Search

Replication-defective amphotropic retrovirus vectors containing either the human beta-globin gene with introns or an intronless beta-globin minigene were constructed and used to study beta-globin expression following gene transfer into hematopoietic cells. The beta-globin genes were marked by introducing a 6-base-pair insertion into the region corresponding to the 5' untranslated region of the beta-globin mRNA to allow detection of RNA encoded by the new gene in human cells expressing normal human beta-globin RNA. Introduction of a virus containing the beta-globin gene with introns into murine erythroleukemia cells resulted in inducible expression of human beta-globin RNA and protein, while the viruses containing the minigene were inactive. The introduced human beta-globin gene was 6 to 110% as active as the endogenous mouse beta maj-globin genes in six randomly chosen cell clones. Introduction of the viruses into human BFU-E cells, followed by analysis of marked and unmarked globin RNAs in differentiated erythroid colonies, revealed that the introduced beta-globin gene was about 5% as active as the endogenous genes in these normal human erythroid cells and that again the minigene was inactive. These data are discussed in terms of the potential treatment of genetic disease by gene therapy.

Experiments were designed to test the effect of introns on gene expression in transgenic mice. Four different pairs of gene constructs, which were identical except that one member of each pair lacked all introns, were compared for expression of mRNA after introduction into the murine germ line by microinjection of fertilized eggs. The expression of two chimeric genes, made by fusing either the mouse metallothionein I or the rat elastase 1 promoter/enhancer to the rat growth hormone gene, was assayed in fetal liver or pancreas, respectively, while two natural genes, an oligonucleotide-marked mouse metallothionein I gene and the human beta-globin gene, were assayed in fetal liver. In each case there was, on average, 10- to 100-fold more mRNA produced from the intron-containing construct. Moreover, mRNA levels were proportional to the relative rates of transcription that were measured in isolated nuclei. However, when the expression of the two mouse metallothionein I gene-based constructs was tested after transfection into cultured cells, little difference was observed. These observations suggest that introns play a role in facilitating transcription of microinjected genes and that this effect may be manifest only on genes exposed to developmental influences.

Adenovirus class B (Ad3 and Ad7) and class C (Ad1, Ad5, and Ad6) late r-strand mRNA's were found to have segmented 5' leaders. These leaders were very similar among serotype within a class but differed in sequence from the leaders on late mRNA's of a different class. However, the leader components of class B viruses mapped at essentially the same map coordinates as those of class C viruses. The 5' coordinates of the main bodies of class B messages to which the tripartite leaders are attached as well as the map positions of several of their early mRNA's were very similar to those of Ad2 transcripts. Infrequent examples of late r-strand polysomal RNAs of Ad3 and Ad7 had, in addition to the three common leader segments, a fourth leader segment derived from RNA encoded at various sites between the second and third leaders. The extra components formed several distinct groups. These molecules are presumably intermediates in the splicing processes that generate mature messages.

The human alpha 1 (I) collagen gene and 48 kilobase pairs of flanking DNA have been isolated on two overlapping cosmids. The alpha 1 (I) gene is 18 kilobase pairs long and contains a single repetitive element of the Alu family; at least 15 repetitive elements are present in the flanking DNA. Analysis of chromatin structure in nuclei isolated from cultured fibroblasts demonstrated a single chromatin domain greater than 65 kilobase pairs in length that contained 9 DNase I-hypersensitive sites. The pattern of hypersensitive sites was also determined in nuclei derived from placental tissue. Five of the DNase I-hypersensitive sites were observed in both placental and fibroblast chromatin including one site near the 5' end and another near the 3' end of alpha 1 (I). An additional two sites located near the 3' end of the alpha 1 (I) gene in fibroblast chromatin are associated with the tissue-specific use of different polyadenylation sites. Two DNase I-hypersensitive sites found only in fibroblast chromatin and one site found only in placental chromatin were located more than 10 kilobase pairs away from the alpha 1 (I) gene and may be related to tissue-specific expression of other genes in the domain. However, the only abundant placental mRNAs from the 65-kilobase pair domain were those transcribed from the alpha 1 (I) gene. These findings suggest that physical linkage does not play a predominant role in controlling coordinate expression of collagen genes.

Cultured dermal fibroblasts from an infant with the lethal perinatal form of osteogenesis imperfecta (type II) synthesize normal and abnormal forms of type I procollagen. The abnormal type I procollagen molecules are excessively modified during their intracellular stay, have a lower than normal melting transition temperature, are secreted at a reduced rate, and form abnormally thin collagen fibrils in the extracellular matrix in vitro. Overmodification of the abnormal type I procollagen molecules was limited to the NH2-terminal three-fourths of the triple helical domain. Two-dimensional mapping of modified and unmodified alpha chains of type I collagen demonstrated neither charge alterations nor large insertions or deletions in the region of alpha 1(I) and alpha 2(I) in which overmodification begins. Both the structure and function of type I procollagen synthesized by cells from the parents of this infant were normal. The simplest interpretation of the results of this study is that the osteogenesis imperfecta phenotype arose from a new dominant mutation in one of the genes encoding the chains of type I procollagen. Given the requirement for glycine in every third position of the triple helical domain, the mutation may represent a single amino acid substitution for a glycine residue. These findings demonstrate further heterogeneity in the biochemical basis of osteogenesis imperfecta type II and suggest that the nature and location of mutations in type I procollagen may determine phenotypic variation.

The sequence of 15,441 nucleotides from the adenovirus-2 genome has been determined and includes the regions between coordinates 0-32% and 89-100%. These regions contain the early (E) transcription units E1A, E1B, E2B, and E4, the genes for polypeptides IVa2 and IX, the COOH terminus of fiber polypeptide, as well as the two virus-associated RNAs and the leader sequences for the major late mRNAs. Analysis of tryptic peptides from the terminal protein and its precursor (Smart, J. E., and Stillman, B. W. (1982) J. Biol. Chem. 257, 13499-13506) has allowed the gene for the precursor terminal protein to be positioned between coordinates 28.0 and 23.5 on the 1-strand. A minimum Mr = 74,500 is predicted. A second, longer open reading frame is also found on the 1-strand between coordinates 22.9 and 14.2 and predicts a polypeptide of at least Mr = 120,000. Many open reading frames longer than 10,000 exist within this sequence although less than half of them can be assigned to previously characterized polypeptides. As with other viral genomes, the available coding information is highly compressed. Intergenic distances are very short and examples are found of genes which overlap either on the same strand or the complementary strand.

Several lines of evidence have suggested that the regulation of type I collagen gene transcription is complex and that important regulatory elements reside 5' to, and within, the first intron of the alpha 1(I) gene. We therefore sequenced a 2.3-kilobase HindIII fragment that encompasses 804 base pairs of 5' flanking sequence, the first exon, and most of the first intron of the alpha 1(I) human collagen gene. A 274-base-pair intronic sequence, flanked by Ava I sites (A274), contained a sequence identical to a high-affinity decanucleotide binding site for transcription factor Sp1 and a viral core enhancer sequence. DNase I protection experiments indicated zones of protection that corresponded to these motifs. When A274 was cloned 5' to the chloramphenicol acetyltransferase (CAT) gene, driven by an alpha 1(I) collagen promoter sequence, and expression was assessed by transfection, significant orientation-specific inhibition of CAT activity was observed. This effect was most apparent in chicken tendon fibroblasts, which modulate their level of collagen synthesis in culture. We propose that normal regulation of alpha 1(I) collagen gene transcription results from an interplay of positive and negative elements present in the promoter region and within the first intron.

To understand the nature of the mutation in type I collagen genes in cells from an infant with the perinatal lethal form of osteogenesis imperfecta (type II), we cloned and sequenced almost 2 kilobases of a normal alpha 1(I) collagen gene and the corresponding region of a mutant alpha 1(I) gene from cell strain CRL 1262. The mutant gene had undergone recombination between two non-homologous introns, which resulted in the loss of three exons coding for 84 amino acids in the triple-helical domain. The deletion predicted the loss of amino acid residues surrounding and including the methionine at the junction between the CNBr peptides alpha 1(I) CB8 and alpha 1(I) CB3, a result confirmed by analysis of the cleavage peptides from the product of the mutant gene. Although large deletions from collagen genes are uncommon causes of the osteogenesis imperfecta type II phenotype, analysis of the de novo change in gene structure in this cell strain suggests that similar rearrangements may have occurred during the evolution of the large collagen genes.

Single base substitutions have been identified in the promoter regions of A gamma-globin genes from individuals with certain types of nondeletion A gamma hereditary persistence of fetal hemoglobin (HPFH). The presence of these mutations is closely associated with the A gamma HPFH phenotype, but proof that they are the nondeletion HPFH determinants is lacking. To test directly whether these base substitutions can result in an increase in A gamma-globin gene transcription, we studied cosmid clones containing the G gamma- through beta-globin gene regions from individuals with Greek-type (G-to-A base substitution at -117) and Chinese-type (C-to-T base substitution at -196) A gamma HPFH in a transient expression assay. When tested as part of a cosmid clone, the Greek HPFH A gamma-globin gene consistently produced about 1.4 times as much RNA as the wild-type A gamma-globin gene when standardized against RNA transcribed from the G gamma genes in cis. The relative strengths of the normal and HPFH A gamma-globin gene promoters were also compared in transient expression assays with plasmids containing the A gamma-globin genes. Pseudo-wild-type A gamma-globin genes containing a short, transcriptionally neutral deletion were used so that two A gamma-globin genes that differed in their promoter sequences could be compared in the same transfection. The plasmid transient expression results indicated a 1.3- to 1.4-fold increase in steady-state RNA levels from the Greek-type A gamma HPFH promoter compared with the wild-type A gamma promoter, while no difference was documented between the Chinese-type A gamma HPFH promoter and the wild-type A gamma promoter.

Regulated expression of the human beta-globin gene has been demonstrated in cultured murine erythroleukemia cells and in mice after retrovirus-mediated gene transfer. However, the low titer of recombinant viruses described to date results in relatively inefficient gene transfer, which limits their usefulness for animal studies and for potential gene therapy in humans for diseases involving defective beta-globin genes. We found regions that interfered with virus production within intron 2 of the beta-globin gene and on both sides of the gene. The flanking regions could be removed, but intron 2 was required for beta-globin expression. Inclusion of beta-globin introns necessitates an antisense orientation of the gene within the retrovirus vector. However, we found no effect of the antisense beta-globin transcription on virus production. A region downstream of the beta-globin gene that stimulates expression of the gene in transgenic mice was included in the viruses without detrimental effects on virus titer. Virus titers of over 10(6) CFU/ml were obtained with the final vector design, which retained the ability to direct regulated expression of human beta-globin in murine erythroleukemia cells. The vector also allowed transfer and expression of the human beta-globin gene in hematopoietic cells (CFU-S cells) in mice.

Replication-competent retroviruses can be modified to carry nonviral genes. Such gene transfer vectors help define regions of the retroviral genome that are required in cis for retroviral replication. Moloney murine leukemia virus has been used extensively in vector construction, and all of the internal protein-encoding regions can be removed and replaced with other genes while still allowing production of virions containing and transmitting the altered retroviral genome. However, inclusion of a portion of the gag region from Moloney murine leukemia virus markedly increases the titer of virus derived from these vectors. We determined that this effect was due to more efficient packaging of the vector RNA into particles and did not depend on protein synthesis from the gag region. We conclude that the retrovirus packaging signal extends into the gag region. We have found that retroviral vectors containing the complete packaging signal allow more efficient gene transfer into a variety of cell types. In addition, these results may help explain why many oncogenic retroviruses have retained gag sequences and often express transforming proteins that are gag-onc hybrids.

In order to study the relationships among mammalian alpha-globin genes, we have determined the sequence of the 3' flanking region of the human alpha 1 globin gene and have made pairwise comparisons between sequenced alpha-globin genes. The flanking regions were examined in detail because sequence matches in these regions could be interpreted with the least complication from the gene duplications and conversions that have occurred frequently in mammalian alpha-like globin gene clusters. We found good matches between the flanking regions of human alpha 1 and rabbit alpha 1, human psi alpha 1 and goat I alpha, human alpha 2 and goat II alpha, and horse alpha 1 and goat II alpha. These matches were used to align the alpha-globin genes in gene clusters from different mammals. This alignment shows that genes at equivalent positions in the gene clusters of different mammals can be functional or nonfunctional, depending on whether they corrected against a functional alpha-globin gene in recent evolutionary history. The number of alpha-globin genes (including pseudogenes) appears to differ among species, although highly divergent pseudogenes may not have been detected in all species examined. Although matching sequences could be found in interspecies comparisons of the flanking regions of alpha-globin genes, these matches are not as extensive as those found in the flanking regions of mammalian beta-like globin genes. This observation suggests that the noncoding sequences in the mammalian alpha-globin gene clusters are evolving at a faster rate than those in the beta-like globin gene clusters. The proposed faster rate of evolution fits with the poor conservation of the genetic linkage map around alpha-globin gene clusters when compared to that of the beta-like globin gene clusters. Analysis of the 3' flanking regions of alpha-globin genes has revealed a conserved sequence approximately 100-150 bp 3' to the polyadenylation site; this sequence may be involved in the expression or regulation of alpha-globin genes.

Murine bone marrow was infected with a high-titer retrovirus vector containing the human beta-globin and neomycin phosphotransferase genes. Anemic W/Wv mice were transplanted with infected marrow which in some cases had been exposed to the selective agent G418. Human beta-globin expression was monitored in transplanted animals by using a monoclonal antibody specific for human beta-globin polypeptide, and hematopoietic reconstitution was monitored by using donor and recipient mice which differed in hemoglobin type. In some experiments all transplanted mice expressed the human beta-globin polypeptide for over 4 months, and up to 50% of peripheral erythrocytes contained detectable levels of polypeptide. DNA analysis of transplanted animals revealed that virtually every myeloid cell contained a provirus. Integration site analysis and reconstitution of secondary marrow recipients suggested that every mouse was reconstituted with at least one infected stem cell which had extensive repopulation capability. The ability to consistently transfer an active beta-globin gene into mouse hematopoietic cells improves the feasibility of using these techniques for somatic cell gene therapy in humans.