Transcriptional regulatory network analysis of developing human erythroid progenitors reveals patterns of coregulation and potential transcriptional regulators.

Deciphering the molecular basis for human erythropoiesis should yield information benefiting studies of the hemoglobinopathies and other erythroid disorders. We used an in vitro erythroid differentiation system to study the developing red blood cell transcriptome derived from adult CD34+ hematopoietic progenitor cells. mRNA expression profiling was used to characterize developing erythroid cells at six time points during differentiation (days 1, 3, 5, 7, 9, and 11). Eleven thousand seven hundred sixty-three genes (20,963 Affymetrix probe sets) were expressed on day 1, and 1,504 genes, represented by 1,953 probe sets, were differentially expressed (DE) with 537 upregulated and 969 downregulated. A subset of the DE genes was validated using real-time RT-PCR. The DE probe sets were subjected to a cluster metric and could be divided into two, three, four, five, or six clusters of genes with different expression patterns in each cluster. Genes in these clusters were examined for shared transcription factor binding sites (TFBS) in their promoters by comparing enrichment of each TFBS relative to a reference set using transcriptional regulatory network analysis. The sets of TFBS enriched in genes up- and downregulated during erythropoiesis were distinct. This analysis identified transcriptional regulators critical to erythroid development, factors recently found to play a role, as well as a new list of potential candidates, including Evi-1, a potential silencer of genes upregulated during erythropoiesis. Thus this transcriptional regulatory network analysis has yielded a focused set of factors and their target genes whose role in differentiation of the hematopoietic stem cell into distinct blood cell lineages can be elucidated.

Simple two-color array-based approach for mutation detection.

The ability to analyze multiple polymorphic/mutation sites rapidly and accurately is pivotal in all areas of genetic analysis. We have applied single nucleotide primer extension (SNE) for detection of multiple point mutations in a micro-array format using two-color, fluorescent dye-tagged dideoxynucleoside triphosphate terminators (ddNTPs). The oligonucleotide primer ending one nucleotide short of the mutation site being probed is bound to the slide and single-base extended in place with two different Cy5/Cy3 dye-tagged terminators using solution-phase, locus-specific, single-stranded complementary templates generated by PCR from genomic DNA. The composite fluorescence produced contains peaks of distinct wave lengths corresponding to each Cy dye-tagged terminator incorporated, resulting in a fluorescent 'fingerprint' for each DNA target. DNA polymerase-catalyzed incorporation of Cy dye-tagged dideoxynucleoside triphosphates was dependent on the particular dyes, the specific ddNTP, the DNA target concentration, sequence of the template, on-slide temperature cycling and washing conditions. Results from analysis of mutations in the human hemochromatosis and connexin 26 genes show that this approach has several advantages over existing methods and is simple, rapid, robust, cost effective and accurate with potential applications in many areas of genetic analysis.

A single-base change at position -175 in the 5'-flanking region of the G gamma-globin gene from a black with G gamma-beta+ HPFH.

Hereditary persistence of fetal hemoglobin (HPFH) is a human hemoglobinopathy characterized by the continued expression of fetal globins during adult life. Both deletional and nondeletional forms have been described. A number of single-base changes in the immediate 5'-flanking region of the fetal G gamma and A gamma have been reported associated with nondeletional forms of HPFH. We now present the nucleotide sequence of a G gamma-globin gene from an American black with G gamma-beta + HPFH. The immediate 5'-flanking region of this G gamma gene has a T-to-C change at -175, C at -158, and a normal C at -202. Additional changes were found in IVS2 and in the immediate 3'-flanking region, some of which may represent gene-conversion events. The sequence change at -175 probably represents a second mutation associated with the G gamma-beta + HPFH phenotype in blacks. This base change alters an octamer sequence known to be of importance in the normal expression of several other genes.

Functional analysis of a beta-globin gene containing a TATA box mutation from a Kurdish Jew with beta thalassemia.

We recently reported a TATA box mutation (ATAAAA to ATACAA) in a cloned beta-globin gene from a Kurdish Jew with homozygous beta thalassemia (Poncz, M., Ballantine, M., Solowiejczyk, D., Barak, I., Schwartz, E., and Surrey, S. (1982) J. Biol. Chem. 257, 5994-5996). We have now introduced this gene into HeLa cells after CaPO4 precipitation of the DNA and studied expression by analyzing globin-gene transcripts with a novel S1 nuclease mapping assay. Quantitative and qualitative comparison with the normal beta-globin gene revealed a promoter-down phenotype in the TATA box mutant, with normal RNA processing, and a normal start site for initiation of the primary transcript. Decreased transcriptional efficiency was confirmed directly by analysis of run-off transcripts using assays in vitro. The patient's phenotype of beta thalassemia major is probably the result of two different mutations since haplotype analysis of the beta-like globin gene clusters in genomic DNA from this patient shows heterozygosity for the Mediterranean-type haplotypes I and VII, with the TATA box mutation on a haplotype I chromosomal background.

A greater than 200 kb deletion removing the entire beta-like globin gene cluster in a family of Irish descent.

We describe a new deletional form of gamma delta beta-thalassemia segregating in two generations of a family of Irish descent. Affected family members present with a beta-thalassemia minor phenotype, normal Hb A2 and Hb F levels. Genomic blotting analyses on DNA from affected family members show heterozygosity for a large deletion beginning at least 15 kb upstream of the 5' endpoint of the gamma delta beta-thalassemia-1 deletion, extending through the entire beta-like globin gene cluster, and continuing for at least 10 kb beyond the 3' endpoint of the deletion associated with the Spanish form of delta beta 0-thalassemia. This deletion is among the largest described so far, and removes at least 205 kb encompassing the entire beta-like globin gene cluster on chromosome 11.

DNA methylation in hereditary persistence of fetal hemoglobin (HPFH-2).

Recent studies show that the region of DNA brought into close proximity to the fetal gamma-globin genes in deletional forms of hereditary persistence of fetal hemoglobin (HPFH) is selectively hypomethylated (and presumably active) in normal erythroid tissue. This region is normally located approximately 100 kilobases (kb) 3' to the fetal genes. The continued expression of fetal hemoglobin in adult life in these forms of HPFH has been ascribed to the effect of this erythroid-specific region in altering local chromosomal structure and allowing transcription. Because transcriptional activity is often associated with hypomethylation, we have examined the methylation status of the gamma-globin genes and the truncated psi beta gene on the HPFH chromosome to determine whether juxtaposition of this erythroid-specific region results in a generalized hypomethylation of the globin gene region upstream of the deletion breakpoint. Genomic DNA purified from nucleated red blood cells (nRBC) from a patient with HPFH-2/beta O thalassemia was digested with Msp I or Hpa II, and the methylation pattern determined on the HPFH chromosome by using secondary cleavage with restriction enzymes which span the deletion breakpoint. These studies show that in nRBC the HPFH-2 chromosome is hypomethylated in the 3'-juxtaposed region (3'JR) and in the region of the gamma-globin genes. In contrast, Msp I sites near the truncated psi beta-globin gene remain methylated, suggesting that only a subset of CpG sites upstream from the 3'JR become hypomethylated in HPFH-2. This subset of sites corresponds to those which are normally hypomethylated when fetal genes are active. The continued high level of fetal globin expression is, therefore, not associated with a generalized hypomethylation upstream from the deletion junction.

The silent carrier allele: beta thalassemia without a mutation in the beta-globin gene or its immediate flanking regions.

A molecular genetic analysis has been performed using as subjects an Albanian family in which the father is a silent carrier, the mother has high Hb A2-beta thalassemia trait, and both children have beta thalassemia. Nucleotide sequence analysis of the daughter's paternal beta-globin gene and its flanking regions failed to reveal any base changes of known functional significance. When introduced into HeLa cells the gene was expressed at normal levels with proper processing of RNA. Haplotype analysis revealed that the affected son and daughter inherited different epsilon gamma delta beta-globin gene clusters from the father. The silent carrier allele is not due to a mutation within the beta-globin structural gene or its flanking regions and as such represents a novel form of beta+ thalassemia.

Detection of a novel DNA polymorphism in the beta-globin gene cluster.

Analysis of DNA from the beta-globin gene cluster in an Albanian family identified a novel RsaI site approximately 550 base pairs 5' to the beta-globin gene. In this family, two chromosomes carrying otherwise identical beta-globin haplotypes were found to differ at the RsaI site. Population screening demonstrated the presence and absence of the site in DNA from individuals of northern European, Mediterranean, Middle Eastern, Southeast Asian, African, and Asian Indian descent, indicating that this site is a DNA polymorphism common in many ethnic groups. The polymorphism is also present in DNA from individuals carrying different beta-globin alleles. Additional nucleotide sequence changes identified in an RsaI (+) genomic clone in the region immediately 3' to the RsaI site suggest a mechanism for the randomization of the site with respect to haplotype.

A large deletion encompassing the entire alpha-like globin gene cluster in a family of northern European extraction.

We describe a new deletional form of alpha thalassemia segregating in three generations of a family of northern European origin. A full-term female girl had hypochromic, microcytic anemia since early infancy associated with delayed language development, slow growth and weight gain. Hematologic studies suggested the presence of alpha thalassemia. Gene-blotting studies showed no abnormal alpha-like globin gene fragments; however, studies of inheritance of informative polymorphic restriction fragments using zeta, alpha and 3'-alpha-hypervariable region (3'-HVR) probes showed evidence for an extensive deletion encompassing the entire alpha-like globin gene cluster. The 3' breakpoint of this deletion maps beyond the 3'-HVR, a region implicated as a hot spot for the generation of other large deletional events within the alpha-like cluster. The 5' breakpoint maps at least 10 kilobases (kb) 5' to the zeta-globin gene. The minimum size estimate for this deletion is greater than 47 kilobases.

Organization of the gene for platelet glycoprotein IIb.

The glycoprotein (GP) IIb/IIIa heterodimer functions as a receptor for fibrinogen, von Willebrand factor, and fibronectin on activated platelets; it is dysfunctional in the bleeding diathesis Glanzmann's thrombasthenia. This receptor is a member of the integrin family, which includes homologous membrane receptors involved in a number of different cell-cell and cell-matrix adhesive interactions. Knowledge of the sequence and organization of the GPIIb and GPIIIa genes will help in understanding evolutionary relationships and functional homologies of this family of adhesion protein receptors and will facilitate analysis of molecular defects responsible for thrombasthenia. Using the GPIIb cDNA as a probe, we have isolated overlapping genomic clones encompassing the entire coding region, the 5'- and 3'-untranslated sequences, and the immediate flanking regions for the GPIIb gene. The gene spans approximately 17.2 kilobases (kb); all but approximately 2.6 kb of intronic DNA sequence has been determined. The GPIIb gene contains 30 exons whose demarcations do not correlate with previously suggested functional domains. Two intron/exon borders have the rare GC splice donor sequence instead of the consensus GT sequence. There are at least seven complete and three partial AluI sequence repeats within the intron sequences. RNase protection, S1 nuclease analysis, and primer extension studies using human erythroleukemia (HEL) cell RNA and platelet RNA map a major transcription start site 32 base pairs (bp) 5' to the beginning of the coding region; however, there are no canonical consensus TATA or CAAT boxes in the region immediately 5' to the proposed cap site. The immediate 5'-flanking sequence of rodent GPIIb demonstrates complete identity near the proposed cap site with its human counterpart, but again, no TATA or CAAT boxes are apparent.

The -175T----C mutation increases promoter strength in erythroid cells: correlation with evolutionary conservation of binding sites for two trans-acting factors.

A point mutation at position -175 has been detected in Agamma as well as Ggamma globin genes in individuals with hereditary persistence of fetal hemoglobin (HPFH). To prove that this single point mutation results in increased promoter strength, we transfected erythroid and nonerythroid cell lines with constructs containing normal and mutant promoters linked to the bacterial chloramphenicol acetyl transferase (CAT) gene. Differences in transfection efficiency were controlled by cotransfection of pRSVgpt. In K562 erythroleukemia cells, the -175 HPFH promoter directed three- to fourfold more CAT activity than its wild type counterpart. However, in HeLa cells the two promoters were similar in strength. The -195 to -165 region of the gamma-globin promoter contains binding sites for two proteins: a ubiquitously distributed octamer binding protein, OBP, and the erythroid-specific protein, GF-1. We find that while the GF-1 binding site is highly conserved among related primate gamma-globin genes, the octamer binding site is not. The evolutionary conservation of GF-1 as well as its erythroid-specific distribution suggest that this protein is important in gamma-globin gene expression. A role for OBP in the regulation of gamma-globin, if any, must have arisen recently in primate evolution.

Interaction of rare illegitimate recombination event and a poly A addition site mutation resulting in a severe form of alpha thalassemia.

The clinical diversity of thalassemia depends on interaction of diverse genetic defects. We have characterized a severe form of alpha thalassemia caused by coinheritance of a rare alpha-globin gene deletion and a nondeletional defect in a southern Italian family. The proband, a 7-year-old girl, exhibited an abnormal hemoglobin electrophoresis pattern with hemoglobin H and hemoglobin Barts, indicating inheritance of H and hemoglobin Barts, indicating inheritance of a severe form of alpha thalassemia. Southern blot analysis of DNA showed normal as well as aberrant alpha-globin gene fragments indicating heterozygosity for a deletional form of alpha thalassemia in the proband and her mother. The coinheritance of a nondeletional form of alpha thalassemia (alpha alpha T) was suspected because of the severity of the proband's phenotype and the presence of normal alpha-globin gene fragments in the father. Selective polymerase chain reaction of the paternal alpha 1- and alpha 2-globin genes in the proband followed by DNA sequence analysis showed an AATAAA to AATGAA mutation in the polyadenylation signal sequence of the alpha 2-globin gene. Genomic DNA mapping and sequence analysis of a unique polymerase chain reaction product generated across the deletion breakpoint of the maternal allele showed a 5,201-bp deletion extending from 870 nucleotides 5' of the alpha 2-globin gene to nucleotide +519 in the alpha 1-globin gene. This deletion is similar to that previously suggested by blotting studies in a Greek family (Pressley et al, Nucleic Acids Res 8:4889, 1980) and removes the entire alpha 2-globin gene and a portion of the 5' end of the alpha 1-globin gene. Sequence characterization of the resultant aberrant truncated alpha 1-globin gene from the proband showed a 27 nucleotide duplication corresponding to the 3' end of the alpha-globin gene IVS-2 region separated by the insertion of a tetranucleotide (GGTT), suggesting that this deletion is caused by an illegitimate recombination event.

Connexin26 mutations associated with the most common form of non-syndromic neurosensory autosomal recessive deafness (DFNB1) in Mediterraneans.

Non-syndromic neurosensory autosomal recessive deafness (NSRD) is the most common form of genetic hearing loss. Previous studies defined at least 15 human NSRD loci. Recently we demonstrated that DFNB1, located on the long arm of chromosome 13, accounts for approximately 80% of cases in the Mediterranean area. Further analysis with additional markers now identifies several recombinants which narrow the candidate region to approximately 5 cM, encompassed by markers D13S141 and D13S232 and including several ESTs and candidate genes, including the connexin26 (GJB2) gene. Analysis of PCR products from our affected patients' DNA shows two frameshift mutations in the connexin26 gene. Deletion of a G within a stretch of six Gs at position 35 of the GJB2 cDNA (mutation 35delG) leads to premature chain termination and is present in 63% of NSRD chromosomes, demonstrating linkage to chromosome 13. Deletion of a T at position 167 of GJB2 (mutation 167delT), also resulting in premature chain termination, was detected in another patient. Four neutral sequence polymorphisms were also identified. These findings are in agreement with a recent study showing that mutations in the connexin26 gene are associated with genetic forms of deafness in three Pakistani families and that GJB2 is DFNB1. Connexin26 is a member of a large family of proteins involved in formation of gap junctions, which are involved in electrical synapses and the direct transfer of small molecules and ionic currents between neighboring cells. The identification of GJB2 as the DFNB1 gene should provide a better understanding of the biology of normal and abnormal hearing, help form the basis for diagnosis and may facilitate development of strategies for treatment of this common genetic disorder.