A common protein binds to two silencers 5' to the human beta-globin gene

Nondeletional type of hereditary persistence of fetal haemoglobin: molecular characterization of three unrelated Thai HPFH

The beta-globin gene clusters of three unrelated Thai families with a nondeletional type of hereditary persistence of fetal haemoglobin (HPFH) were studied using polymerase chain reaction-related techniques. All appeared to have normal nucleotide sequences from the Cap site to position -400 of both the G gamma- and A gamma-globin genes. Two individuals suspected of having a beta-thalassaemia gene linked to the high HbF condition also had a normal beta-globin gene sequence, spanning from position -108 from the Cap site to the polyadenylation site. Deletion of four nucleotides, AGCA, at positions -225 to -222 of one A gamma-globin allele was detected in one subject and was confirmed by dot-blot hybridization. Restriction fragment length polymorphisms in the beta-globin gene cluster showed that the 5' haplotype (-+-++) and the presence (+) of an Xmm 1 polymorphic site at -158 of the G gamma-globin gene are associated with the high F phenotype in these families. Direct sequencing of the 5' hypersensitive-2 (5' HS-2) site of the locus control region (LCR) showed that this Xmn 1 (+) site is also linked to a specific rearrangement of TA repeats (TA)9CACATATACG(TA)10, in HS-2 segment.

Possible factors influencing the haemoglobin and fetal haemoglobin levels in patients with beta-thalassaemia due to a homozygosity for the IVS-I-6 (T-->C) mutation

We have collected haematological, haemoglobin (Hb) and DNA sequence data for 29 patients with a homozygosity for the IVS-I-6 (T-->C) mutation with the intention of identifying factors contributing to the observed variability in the severity of the disease. None of the patients had received blood transfusion therapy for at least 6 months prior to the study. Hb levels varied from 5.0 to 9.9 g/dl. Patients with high Hb F (more than 1.5 g/dl or > 20%) had high total Hb levels (7.5-9.7 g/dl) but some with low Hb F also had high total Hb levels; two had a concomitant alpha-thalassaemia-2 (alpha-thal-2) heterozygosity. An inverse correlation between the Hb F and Hb A2 levels was observed. The majority of the patients were homozygous for haplotype VI (49/58 chromosomes) but haplotypes IV (2/58) and VII (7/58) were also present. The only haplotype IV homozygote had high Hb F levels with high G gamma values and the C-->T mutation at position -158 in the G gamma promoter, while both high and low Hb F levels were observed among patients with haplotypes VI and VII. Analysis of sequence variations in regulatory regions included the 5' hypersensitive sites (HS) 4. 3 and 2 of the locus control region (LCR), the G gamma and A gamma 5' flanking regions, the second intervening sequence (IVS-II), and the 5' beta-globin gene region in two patients with high Hb F (one homozygote each for haplotypes VI and IV), and in two patients with low Hb F levels (one homozygote each for haplotypes VI and VII). Haplotype specific differences were observed in the LCR 5' HS-2 and in the G gamma and A gamma flanking and IVS-II regions; however, no differences were present between the low and high Hb F-producing haplotype VI chromosomes, suggesting a major role for factors which are not linked to the beta-globin gene cluster in mediating gamma-globin gene expression in patients with this type of beta-thal.

Polymorphic pattern of the (AT)X(T)Y motif at -530 5' to the beta-globin gene in over 40 patients homozygous for various beta-thalassemia mutations

Nucleotide sequence analysis of the 5' beta-globin gene flanking region has been carried out for numerous homozygous beta-thalassemia patients with different mutations and of various ethnic backgrounds. Four different rearrangements were found associated with numerous beta-thalassemia mutations. The (AT)X(T)Y repeat motif at -530 showed polymorphic patterns among these patients as follows: All ten IVS-II-1 (G-->A) chromosomes and the two with the -87 (C-->G) mutation are associated with the (AT)9(T)5 rearrangement, while the 30 IVS-I-6 (T-->C), the 16 codon 39 (C-->T), the six codon 8 (-AA) chromosomes, and 12 chromosomes with different promoter mutations had the (AT)7(T)7 motif. Six chromosomes with the promoter mutation at position -29 (A-->G) had the (AT)8(T)6 motif, while an (AT)8(T)4 motif appears characteristic for two IVS-I-5 (G-->A and G-->T). No direct association between any of the (AT)X(T)Y arrangements and an increased gamma gene expression [G gamma and fetal hemoglobin (Hb F)] levels could be demonstrated, suggesting that variations in the (AT)X(T)Y motif are common polymorphisms.

Negative regulation of transcription in eukaryotes

Images

Sequence-specific binding of protein factors to two independent promoter regions of the acidic tobacco pathogenesis-related-1 protein gene (PR-1)

Gel shift mobility analysis, using the proximal 0.3 kb fragment of the tobacco pathogenesis-related protein 1a gene (PR-1a) and nuclear extracts from healthy Samsun NN tobacco leaves, which do not produce PR-1 proteins, showed a broad shifted signal with low mobility. This signal was not detected with nuclear proteins from the interspecific hybrid of Nicotiana glutinosa x Nicotiana debneyi, which constitutively produces the PR-1a protein. Similar shifted signals were detected with both proximal and distal regions of the 0.3 kb fragment using nuclear proteins from healthy Samsun NN tobacco, but not with proteins from the interspecific hybrid. Further experiments, performed using 5' or 3' truncated fragments of the 0.3 kb fragment, identified two independent binding sites: a distal site between -179 and -168 bp from the transcription start site, and a proximal site between -61 and -37 bp. Footprint analysis revealed two protected sequences, a distal region between -184 and -172 bp, and a proximal region between -68 and -51 bp. These results indicate the presence of regulatory factor(s) for expression of the acidic PR-1a gene. The possibility of negative regulation of the gene is discussed.

Inter-ethnic polymorphism of the beta-globin gene locus control region (LCR) in sickle-cell anemia patients

Sequence polymorphisms within the 5'HS2 segment of human locus control region is described among sickle cell anemia patients. Distinct polymorphic patterns of a simple sequence repeat are observed in strong linkage disequilibrium with each of the five major beta s haplotypes. Potential functional relevance of this polymorphic region in globin gene expression is discussed.

Hb F in sickle cell anemia

We have reviewed the methodology for an accurate quantitation of Hb F in the blood of patients with sickle cell anemia, values observed in hundreds of patients of different (racial or ethnic) backgrounds and with differences in severity of the disease, and the various factors that affect the level of Hb F. The latter include sex, age, genetic background or chromosomal haplotypes, variations in the sequences of the locus control region(s) 5' to the epsilon-globin gene, and the presence of an alpha chain deficiency or alpha-thalassemia. Finally, a few remarks about agents effective in increasing the in vivo Hb F synthesis are also included.

Genetic epidemiology of beta-thalassemia in Sicily: do sequences 5' to the G gamma gene and 5' to the beta gene interact to enhance HbF expression in beta-thalassemia?

The present epidemiological study of the molecular characteristics of beta-thalassemia in Sicily was prompted by the disparate phenotypic expression (in clinical status and absolute HbF level) observed in two beta-thalassemic homozygotes who were also homozygous for the beta-like globin gene cluster haplotype III. We suspected that polymorphisms within haplotype III could be the cause for the discrepancy. Based on the association of particular conformations of the (AT)xT(y) motif (-540 5' to the beta gene) with milder forms of thalassemia and sickle cell anemia, 38 homozygous beta-thalassemia patients were studied to define their haplotypes, the -158 site 5' to the G gamma gene (linked to haplotype III) and the structure of the (AT)xT(y) motif. We found that the patient who was phenotypically mild and homozygous for beta-thalassemia, haplotype III, and the -158 C----T mutation was homozygous for the rare (AT)9T5 motif. In contrast, the patient homozygous for beta-thalassemia, haplotype III, and the -158 mutation, but exhibiting a severe clinical course, was homozygous for the (AT)7T7 configuration. Others have suggested that (AT)9T5 is a negative regulatory protein binding sequence, and it is a silent carrier state for beta-thalassemia. The usual configuration (AT)7T7, has considerably less affinity for regulatory protein binding, and it is the most common configuration in Sicilian beta-thalassemics (67 of the 78 chromosomes studied). Within the 38 patients studied, seven were informative because they had various combinations of the (AT)9T5 and (AT)7T7 motif, and the -158 C----T mutation. The results in these patients suggest that only the co-presence of the (AT)9T5 configuration and a C----T change at -158 5' to the G gamma gene is associated with high HbF expression and a mild clinical phenotype. We postulate that these two regions of the beta-like globin gene cluster interact, when endowed with the proper sequences, to enhance the expression of HbF secondary to anemia.

A beta-thalassaemia phenotype not linked to the beta-globin cluster in an Italian family

This paper describes a family of Central Italian origin in which three patients in two generations had either thalassaemia intermedia or a late presenting form of thalassaemia major. Sequence analysis of the patients' DNA revealed that only one of the beta-globin genes was affected by a beta-thalassaemia mutation (the codon 39 nonsense mutation), the other being completely normal, apart from the complex rearrangement (-T +ATA) at position -530 5' to the CAP site of the beta-globin gene, which has uncertain clinical significance. Haematologically, all these patients were characterized by unusually low HbF levels (1.8-7.3%) for a beta-thalassaemia major or intermedia phenotype. The mother of the two patients with thalassaemia intermedia was heterozygous for beta-thalassaemia (codon 39 nonsense mutation), while the father had thalassaemia-like red cell indices, an increased alpha/non alpha chain synthesis ratio, a slight increase of HbF and a low HbA2 level, but showed entirely normal beta-globin gene sequences, apart from the complex rearrangement (-T +ATA) at position -530 5' to the CAP site. One of the thalassaemia intermedia patients married a normal woman and they had a child with thalassaemia major who inherited only the codon 39 nonsense mutation but not the complex rearrangement at position -530. The clinical phenotype of thalassaemia-intermedia or major in the patients from this family may be explained by postulating the inheritance of the double heterozygous state for beta-thalassaemia and for a mutation in a gene coding for an erythroid-specific DNA binding protein which may impair the function of the normal beta-globin gene. Heterozygosity for this postulated mutation (father of the patients with thalassaemia intermedia) may result in the production of a beta-thalassaemia carrier state with normal HbA2 level.

Persistence of repeated sequences that evolve by replication slippage

The evolution of short repeated sequences by replication slippage under the assumption of selective neutrality is modeled using a linear birth and death process. The equilibrium distribution, the distribution of the life expectancy of a repeated sequence when the process starts from two repeats, the age distribution of repeats, the probability of obtaining two genes with i and j copies which diverged t generations ago and the conditional variance of copy number given the repeat number is more than one are computed. The distributions of life expectancy and age are shown to have long tails. Also the statistic which estimates the conditional variance is shown to have a large coefficient of variation. Using these theoretical results, we develop an approximate test of our model and analyze persistent repeated sequences found in the primate beta-globin gene region and Oenothera chloroplast DNA which are polymorphic within species. We found one sequence in Oenothera chloroplast DNA which does not fit to our neutral model.

Binding of a nuclear protein to the rat growth hormone silencer element

The rat growth hormone (rGH) gene is uniquely expressed in a subset of cells from the anterior pituitary. This strongly cell type specific expression is controlled by both cis-acting positive sequences that bind the pituitary specific transcription factor Pit-1 and cis-acting negative regulatory elements that lie upstream of the Pit-1 sites. The negative elements act to prevent expression of the gene in inappropriate cell types. Here we report that the most proximal rGH silencer element is specifically bound by a protein found in a number of rGH non-expressing cell types and which exerts a negative regulatory effect through the recognition of this rGH element in transient transfection assays. The sequence recognized by this protein is similar to sequences of several other negative regulatory elements as well as to the consensus binding site for the transcription factor NF1. However, the 45 KDa molecular weight identified for this protein does not correspond to any of the sizes previously reported for NF1 suggesting that it is likely to represent a new member amongst this family of transcription factors.

Binding of erythroid and non-erythroid nuclear proteins to the silencer of the human epsilon-globin-encoding gene

To clarify the molecular mechanisms involved in the developmental control of hemoglobin-encoding genes we have been studying the expression of these genes in human cells in continuous culture. We have previously reported the presence of a transcriptional control element with the properties of a silencer extending from -392 to -177 bp relative to the cap site of the human epsilon-globin-encoding gene [Cao et al., Proc. Natl. Acad. Sci. USA 86 (1989) 5306-5309]. We also showed that this silencer has stronger inhibitory activity in HeLa cells, as compared to K562 human erythroleukemia cells. Using deletion mutants and cis-cloned synthetic oligodeoxyribonucleotides in transient expression assays, nucleotide sequences responsible for this effect have now been further delimited to 44 bp located from -294 to -251 bp. Gel electrophoresis mobility shift assays and DNaseI footprinting assays demonstrate that these negative regulatory sequences are recognized differently by proteins present in nuclear extracts obtained from HeLa and K562 cells. Two binding proteins are detected in K562 nuclear extracts, while only one is found in extracts from HeLa cells. Possible mechanisms by which these proteins may regulate transcription of the epsilon-globin-encoding gene in erythroid and non-erythroid cells are discussed.

Evolution of the primate beta-globin gene region: nucleotide sequence of the delta-beta-globin intergenic region of gorilla and phylogenetic relationships between African apes and man

A 6.0-kb DNA fragment from Gorilla gorilla including the 5' part of the beta-globin gene and about 4.5 kb of its upstream flanking region was cloned and sequenced. The sequence was compared to the human, chimpanzee, and macaque delta-beta intergenic region. This analysis reveals four tandemly repeated sequences (RS), at the same location in the four species, showing a variable number of repeats generating both intraspecific (polymorphism) and interspecific variability. These tandem arrays delimit five regions of unique sequence called IG for intergenic. The divergence for these IG sequences is 1.85 +/- 0.22% between human and gorilla, which is not significantly different from the value estimated in the same region between chimpanzee and human (1.62 +/- 0.21%). The CpG and TpA dinucleotides are avoided. CpGs evolve faster than other sequence sites but do not confuse phylogenetic inferences by producing parallel mutations in different lineages. About 75% of CpG doublets have become TpG or CpA since the common ancestor, in agreement with the methylation/deamination pattern. Comparison of this intergenic region gives information on branching order within Hominoidea. Parsimony and distance-based methods when applied to the delta-beta intergenic region provide evidence (although not statistically significant) that human and chimpanzee are more closely related to each other than to gorilla. CpG sites are indeed rich in information by carrying substitutions along the short internal branch. Combining these results with those on the psi eta-delta intergenic region, shows in a statistically significant way that chimpanzee is the closest relative of human.

A novel beta-globin structural mutant, Hb Brescia (beta 114 Leu-Pro), causing a severe beta-thalassemia intermedia phenotype

This study describes a patient with a thalassemia intermedia-like phenotype in whom beta-globin gene sequencing detected a novel abnormal hemoglobin (Hb) due to a T-C substitution at codon 114 of the beta-globin gene arising as a de novo mutation. The abnormal variant was designated Hb Brescia after the place of birth of the propositus. Normal sequences were detected at the in trans beta-globin locus. In addition, alpha-globin gene analysis detected a triple alpha-globin locus which was inherited from the father. The T-C change at position 114 of the beta-globin gene results in a leucine to proline substitution (Leu-Pro) in the G-helix. The resulting Hb tetramer is highly unstable and precipitates forming inclusion bodies in the peripheral red blood cells. Moreover, the Leu-Pro substitution interferes negatively with the four alpha 1 beta 1 contact points of the G-helix most likely adversely affecting the alpha beta dimer formation. The very severe phenotype presented by our patient is unusual in a heterozygote for an unstable Hb variant and may be explained by the coinheritance of the triple alpha-globin locus.

Disappearance of the protein of a somatic mutation: a possible example of stem cell inactivation

The low concentration of the hemoglobin variant, Hb Vicksburg (leucine-beta-75 deleted), and a profound deficit of its mRNA led us to postulate that a beta(+)-thalassemia mutation existed in cis to the coding region mutation, suppressing its synthesis. We examined blood from this patient 6, 8, and 10 yr after our initial studies, using methods of analysis unavailable initially. We found 1) mutations causing beta(+)-(-88 C----T) and beta 0-(849 A----G) thalassemia; 2) that the proportion of Hb Vicksburg in erythrocytes fell over time, from 8 to 4%, and ultimately disappeared; and 3) that the mutation causing Hb Vicksburg was not detectable in genomic DNA isolated from blood leukocytes when this variant was present in hemolysate. We postulate that Hb Vicksburg arose from a somatic mutation of a beta(+)-thalassemia gene in an erythroid-committed stem cell. Its gradual disappearance suggests the cycling of stem cells, with inactivation of different clones over time.

Purkinje cell protein-2 regulatory regions and transgene expression in cerebellar compartments

The Purkinje cell protein 2 (Pcp-2) is expressed in cerebellar Purkinje cells and retinal bipolar neurons. To illuminate how Pcp-2 expression is restricted to only two neuronal types and to derive tools to express heterologous genes in these neuronal subpopulations, genomic sequences of the mouse Pcp-2 gene have been cloned and flanking sequences have been evaluated as a source of neuron-specific regulatory elements. An upstream region with homology to other genes expressed in neurons was identified and a hybrid gene containing this sequence was constructed by ligating 0.4 kb of upstream and 0.3 kb of downstream Pcp-2-flanking DNA to lacZ. Transgenic mice bearing this construct exhibited beta-galactosidase in a wide array of neuron types, suggesting that this sequence may play an important role in specifying neuronal expression. Addition of a further 3.1 kb of Pcp-2 upstream sequences restricted expression of beta-galactosidase to a small number of neuron types and most notably to Purkinje cells within parasagitally oriented cerebellar compartments. The presence of elements lying within the 3.1-kb upstream region and acting to specifically restrict Pcp-2 expression is therefore suggested. Moreover, as beta-galactosidase was not expressed in the bipolar cells of these transgenic mice, retinal expression of the endogenous Pcp-2 gene must involve elements in addition to those conferring expression within Purkinje cells.

Hemoglobinopathies among the Gond tribal groups of central India; interaction of alpha- and beta-thalassemia with beta chain variants

We have investigated the frequencies and types of alpha-thal, beta-thal, and Hb variants among nearly 200 inhabitants of villages in the Mandla and Jabalpur districts of Madhya Pradesh in Central India. Over 85% were tribals of the Gond group. alpha-Thal, as -alpha 3.7/and -alpha 4.2/, and the nondeletional Koya Dora mutation were present at the combined frequency of 0.54. There were indications for the presence of other nondeletional types of alpha-thal. alpha-Globin gene triplications were not observed. Four of the six beta-thal alleles observed were in the tribal groups; two (G----C at codon 30 and G----A at IVS-I-1) were found for the first time. The simultaneous presence of an alpha-thal (-alpha/alpha alpha or -alpha/-alpha) greatly improved the clinical and hematological condition of the patients with Hb S-beta(+)-thal (IVS-I-5; G----C). The lower frequency of alpha-thal among the beta-thal heterozygotes (f = 0.32) may indicate that some of the beta-thal alleles in the tribal populations originated from an outside source. Forty-one subjects had SS; all but one had beta S with haplotype #31, while one chromosome had haplotype #17. The presence of an alpha-thal-2 (f = 0.53) in the SS patients did not affect hematological data. The Hb F levels varied between 7.5% and 42.5% with high G gamma values. No difference in Hb F level between males and females was observed. Lower Hb F levels were present in 10 SS patients with an alpha-thal-2 homozygosity (average 16% versus 23.5% for eight SS patients with alpha alpha/alpha alpha) suggesting a decreased formation of alpha gamma dimers in severe alpha chain deficiency. Several younger SS patients (less than 10 years) also had high Hb F levels (32-42%). Variations in the sequence at -530 of the beta-globin gene; i.e. in the so-called silencer sequence, were present in all beta S chromosomes with haplotype #31, but were not considered important for understanding the variability in the Hb F level. gamma-Globin gene deletions (gamma-thal) and triplications were not observed.

Increased protein binding to a -530 mutation of the human beta-globin gene associated with decreased beta-globin synthesis

Although some cases of the syndrome of hereditary persistence of fetal hemoglobin (HPFH) have been correlated with mutations causing a change in the binding of trans-acting factors to DNA sequences flanking the gamma-globin gene, this mechanism has not been described in beta-thalassemias upstream of the canonical promoter of the beta-globin gene. In this report we describe such a change in binding of a protein that may explain a silent carrier phenotype of beta-thalassemia. We have previously demonstrated the binding of a protein (BP1) derived from a nuclear extract of human K562 cells to DNA 5' to the human beta-globin gene in a region having a negative regulatory function. The binding of BP1 in this region can be detected by DNAse I footprinting and by gel mobility shift analysis. We have now compared binding of BP1 to the normal sequence and a mutated sequence (+ATA/-T at -530 bp from the cap site) from the silent carrier of beta-thalassemia. Using mobility shift assays we show that BP1 binds about nine times more strongly to the mutated sequence than the normal sequence. These results suggest the possibility that the decreased expression of the beta-globin gene exhibited by the carrier may be due, at least in part, to tighter binding of a protein which functions as a negative control element or repressor.

Molecular analysis of atypical beta-thalassemia heterozygotes

This paper reviews the molecular pathology of a heterogeneous group of beta-thalassemia heterozygotes which may be referred to as atypical beta-thalassemia. This group includes four different categories of heterozygous beta-thalassemia, which are characterized, respectively, by (1) normal MCV and MCH; (2) normal Hb A2; (3) normal MCV, MCH, and Hb A2 and imbalanced globin chain synthesis only or, (4) the presence of clinical manifestations. The first group is represented by a limited proportion of double heterozygotes for alpha- and beta-thalassemia. The second group includes two categories. One category is double heterozygotes for delta- and beta-thalassemia with the delta-thalassemia mutation in cis or in trans to beta-thalassemia. A number of delta-thalassemia mutations which produce this phenotype by interacting with beta-thalassemia have been described. The other category within the second group is heterozygotes for some mild beta(+)-thalassemia mutations. Within the third group, conclusive evidence for a mutation within the beta-globin gene cluster producing the silent beta-thalassemia phenotype has been obtained solely for a C----T substitution at -101 within the CACCC box of the beta-globin gene. Possible candidates are the complex rearrangements (-T, +ATA; -T, +ATATA) found at position -530 from the cap site. In the group of thalassemic hemoglobinopathies, a series of mutations mostly located in the third exon and producing elongated or truncated molecules have been recently reported. Most of the mutations are silent at the protein level, produce inclusion bodies in peripheral erythrocytes, and show a dominant transmission pattern or occur sporadically.

DNaseI hypersensitive sites 1, 2 and 3 of the human beta-globin dominant control region direct position-independent expression

The human beta-globin dominant control region (DCR) which flanks the multigene beta-globin locus directs high level, site of integration independent, copy number dependent expression on a linked human beta-globin gene in transgenic mice and stably transfected mouse erythroleukemia (MEL) cells. We have assayed each of the individual DNaseI hypersensitive regions present in the full 15kb DCR for position independence and copy number dependence of a linked beta-globin gene in transgenic mice. The results show that at least three of the individual DNaseI hypersensitive site regions (sites 1, 2 and 3), though expressing at lower levels than the full DCR, are capable of position independent, copy number dependent expression. Site 2 alone directs the highest level of expression of the single site constructs, producing nearly 70% of the level of the full DCR. Sites 1 and 3 each provide 30% of the full activity. Deletion of either site 2 or 3 from the complete set significantly reduces the level of expression, but does not effect position independence or copy number dependence. This demonstrates that sites 2 and 3 are required for full expression and suggests that all the sites are required for the full expression of even a single gene from this multigene locus.