The Human ▀-Globin Locus Control Region *
F. Grosveld, M. Antoniou, M. Berry, Ernie de Boer, N. Dillon, J. Ellis, P. Fraser,D. Greaves, 0. Hanscombe, J. Hurst,
M. Lindenbaum, v. Mignotte, S. Philipsen, S. Pruzina, J. Strouboulis, D. Talbot, and D. Whyatt 1    
Hńmatol. Bluttransf. Vol 35

* This work was supported by the MRC (UK). 1 Laboratory of Gene Structure and Ex pression,
National Institute for Medical Research, The Ridgeway, Mill Hill, London NW7 lAA, UK.


The human ▀-globin gene cluster spans a region of70 kilobases (kb) containing five developmentally regulated genes in the order 5'-epsilon,YGY A ,delta,▀-3' (Fig. 1 ). The haematopoietic tissue in the early stages of human development is the embryonic yolk sac and the epsilon-globin gene is expressed. This is switched to the y-globin genes in the foetal liver and the delta- and ▀-globin genes in adult bone marrow (Fig. 2; for review, see [9]). High levels of these genes are expressed in circulating red blood cells (RBC), giving rise to 90 % of the total soluble protein. RBC are derived from a pluripotent stem cell which can differentiate along alternate pathways to erythrocytes, platelets, granulocytes, macrophages and lymphocytes. During the transition to erythroblasts which have lost the capacity to proliferate, the ▀-globin genes become transcriptionally activated achieving messenger RNA (mRNA) levels of more than 25000 copies per cell. A large number of structural defects have been documented in the ▀-globin gene locus (for review, see [9,44]). These defects are responsible for a heterogeneous group of genetic diseases collectively known as the ▀-thalassaemias, which are classified into ▀, delta ▀, y delta ▀, etc. thalassaemia subgroups according to the type of gene affected. In a related condition, hereditary persistence of foetal haemoglobin (HPFH), Y-globin gene expression and hence HbF (fetal haemoglobin) production persist into adult life. These clinically important diseases provide natural models for the study of the regulation of globin gene regulation during development. Most interesting in terms of transcription are the promoter mutations and deletions. The delta ▀ thalassaemias and a number of the HPFHs are associated with an elevated expression of the y-genes in adult life as a result of deletions of varying size. Analysis of these deletions has suggested that they act over considerable distances, to influence differential gene expression within the human ▀-globin domain.

The Locus Control Region

The existance of a region that activates the entire ▀-globin gene cluster first became apparent from the study of a heterozygous Y-▀-thalassaemia (Fig. 1 ) [31]. This patient contained one deletion allele which lacked lOO kb, eliminating the entire upstream region but not the ▀-globin gene [57], which was shown to be completely normal [31, 64]. The other allele was expressed in the patient and it was therefore not a lack of trans-acting factors which silenced the mutant chromosome but an important control region had be missing. A set of developmentally stable, hypersensitive sites, 5' HS 1, 2, 3 and 4, were shown to be present upstream in the deleted region, and these

Fig. 1. Schematic representation of the ▀-globin locus. Boxes indicate the different genes which are all transcribed from left to right. The vertical arrows indicate DNase hypersensitive sites. Thefour arrows mark the LCR containing 5' HS 1, 2, 3 and 4 upstream of the epsilon-gene.


Fig. 2. Schematic representation of the developmental expression patterns of globin genes in human and mouse. The curves refer to the ▀-like genes only

were potential candidates for such a locus control region (LCR; Fig. 1) [17,27 ,60]. Linkage of this region to a cloned ▀-globin gene resulted in erythroid-specific high-level expression of the gene in transgenic mice and in tissue culture cells. This expression is dependent on the copy number of the transgene and independent of the integration site in the host genome [3, 27], a phenomenon which had not previously been observed in transgene expression. This posed two questions: How is independence of the site of integration achieved? And how is the LCR involved in globin gene switching? Position independence and copy number dependence can theoretically be explained by at least two independent mechanisms; either positive activation by the LCR is always achieved, overriding position effects that could be present, or (and) the region contains a locus border element(s) (LEE) that insulate it from neighbouring regions. Matrix attachment sites (MAR) [22, 30] or " A " elements [4, 53] could be LEEs, and we initially speculated these were part of the LCR in addition to activating sequences [27]. However, preliminary experiments indicate that this is not the case and that such a border may be located further upstream. The latter is based on the fact that the DN aseI sensitivity of chromatin is strongly decreased in the sequences 25 30 kb 5' to the LCR (Fig.1) [19, 31, 57]. At least 150 kb of chromatin in the 3' direction is sensitive under the control of the LCR [19], suggesting that such sequences are not present for a considerable distance 3' (Fig.1). The position independence we observe is therefore due to a dominant activation of transcription by the LCR, perhaps by creating very stable interactions between the LCR and the genes. Consequently, positive position effects would only be present in the background and only become apparent in situations where the linked gene is suppressed [12] (see below for discussion). Position effects are not observed at low levels of expression when part of the LCR or mutations in the LCR are used. This indicates that the interaction between the LCR and the promoter is dominant except when the promoter is suppressed [10, 12, 18, 20, 49, 54]. In agreement with the deletion observed in a Hispanic y ▀-thalassaemia (Fig. 1) [13], the main activity of the LCR is associated with HS2, 3 and 4 [10, 18, 20, 54, 61], which can each activate a linked transgene, independent of the site of integration. A number of protein binding sites have been mapped to these fragments, in particular, sites for two erythroid-specific factors and several ubiquitously expressed factors. A number of the binding sites are present in all three active sites (Fig. 3) [43, 45, 55]. One of the shared factors is the erythroid/megakaryocyte-specific factor GATA-1 [38,48], which has been shown to be essential for erythroid development [42]. Deletion of GA T A-1 binding sites prevents erythroid-specific induction of the ▀-glo bin gene [ 11 ], and the protein has been shown to have transcriptional activation properties [38]. However, the presence of GATA-1 binding sites per se is insufficient to give position-independent expression, e.g. the flanking regions of the human ▀-globin gene contain at least six GATA-1 binding sites [11,63], but do not confer integration site-independent expression [32,34,58]. However, all three active 5' HS contain two closely spaced GA T A-1 sites in opposite orientations. This arrangement is also observed in the chicken ▀-globin enhancer, which appears to provide position-independent expression [47]. Possibly an inverted double GATA-1 site is a key component in erythroid -specific, posi tion -independent activation and GA T A-1 can interact with itself or one of the other GA T A proteins to achieve this [ 66] .Classical enhancer activity is only associated with 5'HS2 [41, 61], and not with the others. Dissection of the HS 2 showed that a number of proteins are bound to the core fragment (Fig. 3) [55]. Attention has been focused on a double consensus sequence for the jun/fos family of DNA binding proteins which appeared crucial for HS 2 activity [41, 52, 55]. Several pieces of evidence

Fig.3. Summary of factor binding sites to the minimal fragment of 5'HS2, 3 and 4, which provide position-independent expression in transgenic mice. Individual factors are described in the text. Black boxes indicate erythroid-specific factors; open boxes indicate ubiquitous factors. GT indicates a GT -rich motif [43]

show that the functional activator which interacts with the jun/fos binding site is NF-E2, originally described as interacting with another erythroid-specific gene, that for porphobilinogen deaminase [39, 40]. Two NF-E2 molecules and at least one other protein binding at two nonequivalent sites are involved [56]. However, the presence of this double NF-E2 sequence alone is insufficient to provide high levels of expression [55] and when the jun/fos binding site is removed from the 300 bp core fragment, HS2 retains the ability to activate a linked ▀-globin gene in a copy number dependent fashion, albeit at low levels (Fig. 4) [56]. We therefore conclude that the 5'HS2 NF-E2 region has strong enhancer activity but that it is not necessarily required to obtain position-independent globin gene activation. All the other factors which have been shown to interact with LCR sequences, including the factors H-BP and J-BP, are ubiquitous proteins [56]. This suggests that a combination of erythroid-specific and ubiquitous factors may be required to render the ▀-globin gene independent of its site of integration. The (abundant) ubiquitous factors shared by the three HS of the LCR which have been studied to date are Sp 1 and TEF -2 [23, 65], but a simple multimerized combination of a GATA-l and a Spl/TEF-2 binding site is not functional (S. Philipsen, unpublished results). We therefore think that other, as yet less well characterised factors may be involved in LCR function.

The LCR and Disease

The discovery, characterization and mapping of the LCR has enabled the pursuit of two novel approaches to the study of globin-related diseases. Firstly it allows high-Ievel expression of disease genes such as the ▀ gene which is responsible for sickle cell disease. By linking this gene in combination with human alfaglobin genes several laboratories have succeeded in producing transgenic mice which show sickle cell disease [26, 50, 59]. High levels of human haemoglobin Scan be obtained in mice and the RBC of these mice show a pronounced change in shape when deoxygenated (Fig. 5). We are presently improving this model for two reasons, firstly, to study the effects of sickle cell disease on the progression of infection by different malaria strains, and secondly, to be able to study the progression of sickle cell disease and the treatment thereof by new protocols, in particular the development of gene therapy. The latter has been given new hope by the mapping of the minimal elements that give the full activity of the LCR. The LCR can now be incorporated into retroviral vectors to develop therapy protocols and preliminary experiments (F. Meyer, personal communication) indicate that the LCR will provide high levels of expression in this context in mice.

Fig. 4. S 1 nuclease analysis of HS 2 constructs containing the NF -E2 sites (13) or not (1113) in transgenic mice. Foetal liver RNA (day 13.5) was assayed using a mixed probe S 1 nuclease experiment using the 5' human ▀-globin probe and the mouse ▀ maj probe [56]. Specific activities were 10: 1 for H u ▀ : M ▀ .Protected products are indicated on the left. The 200 series of transgenic mice contains the 1113 construct and the 300 series contains the 13 construct. Copy numbers are shown in parentheses. Lower pnael depicts a quantitation experiment of the S 1 protection analysis using the Hu▀5' probe and a mouse alfa-globin probe as an internal control. The % expression is given as the total Hu ▀-globin signal divided by the total mouse alfa-globin signal (adjusted for specific activities). This was plotted against the copy number. The line represents the result of a linear regression analysis on the data points. The R value, the correlation coefficient, indicates very high correlation with a straight line (R = 1 ). The dashed line in the 1113 graph represents the minimallevel that can be measured accurately

Developmental Regulation of the ▀-Globin Locus

The study of globin gene switching has been assisted by the characterization of deletions and point mutations which affect expression of the y and ▀-genes. Point mutations in the ypromoters have been linked to HPFH phenotypes and these can be divided into two groups (Fig. 6). Those clustered around the distal CAA T box appear to result in the loss of factor binding sites [21, 36], suggesting that this region may contain a binding site for a negative regulator. For example, a 13-bp deletion which removes the distal CAA T box results in a very strong HPFH (60%) [25]. Interestingly, a recently described Japanese HPFH (20%) is associated with a point mutation in the CAA T sequence of the distal CAA T box

Fig. 5. Sickle cell disease in transgenic mice. The top line shows the arrangement of genes and the LCR that was injected into fertilized mouse eggs to obtain transgenic mice. The top panel shows sickled cells from one of the transgenic mice [26]; the bottom panel shows control nontransgenic red blood cells

Fig. 6. Summary of mutations occurring in the y-globin promoter resulting in HPFH phenotypes (see [44])

[21] which reduces affinity for the transcription factor CP 1. The -117 mutation associated with Greek HPFH ( 40% ) has been reported to cause reduced binding of the erythroid-specific factor NF-E3 [36]. These findings suggest a model for y silencing in which factors binding to the distal CAA T box (at -115) compete for interaction with factors bound to upstream promoter sequences preventing the proximal CAA T box (at -87) from forming such interactions. The distal CAA T box is located outside the normal optimal position for CAA T elements, and this is likely to prevent it from functioning as an effective positive promoter element. One would expect this type of silencing mechanism to depend on the topology of the promoter region and it is also likely to be affected by the creation of extra factor binding sites in the upstream sequences. Such sites may partially bypass the competition between the proximal and distal CAA T boxes, resulting in suboptimal transcription. Indeed, a second group of mutations, upstream of -150, result in new or improved binding sites for transcription factors, e.g. Sp1 [16,28] and GATA1 [35, 37]. Activation of y transcription in the nondeletion HPFHs is associated with down regulation ofthe▀-gene. The reduction in ▀ expression (to around 60% in Southern Italian HPFH) is approximately equivalent to the rise in expression of the cis-linked y-globin gene, with only a slight reduction in overall transcriptional output from the locus [24, 62]. This suggests that competition is taking place between the genes and that this is tightly linked to the process of transcription. However, a drastic reduction or loss of ▀ transcription due to point mutations and deletions in the ▀-promoter does not significantly increase y expression (less than 5%; Fig.7) [44]. Clearly, a y-gene exerts a negative effect on the ▀-gene (coupled to transcription) but this effect is not reciprocal. Some ▀-gene deletions show higher levels of y expression (Fig. 7) but these deletions are always large (> 10 kb). Of these, the Ay delta ▀ thalassaemias all have deletions which extend into the region of y transcription. They are uninformative for competition models because enhancers found near the deletion breakpoints may be responsible for the high level, pancellular y expression observed in the deletion HPFHs [1, 15]. Some increased y expression is also observed in the delta ▀- and Dutch ▀ thalassaemias, but the broad range of values between patients with the same deletion and the heterocellular distribution of y-protein among the red cells suggest that the increase in y expression is not solely at the transcriptional level. This is supported by the observation that nontranscriptional defects (e. g. RNA processing) in heterozygous ▀-thalassaemias cause elevated levels of y chains (up to 5 %). Selection of a small proportion of cells expressing y is a likely mechanism for this increase. Deletion of the delta-gene (which is normally expressed at only 2% 3% of the level of ▀) also does not seem to be required for the y-expression observed in the delta ▀-thalassaemias, since it is intact in Dutch ▀-thalassaemia, which has a very similar phenotype. Instead, the requirement appears to be a minimum size of deletion (> 10 kb). Probably these large deletions perturb the chromatin structure of the locus, resulting in a small increase in y transcription which is further amplified by the chain imbalance. In conclusion, the genetic data show that strong downregulation of the ▀-gene can result from an increase of y-gene transcription, while there does not seem to be any significant link between transcription of the ▀-gene and silencing of the y-genes in adult life.

Fig. 7. Schematic representation of the different deletions occurring in the ▀-globin locus in thalassaemias and HPFHs. Black bars indicate the size of the deletion; numbers in parentheses indicate the levels of y-globin expression in heterozygotes

Transgenic Mice.

Attempts to study switching of globin genes have also made use of transgenic mice as a model system. Mice do not possess separate foetal globin genes but instead switch directly from embryonic to adult ▀-globin expression at 11-13 days of gestation (Fig. 2).
The developmental regulation of the human epsilon-gene has been analysed in both embryonic stem cells and transgenic mice. In mice the epsilon-gene is expressed at high levels during the embryonic stage only when linked to the LCR and is completely silenced thereafter [33, 46, 51]. Based on the studies by Cao et at. deletion mutants lacking the -200 to -300 promoter region show a small increase in epsilon expression in adult transgenic mice but the low level indicates that other sequences may also be involved in silencing epsilon (P. Fraser, unpublished observations). The human y-transgene without the LCR is expressed like the mouse embryonic genes [7, 32]. It was initially reported that linkage to the LCR resulted in y expression at all developmental stages and that the y-gene was silenced in adult mice only when the ▀-gene was also present. This appeared to support a competition model where the ▀-gene is required for silencing of the y-gene [2, 14]. However, a different result was obtained when the single y-gene experiments were carried out on animals carrying only one or two copies of the LCR-y-gene con

Fig.8. Microlocus (µLCR) y ▀ and ▀ y constructs [29]. Genes are represented as shaded boxes. All genes are in the same trancriptional orientation, 5' to 3', with respect to each other and the LCR. The dotted LCR lines indicate the situation in multicopy animals to illustrate the distance from a promoter to a 5' and 3' LCR. These distances are indicated by dotted lines below the constructs. Plus and minus symbols indicate high, medium, and very low levels of expression

struct. y expression persisted in the early foetal liver, but was silenced at adult stages, independent of the presence of the ▀-gene [12]. Transcription of the LCRlinked y-gene can therefore also be blocked completely by stage-specific negative regulators acting on the sequences immediately flanking the gene, and this removes the basis of the argument that the ▀-gene would be needed for y silencing. The elements responsible for y silencing have not yet been identified but the mutations associated with the nondeletion HPFHs suggest that at least the sequences around the distal CAA T box are likely to be involved (see above). The availability of a transgenic mouse model for y-gene silencing should allow this to be tested and possibly lead to novel approaches for treating thalassaemia and sickle cell anaemia. If y-gene expression were understood at the level of the transcription factors, it might be possible to develop novel therapies that could specifically interfere with the adult suppression of the y-gene and alleviate the clinical problems associated with severe chain imbalance or sickling. Linkage of the adult ▀-gene to the LCR results in inappropriate expression at the embryonic stage [2,3, 14,29,33], albeit at a low level. Placing a y-gene or a human alfa globin gene between the ▀-gene and the LCR blocks this expression [2, 14, 29], supporting the idea that competition plays a role in preventing premature ▀ expression. However, when the order is reversed and the ▀-gene is placed in the first position, it is expressed at a level similar to that observed for the ▀-gene in the absence of the y- or alfa-gene (Fig. 8) [29]. Silencing of the ▀-gene at the embryonic stage is therefore not caused by reciprocal competition only, but relative distance between the LCR and the genes (i.e. position and polarity) is also important. Polarity in the locus has long been suggested by the fact that the genes are arranged in the order of their expression during development. The order of the genes is conserved among mammals but there is some divergence in the other vertebrate loci. In chicken, the embryonic epsilon- and delta-genes are located at opposite ends of the locus, with the adult ▀-genes between them. However, it is important to note that the chicken ▀-globin LCR may have been split as part of an epsilon translocation such that part of it is located between the ▀- and epsilon-genes [8,47] and that the epsilon-gene contributes only 20 % of the total embryonic haemoglobin compared to 80% for delta [5]. The data reviewed above indicate that developmental regulation of the human

Fig. 9. Model for stage-specific regulation of the genes of the ▀-globin locus. Solid lines indicate activation of genes by the LCR. The symbol O indicates stage-specific negative factors silencing the gene. The location of these is not accurate and there may be more than one factor for each gene

▀-globin locus is a complex process which centres around developmentally specific suppressors and the polarity of the locus (Figs. 9, 10). The earliest gene to be activated, the epsilon-gene, is also the one closest to the LCR. The y- and ▀-genes may be suppressed by competition with epsilon; alternatively, or in addition, the y- and ▀-genes may bind embryonic stage-specific factors which keep their promoters suppressed. The epsilon-promoter is silenced in the foetal liver by one or more suppressor factors, negating its competitive ability (Fig. 9). As a result the y-genes are expressed, and they in turn keep expression of the ▀-gene suppressed by competition. The y-genes are switched off during the period around birth, again by stagespecific negative regulators, and as a consequence the ▀-gene is activated and expressed in the bone marrow. We propose that loop formation between regulatory elements is the crucial parameter to explain the suppression of the late genes by the early genes at early stages but not vice versa. The frequency of interaction between the promoters and the LCR will depend on the effective volume in which these elements operate. This effect would be most pronounced if the LCR and the genes were all present on one structural chromatin loop several times the distance between the LCR and the genes. The fact that the LCR controls DNase hypersensitivity of the ▀-globin locus over at least 150-kb [19] suggests that the entire ▀ locus may be present on one very large chromatin loop. If we assume that to be the case, the frequency of interactions between any of the promoters with the LCR would be proportional to their effective concentration relative to the LCR (Fig. 10). On basis of ring closure probabilities with naked DNA, the effective concentration of two points on the DNA will be related to the volume of a sphere and will be proportional to the power of3/2 of the distance. Applying the rule to the ▀-locus, the ▀-gene is twice as far as the Gy-gene from the HS 2 enhancer of the LCR. Therefore, the ▀-gene occupies an approximately eight-fold larger volume relative to the HS-2 enhancer of the LCR than the Gy-gene, which should give it a three-fold lower frequency of interaction with the LCR (Fig. 10). This effect will work in favour of the proximal gene, decreasing the affinity differences required for competition, but it will work against the distal gene. Distal genes would be incapable of suppressing upstream genes under similar circumstances unless the downstream gene promoter increased its affinity by several orders of magnitude relative to the upstream gene. The transgenic mouse data on the expression of the ▀-globin gene at the embryonic and foetal/adult stages argue strongly against this possibility. Instead, the problem is solved by local suppression of the upstream promoters to allow expression from the downstream gene (Fig. 9). Experiments to substantiate or disprove this prediction are presently in progress.

Fig.l0. Schematic representation of the relative volumes occupied by the Gy- and ▀-genes relative to the LCR. For simplicity of presentation the LCR is shown as a fixed point in the centre of the sphere. Only half the ▀-globin gene outer sphere is shown


1. Anagnou NP, Perez-Stable C, Gelinas R, Costantini F, Liapaki K, Constantopoulou M, Costeas T, Moschonas N, Stamatoyannopoulos G (1990) ONA sequences residing 3' of the breakpoint of the HPFH-3 deletion can modify the developmental regulation of the fetal Ay globin gene. Clin Res 38: 301 A
2. Behringer RR, Ryan TM, Palmiter RO, Brinster RL, Townes TM (1990) Human y- to ▀-globin gene switching in transgenic mice. Genes Oev 4:380-389
3. Blom van Assendelft G, Hanscombe 0, Grosveld F, Greaves OR (1989) The ▀-globin domain control region activates homologous and heterologous promoters in a tissue-specific manner. Cell 56: 969-977
4. Bonifer C, Vidal M, Grosveld F, Sippel AE (1990) Tissue-specific and position independent expression of the complete gene domain for chicken lysozyme in transgenic mice. EMBO J 9: 2843- 2848
5. Brown J, Ingram V (1974) Structural Studies on Chick Embryonic Hemoglobins. J BioI Chem 249:3960-3972
6. Cao S, Outman PD, Dave HPG, Schechter AJ (1989) Identification of a transcriptional silencer in the 5'-flanking region of the human epsilon-globin gene. Proc Natl Acad Sci USA 86: 5306-5309
7. Chada K, Magram J, Costantini F (1986) An embryonic pattern of expression of a human fetal globin gene in transgenic mice. Nature 319: 685- 689
8. Choi O-R, Engel JD (1988) Developmental regulations of ▀-globin gene switching. Cell 55: 17- 26
9. Collins FS, Weissman SM (1984) The molecular genetics of human hemoglobin. Prog Acid Res Mol BioI 31:315-462
10. Collis P, Antonioti M, Grosveld F (1990) Definition of the minimal requirements within the human ▀-globin gene and the dominant control region for high level expression. EMBO J 9:233-240
11. de Boer E, Antonioti M, Mignotte V, Wall L, Grosveld F (1988) The human ▀-globin promoter; nuclear protein factors and erythroid specific induction of transcription. EMBO J 7:4203-4212
12. Dillon N, Grosveld F (1991) Human yglobin genes silenced independently of other genes in the ▀-globin locus. Nature 350:252-254
13. Driscoll C, Dobkin C, Alter B (1989) Gamma/delta/beta thalassemia due to a de novo mutation deleting the 5' ▀-globin gene locus activating region hypersensitive sites. Proc Natl Acad Sci USA 86:7470-7474
14. Enver T, Raich N, Ebens AJ, Papayannopotiloti T, Costantini F, Stamatoyannopotilos G (1990) Developmental regulation of human fetal-to-adult globin gene switching in transgenic mice. Nature 344:309-313
15. Feingold E, Forget B (1989) The breakpoint of a large deletion causing hereditary persistence of fetal hemoglobin occurs within an erythroid DNA domain remote from the ▀ globin gene cluster. Blood 74:2178-2186
16. Fischer K, Nowock J (1990) The T to C substitution at -198 of the Ay globin gene associated with the British form of HPFH generates overlapping recognition sites for two DNA binding proteins. Nucleic Acids Res 18:5685
17. Forrester W, Takegawa S, Papayannopotiloti T, Stamatoyannopoulos G, Groudine M (1987) Evidence for a locus activation region: the formation of developmentally stable hypersensitive sites in globin-expressing hybrids. Nucleic Acids Res 15: 10159 -10177
18. Forrester W, Novak U, Gelinas R, Groudine M (1989) Molecular analysis of the human ▀-globin locus activation region. Proc Natl Acad Sci USA 86: 5439-5443
19. ForresterW, Epner E, Driscoll C, Enver T, Brice M, Papayannopotilou T, Groudine M (1990) A deletion of the htiman▀ globin locus activation region causes a major alteration in chromatin structure and replication across the entire ▀ globin locus. Genes Dev 4:1637-1649
20. Fraser P, Hurst J, Collis P, Grosveld F (1990) DNasel hypersensitive sites 1,2 and 3 of the human ▀-globin dominant control region directs position-independent expression. Nucleic Acids Res 18: 3503-3508
21. Fticharoen S, Shimiza K, Ftiktimaki M (1990) A novel C- T transition within the distal CCAA T motif of the Gy glo bin gene in the Japanese HPFH: Implication of factor binding in elevated fetal globin expression. Nucleic Acids Res 18: 5245
22. Gasser S, Laemmli U (1986) Cohabitation of scaffold binding regions with upstream/ enhancer elements of three developmentally regulated genes of D. melanogaster. Cell 46: 521- 530
23. Gidoni D, Kadonaga JT, Barrera-Saldana H, Takahashi K, Chambon P, Tjian R ( 1985) Bidirectional SV 40 transcription mediated by tandem Sp 1 binding interactions. Science 230: 511-514
24. Giglioni B, Casini C, Mantovani R, Merli S, Comp P, Ottolenghi S, Saglio G, Camaschella C, Mazza U (1984) A molecular study of a family with Greek hereditary persistence of fetal hemoglobin and ▀thalassaemia. EMBO J 11 :2641-2645
25. oilman J, Mishima N, Wen X, Stoming T, Lobel J, Huisman T (1988) Distal CCAA T box deletion in the Ay globin gene of two black adolescents with elevated fetal Ay globin. Nucleic Acids Res 18: 10635- 10642
26. Greaves DR, Fraser P, Vidal MA, Hedges MJ, Ropers D, Ltizzatto L, Grosveld F (1990) A transgenic mouse model of sickle cell disorder. Nature 343:183-185
27. Grosveld F, Blom van Assendelft G, Greaves D, Kollias G (1987) Positionindependent high level expression of the human ▀-globin gene in transgenic mice. Cell 51 :975-985
28. Gumicio D, Rood K, Gray T, Riordan M, Sartor C, Collins F (1988) Nuclear proteins that bind the human y globin gene promoter: Alterations in binding produced by point mutations associated with hereditary persistence of fetal hemoglobin. Mol Cell BioI 8: 5310 -5322
29. Hanscombe 0, Whyatt D, Fraser P, Yannoutsos N, Greaves D, Grosveld F (1991) Importance of globin gene order for correct developmental expression. Genes Dev 5:1387-1394
30. Jarman A, Higgs D (1988) Nuclear scaffold attachment sites in the human globin gene complexes. EMBO J 7:3337-3344 31. Kioussis D, Vanin E, deLange T, Flavell RA, Grosveld F (1983) ▀-globin gene inactivation by DNA translocation in Ythalassaemia. Nature 306: 662- 666
32. Kollias G, Wrighton N, Hurst J, Grosveld F (1986) Regulated expression of human Ay-, ▀-, and hybrid y-▀-globin genes in transgenic mice: manipulation of the developmental expression patterns. Cell 46:89-94
33. Lindenbaum M, Grosveld F (1990) An in vitro globin gene switching model based on differentiated embryonic stem cells. Genes Dev 4:2075-2085
34. Magram J, Chada K, Costantini F (1985) Developmental regulation of a cloned adult ▀-globin gene in transgenic mice. Nature 315:338-340 35. Mantovani R, Malgaretti N, Nicolis N, Ronchi A, Giglioni B, Ottolenghi S (1988) The effects of HPFH mutations in the human y globin promoter on binding of ubiquitous and erythroid specific nuclear factors. Nucleic Acids Res 16:7783- 7797
36. Mantovani R, Superti-Fuga G, Gilman J, Ottolenghi S (1989) The deletion of the distal CCAA T box region of the Ay globin gene in black HPFH abolishes the binding of the erythroid specific protein NFE 3 and of the CCAA T displacement protein. Nucleic Acids Res 17: 6681- 6691 37. Martin D, Tsai S, Orkin S (1989) Increased y globin expression in anon deletion HPFH mediated by an erythroid-specific DNA-binding factor. Nature 338:435438
38. Martin D, Orkin S (1990) Transcriptional activation and DNA binding by the erythroid factor GF-1/NF-E1/Eryf 1. Genes Dev 4: 1886-1898
39. Mignotte V, Eleouet EF, Raich N, Romeo PH (1989) Cis- and transacting elements involved in the regulation of the erythroid promoter of the human porphobilinogen deaminase gene. Proc Natl Acad Sci USA 86:6548-6552
40. Mignotte V, Wall L, deBoer E, Grosveld F, Romeo P- H ( 1989) Two tissue-specific factors bind the erythroid promoter of the human porphobilinogen deaminase gene. Nucleic Acids Res 17: 37- 54
41. Ney PA, Sorrentino BP, Lowrey CH, Nienhuis A W (1990) Inducibility of the HS II enhancer depends on binding of an erythroid specific nuclear protein. Nucleic Acids Res 18:6011-6017
42. Pevny L, Simon MC, Robertson E, Klein WH, Tsai S, D' Agati V, Orkin SH, Costantini F (1991) Erythroid differentiation in chimaeric mice blocked by a targeted mutation in the gene for transcription factor GATA-1. Nature 349:257-260
43. Philipsen S, Talbot D, Fraser P, Grosveld F (1990) The ▀-globin dominant control region: hypersensitive site 2. EMBO J 9:2159-2167
44. Poncz M, Henthorn P, Stoeckert C, Surrey S (1989) Globin Gene Expression in Hereditary Persistence of Fetal Hemoglobin and delta▀ Thalassaemia. Oxford University Press
45. Pruzina S, Hanscombe 0, Whyatt D, Grosveld F, Philipsen S (1991) Hypersensitive site 4 of the human ▀-globin locus control region. Nucleic Acids Res 19:1413-1419
46. Raich N, Enver T, Nakamoto B, Josephson B, Papayannopoulou T, Stamatoyannopoulos G (1990) Autonomous developmental control of human embryonic globin switching in transgenic mice. Science 250: 1147 -1149
47. Reitman M, Lee E, Westphal H, Felsenfeld G (1990) Site independent expression of the chicken ▀A globin gene in transgenic mice. Nature 348:749- 752 48. Romeo PH, Prandini MH, Joulin V, Vignotte V, Prenant W, Valnchenker W, Marguerie G, Uzan G (1990) Megakaryocytic and erythrocytic lineages share specific transcription factors. Nature 334: 447-449
49. Ryan TM, Hehringer RR, Martin NC, Townes TM, Palmiter RD, Hrinster RL (1989) A single erythroid specific DNaseI stiper-hypersensitive site activates high levels of human ▀-globin gene expression in transgenic mice. Genes Dev 3:314-323
50. Ryan T, Townes T, Reilly M, Asahura T, Palmiter R, Hrinster R, Hehringer R (1990) A single erythroid specific DNaseI stiperhypersensitive site activates high levels of human ▀-globin gene expression in transgene mice.
51. Shih D, Wall R, Shapiro (1990) Developmentally regulated and erythroid-specific expression of the human embryonic ▀globin gene in transgenic mice. Nucleic Acids Res 18: 5465- 5472
52. Sorrentino HP, Ney PA, Hodine DM, Nienhaus A W (1990) A 46base pair enhancer sequence within the locus activating regionis required to induced expression of the y-globin gene during erythroid differentiation. Nucleic Acids Res 18:2721-2731
53. StiefA, Winter DM, StratlingWH, Sippel AE (1989) A nuclear DNA attachment element mediates elevated and positionindependent gene activity. Nature 341:343-345
54. Talbot D, Collis P, Antonioti M, Vidal M, Grosveld F, Greaves OR (1989) A dominant control region from the human ▀globin locus conferring integration site independent gene expression. Nature 338:352-355
55. Talbot D, Philipsen S, Fraser P, Grosveld F (1990) Detailed analysis of the site 3 region of the human ▀-globin dominant control region. E MHO J9:2169-2177
56. Talbot D, Grosveld F (1991) The 5' HS2 of the globin locus control region functions through the interaction of a multimeric complex binding at two functionally distinct NF-E2 binding sites. E MHO J 10:13911398
57. Taramelli R, Kioussis D, Vanin E, Hartram K, Groffen J, Hurst J, Grosveld FG (1986) ydelta▀-thalassaemias 1 and 2 are the result of a lOO kpb deletion in the human ▀- globin cluster .Nucleic Acids Res 14:7017-7029
58. Townes T, Lingrel J, Chen H, Brinster R, Palmiter R (1985) Erythroid-specific expression of human ▀-globin genes in transgenic mice. EMBO J 4:1715-1723
59. Trtidel M, Saadaen N, Gare M-C, Hardakdjian-Michau J, Blouquit Y, Gtierquin-Kern J-L, Rotiyer-Fessard P, Vidatid D, Pachnis A, Romeo P-H, Hetizard Y, Costantini F (1991) Towards a transgenic mouse model of sickle cell disease: Hemoglobin SAD. E MHO J 10:3157-3165
60. Tuan D, Solomon W, Li Q, London I (1985) The "▀-like-globin" gene domain in human erythroid cells. Proc Natl Acad Sci USA 82:6384-6388 61. Ttian D, Soloman W, London I, Lee DP (1989) An erythroid-specific developmental-stage-independent enhancer far upstream of the human "beta-like globin" genes. Proc Natl Acad Sci 86: 2554-2558
62. Weatherall DJ, Clegg JH (1981) The thalassaemia syndromes. Blackwell, Oxford
63. Wall L, deHoer E, Grosveld F (1988) The human ▀-globin gene 3' enhancer contains multiple binding sites for an erythroid specific induction of transcription. Genes Dev 2:1089-1100
64. Wright S, deHoer E, Rosenthal A, Flavell RA, Grosveld FG (1984) DNA sequences required for regulated expression of the ▀globin genes in murine erythroletikaemia cells. Phil Trans R Soc Lond B 307: 271 282
65. Xiao J, Davidson I, Macchi M, Rosales R, Vigneron M, Staub A, Chambon P (1987) In vitro binding of several cell-specific and ubiquitous nuclear proteins to the GT -I motif of the SV -40 enhancer. Genes Dev 1:794-807
66. Yamamoto M, Ko L, Leonard M, Heug H, Orkin S, Engel J (1990) Acitivity and tissue-specific expression of the transcrip tion factor NF- E 1 multi gene family. Genes Dev 4: 1650 -1662