Francisco E. Baralle

The structure and evolution of the human β-globin gene family

Argiris Efstratiadis Department of Biological Chemistry Harvard Medical School Boston, Massachusetts 02115 James W. Posakony, Tom Maniatis, Richard M. Lawn* and Catherine O’Connell+ Division of Biology California Institute of Technology Pasadena, California 91125 Richard A. Spritz, Jon K. DeRiel,# Bernard G. Forget and Sherman M. Weissman Departments of Genetics and Internal Medicine Yale University School of Medicine New Haven, Connecticut 06510 Jerry L. Slightom, Ann E. Blechl and Oliver Smithies Laboratory of Genetics University of Wisconsin Madison, Wisconsin 53706 Francisco E. Baralle, Carol C. Shoulders and Nicholas J. ProudfootQ MRC Laboratory of Molecular Biology Hills Road Cambridge CB2 2QH, England Summary We present the results of a detailed comparison of the primary structure of human p-like globin genes and their flanking sequences. Among the se- quences located 5’ to these genes are two highly conserved regions which include the sequences ATA and CCAAT located 31 2 1 and 77 + 10 bp, respectively, 5’ to the mRNA capping site. Similar sequences are found in the corresponding locations in most other eucaryotic structural genes. Calcula- tion of the divergence times of individual @like globin gene pairs provides the first description of the evolutionary relationships within a gene family based entirely on direct nucleotide sequence com- parisons. In addition, the evolutionary relationship of the embryonic e-globin gene to the other human P-like globin genes is defined for the first time. Finally, we describe a model for the involvement of short direct repeat sequences in the generation of deletions in the noncoding and coding regions of B-like globin genes during evolution.

The primary structure of the human ϵ-globin gene

Abstract We describe the complete nucleotide sequence of the human ϵ-globin gene including 387 nucleotides of 5′ flanking sequence and 301 nucleotides of 3′ flanking sequence. The arrangement of coding, noncoding and intervening sequences in this gene is entirely consistent with its identification as the embryonic β -like globin gene.

Localization of the cellular-fibronectin-specific epitope recognized by the monoclonal antibody IST-9 using fusion proteins expressed in E. coli

Here we report on a monoclonal antibody (IST‐9) which distinguishes between human cellular and plasma fibronectin. Using β‐galactosidase‐fibronectin fusion proteins expressed in E. coli we have demonstrated that this monoclonal antibody is specific for a fibronectin segment (ED) which can be included or omitted from the molecule depending on the pattern of splicing of the mRNA precursors. Furthermore, using the same fusion proteins we have been able to localize precisely the epitopes of two other monoclonal antibodies (IST‐1 and IST‐2), specific for the heparin‐binding domain 5 of fibronectin.

Mapping the collagen-binding site of human fibronectin by expression in Escherichia coli.

The collagen‐binding domain of human fibronectin has been expressed as a cro/beta‐galactosidase fusion protein in Escherichia coli. The hybrid polypeptide was recognized by an anti‐(human plasma fibronectin) serum and bound specifically to gelatin‐Sepharose. The collagen‐binding region was subdivided by constructing a series of overlapping bacterial expression plasmids. The fusion proteins produced by these constructs were analysed for gelatin‐binding activity. The results indicate that the binding site lies within an approximately 12.5 kd fragment of fibronectin, and show that the following 14 amino acid sequence is critical for gelatin‐binding activity: Ala‐Ala‐His‐Glu‐Glu‐Ile‐Cys‐Thr‐Thr‐Asn‐Glu‐Gly‐Val‐Met. This sequence links the second type II homology unit with the adjacent type I repeat in the amino‐terminal third of the fibronectin molecule.

Expression of a human alpha-globin/fibronectin gene hybrid generates two mRNAs by alternative splicing.

We have isolated genomic clones for human fibronectin (FN), by screening a human gene library with previously isolated FN cDNA clones. We have recently reported two different FN mRNAs, one of them containing an additional 270 nucleotide insert coding for a structural domain ED. Restriction mapping and DNA sequencing of the genomic clones show that the ED type III unit corresponds to exactly one exon in the gene, whilst the two flanking type III units are split in two exons at variable positions. When an alpha‐globin/FN gene hybrid construct, containing the ED exon, flanking introns and neighbouring FN exons, is transfected into HeLa cells, two hybrid mRNAs differing by the ED exon are synthesized. These experiments confirmed that the two FN mRNAs observed in vivo arise from the same gene by alternative splicing.

Isolation and characterization of cDNA clones for human liver fibronectin.

Cellular and plasma fibronectin dimers are constituted by similar but not identical polypeptides. Their differences are the consequence of internal primary sequence variability due to alternative splicing in at least 2 regions (ED and IIICS) of the pre mRNAs [1–8]. A detailed analysis of human liver fibronectin mRNA in these regions was carried out by isolating cDNA clones and determining their nucleotide sequence. A novel type of IIICS segment (coding for 64 amino acids) was present in the two cDNA clones studied and, as expected from previous Sl mapping studies [6], the ED segment was absent in both.

Donor and acceptor splice signals within an exon of the human fibronectin gene: a new type of differential splicing

We have sequenced that area of a human fibronectin gene clone which codes for a connecting strand separating the last two areas of the type III homology. The gene has a complex exon with two ‘AG’ acceptor sites. One of these can be used (exon subdivision). In addition 93 basepairs inside the exon are sometimes spliced out as an intron. This is the third differential splicing found in the fibronectin gene transcript and it represents a new type of differential splicing.

A family of fibronectin mRNAs in human normal and transformed cells

Previously, two fibronectin mRNAs, generated by alternative splicing of the extra domain (ED) and type III connecting segments (IIICS) sequences, have been described in a human transformed cell line and in human liver, respectively. We now report on a family of fibronectin mRNAs identified by Northern blotting analysis in two normal human fibroblast strains (HEL 299 and Flow 7000) and five transformed cell lines (8387 and HT-1080, fibrosarcomas; G-361, melanoma; JEG-3, choriocarcinoma; and RD, rhabdomyosarcoma). Seven different fibronectin mRNA forms with electrophoretic mobilities ranging between 8.6 and 7.7 kb were identified. Each cell line contains three (HEL 299, Flow 7000 and 8387) or two (HT-1080, G-361, JEG-3 and RD) fibronectin mRNAs species with characteristic size. In all cell lines we detected one fibronectin mRNA form which lacks the ED sequence (ED- fibronectin mRNA) and one or two fibronectin mRNAs containing it (ED+ fibronectin mRNA). These data show that the presence of ED+ and ED- fibronectin mRNAs is a general feature of all cells tested. Moreover, the fibronectin mRNA pattern is characteristic of the cell type analyzed, suggesting the occurrence of specifically programmed splicing mechanisms in each cell line.

GENETIC MARKER IN APOLIPOPROTEIN AI/CIII GENE COMPLEX ASSOCIATED WITH HYPERCHOLESTEROLAEMIA

In 1983 we reported a strong association between hypertriglyceridaemia and a restriction fragment length polymorphism (RFLP) associated with the apolipoprotein AI gene which was not present in normolipidaemic individuas (Rees et al. 1983). This RFLP arises because of the existence of a polymorphic nucleotide in the 3’ non-coding region of the linked apo-CIII gene that creates an additional cleavage site for the restriction enzyme SacI (Shoulders and Baralle 1986). Since our original observation, two groups have reported a significant increase in the incidence of the variant apo-AI/CIII allele (S2) in patients with severe coronary heart disease (CHD), and in survivors of myocardial infarction compard with healthy controls (Ferns et al. 1985; Rees et al. 1985). However, this increased frequency could not entirely be explained by a greater number of hypertriglyceridaemic individuals in these groups. In view of these observations and the fact that Kessling and Humphries (1985) could not confirm our initial findings (Rees et al. 1983), we have tried to clarify the relationship between the S2 allele and hyperlipidaemia by studying newly recruited patients with well-definied forms of hyperlipidaemia.

Variation in the apo AI/CIII/AIV gene complex: its association with hyperlipidemia

This study shows that 33.3% of English patients with primary hyperlipidemia (52/156) had the S2 allele of the apo AI/CIII/AIV complex compared to 6.1% of normolipidemic individuals (3/49). The increased frequency of the allele was statistically significant in each of the hyperlipidemic groups (type IIA, excluding patients with FH, type IIB and IV) examined and was not specifically related to hypertriglyceridemia. This finding may account for the result of several studies which showed groups of patients with CHD had a significantly higher prevalence of the S2 allele than control groups. Our data do not support the notion that the increased frequency of this allele in CHD patients is independent of variations in plasma lipid levels, since we find the frequency of the S2 allele in an apparently healthy hyperlipidemic group of patients is very similar to a hyperlipidemic group with symptomatic premature atherosclerotic disease. This study also shows the BMI of the type IIB and IV hyperlipidemic patients is significantly higher than the type IIA (no xanthomas) group. This may modulate the expression of the defect associated with the S2 allele. When the type IIB and IV hyperlipidemic groups were divided into 2 groups according to their apo AI/CIII/AIV genotype (i.e., S1S1, S1S2 (including S2S2] there was no significant difference in the mean plasma level of total cholesterol, HDL-cholesterol and triglycerides between the 2 groups. In contrast the S1S2 type IIA individuals had higher plasma cholesterol levels.