple pattern comprising a polypeptide doublet with Mrs of 18,000 and 19,000. We confirmed that these species account for EF-1-binding activity by eluting the polypeptides from sodium dodecyl sulfate-polyacrylamide gels and renaturing the appropriate binding activity. Using the purified polypeptides, we mapped seven factor-binding sites that are dispersed across the murine alpha- and beta-globin genes. The murine alpha-globin gene is flanked by at least two EF-1-binding sites. One site is centered at nucleotide (nt) -180 (with respect to the alpha-globin cap site). A fivefold-weaker site is located downstream of the alpha-globin poly(A) addition site, at nt +1049. We mapped five EF-1-binding sites near the murine beta-globin gene. The strongest site was centered at nt -210. Four additional sites were centered at nt -266 (adjacent to the binding site of a factor present in both murine erythroleukemia and Raji cells), -75 (overlapping the beta-globin CCAAT box), +543 (within the second intervening sequence), and -111.

Molecular cloning of the alpha-globin transcription factor CP2.

CP2, a transcription factor that binds the murine alpha-globin promoter, was purified and subjected to amino acid sequence analysis. Oligonucleotide primers derived from the sequence were used to obtain murine and human cDNA clones for the factor. The murine cDNA spans approximately 4 kb and contains two coextensive open reading frames (ORFs) which encode deduced polypeptides of 529 (ORF-1; molecular weight, 59,802) and 502 (ORF-2; molecular weight, 56,957) amino acids, slightly smaller than the purified factor as estimated from its mobility in sodium dodecyl sulfate-polyacrylamide gels (64,000 to 66,000). The human cDNA contains a single ORF of 501 amino acids that is nearly contiguous with murine ORF-2. Indeed, comparison of deduced human and murine amino acid sequences shows that the two polypeptides are 96.4% identical. A strictly conserved region is rich in serine and threonine (17.5%) and in proline (11%) residues (S-T-P domain). This S-T-P domain is immediately amino terminal to a string of 10 glutamines (in the human sequence) or a tract of alternating glutamine and proline residues (in the mouse sequence). Bacterial expression of the full-length (502-amino-acid) murine factor or of a core region comprising amino acids 133 to 395 generated polypeptides with the DNA binding specificity of CP2. These results confirmed the cloning of CP2 and delimited the region sufficient for specific DNA sequence recognition. Antisera produced against the core region recognized polypeptide species with Mrs of 64,000 and 66,000 in immune blots of nuclear extracts prepared from both murine and human cell lines, consistent with the size of the purified factor. Lastly, a data base search revealed that amino acids 63 to 270 of the murine factor are distantly related to a domain in the Drosophila gene regulatory factor Elf-1.

Purification of multiple erythroid cell proteins that bind the promoter of the alpha-globin gene.

Three erythroid cell factors that bind the murine alpha-globin promoter were enriched more than 1,000-fold by conventional and DNA sequence affinity chromatography. Visualization of enriched polypeptides revealed simple patterns suggesting that each binding activity was purified. Two of the purified proteins, alpha-CP1 and alpha-CP2, have been shown previously to interact with distinct binding sites that overlap in the alpha-globin CCAAT box. Affinity purification of alpha-CP1 revealed seven polypeptides with Mrs raging from 27,000 to 38,000. In contrast, purified alpha-CP2 was made up of a polypeptide doublet with Mrs of 64,000 and 66,000. The third purified binding activity, alpha-IRP, interacted with sequences that formed an inverted repeat (IR) between the alpha-globin CCAAT and TATAA boxes. Affinity-purified alpha-IRP was made up of a single polypeptide with an Mr of 85,000. We confirmed that the purified polypeptides corresponded to alpha-CP1-, alpha-CP2-, and alpha-IRP-binding activities by UV cross-linking experiments (alpha-CP2 and alpha-IRP) or by renaturation of binding activity after elution of polypeptides from sodium dodecyl sulfate-polyacrylamide gels (alpha-CP1 and alpha-CP2). The apparent complexity of the polypeptides accounting for alpha-CP1 binding activity prompted a further physical characterization of this factor. Sedimentation of affinity-purified alpha-CP1 in glycerol gradients containing 100 mM KCl showed that all seven polypeptides migrated as a complex that cosedimented with alpha-CP1-binding activity. In contrast, when sedimented in glycerol gradients containing 500 mM KCl, alpha-CP1 dissociated into at least two components. Under these conditions, alpha-CP1-binding activity was reduced or lost. Activity was reconstituted, however, by combining fractions that were enriched in the two components. These results were confirmed by experiments in which we showed that alpha-CP1-binding activity can be recovered only by combining distinct sets of polypeptides that were isolated and renatured from sodium dodecyl sulfate-polyacrylamide gels. Our results suggest that the seven polypeptides visualized after affinity purification of alpha-CP1 interact to form a heterotypic complex (or set of complexes) required for alpha-CP1-binding activity.

Promoter elements and erythroid cell nuclear factors that regulate alpha-globin gene transcription in vitro.

We have previously purified four factors (alpha-IRP, alpha-CP1, alpha-CP2, and NF-E1) that interact with the promoter of the alpha-globin gene. One of these (NF-E1) is a tissue-restricted factor that has recently been cloned. The binding sites of these factors identify DNA sequence elements that might mediate the tissue-specific and inducible transcription of the alpha-globin gene. This possibility was tested in a series of in vitro transcription experiments. An examination of 5' truncated templates and synthetic promoters constituted from individual factor-binding sites apposed to the alpha-TATAA box showed that the binding elements of three factors (alpha-CP1, alpha-IRP, and NF-E1) mediate four- to sixfold activation of transcription in vitro. In contrast, one element (alpha-CP2) stimulated transcription less than twofold. The 5- to 10-fold stimulation of these latter templates upon addition of a DNA sequence affinity-purified factor suggests that alpha-CP2 is functionally limiting in nuclear extracts. Additional experiments further tested the effect of supplementing extracts with factors purified from erythroid cell nuclear extracts or, in the case of NF-E1, enriched from a bacterial cDNA expression system. Each factor tested stimulated transcription in vitro in a binding-site-dependent manner. Our results provide a comprehensive functional view of the murine alpha-globin promoter and suggest possible mechanisms for activation of alpha-globin gene transcription during induced differentiation of murine erythroleukemia cells.

Identification and characterization of multiple erythroid cell proteins that interact with the promoter of the murine alpha-globin gene.

The proteins responsible for erythroid-specific footprints extending to -180 on the mouse alpha-globin gene were identified, enriched, and characterized from extracts of murine erythroleukemia (MEL) cells. Three proteins accounted for most aspects of the footprints. The binding sites of two proteins, termed alpha-CP1 and alpha-CP2, overlapped in the CCAAT box. Further characterization of these two CCAAT binding proteins showed that neither interacted with the adenovirus origin of replication, a strong CCAAT transcription factor-nuclear factor 1 binding site. A third protein, termed alpha-IRP, interacted with two sequences that formed an inverted repeat (IR) between the CCAAT and TATAA boxes. Interestingly, the binding domain of one of the CCAAT factors, alpha-CP1, overlapped one alpha-IRP binding site. alpha-CP1 thus overlapped the binding domains of both alpha-CP2 and alpha-IRP. The IRs included GC-rich sequences reminiscent of SP1-binding sites. Indeed, alpha-IRP bound as well to the alpha-promoter as it did to SP1 sites in the simian virus 40 early promoter. These results suggest that alpha-IRP may be related to the transcription factor Sp1. We determined the level of each alpha-globin-binding activity before and after induced erythroid differentiation of MEL cells. We found that differentiation caused alpha-CP1 activity to drop three- to fivefold, while alpha-IRP activity decreased slightly and alpha-CP2 activity increased two- to threefold.

Induced differentiation of murine erythroleukemia cells: cellular and molecular mechanisms.

Study of inducer-mediated differentiation of murine erythroleukemia cells provides insights into the cellular and molecular mechanisms implicated in cell differentiation. The loss of proliferative capacity is revealed to be a complex multistep process during which the cells progress through a series of stages, including a precommitment "initiation" stage, a stage suggestive of the accumulation of commitment-related factors, and, finally, a stage of expression of the characteristics of the differentiated state. Cell cycle arrest in G1 phase of the cell cycle may, in part at least, be related to down-regulation of protein p53 synthesis. Expression of induced differentiation is accompanied by an acceleration of transcription at the globin loci, and possibly by posttranscriptional modulation of globin mRNA accumulation, as well. Cells at the stage of erythroid cell development represented by the transformed, differentiation-arrested MELC, have acquired a unique DNA structure and chromatin configuration around the globin genes which distinguish them from other, nonerythroid cells; additional complex changes in chromatin configuration accompany, and probably precede, inducer-mediated acceleration of globin gene transcription during terminal differentiation. Passage through G1 and early S phase of the cell cycle, in the presence of inducer, is critical for subsequent globin gene expression and may be important in establishing the chromatin reconfiguration required for gene expression.

Nucleosome disruption precedes transcription and is largely limited to the transcribed domain of globin genes in murine erythroleukemia cells.

We used micrococcal nuclease to separate murine erythroleukemia cell (MELC) chromatin into soluble and insoluble fractions which differ in gene content and chromatin structure. Genes that are not expressed in the erythroid lineage, such as the Ig alpha and albumin genes, distribute preferentially into the soluble rather than the insoluble fraction, and are organized into nucleosomes in both fractions. Both alpha 1- and beta maj-globin genes are enriched in the insoluble fraction and are organized into structures that are partially devoid of nucleosomes in uninduced MELC, when the genes are transcriptionally inactive. Following chemical induction of MELC and the onset of globin gene transcription, globin gene enrichment and nucleosome disruption in the insoluble chromatin fraction increase. Using seven DNA subclones that span the beta maj-globin gene we show that insolubility and nucleosome disruption are largely limited to DNA sequences lying within the transcribed domain. Non-transcribed, flanking sequences are soluble and organized into nucleosomes. In addition, the globin genes found in insoluble, non-nucleosomal chromatin contain previously engaged RNA polymerases which can elongate globin RNA chains in vitro in a pattern qualitatively and quantitatively similar to intact nuclei. These results are discussed in terms of a model for globin gene activation during erythropoeisis.

Gene expression in murine erythroleukemia cells. Transcriptional control and chromatin structure of the alpha 1-globin gene.

The transcriptional activation and chromatin structure of the alpha 1-globin gene was analyzed during induced erythroid differentiation in murine erythroleukemia cells (MELC). In uninduced MELC, a low level of alpha 1-globin, coding-strand-specific transcription is detectable. Hexamethylene bisacetamide (HMBA)-mediated MELC differentiation is associated with a 10 to 20-fold increase in the rate of alpha 1-globin gene transcription. In induced MELC, alpha 1-globin gene transcription initiated predominantly near the cap site, occurs only off the coding strand, and might terminate, or attenuate, in a region 50 to 250 base-pairs 3' of the polyadenylation site. Before transcriptional activation of the gene, chromatin surrounding the gene displays overlapping DNase I and S1 nuclease sensitive sites, which map to a region 100 to 200 base-pairs 5' of the cap site. After induction, the nuclease sensitivity of these pre-established, overlapping sites increases. In addition, induction generates novel, non-overlapping DNase I and S1 nuclease sensitive sites, which map to regions centered 300 base-pairs 5', and approximately coincident with the cap site, respectively. We compared the time-course of alpha 1-globin transcriptional activation to the chromatin structure changes. A twofold increase in gene transcription is detected within two cell cycles (approximately 24 hours) of exposure of cells synchronized in the G1/early S-phase to inducer. Transcription rates continue to increase for at least 48 hours in MELC cultured with HMBA (the latest time assayed). Chromatin structure changes appear nearly complete after two cell cycles.

Induced transcription of the mouse beta-globin transcription unit in erythroleukemia cells. Time-course of induction and of changes in chromatin structure.

The transcription of the beta-globin genes in mouse erythroleukemia cells has been examined by hybridizing labeled RNA obtained from isolated nuclei after chain elongation in the presence of [alpha-32P]UTP. There is induction of at least 30-fold of beta maj globin transcription after cells are treated with either dimethylsulfoxide or hexamethylene bisacetamide. The induction requires 36 to 48 hours to be maximal, during which time the cells double about three to four times. During this time, a site in the beta maj DNA region becomes hypersensitive to DNase. The development of this hypersensitive site is co-ordinate with the transcriptional increase. The induced transcripts in the beta-globin region are alpha-amanitin-sensitive (and therefore are RNA polymerase II products). An examination of weak transcriptional signals to DNA fragments upstream of the beta maj globin gene in uninduced mouse erythroleukemia cells and in cells that do not make globin is also reported. The low level of hybridization to the upstream regions in uninduced erythroleukemia cells, in L cells (a fibroblast) and in a strain of erythroleukemia cells that no longer make globin are not equally sensitive to alpha-amanitin as in the induced signal. These experiments help define the inducible transcription unit for beta maj globin mRNA production.

Induction of globin gene expression during erythroid cell differentiation.

We can provide increasing insight, albeit still incomplete, into the changes in MELC that accompany globin gene expression induced by polar chemicals, such as DMSO, and other agents. These transformed, CFUe-like erythroid precursor cells exhibit in their uninduced state, a DNA methylation pattern and globin gene (formula; see text) chromatin configuration (DNase I sensitivity) that is compatible with actual or potential gene transcription. Such features may reflect alterations in chromatin configuration that have occurred at a stage prior to leukemic transformation, during the differentiation of earlier erythroid precursor cells and associated with the restriction in developmental potential characteristic of progression to the CFUe (or MELC) stage of erythropoiesis. Uninduced MELC display a low level of globin gene transcription, producing globin mRNA or mRNA precursors whose processing or stabilization is the target of action of hemin. The major increase in MELC globin gene transcription that is initiated by DMSO, HMBA, or butyric acid, is accompanied by, and perhaps preceded by, an increase in DNase I hypersensitivity in the regions 5' to the active globin genes. This suggests that reorganization of chromatin structure in the globin gene domains is associated with accelerated globin gene transcription and may be characteristic of a developmental transition during terminal differentiation in the erythroid cell lineage.

Induction of transformed cells to terminal differentiation.

HMBA induces MEL cells to terminal erythroid differentiation. HMBA causes a decrease in diacylglycerol concentration, a decrease in Ca+2 and phospholipid-dependent protein kinase C activity (within 2 hr). There is an early (within 1-2 hrs) suppression of c-myb and c-myc gene transcription and an increase in c-fos mRNA (within 4 hrs). During the early or "latent" period there is no detectable commitment of MELC to terminal cell division or expression of differentiated genes such as alpha 1 or beta maj globin genes. HMBA-induced commitment to terminal differentiation is detected by 12 hrs and over 95% become committed cells by 48-60 hrs. Commitment is associated with persistent suppression of c-myb gene transcription and elevated levels of c-fos mRNA, whereas the level of c-myc mRNA returns to that of uninduced cells. By 36-48 hrs, transcription of the alpha 1 and beta maj globin genes increases 10-30 fold, and that of rRNA genes is suppressed. Changes in expression of c-myb, c-myc, c-fos and p53 genes that occur early during HMBA-induced differentiation may be important in the multistep process involved in commitment of MEL cells to terminal differentiation. Continued suppression of c-myb gene expression may be required for terminal differentiation of these cells.

Purification and characterization of a polyhook protein from Caulobacter crescentus.

A polyhook-producing strain of Caulobacter crescentus was isolated, and the polyhook protein was purified. The antigenicity and morphology of the polyhook structure are similar to the wild-type hook except that the mutant strain produces a hook structure at least 10-fold the length of wild-type hooks (1.0 versus 0.1 micrometers). The molecular weight of the polyhook protein, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, is 72,000, and the protein has a pI of approximately 6.1. Antibodies prepared against the polyhook protein were used to show that this protein is antigenically distinct from the Caulobacter flagellins. Amino acid analysis of the polyhook protein revealed compositional similarities to other gram-negative, bacterial hook proteins.

Reconstitution and purification of flagellar filaments from Caulobacter crescentus.

Filaments from isolated flagella of Caulobacter crescentus have been purified by successive dissociation and reconstitution. After the second and third reconstitutions from subunits in 0.8 M sodium citrate, filament preparations contained only two proteins, flagellin A (26,000 daltons) and flagellin B (28,000 daltons). There was some enrichment for flagellin A during reconstitution by this procedure, since isolated flagella contained flagellin A and flagellin B in a ratio of approximately 3.8:1 and filaments after the third reconstitution contained the two proteins in a ratio of 5.0:1.

Physical characterization of the purified CCAAT transcription factor, alpha-CP1.

We have used DNA sequence affinity chromatography previously to purify a murine erythroid cell nuclear factor termed alpha-CP1. This promoter selective transcription factor is a heterotypic CCAAT factor composed of at least seven polypeptides with Mr values that range from 27,000 to 38,000. Peptide mapping experiments reported here show that these seven polypeptides fall into three distinct classes (alpha, beta, and gamma). In addition chemical cross-linking, sedimentation, and gel filtration studies suggest that alpha-CP1 is a heterotrimeric factor composed of one polypeptide from each class. A core component of the factor (alpha beta) is stable at moderately high ionic strengths, whereas the gamma polypeptides are more weakly associated with the particle. The native factor binds tightly to the alpha-globin CCAAT box (Kd = 5.71 x 10(-11 M), and mutational studies show that the DNA recognition site resides in a sequence decamer. DNA binding is significantly stabilized, however, by apparently nonspecific sequences 3' of the CCAAT recognition motif. Finally, the DNA binding domain of purified alpha-CP1 is moderately stable to protease digestion, a feature characteristic of heterotypic CCAAT factors. The proteolyzed factor has a slightly higher affinity for the CCAAT box (Kd = 2.8 x 10(-11) M), and its footprint cannot be distinguished from that of the intact factor. In contrast protease treatment abolishes the ability of alpha-CP1 to activate alpha-globin gene transcription in vitro. These latter results show that the DNA binding domain of alpha-CP1 is readily distinguished from the domains required to mediate activation of gene transcription.

Modulation of gene expression during terminal cell differentiation.

Inducer-mediated differentiation of MELC provides a model to define the cellular and molecular mechanisms involved in terminal cell differentiation (Fig. 2). (Formula: see text) Fig. 2. HMBA-mediated cellular and molecular changes during induced MELC terminal erythroid differentiation. See text for detailed discussion. HMBA-induced differentiation is accompanied by an acceleration of transcription at the alpha 1 and beta maj globin loci. MELC are virus-transformed cells blocked in erythroid lineage development. MELC have acquired a pattern of DNA hypomethylation, nuclease sensitivity, and partially disrupted nucleosome configuration of the chromatin containing the transcriptionally inactive alpha 1 and beta maj globin domains, which distinguish them from the domains of nonerythroid transcribed genes, e.g. immunoglobulin or albumin, or ribosomal RNA genes which are actively transcribed in uninduced cells. HMBA-mediated MELC differentiation is associated with changes in chromatin configuration, characterized by an increased disruption of the nucleosome structure across the structural nucleotide sequences of alpha 1 and beta maj globin genes and the appearance of nuclease hypersensitive sites 5' upstream from the alpha 1 and beta maj genes. The alteration in chromatin structure appears to precede increased transcription of these genes. Inducer-mediated loss of proliferative capacity involves a complex multistep process during which the cells accumulate factor(s), probably mRNA(s), required for the synthesis of proteins which are, in turn, required for expression of the commitment process and activation of globin gene transcription. Characterization of these genes whose expression is modulated in HMBA-induced commitment to terminal cell division is in process.

Regulation of flagellin synthesis in the cell cycle of caulobacter: dependence on DNA replication.

Synthesis of the two filament proteins (flagellin A and flagellin B) of the Caulobacter creascentus flagellum was measured during the cell cycle. Synchronous cells were pulse-labeled with 36S-methionine, and flagellin proteins were isolated from crude extracts by radioimmune precipitation. The results showed that both proteins are maximally induced during the G2 period and that their induction requires de novo transcription. Flagellin A, however, continues to be made in the progeny swarmer cells after flagellin B synthesis has stopped. This discoordination in flagellin A and B synthesis and the relative abundance of the two proteins may result in part from the longer functional half-life of the flagellin A messenger RNA. Analysis of temperature-sensitive DNA chain elongation mutants suggests that the periodicity of flagellin A and B synthesis in the cell cycle is controlled by a late cell cycle event, presumably the completion of chromosome replication.

Murine erythroleukemia cell differentiation: DNase I hypersensitivity and DNA methylation near the globin genes.

The sensitivity to digestion by DNase I of chromatin containing the alpha- and beta(major)-globin genes and the pattern of DNA methylation near these genes was examined during hexamethylenebisacetamide (HMBA)-mediated erythroid differentiation of murine erythroleukemia cells (MELC). In uninduced and induced cells, the chromatin regions containing the alpha- and beta-(major)-globin genes are more sensitive to digestion by DNase I than is the region containing an immunoglobulin gene (Igalpha) not expressed during erythroid differentiation. However, at low concentrations of DNase I, a 6- to 10-fold increase in site-specific cleavages was generated in chromatin regions near both the alpha- and beta(major)-globin genes in cells induced to differentiate by HMBA. The DNase I hypersensitive site near the beta(major)-globin gene maps to a small region near the 5' terminus of the gene. No detectable change in the pattern of DNA methylation around either the alpha- or beta-globin genes was observed during HMBA-mediated erythroid differentiation. Of the potentially methylated sites assayed and mapped near the beta(major)-globin gene, one site is fully methylated, one is partially methylated, and one is unmethylated both in uninduced and induced cells. Many (but not all) sites assayed near the alpha-globin genes are unmethylated in both uninduced and induced cells. These results show that specific alterations of chromatin structure occur during MELC differentiation and suggest that these changes may not involve alterations in the pattern of DNA methylation.

Hexamethylenebisacetamide-resistant murine erythroleukemia cells have altered patterns of inducer-mediated chromatin changes.

We determined inducer-mediated changes in chromatin structure near the globin genes in a variant line of murine erythroleukemia cells (MELC). The variant cell line, R1, was derived from the inducer-sensitive DS19 cell line by selection for inducer-resistance. R1 cells are resistant to induction of erythroid differentiation by hexamethylenebisacetamide (HMBA) whereas the parental line is HMBA-sensitive. Uninduced MELC (both inducer-sensitive DS19 cells and inducer-resistant R1 cells) have DNase I-sensitive sites in chromatin containing the alpha 1- and beta maj-globin genes. These nuclease-sensitive regions are located within the beta maj-globin second intervening sequence (IVS2) and near the alpha 1-globin gene 5' cap site. Culture with HMBA causes changes in chromatin structure in both parental and variant cell lines. In DS19 cells, the DNase I-sensitive site within the beta maj-globin IVS2 becomes more resistant to nuclease cleavage, and a new DNase I-sensitive region develops near the beta maj-globin cap site. In addition, the nuclease-sensitive region adjacent to the cap site of the alpha 1-globin gene increases, and a novel 5' nuclease-sensitive site is also established. In R1 cells, HMBA-mediated changes in chromatin structure are incomplete. The DNase I-sensitive site within the beta maj-globin IVS2 becomes more resistant to nuclease cleavage, but the nuclease sensitivity near the beta maj-globin cap site does not increase to the extent observed in DS19 cells. The pattern of nuclease sensitivity near the alpha 1-globin gene is essentially unchanged after culture of R1 cells with HMBA. Thus, in R1 cells, resistance to HMBA-induced expression of globin genes is associated with failure to detect inducer-mediated changes in chromatin structure 5' to the cap site of the alpha 1- and beta maj-globin genes. These results also suggest that the increased nuclease resistance of a site in the beta maj-globin IVS2 does not depend on the establishment of a DNase I-sensitive region near the beta maj-globin gene cap site.