T to C Substitution at (cid:1) 175 or (cid:1) 173 of the (cid:2) -Globin Promoter Affects GATA-1 and Oct-1 Binding in Vitro Differently but Can Independently Reproduce the Hereditary Persistence of Fetal Hemoglobin Phenotype in Transgenic Mice*

The T to C substitution at position (cid:1) 175 of the (cid:2) -globin gene has been identified in some individuals with non-deletion hereditary persistence of fetal hemoglobin (HPFH). In this study, the HPFH phenotype was reestablished in transgenic mice carrying the (cid:3) (cid:1) LCRA (cid:2) (cid:1) 175 (cid:4)(cid:5)(cid:6)(cid:5) construct, which contained a 3.1-kb (cid:3) (cid:1) LCR cassette linked to a 29-kb fragment from the A (cid:2) - to (cid:5) -globin gene with the natural chromosome arrangement but with the (cid:1) 175 mutation, which provided evidence for this single mutation as the cause of this form of HPFH. The HPFH phenotype was also reproduced in transgenic mice carrying the (cid:3) (cid:1) LCRA (cid:2) (cid:1) 173 (cid:4)(cid:5)(cid:6)(cid:5) construct, in which the (cid:1) , genomic DNA. The footprint assay revealed that the (cid:1) 180 and (cid:1) 170 G residues were protected in cells of mice containing (cid:4) (cid:2) LCRA (cid:1)(cid:5)(cid:3)(cid:2)(cid:3) ( lanes 5 and 6 ), suggesting that proteins bind to this DNA region, in which an Oct-1 binding site partially overlaps with a GATA-1 binding site. However, a significant decrease in the proteins binding to the corresponding DNA region was observed in the cells of mice carrying (cid:4) (cid:2) LCRA (cid:1) (cid:1) 175 (cid:5)(cid:3)(cid:2)(cid:3) or (cid:4) (cid:2) LCRA (cid:1) (cid:1) 173 (cid:5)(cid:3)(cid:2)(cid:3) ( lanes 1–4 ). The protein binding status of the upstream GATA-1 binding site ( A ), where mutations were not involved, could not be detected because only guanine-specific sequencing reactions were performed in this study.

The human fetal globin genes (G␥ and A␥) are mainly expressed in fetal liver. The switch from fetal to adult (␦ and ␤) globin gene expression occurs in the perinatal period concomitantly with the establishment of the bone marrow as the main site of erythropoiesis. In some individuals, however, this switching is incomplete, and ␥-globin expression persists in adult erythrocytes. Such a condition, found in syndromes of hereditary persistence of fetal hemoglobin (HPFH) 1 and (␦␤) 0thalassemias, can result from deletions within the ␤-globin gene cluster (1). However, in some syndromes known as nondeletion HPFH, point mutations in the ␥-globin gene promoter seem to be responsible for the continuance of ␥-globin expression (1). It has been revealed that most of the point mutations lie in or near the trans-acting factor binding motifs within the promoter and that the mutations usually affect the binding of some trans-acting factors in vitro (2).
Studies in transgenic mice carrying either Ϫ117 A␥ HPFH cosmid or Ϫ117 A␥ HPFH yeast artificial chromosome showed that transgenic expression could mimic the developmental expression program observed in HPFH, providing powerful proof that the single point mutation was the cause of the Greek form of HPFH (3,4). Berry et al. (3) speculated that this change of ␥-globin expression during development could be correlated with the loss of GATA-1 binding to the ␥-promoter region around Ϫ117 and deduced that GATA-1 might act as a negative regulator of the ␥-globin gene in normal adults. However, earlier in vitro experiments on a black form of HPFH point mutation, a T to C substitution at position Ϫ175 of the ␥-globin promoter, showed no significant difference in the GATA-1 affinity for the promoter. Nevertheless, this mutation dramatically decreased the affinity of the Oct-1 transcription factor, which binds to an area partially overlapping the GATA-1 binding site (5)(6)(7)(8)(9).
To test whether the Ϫ175 T to C substitution might be the cause of the black form of HPFH, we introduced two ␤-globin minilocus constructs into mice. The first was ЈLCRA␥␤␦␤, which contained a 3.1-kb ЈLCR cassette linked to a 29-kb fragment from the A␥-to ␤-globin gene with the natural chromosome arrangement. The second was ЈLCRA␥ Ϫ175 ␤␦␤, which was similar to ЈLCRA␥␤␦␤ except for a T to C substitution at position Ϫ175 of the A␥ promoter. Normal developmental expression of ␥and ␤-globin genes was observed in transgenic mouse lines with ЈLCRA␥␤␦␤, whereas persistent expression of the ␥-globin gene in adults of transgenic mouse lines with ЈLCRA␥ Ϫ175 ␤␦␤ was noted, suggesting that this single mutation is responsible for the HPFH phenotype.
To further investigate the relationship between GATA-1 or Oct-1 binding and the developmental regulation of the ␥-globin gene, we substituted the Ϫ175 T to C A␥-globin gene with a Ϫ173 T to C A␥-globin gene in the construct. The Ϫ173 T base has been reported to be involved in the conserved GATA-1 binding motif (5)(6)(7)(8)(9). By electrophoretic mobility shift assays (EMSAs), we found that the Ϫ173 mutation decreased GATA-1 binding dramatically but reduced Oct-1 binding only slightly. Despite these differences in transcription factor binding ability, the HPFH phenotype was reestablished in mice carrying ЈLCRA␥ Ϫ173 ␤␦␤. These results suggested that the inability of either the GATA-1 or Oct-1 to bind this promoter region could result in the HPFH phenotype. We then tested the protein binding status within this promoter region in adult erythrocytes of mice carrying normal or mutant ␥-globin genes by chromatin immunoprecipitation (ChIP) and in vivo footprint analysis. The results show that either the Ϫ175 or the Ϫ173 mutation significantly decreased protein binding to both the GATA-1 and Oct-1 binding sites in adult erythrocytes of transgenic mice. Taken together, these results suggest that the GATA-1 and Oct-1 might act together as part of a repressor complex that silences ␥-globin gene expression in normal adult erythrocytes and that abrogation of binding by either disrupted the repressor complex and resulted in HPFH.

EXPERIMENTAL PROCEDURES
Cosmid Constructs-The plasmid pЈLCR, containing a 3.1-kb cassette consisting of the intact core sequences of HS1-4, is a gift from Dr. Q. Li. The cosmid construct LCRA␥␤␦␤ (10) is a gift from Dr. G. Stamatoyannopoulos. The 3.1-kb ЈLCR was released as a HindIII fragment from pЈLCR and inserted into pBA␥ that contained a 3.3-kb HindIII fragment with A␥-globin gene and flanking sequences, generating pЈLCRA␥. A XhoI-NotI fragment, which contained the 3.1-kb ЈLCR, the promoter, and a 5Ј part of the A␥-globin gene, was released from pЈLCRA␥ and replaced the initial XhoI-NotI fragment of cosmid LCRA␥␤␦␤, generating cosmid ЈLCRA␥␤␦␤ (Fig. 1). An engineered T to C substitution at position Ϫ175 was introduced into the 0.7-kb ApaI fragment of plasmid pBA␥ by PCR and then inserted back into the parental plasmid pBA␥, generating plasmid pBA␥ Ϫ175 . The mutation was confirmed by DNA sequencing. Subsequently, the mutated HindIII-XhoI fragment of plasmid pBA␥ Ϫ175 was linked to the 3.1-kb ЈLCR cassette and then inserted as an XhoI-NotI fragment back into the ЈLCRA␥␤␦␤, generating construct ЈLCRA␥ Ϫ175 ␤␦␤ (Fig. 1). Similar strategies were adopted to create construct ЈLCRA␥ Ϫ173 ␤␦␤ (Fig. 1).
DNA Purification for Microinjection-The 35-kb KpnI-NotI DNA fragments, with and without the point mutations, were released from the cosmid constructs ( Fig. 1) and purified by centrifugation through NaCl gradients (11). Gradient solutions were prepared by diluting 25% (w/v) NaCl (S-5886; Sigma) in 10 mM Tris, pH 8.0, 5 mM EDTA. Digested DNA was diluted with water to a total volume of 2 ml and then gently layered on the gradient solution. Tubes (Beckman SW40 Ti) were cen-trifuged at 25,000 rpm (80,000 ϫ g) at 20°C for 18 h. After centrifugation, an 18-gauge needle was inserted into the bottom of the tube, and 9 drops (ϳ500 l) were collected per fraction. Aliquots of 10 l from each fraction were mixed with 10 l of water and DNA loading buffer, and the samples were loaded onto an agarose gel. After electrophoresis, the ethidium bromide-stained gel was viewed, and fractions containing only the desired band were identified and pooled. The pooled fractions were desalted with the aid of NAP-10 columns (Amersham Biosciences) equilibrated with microinjection buffer (0.22-m filtered 10 mM Tris, pH 7.4, 0.1 mM EDTA). The desalted pool was centrifuged for 1 h at 10,000 ϫ g, and all but 100 l of the centrifuged fractions at the bottom were transferred to a clean tube (12). An aliquot was used to determine DNA concentration by UV absorbance with a spectrophotometer. The transgenes were diluted with microinjection buffer to a final concentration of 0.2 ng/l.
Production of Transgenic Mice and DNA Analysis-Purified DNA fragments were injected into the pronuclei of KM fertilized eggs and then transferred to KM pseudopregnant foster mothers (13). Transgenic mice were identified by PCR and Southern blot analysis of tail skin DNA. Transgenic founder mice were bred with nontransgenic KM mates to establish transgenic lines. The structures of the integrated constructs were analyzed by Southern blot hybridization, using the ␥and the ␤-globin gene and the ЈLCR as the probes. Copy numbers were determined by comparing the intensity of hybridization bands from the transgenic lines with that of human genomic DNA by PhosphorImager analysis.
Globin Gene Expression Analysis-For each mouse line, transgenic males were mated with nontransgenic KM females to generate timed pregnancies. The morning when the mating plug was observed was designated day 0.5. RNA samples were prepared from the yolk sac of mouse embryos on day 11.5; from fetal livers on days 13.5, 16.5, and 18.5; and from the blood of newborn pups as well as adult animals (9). Total RNA was isolated with TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Human and murine globin RNAs were analyzed by quantitative RNase protection assays (14). Four plasmids containing the pT7 promoter and a fragment of each of the four globin genes were used as probe templates: 1) pT7A␥, made by cloning a NcoI-PvuII fragment of the human A␥-globin gene into pBluescript and linearized with BstEII, giving a protected fragment of 170 nucleotides derived from exon 2 of the human A␥ mRNA; 2) pT7␤, made by inserting a BamHI-BalI fragment of the human ␤-globin in pGEM and linearized with BsaI, giving a protected fragment of 205 nucleotides derived from exon 2 of the human ␤ mRNA; 3) pT7M␣, made by inserting a PstI-BamHI fragment, corresponding to the 5Ј end of mouse ␣-globin gene, into pBSks and linearized with HindIII, giving a protected fragment of 128 nucleotides derived from exon 1 of the mouse ␣ mRNA; and 4) pT7M, made by cloning a mouse gene AvaII-XbaI fragment into pBluescript and linearized with EcoRI, giving a protected fragment of 151 nucleotides. The RNA probes labeled with [␣-32 P]UTP were obtained by in vitro transcription of the plasmids with pT7 polymerase. 10 g of total RNA was hybridized overnight at 47°C with 1 ϫ 10 6 cpm of each probe, followed by digestion with RNase A and T1 for 30 min at 37°C using the ribonuclease protection assay kit (Roche Applied Science). The protected fragments were separated on 5% FIG. 1. The constructs used for microinjection. The original cosmid construct LCRA␥␤␦␤ (10), in which a 2.6-kb LCR cassette linked to a 29-kb fragment from A␥-to ␤-globin gene was inserted into cosmid pHC79, was used to produce the other cosmid constructs. The 2.6-kb LCR cassette was replaced by the 3.1-kb ЈLCR cassette to produce the cosmid construct ЈLCRA␥␤␦␤ as described in the text. The wild-type A␥-globin gene in ЈLCRA␥␤␦␤ was replaced by either an A␥-globin gene with the Ϫ175 T3 C or an A␥-globin gene with the Ϫ173 T3 C to produce the cosmids ЈLCRA␥ Ϫ175 ␤␦␤ and ЈLCRA␥ Ϫ173 ␤␦␤, respectively, as described in the text. Only the inserted fragments (not the vector in the constructs) are shown. The 32-kb NotI/ KpnI fragments were recovered from the cosmid constructs and purified for microinjection.
acrylamide/8 M urea gel and autoradiographed. The intensities of protected fragments were quantitated using the Amersham Biosciences PhosphorImager.
EMSA-Human erythroleukemia K562 cells, mouse erythroleukemia (MEL) cells (MELGM979 and MEL585), and human cervical carcinoma cells (HeLa) were cultured as described previously (5). In some experiments, K562 cells were induced to differentiate by the addition of 25 M hemin to the medium for 4 days before nuclear isolation, and MEL cells were induced by the addition of 2% Me 2 SO to the medium for 2 days before nuclear isolation. The nuclear extracts were prepared, and EMSAs were carried out as described previously (5). The 44-bp ApaI-AvaII fragments (Ϫ201 to Ϫ158) of the ␥-globin promoter with the normal sequence or the Ϫ175 or Ϫ173 mutation was used as probe. The probes were labeled with Klenow polymerase and [␣-32 P]-dCTP (normal probe, 5Ј-GCTTCCCCACACTATCTCAATGCAAATATC-TGTCTGAAACGGTC-3Ј; Ϫ173 mutation probe, 5Ј-GCTTCCCCACAC-TATCTCAATGCAAATACCTGTCTGAAACGGTC-3Ј; and Ϫ175 mutation probe, 5Ј-GCTTCCCCACACTATCTCAATGCAAACATCTGTCTG-AAACGGTC-3Ј). The EMSAs were carried out as described previously (5). Binding reaction mixtures (20 l) contained 20 mM HEPES, 60 mM KCl, 1 mM MgCl 2 , 0.1 mM EDTA, 0.5 mM dithiothreitol, and 12% glycerol. 2 g of nonspecific competitor DNA poly(dI-dC) was added to reduce nonspecific binding. 10 g of nuclear extracts was added, and the reaction was incubated at 30°C for 5 min before the addition of labeled probe (1 ϫ 10 4 cpm, 1 ng). Complete binding reactions were incubated at 30°C for 15 min. The sequences of competitors for GATA-1 and Oct-1 binding were 5Ј-CGAGGCAAGAGATATATCTTAGAGGGA-GT-3Ј and 5Ј-TGCTCATGAATATGCAAATCCTGTGTGTCT-3Ј respectively. Electrophoresis was carried out at 25°C in 4% polyacrylamide gels, using 0.5ϫ Tris-borate EDTA buffer. The gels were then dried and analyzed by autoradiography. To assess quantitative differences in binding affinity, the intensities of retarded bands were determined by an LKB (Amersham Biosciences) UltroScan XL laser densitometer equipped with GelScan 2.1 software.
ChIP and Ligation-mediated PCR (LM-PCR)-ChIP and LM-PCR were performed as described by Kang et al. (15), with the following modifications. Bone marrow cells (5 ϫ 10 6 ) of the transgenic mice were cross-linked by adding formaldehyde to a final concentration of 1% and incubated at room temperature for 10 min. Cells were washed with phosphate-buffered saline and resuspended in ChIP lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.0) with protease inhibitors. After incubation on ice for 10 min, the cells were sonicated to shear DNA to lengths between 200 and 1000 bp on ice. The lysate was transferred to a 15-ml conical tube and diluted with dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH 8.1, 16.7 mM NaCl, and protease inhibitors). An aliquot of 500 l of protein A-Sepharose beads was added to the diluted nuclear lysate and incubated at 4°C for 2 h while rotating. The beads were pelletized at 2000 rpm for 10 min. The supernatant was divided into three aliquots of 25 l. Anti-GATA-1 and anti-Oct-1 antibodies (Santa Cruz Biotechnology) were added to two aliquots, respectively, whereas no antibody was added to the third aliquot, and then the aliquots were incubated at 4°C overnight while rotating. Dimethylsulfate (DMS) treatment of the immunoprecipitated chromatin fragments was performed using the Maxam and Gilbert guanine-specific sequencing reaction with 0.1% DMS for 45 s at room temperature. An aliquot of the precipitated DNA was also analyzed by PCR with primers specific for the human ␥-globin promoter region (forward primer, 5Ј-TGGAATGACTGAATCGGAAC-3Ј; reverse primer, 5Ј-GTGTGTGGAACTGCTGAAG-3Ј). Approximately 2 g of DMS-treated DNA was annealed to 0.6 pmol of human ␥-globin promoter specific primer 1 (5Ј-AACCTCAGACGTTCCAGAA-3Ј) by denaturing at 96°C for 10 min, followed by annealing at 47°C for 30 min. Primer extension was performed at 53°C for 1 min, 55°C for 1 min, 57°C for 1 min, 60°C for 1 min, 62°C for 1 min, 66°C for 1 min, 68°C for 1 min, and 72°C for 5 min. Ligation of the blunt end of the extension product with the linker pair (AЈ and B) was performed at 16°C overnight. After standard phenol/chloroform extractions and ethanol precipitation, the ligated DNA was resuspended in 30 l of H 2 O. 18 amplification cycles were performed using the linker primer AЈ and primer 2 (5Ј-GTGTGTGGAACTGCTGAAG-3Ј). The first cycle had a 3-min denaturing step at 95°C, and the remaining denaturing steps were 1 min. All cycles had annealing steps at 55°C for 2 min. The extension time began with 3 min and was increased by 5 s with each cycle at 72°C. The 5Ј end of a primer 3 (5Ј-GCTTCCTTTTATTCTTCATCCC-3Ј) was labeled with [␥-32 P]ATP. A single-strand amplification of the PCR product was performed for 2 cycles using the labeled primer 3. The first cycle was initiated with denaturation at 95°C for 3 min, and the second cycle was initiated with denaturation for 1 min. Annealing steps were carried out at 55°C for 2 min, and the extension steps were carried out at 72°C for 10 min. The products of the single-strand amplification were stored at 4°C before running sequencing gel.

Normal Developmental Expression of ␥-Globin Gene Was
Observed in Transgenic Mice with ЈLCRA␥␤␦␤-Two transgenic lines with intact integration of the construct ЈLCRA␥␤␦␤ were established. The levels of human A␥-and ␤-globin transcripts as well as endogenous mouse ␣and -globin transcripts were measured by RNase protection analysis ( Fig. 2; Table I).
In line N1, the highest A␥-globin mRNA detected was ϳ17% of mouse ␣plus -globin mRNA (per transgene copy) on day 13.5, declining significantly to 0.9% in newborn pups and to 0.4% in adult animals. This result suggests active expression of the A␥globin gene during the fetal stages and subsequent silence after birth. The human ␤-globin gene was undetectable at the embryonic stage but became active in the early fetal liver, rising rapidly to maximum levels and remaining there in the adult stage. The pattern of developmental expression of human ␥and ␤-globin genes in line N5 was very similar to that in line N1. These results indicate that transgenic mice carrying ЈLCRA␥␤␦␤ can mimic the normal developmental expression programs of both the human A␥-and ␤-globin genes. This finding is similar to that from a previous study on transgenic mice carrying LCRA␥␤␦␤, a construct that was the same as ЈLCRA␥␤␦␤ except for the use of a 2.6-kb LCR cassette (10).
The HPFH Phenotype Was Reproduced in Transgenic Mice Carrying ЈLCRA␥ Ϫ175 ␤␦␤-Four transgenic lines with intact integration of the construct ЈLCRA␥ Ϫ175 ␤␦␤ were established. The expression of the human and murine globin mRNAs during mouse development is shown in Fig. 3 and Table II.
The human ␥/mouse ␣ ϩ mRNA/copy ranged from 8% to15% in adults, only a moderate decline when compared with the highest A␥-globin mRNA expression, which ranged from 18% to 29% in the fetal stage. That is, A␥-globin gene expression in the adult stage was between 28% and 71% of the maximal fetal-stage expression. Although transgene expression varied between lines, the 29 -72% reduction of ␥-globin gene transcription in adult stage as compared with that in the fetal stage for mice carrying ЈLCRA␥ Ϫ175 ␤␦␤ differed significantly from the 97.7-98.2% reduction occurring in mice carrying ЈLCRA␥␤␦␤. These results demonstrated that the A␥globin gene with the Ϫ175 mutation still remained active Ϫ173 or Ϫ175 Mutation of ␥-Globin Gene and HPFH Phenotype during the adult stage and that the point mutation was responsible for the HPFH phenotype.
The Ϫ175 and Ϫ173 Mutation Affected GATA-1 and Oct-1 Binding Differently in Vitro-The region from Ϫ170 to Ϫ190 of the A␥-globin gene promoter contains two GATA-1 binding motifs and one Oct-1 binding motif (Fig. 4A). EMSAs of the normal ApaI-AvaII fragment (Ϫ201 to Ϫ158) of the ␥-globin gene promoter and its Ϫ173 mutant counterpart were performed using nuclear extracts from uninduced and hemin-induced erythroid K562 cells, uninduced and Me 2 SO-induced erythroid MEL cells, as well as nonerythroid HeLa cells. Two MEL cell lines, MELGM979 (which expresses both adult and embryonic globins) and MEL585 (which expresses adult globin only), were used for nuclear extract preparation. Very similar EMSA profiles were observed between the two lines (data not shown). Two common mobility shifted bands (A and B) were produced when the normal fragment reacted with all the nuclear extracts, with the exception of the HeLa cell extract. (Fig.  4B). Compared with the normal probe, the intensity of band A, produced by the Ϫ173 mutation probe, decreased by 19 Ϯ 10%, and the intensity of band B, produced by the same mutation probe, decreased by 85 Ϯ 8%. The competition assay demonstrated that band A resulted from Oct-1 binding and that band B resulted from GATA-1 binding (Fig. 4C). This result indicated that the Ϫ173 mutation dramatically decreased the binding affinity of GATA-1 to the mutant promoter but only slightly reduced Oct-1 binding.
EMSAs of DNA fragments with the Ϫ175 mutation have been carried out exhaustively in previous studies (5-9). Our results here showed that when compared with the normal fragment, the binding affinity of GATA-1 to the Ϫ175 mutation promoter decreased by 5 Ϯ 2%, whereas the binding affinity of Oct-1 decreased by 93 Ϯ 6% (Fig. 4D). This is consistent with the results of most of the previous studies, i.e. the Ϫ175 mutation decreased GATA-1 binding slightly but reduced Oct-1 binding greatly, which was just the opposite of the effects of the Ϫ173 mutation.
The HPFH Phenotype Was Also Reproduced in Transgenic Mice Carrying ЈLCRA␥ Ϫ173 ␤␦␤-To further investigate whether Ϫ173 T to C substitution of the A␥ gene could reestablish the HPFH phenotype in vivo, we produced the construct ЈLCRA␥ Ϫ173 ␤␦␤. Three transgenic lines carrying the construct were established, and developmental expression of the globin genes was analyzed ( Fig. 5; Table III). The human ␥/mouse ␣ ϩ mRNA/copy ranged between 5% and 14% in   Ϫ173 or Ϫ175 Mutation of ␥-Globin Gene and HPFH Phenotype adults, whereas the highest A␥-globin mRNA expression in the fetal stage of the transgenic animals was between 17% and 25%. The A␥-globin gene expression in the adult stage was between 25% and 65% of the maximal fetal-stage expression level. Despite expression variation between lines, the results showed that the A␥-globin gene with the Ϫ173 mutation still remained active in the adult stage of mouse development. The Ϫ173 mutation resulted in the HPFH phenotype in transgenic mice, similar to the manifestation of the A␥-globin gene bearing the Ϫ175 mutation (Fig. 6).
Both the Ϫ175 and Ϫ173 Mutation Significantly Decreased Protein Binding to the A␥ Promoter Region in Vivo-Bone marrow cells were collected from a transgenic mouse line carrying ЈLCRA␥␤␦␤ (line N1), a transgenic line carrying ЈLCRA␥ Ϫ175 ␤␦␤ (line 518), and a transgenic line carrying ЈLCRA␥ Ϫ173 ␤␦␤ (line 318), respectively. After obtaining sheared chromosome fragments from the cells, the specific fragments were precipitated with the anti-GATA-1 or anti-Oct-1 antibodies. From an aliquot of the precipitated fragments, the ␥-promoter region was amplified using specific primers. Amplification fragments were explored in all of the samples precipitated with the anti-GATA-1 and anti-Oct-1 antibodies, but not in the samples without antibody precipitation (data not FIG. 4. Binding of GATA-1 and Oct-1 to normal/mutant ␥-globin promoter region. A, sequence of the Ϫ158 to Ϫ201 region in human A␥-globin promoter. The recognition sequences of GATA-1 are boxed, and the Oct-1 motif is underlined. B, EMSAs of the normal (N) 44-bp ApaI-AvaII fragment and its Ϫ173 T to C counterpart (173), using nuclear extracts (10 g) from erythroid K562, hemin-induced K562(ϩ), MEL, Me 2 SO-induced MEL(ϩ), and nonerythroid HeLA cells. Similar results were observed when using extracts from K562, K562(ϩ), MEL, and MEL(ϩ) cells, but not from HeLa cells. The positions of bands A and B are indicated at the left. The marked reduction of band B and slightly weakened band A were seen when the Ϫ173 mutant fragment was used as probe. C, competition assays. The nuclear extract was obtained from Me 2 SO-induced MEL cells. The left lane shows the binding of 10 g of nuclear extract to 1 ng of the normal fragment without specific competitors. Lanes aϪc, with 25, 100, or 250 ng of competitor of Oct-1 binding or GATA-1 binding, respectively. Competition assay demonstrated that mobility-shifted band A resulted from Oct-1 and that mobility-shifted band B resulted from GATA-1 binding. D, EMSAs of the normal ApaI-AvaII fragment and its Ϫ175 T to C counterpart (175) as well as its Ϫ173T to C counterpart (173). A slightly decreased band A intensity and a greatly reduced band B intensity were observed when the Ϫ173 mutation fragment was used as probe, whereas a greatly decreased band A intensity and a similar band B intensity were noticed when the Ϫ175 mutation fragment was used as probe.  Ϫ173 or Ϫ175 Mutation of ␥-Globin Gene and HPFH Phenotype shown). Previous in vitro experiments have identified a GATA-1 binding site and an Oct-1 binding site within the Ϫ233 region, a GATA-1 site within the Ϫ117 region, and two GATA-1 sites and an Oct-1 site within the Ϫ175 region of ␥-promoter ( Fig. 7A) (16). The ability to obtain specific chromatin fragments was probably due to the combination of the antibodies with the corresponding proteins bound to the alternate binding sites in addition to the binding sites within the Ϫ175 region. DMS treatment of the immunoprecipitated chromatin fragments and linker-LM-PCR revealed the footprints of protein binding (Fig. 7B). The results showed that both of the G residues at the Ϫ180 and Ϫ170 positions of ␥-globin promoter were protected in adult bone marrow cells derived from transgenic mice carrying the ЈLCRA␥ Ϫ ␤␦␤ construct, suggesting that proteins bind to this promoter region of the chromatin. Although these in vivo footprinting experiments could not confirm which proteins bound to this DNA region of the chromatin in adult erythrocytes of the transgenic mice with ЈLCRA␥ Ϫ ␤␦␤, the validation of GATA-1 and Oct-1 binding to this region by previous in vitro experiments suggested that the binding proteins include GATA-1 and Oct-1. The footprints also revealed loss or significant reduction of protection on the Ϫ180 and Ϫ170 G residues in cells derived from transgenic mice carrying either ЈLCRA␥ Ϫ175 ␤␦␤ or ЈLCRA␥ Ϫ173 ␤␦␤, suggesting that both the Ϫ175 mutation and the Ϫ173 mutation resulted in significantly decreased protein binding to the corresponding region of the chromatin. DISCUSSION Several in vitro experiments have been performed to explore the mechanisms of Ϫ175 T3 C HPFH. Two consensus recognition motifs for the DNA-binding protein GATA-1 are known to be near the Ϫ175 site within the ␥-globin gene promoter. An Oct-1 binding motif is found between the GATA-1 sites and overlaps the downstream GATA-1 site. The Ϫ175 HPFH mutation modifies the downstream GATA-1 site but also alters the last base in the Oct-1 site (Fig. 4A). Whereas there is general agreement that the Ϫ175 T3 C mutation significantly reduces Oct-1 binding, the affinity of GATA-1 for the Ϫ175 region has been reported to be unaffected by the mutation or even slightly increased by it (5,7,17). Detailed mutational analysis suggested that increased activity of the Ϫ175 mutation promoter was not due to decreased Oct-1 binding but was dependent on GATA-1 binding (7,8). O'Neill et al. (18) reported that their data were consistent with a competition model between Oct-1 and GATA-1 for binding in this region, whereas Martin et al. (7) suggested that the Ϫ175 mutation altered the mechanism in which GATA-1 binds to the mutant promoter rather than the affinity of GATA-1 for the promoter. HMG-1 (high-mobility group-1), a ubiquitous and abundant nuclear DNA-binding protein, has also been demonstrated to bind strongly to one of a pair of downstream GATA-1 motifs; the Ϫ175 mutation significantly decreased HMG-1 binding (19), but the physiological effect of this decline is unclear.
In most transient expression assays, only a 3-to 4-fold increased expression of the Ϫ175 mutant ␥-globin gene above that of the wild-type gene was observed in K562 cells, far less than the Ͼ100-fold increased expression of the mutant gene in vivo (20 -22). Additionally, several putative HPFH mutations such as Ϫ117, Ϫ196, and Ϫ202 mutation, did not result in striking effects on promoter activity in transient K562 and MEL cell assays (7,(23)(24)(25). These findings demonstrated that in vitro experiments could not accurately reflect the in vivo conditions.
Because the Ϫ117 mutation reproduced the HPFH phenotype in transgenic mice (3,4) and the Ϫ175 mutation affected GATA-1 binding differently from the Ϫ117 mutation, it is important to test whether the Ϫ175 mutation of ␥-globin promoter also reproduces the HPFH phenotype. In a recent study, transgenic mouse lines carrying the LCR-382A␥ construct with or without the Ϫ175 mutation were produced (26). Although a high level of ␥-globin expression was observed in the adult animals carrying LCR-382A␥ with the Ϫ175 mutation, only a 1.3-fold higher level of ␥-globin expression compared with the LCR-382A␥ transgene was observed because the LCR-382A␥ adult transgenic mice expressed high levels of ␥-globin themselves. The loss of developmental control of ␥-globin gene expression in LCR-382A␥ was probably due to the removal of a possible silencer sequence between Ϫ382 and Ϫ730 of the A␥ promoter (27), the lack of the region downstream of the A␥ gene suspected to have a role in A␥ gene silencing (28), or from the absence of ␤-globin gene competition for interaction with the LCR (10). To exclude these possibilities, we produced constructs containing a 3.1-kb LCR cassette that includes the intact core sequences of the four HS sites, flanked by a 29-kb fragment from the A␥-to ␤-globin gene with normal chromosome organization. A similar cosmid construct, containing a 2.6-kb LCR cassette flanked by the 29-kb fragment, has been demonstrated to confer correct developmental control of the globin genes in transgenic mice (10). In our study, human and endogenous globin mRNAs of mice carrying one or two copies of the transgenes were analyzed. The ␥-globin expression displayed normal developmental control in mice carrying ЈLCRA␥␤␦␤, whereas continuance of ␥-globin expression was observed in adult mice carrying ЈLCRA␥ Ϫ175 ␤␦␤. These results provided proof that the Ϫ175 T to C substitution caused this form of non-deletion HPFH.
Based on the observation that the Ϫ117 mutation reproduced the HPFH phenotype in transgenic mice as well as dramatically decreasing GATA-1 binding, Berry et al. (3) proposed GATA-1 might act as a negative regulator of the ␥-globin gene in normal adults (3). However, an increased binding of CP1 and CDP and a decreased binding of NF-E3 in the CCAAT region have also been observed (15,22,29,30), making it difficult to evaluate the role of GATA-1 in the developmental regulation of ␥-globin gene expression. Establishment of the HPFH phenotype in transgenic mice by the Ϫ175 mutation at least suggests that the decreased GATA-1 binding is not necessary for active expression of ␥-globin in adults in this form of HPFH. However, it cannot be excluded that GATA-1 binding plays various roles in different regions of the promoter.
To further investigate the relationship between transcrip-  (Tables I-III). The results from three lines carrying ЈLCRA␥ Ϫ173 ␤␦␤ are shown at the left, those from four lines carrying ЈLCRA␥ Ϫ175 ␤␦␤ are shown in the middle, and those from two lines carrying ЈLCRA␥␤␦␤ are shown at the right.
tional factor binding and developmental regulation of ␥-globin expression, we introduced a mutation at position Ϫ173. Contrary to the Ϫ175 mutation, the Ϫ173 mutation slightly decreased Oct-1 binding but greatly reduced GATA-1 binding in vitro. Nevertheless, when we introduced the ЈLCRA␥ Ϫ173 ␤␦␤ construct into mice, the HPFH phenotype was also reproduced.
Because the Ϫ175 mutation greatly decreased Oct-1 affinity for the promoter region, it is reasonable to speculate that this change is responsible for overexpression of the mutant ␥-gene in adults. In fact, Oct-1 has been found to possess an intrinsic silencing activity in its alanine-rich C-terminal domain (31). The role of Oct-1 as a transcriptional repressor has been well studied in promoters of the thyrotropin ␤ subunit gene, pituitary-specific transcription factor pit1/ghf1 gene, von Willebrand factor gene, and B-cell-specific B29 (Ig␤) gene (31)(32)(33)(34). Recently, a detailed study of the 5Ј-flanking region of human gonadotropin-releasing hormone receptor gene revealed that the POU domain transcription factor, Oct-1, played an important role in the transcriptional repression of the gonadotropinreleasing hormone receptor gene via binding to an octamer sequence (35). Whether Oct-1 also acts as a transcriptional repressor for ␥-globin regulation in adults still remains unclear. However, the fact that the Ϫ173 mutation that only slightly decreased the Oct-1 affinity for the promoter region resulted in the HPFH phenotype in transgenic mice reduced the possibility that the change in Oct-1 binding directly increased ␥-globin expression of the Ϫ175 HPFH.
Because both the Ϫ175 and the Ϫ173 mutation, which affected binding of GATA-1 and Oct-1 to the promoter region in vitro differently, resulted in HPFH phenotypes, we believe that the continued expression of ␥-globin gene in adults is not simply due to the decreased affinity of GATA-1 or Oct-1 alone to the promoter region. Therefore, it is reasonable to assume that Oct-1 and GATA-1 binding are both necessary for silencing of the ␥-globin gene in adult erythrocytes.
Finally, we analyzed the protein binding status within this promoter region in adult erythrocytes of transgenic mice carrying normal or mutant ␥-globin genes. After the sheared chromosome fragments were precipitated with anti-GATA-1 or anti-Oct-1 antibodies and subjected to PCR amplification using primers specific for the ␥-promoter region, the amplification fragments were explored in all samples derived from bone marrow cells of transgenic mice. This means that the ChIP assay alone could not identify changes of GATA-1 and Oct-1 binding to the Ϫ175 region of the ␥ promoter. Because previous in vitro experiments have identified a GATA-1 binding site and The footprint assay revealed that the Ϫ180 and Ϫ170 G residues were protected in cells of mice containing ЈLCRA␥␤␦␤ (lanes 5 and 6), suggesting that proteins bind to this DNA region, in which an Oct-1 binding site partially overlaps with a GATA-1 binding site. However, a significant decrease in the proteins binding to the corresponding DNA region was observed in the cells of mice carrying ЈLCRA␥ Ϫ175 ␤␦␤ or ЈLCRA␥ Ϫ173 ␤␦␤ (lanes 1-4). The protein binding status of the upstream GATA-1 binding site (A), where mutations were not involved, could not be detected because only guanine-specific sequencing reactions were performed in this study.
an Oct-1 site within the Ϫ233 region and a GATA-1 site within the Ϫ117 region in addition to the two GATA-1 sites and the Oct-1 site within the Ϫ175 region of the ␥-promoter (16), the obtainment of specific chromatin fragments was probably due to the combination of the antibodies with the corresponding proteins bound to the Ϫ175 region and/or the alternative sites. Subsequently, a DMS footprint assay was performed. The result showed that both the G residues at Ϫ180 and Ϫ170 positions of the ␥-globin promoter were protected in adult bone marrow cells derived from mice carrying ЈLCRA␥␤␦␤ construct, suggesting that proteins bind to this promoter region of the chromatin. Whereas in vitro experiments demonstrated that the Ϫ175 mutation and the Ϫ173 mutation affected Oct-1 and GATA-1 binding differently, our in vivo footprint analysis revealed that both Ϫ175 and Ϫ173 mutation significantly decreased overall protein binding to the corresponding region of chromatin. Although the in vivo footprint experiment could not confirm which proteins bound to this DNA region of the chromatin in adult erythrocytes of mice with ЈLCRA␥␤␦␤, we could assume that GATA-1 and Oct-1 were included because the previous in vitro experiments proved binding of the two transcription factors to this region. Perhaps a multi-protein complex containing GATA-1, Oct-1, and other protein factors binds to this promoter region and contributes to the formation of a repressive chromatin structure, silencing ␥-globin gene expression in adult erythrocytes. Both the Ϫ175 and Ϫ173 mutations may block the formation of such a complex, thereby resulting in the continued expression of the ␥-globin gene in adult erythrocytes. Additional experiments are needed to prove and determine the presence and composition of this multiprotein complex in adult erythrocytes.