Molecular Cloning of the Human α-Globin Gene Family Citation Lauer, Joyce Ellen (1981) Molecular Cloning of the Human α-Globin Gene Family. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/tdac-ny02. https://resolver.caltech.edu/CaltechTHESIS:09082025-201021359 Abstract During human development a series of α-like and β-like subunits of hemoglobin are produced. Five β-like polypeptides, embryonic (Ɛ), fetal ( G γ, A γ) and adult (δ, β), and two α-like polypeptides, embryonic (z;) and adult (α), have been identified. A structural analysis of the gene family encoding the human α-like globins is described. An improved gene isolation procedure was developed. This procedure makes it possible to isolate genes and their flanking sequences, and thus to directly determine the linkage relationship between genes, without requiring partial purification of genes. Large random genomic DNA fragments resulting from physical shear of DNA were joined to phage lambda vectors using synthetic DNA linkers. The resulting recombinant DNA molecules were packaged in vitro into viable phage to yield a collection of cloned overlapping DNA fragments which were then amplified to produce a permanent library of genomic DNA sequences. This library can be screened repeatedly using nucleic acid probes and an in situ plaque hybridization procedure in order to isolate many different genes. Clones containing the duplicated human α-globin genes (a1 and a2) were isolated from a library of human DNA. Also present on these clones are an α-like pseudogene (ψα1) and an embryonic α-like gene (ς1) which were identified by blot hybridization experiments and DNA sequence analysis. Genomic blotting using the ς1 coding sequence as probe identified a second embryonic gene (ς2). The ς2 gene was isolated by cloning a DNA fragment which overlaps the clones containing the other α-like genes. All five genes are transcribed from the same DNA strand and are arranged in the order 5'-ς2-ς1ψα1-α2-α1-3'. Comparison of α1 and α2 by restriction mapping and heteroduplex analysis of DNA fragments containing al or α2 plus 5' flanking sequences demonstrated that each gene is located within an approximately 4 kb region of homology interrupted by two short regions of nonhomology. The association of these large blocks of homology with genes which are thought to have duplicated long ago suggests the existence of a mechanism for sequence matching. Two types of deletions invariably occur during propagation of clones containing α1 and α2. The breakpoints of these two types of deletions are located within the two blocks of al-a2 homology. The positions and the precise lengths of these deletions indicate that deletion occurs by homologous but unequal crossing-over between corresponding regions of al and a2. The lengths and positions of these deletions are indistinguishable from those of the two types of deletions which are associated with α-thalassemia 2, suggesting that this common genetic disease results from homologous but unequal crossing-over between regions within and/or surrounding the adult α-globin genes. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Biology) Degree Grantor: California Institute of Technology Division: Biology Major Option: Biology Thesis Availability: Public (worldwide access) Research Advisor(s): Maniatis, Thomas P. Thesis Committee: Unknown, Unknown Defense Date: 24 July 1980 Funders: Funding Agency Grant Number National Institutes of Health (NIH) UNSPECIFIED Jean Weigle Memorial Fund UNSPECIFIED Record Number: CaltechTHESIS:09082025-201021359 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:09082025-201021359 DOI: 10.7907/tdac-ny02 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17672 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 12 Sep 2025 10:16 Last Modified: 12 Sep 2025 10:22 Thesis Files PDF - Final Version See Usage Policy. 32MB Repository Staff Only: item control page

MOLECULAR CLONING OF THE HUMAN ~GLOBIN GENE FAMILY Thesis by Joyce Ellen Lauer In Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1981 (Submitted July 2 4, 198 0) ii Quoted by Denise Levertov in With Eyes at the Back of Our Heads The true artist: draws out all from his heart, works with delight, makes things with calm, with sagacity, works like a true Toltec, composes his objects, works dextrously, invents, arranges materials, adorns them, makes them adjust The carrion artist: works at random, sneers at the people, makes things opaque, brushes .across the surface of the face of things, works without care, defrauds people, is a thief iii Acknowledgements I am grateful to Tom Maniatis for kindness, guidance, and conversations, and for teaching me about DNA and cloning. It has been a wonderful privilege to be a member of this lab. iv Acknowledgements I was supported by a training grant from the National Institutes of Health and received funds for the preparation of this thesis from the Jean Weigle Memorial Fund. V Abstract During human development a series of a-like and 13-like subunits of hemoglobin are produced. Five 13-like polypeptides, embryonic (e:), fetal (G y, Ay) and adult (c, 13), and two a-like polypeptides, embryonic (z;) and adult (a), have been identified. A structural analysis of the gene family encoding the human a-like globins is described. An improved gene isolation procedure was developed. This procedure makes it possible to isolate genes and their flanking sequences, and thus to directly determine the linkage relationship between genes, without requiring partial purification of genes. Large random genomic DNA fragments resulting from physical shear of DNA were joined to phage lambda vectors using synthetic DNA linkers. The resulting recombinant DNA molecules were packaged in vitro into viable phage to yield a collection of cloned overlapping DNA fragments which were then amplified to produce a permanent library of genomic DNA sequences. This library can be screened repeatedly using nucleic acid probes and an in situ plaque hybridization procedure in order to isolate many different genes. Clones containing the duplicated human a-globin genes (al and a2) were isolated from a library of human DNA. Also present on these clones are an a-like pseudogene (\jJ al) and an e mbryonic a-like gene (z; 1) which were identified by blot hybridization experiments and DNA sequence analysis. Genomic blotting using the z; 1 coding sequence as probe identified a second embryonic gene (z; 2). The z; 2 gene was isolated by cloning a DNA fragment which overlaps the clones containing the other alike genes. All five genes are transcribed from the same DNA strand and are arranged in the order 5'-r; 2-1; 1-\/Ja l- a2-al-3'. Comparison of al and a2 by restriction mapping and heteroduplex analysis of DNA fragments containing al or a2 plus 5' flanking sequences demonstrated that each gene is located within an approximately 4 kb region of homology interrupted by two short regions of nonhomology. The association of these large blocks of homology with vi genes which are thought to have duplicated long ago suggests the existence of a mechanism for sequence matching. Two types of deletions invariably occur during propagation of clones containing al and a2. The breakpoints of these two types of deletions are located within the two blocks of al-a2 homology. The positions and the precise lengths of these deletions indicate that deletion occurs by homologous but unequal crossing-over between corresponding regions of al and a2. The lengths and positions of these deletions are indistinguishable from those of the two types of deletions which are associated with a-thalassemia 2, suggesting that this common genetic disease results from homologous but unequal crossing-over between regions within and/or surrounding the adult a-globin genes. vii Table of Contents I:rrtroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 1: The isolation of structural genes from libraries of eukaryotic DNA Chapter 2: . . . . . . . . . . . . . . . . . . . . . . . 12 The chromosomal arrangement of human Cl-like globin genes: Sequence homology and Cl-globin gene deletions . . . . . . . . . . 28 Appendix: Isolation of the i; 2 globin gene . . . Chapter 3: 1 The molecular genetics of human hemoglobins . . 41 . . . . . . . . . . 45 1 Introduction The human globin gene family is a paradigm for studying differential gene activity during development and the molecular basis of inherited disorders in gene expression. Globin polypeptides and genes from a number of species have been characterized (Tilghman et al., 1977; Lawn et al., 1978; Dodgson et al., 1979; Lacy et al., 1979) and cloned globin genes from any of these species can be manipulated in various ways in order to study gene expression (Mulligan et al., 1979; Weil et al., 1979; Wigler et al., 1979; Manley et al., 1980). Unique to the human globin gene system, however, is the existence of a large number of well-characterized natural mutants in globin gene expression (for review, see Maniatis et al., 1980, which is presented as Chapter 3 of this thesis). Mutants analyzed thus far include examples of defects in transcription, RNA processing and translation, although in most cases the exact DNA sequence alteration responsible for the defect remains to be identified. Particularly intriguing is a class of mutants, of which numerous examples are known in both the a-like and 8-like globin gene families, in which deletion of DNA in one region of a cluster of linked globin genes disrupts the regulation of a globin gene located many kilobases (kb) distant from the deletion. Whereas DNA sequence alterations which affect globin gene transcription or globin translation are perhaps analogous to well-understood prokaryotic examples, this class of deletion mutants sugges ts the existence of mechanisms of gene regulation unique to eukaryotes (for discussion and review, see Maniatis et al., 1980). As a first step towards studying the various aspects of eukaryotic gene regulation, we have isolated the human a-like and 8-like globin gene clusters. This thesis describes our approach to gene isolation, and structural characterization of the a-like globin gene cluster. 2 The Ontogeny of Human Globin Gene Expression The a-like and 8-like subunits of hemoglobin (Hb) are encoded by a small group of genes which are expressed sequentially during development (Weatherall and Clegg, 1979). The earliest embryonic hemoglobin tetramer, Gower 1, consists of e: ( 8like) and z; (a-like) polypeptide chains. Beginning at approximately eight weeks of gestation, the embryonic chains are gradually replaced by the adult a-globin chain and two different fetal 8-like chains, designated G y and Ay. The y chains differ only in the presence of glycine or alanine at position 136, respectively. During the transition period between embryonic and fetal development, Hb Gower 2 (a2 e: 2) and Hb Portland (z; y 2) are detected. Hb F (a2 y 2) eventually becomes the predominant Hb tetramer 2 throughout the remainder of fetal life. Beginning just prior to birth, the y-globin chains are gradually replaced by the adult 8- and o-globin polypeptides. At six months after birth, 97-98% of the hemoglobin is Hb A (a S ), while Hb A (a o 2) accounts for 2 2 2 2 approximately 2%. a-Like Globin Gene Number The existence of multiple human a-like globin genes was established on the basis of genetic studies, analysis of the structure and function of embryonic hemoglobins, and the isolation of the a-like globin gene cluster by molecular cloning. Genetic and structural studies of a-globin polypeptide chain variants provided the first evidence for duplication of the human a-globin genes (Lehmann and Carrell, 1968; Brimhall et al., 1970; Lie-Injo et al., 197 4; for reviews see Weatherall and Clegg, 1976; Bunn et al., 1977; Weatherall and Clegg, 1979). The hypothesis of a-globin gene duplication was strongly supported by cDNA/DNA solution hybridization experiments (Kan et al., 1975). In most cases the two genes are expressed at approximately the same level, since in individuals heterozygous for an a-chain variant, the variant polypeptide usually represents approximately one-quarter of the a-globin protein (Lehmann and Carrell, 1968). Since only one a-globin amino acid sequence is found in 3 normal individuals (Dayhoff, 1972), the two a-globin genes must encode identical polypeptides. Until relatively recently, it was thought that there was only one a-type globin polypeptide, which was produced throughout all stages of human development. This contrasted to the 8-like globin gene family, where two switches in globin gene expression, from embryonic to fetal and from fetal to adult, had been identified. The misconception about the number of a-type globin polypeptides resulted from the hypothesis that the earliest embryonic hemoglobin, Gower 1, consisted of a tetramer of epsilon chains (e:4) (Huehns et al., 1961, 1964). The embryonic z;-globin polypeptide chain was first identified as a component of Hb Portland (z; y ) (capp et al., 1967, 2 2 1970). Peptide comparisons suggested that z; is an a-like chain (Kamuzora and Lehmann, 1975; Huehns and Farooqui, 1975). On the basis of the oxygen dissociation behavior of embryonic blood, it was proposed that the structure of Gower 1 is in fact z; 2 e: 2 (Huehns and Farooqui, 1975). Peptide analysis of purified Gower 1 confirmed this hypothesis (Gale et al., 1979). Although only one z;-globin polypeptide sequence has been identified (J. Clegg, personal communication), other evidence indicates the existence of two highly homologous z;-globin genes (see below). Genetic Disorders in ct-Globin Gene Expression Inherited disorders in a-globin gene expression (a-thalassemias) usually result from a-globin gene deletion. One or both a-globin genes on a chromosome can be deleted, and diploid combinations of the "single-gene" and "zero-gene" chromosomes cause a variety of syndromes whose severity reflects the number of a-globin genes affected (for review, see Maniatis et al., 1980). The apparent cause of the deletions which generate the "single-gene" or a-thalassemia 2 chromosome, and the association of "zero-gene" chromosomes with alterations in z;-globin gene regulation, are described below. 4 Gene Isolation by Molecular Cloning Globin genes were chosen as a model system in which to study the regulation of gene expression in eukaryotes. The initial phase of the research described in this thesis was the development of a modified procedure which makes it possible to efficiently isolate genes and th.eir flanking sequences and to determine the linkage relationship between genes. In this procedure, DNA is fragmented randomly and high molecular weight (16-20 kb) DNA fragments are isolated. These are inserted into a bacteriophage lambda vector using DNA linkers and the recombinant DNA molecules are packaged in vitro into viable phage to yield a collection of cloned overlapping DNA fragments. These recombinants are amplified in order to establish a permanent library of DNA segments representing the entire genome. For such a library to be truly representative, it is desirable that the initial DNA fragmentation be as random as possible, and this is best accomplished by physically shearing the DNA. Sheared DNA molecules are likely to have single- stranded tails which mus t be r emoved using S1 nuclease before linkers can be attached by flush-end ligation. I constructed a library of Drosophila melanogaster DNA by this method. The construction a~d characterization of t he Drosophila library are described in Chapter 1 (Maniatis et al., 1978). Trimming DNA molecules using Sl nuclease is a relatively inefficient process, leading to low cloning effic iency. As a r esult, subsequent library constr uction in our lab employed combined p,artial digestion with several restriction enzymes that generate flush ends which can be efficiently ligated to DNA linkers. The human DNA library from which a-globin genes were isolated was constructed by the partial digestion method (Lawn et al., 1978; Maniatis et al., 1978). There are several advantages to the library approach to gene isolation: 1) If numerous gene-specific probes or a mixed (e.g., cDNA) probe is available, a number of genes can be isolated following a single cloning step; 2) Because the genomic DNA 5 inserts are large (16-20 kb), one can isolate linked genes for which no probe is available by their presence in the same clone as a gene for which there is a probe. The isolation of r;; 1 and of 1\Jal illustrate this (see below); 3) By using the terminal fragment of a previously-isolated clone, one can rescreen a library in order to isolate an adjacent region of DNA. The isolation of r;;2 by screening with the terminal fragment of )..HaG 1 illustrates this. (Although problems with the original human library necessitated construction of a second, Bgl II library, this point is still valid.) Structural Analysis of the Human a-Like Globin Gene Cluster Structural analysis of the human a-like globin gene cluster, as derived from cloned DNA, is presented in Chapter 2 (Lauer et al., 1980, and Appendix). The results of this analysis may be summarized as follows. 1. The a-like genes are linked in the order 5'-r;;2-r;;l-lj)al-a2-al-3' which parallels the 5' to 3' embryonic-fetal-adult order of genes within the S-like globin gene cluster (Fritsch et al., 1980; Figure 1). 2. 1jJ al is a pseudogene for which no globin polypeptide can be identified; it probably resulted from gene duplication followed by divergence. Analysis of the complete nucleotide sequence of lj)al (Proudfoot and Maniatis, 1980) reveals sequence alterations which would prevent the production of an a-globin polypeptide. 3. Fine-structure mapping indicated extremely close sequence homology, except for a small insertion, between the coding, intervening and immediately-adjacent flanking regions of al and a2. Furthermore, the repeated positions of certain more distant restriction enzyme sites, and the invariable occurrence of two types of deletions during propagation of elones containing al and a2, suggested the existence of additional regions of homology within the several kb 5' to each gene. To verify this, I collaborated with J. Shen to examine heteroduplexes between DNA fragments containing al or a2 plus 5' flanking sequences. Each gene is located within an approximately 4 kb region of homology interrupted by two short regions of non- 6 homology. The association of these large blocks of homology with genes which are thought to have duplicated long ago suggests the existence of a mechanism for sequence matching. 4. The breakpoints of the two types of deletions which occur in cloned DNA are located within the two blocks of al-a2 homology. The positions and the precise lengths of these deletions indicate that deletion occurs by homologous but unequal crossing-over between corresponding regions of al and a2. The lengths and positions of these deletions are indistinguishable from those of the two types of deletions which are associated with a-thalassemia 2 (Embury et al., 1980), suggesting that this common genetic disease results from homologous but unequal crossing-over between regions within and/or surrounding the adult a-globin genes. 5. Using a fragment of the embryonic 1; 1 gene to probe genomic blots, I identified a second embryonic gene, 1; 2, located nearly 12 kb 5' to 1; 1. Hybridization-melting experiments indicated very close sequence homology between these two genes, which recalls the high degree of cd-a2 homology. 6. 1; 1 had been identified by the correspondence between its DNA sequence and the sequence of I; polypeptide. Using my genomic blotting map of 1; 2, this second gene was shown to be functional by blotting DNA from an infant with the a-thalassemia syndrome, hydrops fetalis (Pressley et al., 1980). In this individual, there is a homozygous deletion of al, a2, iµal and I; 1, but I; 2 remains and I; polypeptide is produced in large amount. Ordinarily 1;-globin is not detectable after 10 weeks gestation. The persistence of 1;-globin production in cases of hydrops fetalis appears to be the a-locus analogue of persistent expression of y-globin caused by the 8-locus disease HPFH. In this syndrome, a DNA rearrangement many kb 3' to the y genes removes a control which would ordinarily switch off these genes during the normal developmental program (for review, see Maniatis et al., 1980). 7 References Brimhall, B., Hollan, S., Jones, R., Koler, R., Stocklen, z. and Szelenyi, J. (1970). Multiple alpha-chain loci for human hemoglobin. Clin. Res. 18, 184. Bunn, H. F., Forget, B. G. and Ranney, H. M. (1977). Human Hemoglobins. Philadelphia: W. B. Saunders Co. Capp, G., Rigas, D. and Jones, R. (1967). Hemoglobin Portland 1: a new human hemoglobin unique in structure. Science 157, 65-66. Capp, G., Rigas, D. and Jones, R. (1970). Evidence for a new hemoglobin chain (r,;chain). Nature 228, 278-280. Dayhoff, M. O. (1972). Atlas of Protein Sequence and Structure. Washington, D.C.: National Biomedical Research Foundation. Dodgson, J. B., Strommer, J. and Engel, J. D. (1979). Isolation of the chicken 8-globin gene and a linked embryonic 8-like globin gene from a chicken DNA recombinant library. Cell 17, 879-887. Embury, S. H., Miller, J. A., Cozy, A. M., Kan, Y. W., Chan, V. and Todd, D. (1980). Two different molecular organizations account for the single a--globin gene of the a-thalassemia 2 genotype. Manuscript submitted. Fritsch, E. F., Lawn, R. M. and Mania t is, T. (1980). Molecular cloning and characterization of the human 8-like globin gene cluster. Cell 19, 959-972. Gale, R., Clegg, J. and Huehns, E. (1979). Human embryonic hemoglobins Gower 1 and Gower 2. Nature 280, 162-164. Huehns, E., Dance, N., Beaven, G., Keil, J., Hecht, F. and Motulsky, A. (1964) Human embryonic hemoglobins. Nature 201, 1095-1097. Huehns, E., Flynn, F., Butler, E. and Beaven, G. (1961). Two new hemoglobin variants in a very young human embryo. Nature 189, 496-497. Huehns, E. and Farooqui, A. (1975). Oxygen dissociation properties of human embryonic red cells. Nature 254, 335-337. 8 Kamuzora, H. and Lehmann, H. (1975). Human embryonic hemoglobins including a comparison by homology of the human z:; and a chains. Nature 256, 511-513. Kan, Y. w., Dozy, A., Varmus, H., Taylor, J., Holland, J., Lie-lnjo, L., Ganesan, J. and Todd, D. (1975). Deletion of a-globin genes in hemoglobin H disease demon- strates multiple o-globin structural loci. Nature 255: 255-256. Lacy, E., Hardison, R. c., Quon, D. and Maniatis, T. (1979). The linkage arrangement of four rabbit 8-like globin genes. Cell 18, 1273-1283. Lauer, J., Shen, c.-K. J. and Maniatis, T. (1980). The chromosomal arrangement of human a-like globin genes: sequence homology and a-globin gene deletions. Cell 20, 119-130. Lawn, R. M., Fritsch, E. F., Parker, R. c., Blake, G. and Maniatis, T. (1978). The isolation and chara cterization of linked o- and 8-globin genes from a cloned library of human DNA. Cell 15, 1157-1174. Lehmann, H. and Carrell, R. (1968). Differences between a- and 8-chain mutants of human hemoglobin and between a- and 8-thalassemia. Possible duplication of the a-chain gene. Br. Med. i!..· 4, 748-750. Lie-lnjo, L., Ganesan, J., Cl~gg, J. and Weatherall, D. (1974). Homozygous state for Hb Constant Spring (slow moving Hb X components). Blood 43, 251-259. Maniatis, T. , Fritsch, E. F., Lauer, J. and Lawn, R. M. (1980). The molecular genetics of human hemoglobins. Ann. Rev. Genetics, in press. Maniatis, T., Hardison, R. c., Lacy, E., Lauer, J., O'Connell, c., Quon, D., Sim , G. K. and Efstratiadis, A. (1978). The isolation of structural genes from libraries of eucaryotic DNA. Cell 15, 687-701. Manley, J. L., Fire, A., Cano, A., Sharp, P. A. and Gefter, M. L. (1980). DNA- dependent transcription of adenovirus genes in a soluble whole-cell extract. Proc. Natl. Acad. Sci. USA, in press. 9 Mulligan, R. c., Howard, B. H. and Berg, P. (1979). Synthesis of rabbit S-globin in cultured monkey kidney cells following infection with a SV40 S-globin recombinant clone. Nature 277, 108-114. Pressley, L., Higgs, D., Clegg, J. and Weaterhall, D. (1980). Gene deletions in athalassemia prove that the 5' I.; locus is functional. Proc. Natl. Acad. Sci. USA, in press. Proudfoot, N. J. and Maniatis, T. (1980). The structure of a human a-globin pseudogene and its relationship to a-globin gene duplication. Cell, submitted. Tilghman, S. M., Tiemeier, D. C., Polsky, F., Edgell, M. H., Seidman, J. G., Leder, A., Enquist, L. W., Norman, B. and Leder, P. (1977). Cloning specific segments of the mammalian genome: bacteriophage >.. containing mouse globin and surrounding sequences. Proc. Natl. Acad. Sci. USA 74, 4406-4410. Weatherall, D. and Clegg, J. (1976). Molecular genetics of human hemoglobin. Ann. Rev. Genetics 10, 157-178. Weatherall, D. J. and Clegg, J. B. (1979). Recent developments in the molecular genetics of human hemoglobin. Cell 16, 467-479. Weil, P. A., Luse, D. S., Segall, J. and Roeder, R. G. (1979). Selective and accurate initiation of transcript ion at the Ad2 major late promotor in a soluble system dependent on purified RNA polymerase II and DNA. Cell 18, 469-48 4. Wigler, M., Sweet, R., Sim, G. K., Wold, B., Pellicer, A., Lacy, E., Maniatis, T., Silverstein, s. and Axel, R. (1979). Transformation of mammalian cells with genes from procaryotes and eucaryotes. Cell 16, 777-785. 10 Figure 1. Linkage Arrangement of Human 8-Like and a-Like Globin Genes. The positions of the embryonic (e:), fetal (Gy, Ay) and adult (cS, 8) 8-like globin genes and of the two 8-like pseudogenes (\j/81, \j/82) are shown on the upper line. The positions of the embryonic (r;; 2, r;; 1) and adult (a2, al) a-like globin genes and of the a-like pseudogene (\jlal) are shown on the lower line. For each gene, the black and white boxes represent the coding (exon) and noncoding (intron) sequences, respectively. C, i.... '/1/32 - ,, 1/1011 o,2 ---«Jt---------------io:i~--m'----1W ,2 m-- orl G] A] '/1/31 cJ /3 ---------io:=J'-------<o:::::::J------'c=Jl-----------lo:::::::J--------<n=::J----------- I I I I I I £ 10 20 30 40 50 60 I 0 Kb ~ ~ 12 Chapter 1 The isolation of structural genes from libraries of eukaryotic DNA Cell 688 14 "'J I I ( )c ) \ I ··"' I ·-·,. .....~ .. c,. .. ) .. , -c.'O)»'ICO:> ,x,,,- I u '"1 ~ l r. O"f',,., t" ,.o, "• "• ,.. , I r:- •··~I ,.,i l ,qn,,o,, .:~~~v- .:i,,, ,,r,Y'n1..t.r i", l In .,t, r (Ml l \,( • •1<1 n~ ,_,,, '"'" ' i),w\ St• ~~ ...... Figure 1. Schematic Diagram Illustrating the Strategy Used to Construct L1branes of Random Eucaryot1c ONA Fragments al., 1977). For th is purpose , high molecular we ight eucaryot ic DNA was fragmented and fract ionated by size to obta in molecules of approximately 20 kb . Th is reduces the number of plaques requ ired for screening an ent ire genome and m inimizes the poss ibility of small eucaryotic DNA fragments jo ining to each other and subsequently to the vector . Although a nonlimit Eco RI digestion is the most direct method of obtain ing large DNA fragments w ith Eco RI cohes ive ends (Glover et al ., 1975) , we were concerned that a nonrandom distribution of Eco RI sites within and adjacent to particular structu ral genes mi ght result in the ir se lect ive loss from the population of size-fract io nated DNA . We therefore used two methods to obta in random eucaryot ic DNA fragments , both of wh ich gene rate mo lecules with blunt ends that can be joined to Eco RI linkers and subsequently inserted into the clon ing vector . In one method, DNA was fragmented by shearing and the ends were trimmed using S1 nuclease . Un fortu nately, this met hod o f preparing bluntended DNA fragments is rather 1ne1l1c,enl (see also Scheller et al., 1977b ; Seaburg et al. , 1977). An alternative , and in fact more efficient , app roa<:h is to pe rform a nonl imit restrictio n endon uclease d igest ion with two enzymes tha t cleave frequently and generate blunt-ended molecules- that is, Hae Ill (GGCC ) and Alu I (AGCT) (Roberts , 1976) . The combined activities of these two enzymes under cond it ions of nonlimit digest ion should generate a collection of large fragments approaching the random sequence representation of sheared DNA more closely than the products of a nonl imit Eco RI digest ion ; the reason for th is is tha t a recogn ition site for each of the two enzymes on DNA shou ld occur once every 256 (4') nucleotides , whereas Eco RI recogn ition sites should be pre sen t an average of once every 4096 (4• ) nucleotid es (uncorrected for base composition ). The great er the number of poss ible cleavage sites, the larger the number of possible ways of generat ing a 20 kb fragment by nonlim it d igestion of a h igher mol ecular we ight molecule , and thus the more rand o m the collection of resulting fragment s. The results of an electropho reti c ana lys is of non lim 1t Hae Il l and Alu I d igests of rabb it DNA are shown in Figure 2 . Three react ions w ith d ifferent enzyme to DNA ratios were performed se paratel y for each enzyme , and the digestion products con ta in ing a substant ial fract ion of 20 kb fragments were pooled and fract ionated on a 10- 40% sucrose gradient. Modification of Eco RI Sites In Eucaryotlc ONA Since the synthetic Eco RI linkers attached to eucaryot ic DNA must be cleaved by Eco RI to generate cohes ive ends , the Eco RI sites w ith in t he DNA fragments must fir st be rendered re sistant to cleavage . Th is was ac com plished by rea ct ing the eucaryot 1c DNA w ith the Eco RI modificat 1o n-meth ylase in the presen ce of S- adenosy l-L-methi on ine (Greene et al., 1975). Small al iquots of the methylation reaction m ixture were taken before and after the add it ion of met hylase and mixed w ith I. DNA to monitor the extent of methylation ; following ele ctrophores is , the appearance of d iscrete I. DNA bands reveals incomplete methylation . The results of a typica l assay are shown in Figure 3. The mixture of 1. and eucaryot ic DNAs taken before the add ition of met hylase is digested to complet ion , while the methylated DNA Is totally res istant to Eco RI cleavage . Joining of Synthetic ONA Linker• lo Eucaryotlc ONA Duple x DNA linker molecules bear ing rest riction endonuclease recogn ition sites have been chemi cally synthes ized (Bah l et al., 1977 ; Scheller et al., 1977a) and used to insert a number of d ifferent DNA molecules into plasm ids (Heyneker et al. , 1976 ; Scheller et al. , 1977b ; Shine et al., 1977 , Ullrich et al. , 1977) . In each case , the linkers were first blunt-end ligated to the DNA of interest and then d igested with the appropr iate restr ict ion en zyme to generate molecules with cohesive term ini that could be jo ined to a vector w ith complemen tary ends . We attached Eco RI dodecameric linkers to in vitro methylated eucaryotic DNA using T4 ligase and assayed for blunt-end ligat ion by the fo rmat ion Isolation of Structural Genes from Libraries 689 Pool R ).. R T4 ~ 9 15 ~ 9 1 2 ◄ Origin ◄ 9 3 4 ◄ 7 ◄ 2 3 4 5 6 7 b 9 6 10 11 12 13 Figure 2. Agarose Ge l Electrophoresis of Non limit Hae Ill and Alu l Digests of Rabb it Liver ONA High molecular weight rabbit ONA . was digested with d ilferenl amounts of Hae Ill or Alu I and etecl rophoresed on a horizontal 0 .5% agarose gel containing 0.5 ~g / ml ethid ium brom ide , and the ONA was photographed on a short wave ultra violet tr ■ ns 1llu­ m Inator (S harp et al . 1973). (1 . Pool ) the pooled DNA lrom the Hae 111 and Alu I d igests of slots 6- 8 and 1~12 . (2 and 4. R) und igested rabbit liver DNA . (3, -' ) undigested wild -type -' ONA (49 .4 kb ) (B lattner et al .. 1977 ). (5 , T4 ) und igested T4 DNA (171 kb) (Ki m and Dav idson, 1974): (6-8 , Hae 111) Hae 111-d igested rabb it liver ONA : (9 end 13 , 9) size markers , Eco Al-digested Cha ron 9 DNA ; (10-12 . Alu I) Alu I-d igested rabb it DNA . The ap proxi mate sizes of the markers in kb are ind icated on the side of the figure of linker ol igomers . After ligat ion , the DNA was sedimented through a 10-40% sucrose gradient (or passed over a Sepharose 28 column ) to remove unincorporated linker ol1gomers . Without this step. the d igestion of linkers attached to eucaryot ic DNA is difficult , since linker ol 1gomers not incorporated into eucaryot ic DNA compete for Eco RI. Eco RI d ig estion o f the DNA thus ~urified p rov ide s large frag ments w ith Eco RI cohesive ends that can be joined to the vector DNA . Ligation of Eucaryotlc ONA to Charon 4 ONA C haron 4 conta ins th ree Eco RI cle avage sites (Figure 1) . Digestion w ith Eco RI produces two internal fragments with genes nonessent ial for phage viabil ity and tw o end fragment s. To max imize the eff icien cy of in vitro recombinat io n and to minim ize the number of nonrecombinant phage in the library , we removed the internal fragments by sucrose gradien t centr ifugation . After annealing the cohesive ends of Charon 4 DNA and Ec o RI d igestion, excellent separation o f t he cohered A DNA arms (31 kb ) and the internal fragments (7 and 8 kb ) can be achieved . To determine the ratio of vector ONA to eu caryot,c DNA th at produces the Figure 3 . Assay !or Met hylat,on of Eco RI Sites in Eucaryotic ONA A mocture of eucaryot ic DNA and Charon 4A DNA was reacted wit h Eco Al methylase as described in Experimental Procedures and anal yzed by agarose gel electrophores is (1) ONA mixtu re plus Eco RI methylase , mmus Eco RI nuclease . (2) ONA m,xture plus Eco RI methyl ase , ptu!t Eco Al nuclease , (3) ONA mixtu re minus Eco RI methy tase . minus Eco RI nuclease . (4) ONA mixture minus Eco Al melhylase pl us Eco RI nuclease The arrows indicate the posItIons of phage A Eco RI fragments smallest number of background plaques without reducing the absolute yield of recomb inants , we ligated varying amounts of eucaryot ic DNA with a constant amount of purif ied>- DNA arms . packaged the DNA into phage and determined the number of plaque -forming un its (pfu) . Using the establ ished opt imal ratio , ligat ion reactions were performed at high DNA concentrations to min im ize intramolecu· lar joining and to max im ize the formation of concatemeric DNA recombinants , the substrate for in vitro DNA packag in g . In Vitro Packaging of ONA Into Phage Particles The numbe r o f independently der ived phage recombinants (library size ) requ ired for a 99% probabil ity of finding an y g iven sing le-copy sequence in the library can be calculated ,f the average size of the eu c aryot,c DNA ,nserts is known (Clarke and Carbon , 1976) . Thus 7 x 10' recombinants are requ ired for a mammalian DNA library o f 20 kb DNA inserts . Using the CaCI, transfect ,o n procedure of Mandel and Higa (1970 ), approx imate ly 2-10 x 10 3 pfu / µg of cleaved and religated >.. vector DNA are obta ined (in contrast to 10• pfu / µg of intact >.. DNA ). Thus approx imately 100-400 µg of DNA fragments attached to synthet ic linkers are needed tor the construct ion of a complete library , an amount difficult to obtain . Fortunately. efficiencies of 2 and 0 .15 x 10' pfu / µg have been achieved for intact and rel igated >- DNAs . respect ively , using in Cell 690 vitro packaging procedures (Hohn and Murray, 1977 ; Sternberg et al ., 1977) . To use this technique for mammal ian ONA cloning , it was necessary to demonstrate that the procedure does not alter the biological containment features of Charon 4A . Most in vitro packaging procedu res invo lve the temperature induction of A lysogens wh ich carry amber mutati ons in different genes required for packaging . Ea ch lysogen alone is incapable of produc ing viab le phage part ic les . but m i xed lysates of the two stra ins complement in vitro to conve rt A DNA into a plaque-forming particle . If hybrid phage ONA carrying eucaryot ic sequences is added to a mixed extract , there is a poss ib il ity that endogenous prophage DNA will recomb ine w ith the hybrid ONA in vitro or during subsequent in vivo ampl ificat ion to produce DNA carrying the wild-type markers of the pro phage . Sternberg et al. (1977) have developed a pack aging system that minim izes these problems . Fi rst . the prophage o f their strains carry the >.b2 mutation, which removes part of the attachment site and therefore preven ts prophage exc ision after induc tion . Second . the lysogens are recombination-def ic ient (the prophage is red and the host is recA ). To red uce furt her the c hance o f prop hage DNA pac kag ing and recomb1nat 1on we ultra vi olet -i rrad iated the cells pr ior to the ir use in in vitro packag ing reactions . Ho hn and M urra y (1977) found tha t both recomb inat ion and prop hage packag ing in the ir ext racts could be suppressed by irrad iat ion w ith ultraviolet light. To obta in EK2 cert ification for the system of Sternberg et al. (1977), we examined extracts derived from ult ravi olet -i rradiated cel ls for recomb inat ion and fo r the presence of in vitro packaged prophage . A procedure for the preparation and test ing o f in vitro packag in g extra cts for EK2 experiments is presented in Experimental Procedures . Fo llow ing in vitro packag ing of recomb inant DNA , the resulting ph age part icle s were separated from cellular d eb r is by sed imentat ion on a CsCI step grad ient. Th is proced ure concentrates the p hage and remo ves materi al present in the ex tracts wh ic h inh ib its the g ro wt h o f bacterial cells . Ampllflcatlon of Libraries An essent ial fea tu re o f our st rateg y for gene isolation is to establish a permanent library that can be repeated ly screened . To ach ie ve t his . it is neces sary to amplify the in vitro pac kaged recomb inant phage and to store the lib rary in the fo rm of a plate lysate . There is, of course , a r isk that a particular recomb inant phage w ill ex hib it a growth disad•an tage and w ill be elim inated fro m the library during amplificatio n . It is therefore important to minimize competitive growtt1. We acco mp rtstled this by plat ing the in vitro packaged phage on aga r plates at 16 low density (10 ,000 pfu per 15 cm diameter plate) . In th is manner , the phage are amplified approx imately 0 .1 to 1 x 10• fo ld and recovered as a plate lysate . Efficiency of Cloning Eucaryotlc ONA The eff ic iency o f eucaryot ic DNA cloning under ou r condit ions depends primarily upon the qual ity of the in vitro packag ing ex tracts (which var,es be tween preparat ions) and the fract ion of DNA fragments bearing Eco RI linkers The latter is deter m ined by the fraction of molecu les w ith two blunt ends which , in turn , depends upon the method of preparat ion . The efficiency of our extracts pre pared for EK1 expe riments varies from 2-20 x 10' pfu / µg of intact A DNA . Extracts prepared for EK2 exper iments are consistent ly less eff icient, varying from 0 .4-5 x 10' pfu/µg of intact>. DNA . When using cleaved and religated DNA , this effic iency drops . Even the lowest eff ic iency observed (3 .8 x 10' pfu / µg of DNA in the case of the rabbit library). however , Is h igher than that reported for clon ing rel igated A DNA by transfec t 1on (2-10 x 10' pfu / µg : Thomas . Cameron and Dav is. 1974 : Hohn and Murray , 1977). Analys is of the data o f Tab le 1, wh ic h describe s the character iz ation of the li brarie s . revea ls tha t DNA fragment s produced by nonl 1m1t endonucle ase d igestion (rabb it library ) are cloned more eff ic iently than t hose produced by shear ing fol lowed by S1 nuc lease treatment (Drosop hi la and silkmo th libraries). When the number of plaques formed per µg of eucaryot ic DNA is normal ized to the particular efficiency o f the in vitro packag,ng extract used to construct each ltbrary , it becomes evident that the rabbit DNA was c loned 15 .8 and 3 .4 times more efficiently than the Drosoph ila and stlkmoth DNAs , respect ive ly . (We cannot . of course , rule out the unl ikely poss ib ilit y that the cloning eff iciency is genome-spec ific ). Fraction of Clone ■ Containing Eucaryollc ONA To test whether a library contain ing the entire complement o f genomic DNA can be constructed , we have measured the number of recombinants carry ing eucaryot ic sequences . the average size of eucaryotIc DNA inserts and the single-copy DNA sequence representat ion in one of the libraries (Drosoph il a). To estimate the number of recombinant phages in each library. it was necessary to determine the number of background (nonrecomb inant) phage present . The numbe r of background plaques was minim ized b y separati ng the annealed cohesive ends lrom the internal k DN A fragments . When the pur ified cohesive ends of Charon 4 we re ligated w ithout the add ition of eucaryot1c DNA . however, a small number of pfu (on the order of a few percent Isolation of Structural Genes from Libraries 691 17 Table 1. Characterization of libraries 2 Library Efficiency ol Extract PlaQues per ,.g of Eucaryotic DNA 6 3 Relative Efficiency ol Clon ,ng• % Blue Plaques after Ampl 1ficat,on Total Number of Independent Recombinant Phage Recovered Mean length ol Eucaryotic DNA Number of Recombinant Phage Required for a "Complete "' library" Drosophila 1 X 10 1 6.0, 10' 3 .0 6 .0 x 10• 16 kb 4 8 • 10' Silkmo1h 2 X 101 5.6 X 101 4.6 7.0 2 .8 X 101 19 kb 2 4 X 101 Rabbit 4 X 10' 3.B x 10• 15 .8 2 .6 7 .B x 101 17 kb e 1 x 10• • This number was determined by d1v1ding the number of column 2 by thal in column 1 and normalmng to the value calculated tor the Drosophila library . b Calculated as described by Clarke and Carbon (1976) using the values in column 6 and assuming genome siz es of 1.65 )( 10• bp for Drosophila (Rudkm, 1972). and 1 x 101 and 3 x 10' bp for silkmoth and rabbi!, respectively (J . Yeh. l . V1lla-Komarofl and A Efslrat1ad1s , unpublished results) . of the final number of plaques in the libraries) was recovered from the in vitro packaging reaction . The number of non recombinant phage in the libraries can also be estimated by an indicator plate assay . One of the internal Charon 4 fragments carries the E.coli lactose operator-promoter reg io n and the gene for t,-galactosidase. The presence of this fragment in library phage DNA can be detected by plating the phage on a lawn of lac E. coli grown on an Xgal indicator plate. Phage carrying an intact /3-galactosidase gene produce blue plaques under these conditions (for a discussion of this assay , see Blattner et al. , 1977) . The number of blue (nonrecombinant) plaques observed after ampl i fication is small (Table 1), indicating that most of the library phage carry eucaryotic DNA. An independent, nonquantitative estimate of the degree of nonrecombinant phage contamination of the libraries can be obtained by determining the amount of internal Charon 4 DNA fragments in library DNA. This was accomplished by growing an aliquot of each library in liquid culture, pur ifying recombinant phage DNA and digesting the DNA with Eco RI. Figure 4 (lanes ·2 and 4) shows the results of such an analysis of Drosophila and rabbit library DNAs . In addition to the left and right arms o f the Charon 4 DNA, a characteristic smear of rest riction endonuclease-digested eucaryotic DNA can be observed in the two library DNAs (compare to lanes 1 and 5 oi Figure 4) . The gel was intentionally overloaded to show th at a small number of contaminating internal phage ONA fragments are present in both libraries . A similar result was obtained with the silkmoth library (not shown) . All our assays together clearly demonstrate that most of the phage in each library contain eucaryotic DNA . average size of the cloned eucaryotic DNA inserts. which we estimated for each of the three libraries by CsCI sedimentation equilibrium analysis . Since the amount of protein in different>. phage is constant and the buoyant density depends upon the DNA/ protein ratio, the distribution of DNA sizes in a >- phage population can be determined by measuring the distribution of phage in a CsCI density gradient (Weigle , Meselson and Paigen, 1959 : Davidson and Szybalsky, 1971 : Bellet , Busse and Baldwin , 1971) . Figure 5 shows the results of a CsCI density gradient analysis of rabbit library phage. The density of each fraction was determined by its position in the gradient relative to two >. phage density markers. The average size of the rabbit DNA inserts calculated from the midpoint of the curve of Figure 5 is 17 kb , resulting in recombinant Charon DNA molecules whose average size is 97% of that of wild-type>. DNA . Similar analyses of the Drosophila (W . Bender and D. S Hogness. personal communication) and silkmoth libraries yielded insert sizes of 16 and 19 kb. respectively . Knowing the approximate size of the eucaryotic DNA fragments carried in each library and the complexity of each haploid genome (Table 1). we can calculate the number of independent recombinant phage needed to find any given single-copy sequence in the library with a probability of 0 .99 (Clarke and Carbon ; 1976) . assuming that the entire genome consists of single-copy DNA se quences . Thus a 99% complete library of Drosophila, silkmoth or rabbit DNA would consist of 4.8 x 10', 2 .4 x 10' or 8 .1 x 10 5 recombinant phage, respectively . By comparing the theoretical library sizes to the actual ones (Table 1), we conclude that our libraries are " complete. " • Mean length of F.oc.aryolle DNA In the libraries The nu mber ol i11dependen\ phage recombinants required for a compfete library depends upon the Sequence Representation In library DNA To determme whether a significant fraction of single-copy sequences is lost during amplification of Cell 69 2 18 Dm Dml Ch4 RL .. " R .. .. Tota:k~ngth " •♦ ,. Ch 4A Ch 14 -1, -1, :, Cl .,. "' § 0 c ~ ,o ~ "- 1.0 Inse rt leng th (kb) Figure 5 Average Size of Eucaryot,c ONA Fragments Rabb it Library 2 3 4 5 Figure 4 Presence of Internal Charon 4 DNA Fragmenls m the Drosophila and Rabb it L1branes Aliquots of the Drosophila and rabb it libraries were ea ch grown ,n liquid culture . and ONA was isolated as described m Experimenta l Procedures . The ONA samp les were digested w,t h Eco Al and analyzed on a O 5% agarose ge l (1. Om ) Orosoptula DNA , (2 . 0ml) Orosoph,la lo brary DNA . (3 . Ch4 ) Charo n 4 DNA , (4 . RL ) rabb it library DNA , (5 , A) rabb it liver DNA the libraries . we measured the sing le-copy complexity of library DNA using the procedure of Galau et al. (1976) . Tritium -labeled single-copy tracer ONA was prepared from Drosophila DNA and driven with both sheared library and embryo DNAs . As shown in Figure 6 , the reassociat io n rate and extent of reaction of the tracer are identical in both cases . We conc lu de that within the sens 1t1vity of our measurements , the Drosophil a library contains the ent ire complement of single-copy sequences present in genom ic DNA . Phage libraries produced by non/Im rt Eco Rr d igestion or sea urchin DNA (D Ande i son arn:1 E . Davidson , personal communication) and Drnsophila DNA (R. Robinson and N. Davidson , personal communication) have been shown to be nearly complete by this criterion. Screening Libraries for Structural Gene Sequence s Assuming an average size of 17 kb for the inserts in 1n the The dens ity d istribution of phage lrom the rabbit library was determined by CsCI eQu1l ibr1um cenlr 1lugat1on Gradient Ir act ions were tnered , and the dist r ibution of library phage ( +-) was com puter fitted to a Gauss ian curve (solid Ima) The arrows md1cate the pos itions ol the Charon 4A (93 6°,H) and Charon 14 (83 .7%A ) phage density markers . Phage ONA lenglh (upper abscissa ) was calculated lrom the density relat ive to marker phage (Davidson and Szybalsk r. 1971) The insert ~ngth (lower abscissa) was determined by subtracting the length of the Charon 4A ONA arms (31 kb ) from the tota l length of the phage DNA The halt -mu ,mal band width of the curve Is significantly greaIer than that observed tor the marker phage (dala no! shown ). md1catmg that a heterogeneous size papulat1on of rabb it ONA mserts 1s present in the library This heterogeneity is not predicted by the relal1onsh 1p between insert size and packaging elt 1c1ency reparted by Stern• bergetel (1977 ). the rabbit library and a genome size of 3 x 10' bp , only 1 in 180 ,000 plaques will carry a particular single -copy sequence . Opt imal screening cond itions are therefore necessary to tdenttfy clones carrying such a sequence . We examined a number of variables in the plaque hybr1d1zat ion procedure of Benton and Davis (1977) , including the type of medium used in the agar plates (L broth or NZCY) , the strain of host bacteria (KH802 or OP50SupF). the concentration or plating bacteria and the method of preparing filters . Most of the variables had little effect on the intensity or the number of positive signals observed . The most significant differences resulted from varying the concentration of plating bacteria ; the best signals were obtained Isolation of Structural Genes from Libraries 19 69 3 A Om fMBR Yl1 B ~" - v~-• ~ft -,~ ~: ,a~ - - - ~ - - - ~10' ,a' - _ _ . J _ _ __ - ~ - ,a' EOUI VALEN l C0t Figure 6. Single-Copy Sequence Representation in the Orosoph· ila Library A comparison of the reassociation kinet ics of Oroaophlla embryo and Drosophila library ONAs with 1 H•l abe~d sing le-copy trace r (1 x 10' cpml ,..g). Drosophila library DNA was prepared from phage grown in liqu id culture (5 x 10' told ampl ification) Single-c opy trace r ONA was prepared from Drosophila pupae ONA , and reacted with sheared Drosophila embryo DNA and Drosoph il a library ONA accord ing to the procedures of Galau et al. {1976). For each poin t, 3 ~g of embryo ONA or 7.5 ~g of library DNA were reacted with approximately 800 cpm of single-copy tracer The library ONA was at 2 .5 times the concentration of the embryo ONA on the assumption that 40% of ea ch hybri d phage molecule consists of eucaryot 1c ONA . The solid ltne is a computer fit ot the embryo data describing a single second-o rder component . using 3 .1 x 10• exponent ially grow ing bacter ia per 15 cm plate . The best correspondence between positive signals on duplicate filters occur red when the filters were sequent ially applied to the plates , stack ing filters often resulted in fa ilure to observe duplicate signals . To determ ine the optimal plaque density for screening , we plated various amounts of a mixture of Charon 4A and a Charon 16·p/:IG1 hybrid (p f:iG1 is a /3-globin cDNA plasmid ; Maniatis et al., 1976) at a constant ratio of 250 :1. As many as 20,000 pfu per 15 cm plate could be sc reened for glob in w ith no apparent loss of signal on autorad iograms . The most serious prob lem encountered in screen ing libraries is nonspecific background hybrid1zat1on to nitrocellulose filters . Using conditions adapted from those developed for Southern transfer experiments (Jeffreys and Flavel l, 1977a) , however , we reproduc ibly obse rved low background . Al l three libraries have been successfully screened using gene-specific hybrid izat ion probes . Several d ifferent phage recomb1nants that hybridize to different Drosophila cDNA plasm id clones have been se lected from the Drosop hila library (W . Bender and D. S. Hogness, personal commun ication) . The frequencies at wh ic h these clones were detected and the single-copy complexity m easurement of Figure 6 indicate that most if not all Drosophila structural gene sequences are present in the library. The sifkmoth library was screened for genes , sequentially expressed during oogenesis , w hi ch Figure 7 Scr eening ol Sdkmot h Library for Chorion Ge ne Se· quences (A and B) Autor1d 1ograms of duplicate nitrocellulose tilters show ing 1pecific hybrid 1z1t1on ot cho rion cONA to sllkmoth library phage clones . 5000 phage we re plated onto a 10 cm petri dish and incubated for 1e hr at 37"C , and the phage ONA was trans ferred to two nitrocellulose filters applied in succ ession Agarose rathe r than agar was used in the 0 .7% top agar layer . The tilte rs were prepared for hybr 1d1zat1on as described 1n Expeomental Procedures and hybridized for 48 hr to 50 ngtml HP- cONA (spec act . 2 .5 x 10' cpm / µg ) prepared from tota l chorion mRNA FIiters were washed . dried and exposed to X-ray him !or 48 hr using a single intensifier screen Five stron g and two weak pos1t1ve signals appear on bo th filters 1dent1fy1ng phage clones bea ring chorron gene sequences . encode the approx imately 100 eggs hell (chorion) prote ins of the deve loping oo cyte (for a review of th is system, see Kafatos et al ., 1978) Although the exact number of chor ion genes Is not known , prelim inary evidence suggests that each gene canno t be present in more tha n a few copies (J . Ye h , W . C. Jones and A. Efstrat iad 1s, unpubli shed results ). With an ave rage insert size of 19 kb tn the silkmoth library and a genome size o f 1o• bp , we expected to observe one posit ive signal per 530 plaques using cDNA transcr ibed from total chonon mRNA as probe , assum ing that the chorion genes are un ique . If some of the genes are closely linked , pos itive signals w ill be less frequent . Figure 7 (A and B) shows an autorad ,ogram of dupl ic ate filters prepared from a plate conta ini ng 5000 plaques . 100% of the dupl ic ate pos itive s,gnals proved real when individual plaques we re pi c ked onto a bacterial lawn and rescreened . Whe n on ly single (i nstead of dupl icate ) filters were prepared , 88% of the initial pos itives proved rea l upon rescreening To date , screen ing of 350 ,000 plaques has yielded 350 independent isolates , about 53% of the number expected if the genes are unlinked . The rabbit library was screened fo r globin sequences using cDNA prepared from tota l globin mRNA . Figure BA shows an autorad iogram of a filter prepared from a 15 cm agar plate carry ing 10,000 plaques . In the example shown , two posi tives were observed on one filter (a rare event). demonstrat ing two types of s,gnals . One signal reflects the plaque morphology , while the other conta ins a head and a comet-l ike ta il. The latter frequently observed morpholog y could result from Cell 694 A 20 B Figure 8. Screening the Rabbit library for Globin Sequences (A) Autorad iogram of a nitrocellulose filter prepared lrorn a 15 cm plate conta1n1ng 10 ,000 recom bi nant plaques The fil ter was hybr 1d,zed to 2 .4 ng / ml »P-glob1n cONA (5 x 10' cpm / µg ) fo r 36 hr , was hed as described m E11;pe nm ental Proced ures. dried and e,;posed to preflashed X-ray him fo r 48 hr at - 70"C usmg a single intensifier screen . Arrows md,cete the loca!lons of two pos,t 1ve signals . (8 ) Autor ad 1ogram of a filter prepared lrom a plate conta 1n1n g 500 plaq ues obtained by plating a nu mber at plaques from the area on tne plate (A) correspond ing to the locat 1on of one of the two pos 1t1ve signals shown m (A) The hlter was hybrid ized w1lh 25 ng t ml of nic k• trans lated p,(1G1 ONA (5 x 10' cpm / µg ) fo r 12 hr . end expo sed to prel lashed X-ray l1l m tor 24 hr at 7CY"C usmg two 1ntens1l1er screens the spreading of phage DNA from the plaque during filter applicati on or subsequent handl ing . Both spots were show n to be true posIt 1ve s by the ir appearance on duplicate filters and by rescreening (Figure BB ). A total of four independent /3-globin clones were recovered from 750 , 000 plaques screened with globin cDNA . To id entify clones carrying /3-glob in sequences, each clone was hybridized to in vitro labeled pµG1 DNA , a rabb it µ-glob in cDNA plasmid (Maniatis et al., 1976). With a genome size of 3 x 10' bp and c loned inserts of ilpproxi mate ly 17 kb, we expected to recover four to five clones of the adult ,B-globin sequence from 750 ,000 plaques , a number close to that actually recovered . In th is calculation, we assumed tha1 cro ss-hybrid1zat 1on of ad ult µ-glo bin probe to embryonic /:l·h ke rabb it globin genes is inadeq uate to allow detection of these genes in a total genome screen . Chan1c terlzatk:> n of Clonets 'Thal Hybridize to Rabbit Globln Probea To d em onstrate that the fou1 clones r.yf:rrid izin g to the /3-globin plasm id actua lly carry the µ'--g fob,n gene sequence , we d igested DNA from each clone with Eco RI , fractionated the products on a 1 .4% agarose gel, transferred the DNA to a nitrocellulose filte r and hybrid ized them to In vitro labe led globin cDNA . Figure 9A shows the Eco RI cleavage pattern of DNAs from the four µ- globin clones . As expected for DNA fragments generated by random cleavage , each clone conta ins common and unique Eco RI fragments . Presumably the common set conta ins the /i- glob1n gene and it s adjacent se quences . while the un ique fragments lie further from the gene in the 5' or 3' direction Figure 9B shows the resulting autoradiogram of the hybrid izat ion expe ri men t. In all four clones . two fragments of approx imately 2600 and 800 bp hybr idize to the probe . These fragments are . respect ively , the sizes of the 5' and 3' /J- glob in Eco RI fragments fo und in genomic DNA (Jeffreys and Flave ll . 1977a , 1977b) . The ident i fication of these bands Is confirmed by detailed restriction mapping and DNA sequence ana lysis of the cloned DNA (our unpubl ished results ). In addition to these two fragments . Rµ G2 (lane 2) contains one Eco RI fragment and RµGS (lane 4) conta ins three fragments that hybridize weakly to glob in cDNA . These data and the failure ot these two clones to hybridize n-globin probe ,.ndicate th at the add itional Eco RI fragments correspond to /J-l;lce sequences closely linked to the adult µ-glob1n g ene. The 6 .3 kb Eco RI fragment of RµG2 and R/J GS . which hybridizes weakly compared with the 2 .6 kb fragment . may correspond lo IM fain! 6 .9 kb Eco RI fragment detected in genom ,c DNA (Jeff reys and Fla vell, 1977a) . Preliminary restriction mapping data from Rµ G2 indicate that the µ-like sequence Isolat ion of Structural Genes from Libraries 695 21 A 1 2 B 3 4 1 i 2 3 4 ◄ -6.3 -3.5 -2.6 -1.9 ◄ -1.08 1t l., irJu· Figure 9 Eco RI Cleavage Patterns of ONAs from Rabbit fi-Glob1n Phage Clones ONAs from rabb it ,B- gtob 1n phage clones were digested with Eco RI, tract1onated on a 1 4% agarose gel . lransferred to a nitrocellulose ftlter (Southern, 1975 ) and hybr1d tzed to 11 P-glob1n cONA (1 x 10• cpm / ,ug) (A) Ethidium bromide-stained ge l. (A) 1, CH4A-R /3G1 . (21 >,Ch4A-R/JG2 . (3) 1,Ch4A-R/!G3. (4) 1,Ch4A-R/!G5 . (B) Autorad togram of the nitrocellulose tilter prepared from the gel shown in (A) . The sizes of various Eco RI fragments are 1nd1cated in kb The arrows indicate DNA fragments bearing fl-li ke globm seq uences that are not a par1 ol the adult 11-globIn gene . carried on the 6.3 kb fragment is approximately 9 .1 kb away from the adult /j-globin gene in the 5 ' direction (data not shown ). The /j-like sequences in R /3G5 have not yet been mapped . Discussion This paper shows that ii is possible to iso late structural genes d irectly lrom large eucaryotic genomes by screening libraries of DNA fragments cloned in phage II . The overall efficiency of the procedure described yields a collection of recombinants large enoug h to represent the ent ire genome ol a mammalian cet\. Such a co!lect.10n ca n be amplified by a !actor of 10•. w ill\\ no ap·parnnt loss of sequence co mplexity. to produce a ltbrary of eucaryoUc DNA that can !i:e sciee.ned r epea.tedly using dif1erent probes. This rapid ,net/Jod of gene 1.SOlafion provides many advantages over eK isting tec hniques . For exam ple , all the members of a famil y of evolutionari ly or developmentally related genes can be isolated in a single step by screening a library with a mixed probe . Furthermore , isolation of a set of overlapp ing clones, all of wh ich conta in a given gene , permits the study of sequences extending many kilobases from the gene in the 5 ' and 3 ' directions . Moreover , even more distant reg ions along the chromosome can be obtained by rescreening the library us ing term,nal fragments of the initially selected clones , allowing the isolation of linked genes . The power of th is approach is clearly illustrated by the isolation of globin and chorion genes . The mammaltan globin genes const it ute a relatively simple family compri sed of at least two o-like and at least four /j- like embryon ic and adult genes . wh ich are expressed during erythropo ,es is ,n d i fferent cell populations at d ifferent developmental times (Cl1ssold , Arnstein and Chesterton . 1974 : Melderis, Steinhe1der and Ostertag . 1974 : Steinhe ioe,, Mel.deris and Ostertag , 1975. 1977) . The gene fam ,ty thal co des for the chorion proteins of the sdlt<. moth Antherae a polyphemus is considerably more complex (Kafatos et al. , 1978) . Approximately 100 different chorion genes are sequentially expressed during oogenesis . We have used the procedure of cDNA cloning to purify to homogeneity sequences that correspond to individual members of the chorion gene family (Maniatis et al. . 1977 : Sim et al. , 1978 : G . K . Sim . unpublished results) . Cell 696 Some of these cDNA plasmids have been classified according to the time of their expression during development (Kafatos et al ., 1978) . Using these cDNA plasmids to rescreen the phage recombinants that hybridize to total chorion cDNA, 11 will be possible to isolate and study genes that are coord inately and / or sequentially expressed during choriogenesis . Since the libraries were prepared from random DNA fragments, independent isolates of a given gene will carry DNA sequences that extend for various distances away from the gene in both directions . This is clearly illustrated by the analysis of the /3-globin clones (Figure 9) . In some cases , the cloned DNA extends far enough in one direc tion to include a linked gene . Although the possibility of linkage between different members of the rabbit a- or /3-globin gene families has not been studied , evidence exists that such linkage occurs in other mammals , including mouse (Gilman and Smithies , 1968) and human (Clegg and Weatherall , 1976). For example , in humans the (3-, ti- and -y-globin are genetically linked (Huisman et al., 1972 ; Clegg and Weatherall, 1976) on chromosome 11 (Deisseroth et al., 1978). In most human populations , the u-globin genes are present in two copies per haploid genome (Lehmann, 1970 ; Hollan et al ., 1972) and are located on chromosome 16 (Deisseroth et al. , 1977) . Thus the close phys ical linkage between rabbit glob in genes reported here is probably a general characteristic of mammalian globin genes . Clones bearing such genes can be used to study the precise organization of linked genes and the pos sible relat ion between linkage and control of gene expression . Moreover , a permanen t library of clones bearing overlapping sequences w ill facil 1\a\e the isolation of the many linked genes \hat constitute a complex genetic locu s. Cloned segments of eucaryotic DNA can also be used to study the fine structure of genes . Most current methods of mammal ian gene isolation involve partial purification o f geno mic DNA frag ments generated by a limit restriction endonuclease digestion pr ior to clon ing (Tilghman et al ., 1977 ; Tonegawa et al., 1977) . If the restriction endonuclease used to fragment genomic DNA cleaves w ithin the coding sequence or noncod ing intervening sequences of the gene of interest , the gene must be cloned in pieces . The rabbit n- and (3-globin genes , w hich carry a single Eco RI site w ithin the coding sequence (Maniatis et al. , 1976 , Salser et al., 1976 ; Liu et al., 1977), and the chicken ovalbumin gene , wh ich carries at least one Eco RI site within each of three interven ing sequences (Breathn ach . Manda.I and Cnambon, 1977 ; Weinstoclt et al ., 1977 ·, Lai et al., 1978), illustrate this problem . If more than one Eco RI site is located 22 within an intervening sequence, a portion of the chromosomal gene structure will remain un1dent1fied . The procedure described here avo ids these problems by cloning large p ieces of randomly fragmented DNA that should carry intact genes includ ing their intervening sequences . E1perlment1I Procedure, Materlal1 ONA polymerase I and T 4 polynucleot1de kinase were purchased from Boehringer Mannheim . T4 ligase wa s purchased from Be thesda Research Laboratories Eco RI methylase prepared ac cording to Greene et al (1 975) was provided by John Rosenberg DNA polymerase from avian mye loblastos1s virus (AM V reverse transcnptase ) was provided by Or . J . W. Beard and the Office of Program Resources and Log,st1cs (Viral Cancer Program . NIH ) Eco Al was prepared accord mg to the procedure of Greene el al (1975) . Hae Ill and Alu I were prepared as described by Roberts et al. (1976) and Rober1s P 976), respect ively Prote1nase K was purchased from EM Labs . Pancreat ic DNAase I was purchased from Worthington B1ochem1cals NZ amme was pvrchased from Humko•Shefhe ld (Llnnhurst . New Jersey) Nitrocellulose t ilters were purchased from M1ll1pore S- adenosyl- L-me th1on1ne was purchased from Sigma o- 11 P-deo;.;ynucleos 1de tnphospha tes were purcnased from New England Nucl ea r ( 300 C1/mmole ) or ICN (120-200 C1 t mmole ). Ollgo(T) 11 • 1• was purchased from Colle b· orat,ve Research . Preparetlon of Bacteriophage A DNA Charon phage were grown essen tially as desc ribed by F A Bla ttn er ,n the detalled prot oc ol that accomp anie s the Charon A phages Phage we re purified as described 1n the abo ve prot ocol and by Yamamolo el al (19701 For preparation ot phage DNA , purified phage were dialyzed against 10 mM Tros-CI (pH 8), 25 mM NaCl , 1 mM MgSO, . broughl lo 0.2% sos . 10 mM EDTA, healed lo 65"C to, 15 min and digested for 1 hr al 37··C w ith SO µg / ml Protemase K . The ONA was extracted severa l times with phenol . ether-extracted and dialyzed extensively aga inst TSE (5 mM NaC l. 10 mM Tris-Cl (pH 8) . 1 mM EDTAI To prepare the end fragments of Charon 4 (see Figure 1). the cohes ive ends IN'ere annealed by 1ncubat1on for 1 hr ,n 0 1 M TrisCl (pH 8.01 . 10 mM MgCI, al 42"C . o,1h,othre,10I (OTT ) was aoded to 1 mM along wit h an e xcess of Eco Rt and lhe react ion mix wa s incubated for 3 hr at 37"C An aliquot was run on a 0 5% agarose gel to verify that d 1gest1on was co mplete The ONA was e xtracted with pheno l and then with et her 50- 70 ~g ot Eco Al-cleaved phage DNA we re layered onto a 10-40% linear su c rose grad ient (1 M NaC l, 20 mM Tros-CI (pH 8 0) . 10 mM ED TA] ,n a Beckman SW2 7 centri1uge tu be The gradient was centrifuged (at 27 ,(X)() rpm for 24 hr at 20 C) (Neal and Florin 1, 1972 ). and 0 .5 ml fractions were collected using an ISCO ultra violet flow cell . Fractions were analyzed on an agarose gel and those conta1nmg lhe 31 kb annealed end fragments were pooled To e;.;am1ne the restriction endonuclease cleavage patterns ol ONAs from md1v1dual plaques . ONA was prepared as desc ribed above from phage grown 1n 4 ml cultures Enough ONA was obtained to perform several restnct10n endonuclease digestions ONl-hlla DNA Drosophila embryos (Canton S w 1k' -type ) aged 6-16 h, were conected , washed . frozen on dry ,ce and stored at - 70 C. ON.A was prepared according to Brutlag et at. (1977) with mod 1llc a• t1ons DNA from the CsCI gradient was dialyzed , d igested w ith Prote1nase Kand phenol-extracled . The DNA was then broughl to 5 M NaCl and chilled to ere to mm ,m,ze the generation of molecules with s1ngle•stranded tails during shea rmg (Pyerit z, Schlegel and Thomas . 1972). The DNA was sheared by slowly Isolation of Struct ural Genes from Libraries 69 7 23 drawing it into a chilled 5 mt plastic syringe with a 20 gauge 1 in needle and expelling it as hard as poss ible into a SO ml conical polyethylene tube on ice . The number of passes through the needle prior to the add ition of S1 requ ired to generate 20 kb fragments follow ing the nuclease treatment varted from prepara tion to preparat ion . For th is particular ONA preparat ion , three passes through a 20 gauge needle produced a mean size al 30 kb . The DNA was dialyzed aga,nst 0.5 M NaCl , 10 mM Tr,s-CI (pH 8.0) . 1 mM EDT A. Sod ium ecetate (pH 4.5) and ZnSO, were added to final concentrati ons ol 50 mM and 2 mM , respectively . An amount of S1 nuclease sutf1c1ent to convert an equ ivalent amount of single·stranded A. ONA to sma ll fragments as assayed by agarose gel electrophoresis was added . Following incubation for 1 hr at 37°C, the react ion mixture was extracted repeatedly with phenol and then with ether . Sllkmoth DNA DNA was isolated from silkmoth Antheraea polyphemus pupae as previously described (Efstratiadis et al. , 1976 ). Shearing and Sl nuclease d igestion were performed as described abo ve A•bbtt LI .. , DNA The liver of the New Zealand rabb it was removed and frozen in small pieces in liquid nitrogen . ONA was isolated using a mod ifl· cation of the Blin and Staff ord (1976 ) procedure . After Protemase K d igestion . phenol extract ion and dialysis. solid CsCI (0 .95 g / ml) and eth td ium bromide (1/10 vol of a 5 mg / ml solut ion ) were added (final density 1.65 g / cm'). The solution was centr ifuged (i n a T i60 rotor at 45 ,000 rpm tor 60 hr at 20"C) to separate ONA from RNA and polysaccharides, and the DNA was collected as a viscous band by puncturmg the side of the tube w tt h a needle Eth id1um bromide was removed by several extractions w it h ,sopropanol equilibrated with saturated CsCI, fol lowed by exhaust ive dia lys is against TSE . The molecul ar we ight of the ONA was est imated by electrophoresis on neutral (Sharp , Sugden and Sambrook , 1973) and alkaline (Mc0onne1 , Simon and Studier , 1977) 0.5% agarose gels , using bacteriophage A. Charon -4 ONA (46.200 bp ; Blattner et al. , 1977) and bacteriophage T4 DNA (171 ,000 bp . K,m and Davidson , 1974) as molecu la r weight standards Both the duplex and single-stranded lengths of the rabbit ONA molecules were estimated to be > 100,0<X) bp or nucleotides . Partial endonuctease digest ion conditions were establ ished lor the restr iction enzymes Hae Ill and Alu I by performing a serial dilution of each enzyme in the presence of 1 ,-ig al rabbit liver DNA in 1X restrict ion enzyme buller (6 mM Tris-Cl (pH 7.5). 6 mM MgCti , 6 mM t3-mercaptoethanol} . Reactions were Incubated for 1 hr at 37"C , and the exten t of digestion was e1timated by electrophoresis on a 0 .5% neutral agarose gel using Eco RI • df08Sted Charon 4 ONA as a molecu l8r weight standard . On the basis of th,s informatio n, six large scafe d igests (330 "'9 ONA per reaction ) were performed w it h 0 .5, 1 and 2 t imes the est imated amount of enzyme yteld ing the maximum proport ion of 20 kb fragments The six d igest s were pooled , phenol--extracted and concentrated by ethanol precipItat,on laol•tlon or 20 kb Euc•ryotlc DNA 250- 300 µ.g of sheared or enzymatically cleaved DNA in 0 .5 ml of 10 mM Tr is-Cl (pH 8.0). 10 mM EDTA were heated at 88°C !or 20 min and sed imen ted thro ugh a 10-40% linear sucrose gradient as described abow Ahquot11 o1 trachons were analyzed by eJect,opho"'"• on a 0 .5% agarOM gel using Eco Rl-011,esteo Charon 4A ONA as a mo~cular weight standard The i'<actlons conta ining 19-20 kb DNA were pooled . dialyzed againsl TSE. concentrated by ethanol p r&cipi'tetion and fet.u!ipe,nded Kl TSE . Ec:o R I _ , _ , , 0-tl:uc,r,ryollc DN A 115 pi g ot 20 kb ON~ were brought to a volume of 1 m l in 0 .1 M Tns- Ct (pH 8.0), 10 mM EDTA, 6 µm S-adeno1yl- L-meth10nine . Eco RI methylase (20 units In 1 ml) was added . Two 10 µ I ailquots were taken before and after the addit ion of the enzyme and mixed w ith 0 .5 µg al A ONA . The eucaryot1c ONA and the tour io ,-ii contro l reactions were mcubated tor 1 hr at 37"C . Each 10 µ.I aliquot conta ining >. DNA was mixed with 25 µ.l al a butter containing 0.2 M Tr,s-CI (pH 7.5). O.t M NaCl . 20 mM MgCI, . and 2 mM OTT . Two al the aliquots (one each withdrawn before and after the add11ton of methylase ) were m11i:ed w ith Eco RI and incubated for 1 hr al 37'"C . The other two corresponding aliquots were incubated w itho ut the add1t 1on of Eco RI (see Figure 3 ) The methylated DNA was phenol-extracted . ether -extracted , ethano l• prec ipitated , rediss olved in 100 µ.I ol 5 mM Tri s-Cl (pH 7 .5 ), d ialyzed aga inst !he same butte r m a Schleicher and Schuetl c0Uod1on bag . and evaporated to 40 1Jl under rntrogen Cov•lent Joining of Eco At Llnk•ra to Euc•ryollc DNA The synthes is o f dodecamer linke rs produ ces mo lecules with 5' hydroxyl ends (Scheller et al . 1977a ). Smee the hgase requires 5' phosphate ends . the first step m th e jom1ng reaction 1s to phosphorylate the linker. 5 µg of dodecamer linker in 1O µ.I of 66 mM Tris-Cl (pH 7.6), 10 mM MgCI, , 1.0 mM ATP , 1.0 mM aperm idine , 15 mM OTT , 200 µg / ml gelatm were added to 2 µ.I (10 un its) of T 4 kma!.8 and incubated at 3rC for 1 hr. This react ion mixture was then added d irectly to 100 µ.I of eucaryol 1c ONA ( 100 µ.g ) in the same buffer . 5 µ.I (5 units) of T4 hgase were added and the reaction was mcubated al room lemperat ure tor 6 hr . A 5 "'' aliquot of the react ion mixtu re was ele ctrophoresed on a 12% polyacryIam 1de Trts- borate - EDTA gel (Man 1at1s, Jelt rey and van de Sande , 1975) and the DNA was vi sualized by el h1d 1um brom ide staining . A successful hgat 1on was ev idenced by the presence of a series of linker ohgomers , from dimers up to 14 •me rs The ligation react ion m ixtu re was diluted to 500 µ I wit h 10 mM EDTA , mcubated !or 15 m m at 68 'C, laye red onto a 10-40% tmear sucrose gradien t and ce n trifuged as described above Tne g ra• dient fract io ns contammg 20 kb ONA were 1dent 1fIed by agarose gel electrophoresis. pooled . dial yzed aga mst TSE and ethano lprec ipi tated The DNA (25 µg ) was resuspe nded ,n 100 ,.1 ot 5 mM NaCl and brought to 1X Eco RI buller (0 1 M Tris -Cl (pH 7.5), 50 mM NaCl . 10 mM MgCI, . 1 mM DTT J, and 150 units ol Eco RI were added . 3 ,-ii of the reaction were mixed w ith 0.2 µ.gal>. ONA and both tubes were incubated tor 1 hr at 37"'C . The small scale reaction contammg >. DNA was electrophorned on a 0 .5"- aga· rose gel , in which a complete d1gesI of the linkers attached to rabbit ONA was evidenced by a limit digestion pattern a t A. ONA . The Eco RI linkers were attached to 20 kb Drosoph ila and 1llkmoth ONAs using similar procedures . except !hat linker ollgo• mars were removed on a Sepharose 28 column rather than on a sucrose gradient . llg•llon of CohHlv• Endo High cloning effic iencies were obtained using a 2 fold molar excess of Charon 4 arms to 20 kb eucaryotIc DNA fragments For example , In the case ol the rabb it library, the reaction mixture contained 20 .5 ,-ig ol euc1ryotIc ONA and 55 µ.g o f purified Charon 4A arms in 300 µ.I of ligase buffer . The cohesive ends o f the phage were annealed for a second time m MgC~. Tris and gelt, t1n tor 1 hr at 42"C before ad ding ATP . OTT . euc aryot1c DNA and 19 ,-ii (19 units ) l1gase . The m ixture was incubated al 12"'C for 12 hr . An allquot was heated to 68"C , cooled and elec trophoresed on a 0.3% agarose gel w ith Eco Rl--d109sted Charon 4A DNA as a molecular weight standard . Successful ligation was evidenced by IM absence or CtlarOII 4A end lragmenls ( - 12 and 19 kb ) and the preaence of concetemeric ONA molecules larger than intact Cllaraa 4A ONA. In Yllto Pack•glng or Recombinant DNA Into Ph•g• P•rtlcl. . In vitro packaging extracts were prepared u described by Stern berg et 11. (1977 ), except thal both types (A and B) ol extracts which they describe were handled as 80 µ. I aliquots in 1.5 ml polypropylene tubes , wh ich were frozen In hquid nitrogen and atored at - 70"C . Proportionate amounts ol lysozyme . then bu ffe r B (Sternberg et al. , 1977) and glycerol were added to each tube or Cell 698 resuspended B extract cells and thoroughly stirred mto the extremely viscous suspem1ton using a glass m1crop1pel. Preparatton and Te1tlng of In Vitro Packaging Extract• for EK2 Experlmenta N\H regulat ions require tha t ,n vitro pa,c.M.ag ing extracts " be irradiated with ultrav,ole1 hght to a dose of 40 phage lethal h1ts " before they can be used in EK2 level recombinant ONA experi ments . Wave~ngths between 250 and 280 nanometers are !he most etteclive to r photoinact1vat,on (Hollaender , 1955 ) We use as an ult rav10let hght source four 18 1n 15 watt germicidal lamps. (GE G15T8), w hrch emit 1 3 x 10' erg / cm• - n (2200 µW / cm't sec) at a distance of 20 cm . The k,llln g effic iency of this hght source was calibrated as follows . 1 liter ol NZCYM broth (1% NZ amine , 0 .5% NaCl. 0 .5% yeast extract , 0 .1% casamino acids . 10 mM MgSO.) was inoculated with 20 ml of an overn ight culture of ~Ci857Sam7 (a heet-1nduc1ble >. lysogen ) The culture was grown at 32-'C to an 00_. ol O 3, transferred to a 42°C shaking water bath for 20 min to induce the lysogen and t hen incubated with shaking tor 2.5 hr at 37"C. The cells were transferred to an ename l dish (32 cm x 18 cm x 5.5 cm) and 1rrad1ated at a distance ol 20 cm w ith constant m1x1ng on a rotary shaker platform . 1 ml aliquots were taken atter 5. 10. 20, 30 . 40 and 60 min of irradiation . The ce lls were lysed by the addition of two drops of chloroform and the ONA was removed by adding 10 ,.,.1 of a 1 mg / ml ONAase I solution . The mixture was centr ifuged for 2 min in a Brinkman Eppendorf centr ifuge , and the number of surviving phage was determined by t itenng the supernatant on OP50SupF al 42'C under yellow llght to avoid photoreactivat1on . Our yellow light source consists of a commercial "Gold " fluorescent lamp (GE ) with a gold plex1glass filter . A plot of t he fog of survi val versus time of 1rradiat1on extrapolates lo 40 leth al hi ts at approximately 30 m tn o f ultravtolel irrad iation Ce lls used m preparing EK2 extracts were ultrav1olet-i rrad 1ated (as above) after heat mduct1on and incuba tio n al 37°C end pr1or 10 pelleting (see Sternberg et al., 1977) Before any recombinant ONA was packaged . every prepa,11 1on or extracts derived from ult1av1ofet -1rrad1ated cells was iested to, recomb ination and the p.cesence of prophage All of !he steps de:scribed below were pe.rfor mod under yellow llghl Extracts were tested for prophage. ex.c1s1on and packaging by performing " mock" packag ing reactions without exogenous ONA and plating on OPSOSupF , which provides the appropriate suppressor lo, the S amber 7 mutation of the prophage (Sternberg et al . 1977). Plates were incubated at 42"C to inactivate the prophage repressor. No more than one tenth (30 µI ) of one packagmg reaction could be assayed on a single 10cm plate because the concentrated packaging mixture kills bacterial cells and thus masks the presence of prophage Accord ing to NIH regu lation s, "the ratio of plaque -forming units w ithout add 1t1on,of exogeno us >. ON A (endogenous virus) to plaq ue-fo rm ing umts w it h exogenous ONA. (e xogeno us vr rusJ mu s1 be iess than 10 • -" E,c;ogeno us v1 rat ONA means recombma.n.t DNA . We cons{sten!ty t,nd \hat th is ratio 1s -. 10· • for recombin ant A DN A and c: 10 • using intact Charon 4A ONA 1n tne assay . Without ultraviolet irrad 1at1on , the above ratio is > 10 • . NIH regulations further req uire that when the EK2 veclor 1s packaged , " the ratio of am ' phage fr« omb~ allU1 lo totv J phage mosl be less tha n 10-• ... Wo added Ch•'·""" <till Dt-lA to the packagin g extracl SM ~111'0 th e packagoa DNA"" Su· and Su atram s The ratio of pfu on Su to pfu on Su · 1s a measure of the frequtll'\¢ 'f of r-,.combinat10n wdh prophage ONA We ftnd thal th is raho is < 110 • , io fact. ~e hue. n,e,.e, otlM' ~ e pla:q1J~ ret ult""9 from. rec.om b1nat1o n On the basis: of these and $'.m~w , ciatV: obtained by N . Sternberg and L. Enqu1st (unpublished observa tions) , the NIH has approved 1n vitro packagmg to r EK2 level experiments . Peckaglng ONA !Dr U,rar1H To constr uct , to r example, the rabbi! library , 26 packaging reac· tions were performed. each contaming 2 .5 ,.,.g of recombinant 24 ONA After ONAase digestion and chloroform treatmenl. the packaged phage were purified and concentrated on a CsCI step gradient . The reactions were pooled . mixed with solid CsCI (0 .5 g t ml ). brought lo a final volume ol 30 ml w1lh O 5 g t ml CsCI 1n SM {O 1 M NaCl. 0 05 M Tros-CI (pH 7.5). 10 mM MgSO, , 0 01% gelatin) , layered onto CsCI gradients (each one composed ol 1.5 ml steps ol 1.45 , 1 .5, 1.7 g t ml CsC I in SM , in an SW41 tube ) and centrifuged (at 32 ,000 rpm !or 1 5 hr at 4"C ) 200 µI lract1ons were collected and phage were located by spollmg ddut1ons of 10 'to 10 'onto a lawn ot DP50SupF Library Ampllllcatlon The fractions containing phage were pooled and d1alvzed aaarnsl 0 .1M NaCl , 50 mM Tros - CI (pH 7.5). 10 mM MgSO, Ge laton was added to the d ialysis bag to a concentration ot O 02% lo stab ili ze the phage In vitro packaged phage tro m l he rabb it library were plated onto a fresh overn ight of DP50SupF at a dens,ty of 10 .000 phage per 15 cm d iameter plate (75 plates lota t) The phage were preadsorbed w1lh bacteria tor 20 mm at J7·C, mixed w1lh 6 .5 mt of NZCYM-DT (NZCVM med ium w rth O 01'% d 1amtnop1mehc acid and 0 .004% thym1dme . Blattner et al . 1977 ). 0 5% top agar and plated . The plates were incubated for 14 ht al 37"C (Plat ing was camed out in yellow lighl and the plales were wrapped 1n aluminum toll while growing to prevent photoreacl1vat1on J The top agar was scraped into e sterile beaker and the plates were rinsed once w ith 3.75 ml ol SM The 285 ml lysate from 38 plates was mixed w rt h 10 ml of chloroform and st irred slowly al room temperature for 20 mm The lysate was lranslerred lo a 1 liter centrifuge bottle and centr ifuged (ma Be ckman J6 centrduge al 52.0CX> rpm for 20 mm at 4°C) to remove the top agar 8crHnlng th• LlbrarlH Amphlted llbranes were screened us,ng t he 1n situ plaQue hybn d1zat1on techn1Que ol Benton and Davis {197 7) . 10.000 recomb inant phage were plated on 3 1 x 10• e11:ponent1a r ph ase oacter1al ce lls on 15 cm NZCVM petn dishes To prevent top agar from adhenn" to the n 1troceltulose filter when 1t was lifted from the plate (which tends to increase the ba c kground hybnd 1za t1 on ). plates were dried m a 37"C incubator tor severa l hours or set on edge overnight to drain excess IIQu1d The use of O 7% agarose rathe r than agar 1n the lop agar layer also mm1m 1zed this problem The plates were incubated at 37"C for 14 - 16 hr . at which time the plaques were m contact but lys1s was no! confluent Plates were refr109rated for an hour or longer belore the filters were applied Nitrocellulose litters were cut from rolls (HAMP 000 10) or circles (HAWP 142 50) of M1ll 1pore filter paper (type HA . pore size O 45 µm) to flt eas ily over the agar plate Phage and ONA were adsorbed to these fallers 1n duplicate by placing 1wo fillers on e ■ ch plate seQ uent,ally , 2 min lo r the fust filte r and 3 mm for the second . at roo m temperalure The ONA was denat ured and bound to the l1tt1rs as described by Benton and Da vi s (19 77) To prepare !he filters tor hybrtd 11at1on to a labeled probe . the y were wetted 1n about 10 ml per lllter ol 4X SET 11X SET = 0 15 M NaCl. 0.03 M Tris - Cl (pH 8). 2 mM EOTA J at room temperature tor 30 min . washed tor 3 hr at 68 'C 1n about 10 ml per l 11ter ol 4X SET . 10X Oenhardt s so lution (10X Denhardl s solu1 1on =- 0 2°.,. bovine serum album ifl. 0 .2% poty\lmylpyrolldone , 0 2°,o F1coll , Oenhardl . 1966} a1\d. Q "I '¼ SOS , and prehybrid1zed w ith continuous ag 1tal1on to, lltl '68st 1 hr &\ 68'-C tn about 4 ml per filler ol 4X SET . 1OX Oenharcn ·, so lu tion , 50 µg /m( denalured salmon sperm ONA . 10 ;.gi ml pc ly(A f. O l'fo SOS and O t% sod ium pyrophosphate De· natu:r«:-1 E cGlt QttM. {10-50 µ.g tmll was included ,n rne prehybr1 ,d1,z 1t~o m ~,c w l'-:41M \JS1ng l at>e>t.-d pJa&m1d ONA as probe The pt"ehyO,id 1zaf1on and subsequ~nt l'lyb11d1zat 1on were earned out m a therm atfy sealed plas tic bag The fillers were hybridized with ag,tat1on 11 68"C in prehybr1d 1zat ,on solution conta,ning HP-la· beled hyb11d 1zat1on probe at lhe concentrat ions and lor the times 1nd1cated in the figure legends Atter hybnd1zat1on . the filters were washed once with ag11at1on in about 15 ml per ftlte r of 4X SET , lOX Denhard!'s solution , 10 ;.gi ml poly (A). 0 .1% SOS . 0 1% sod ium pyrophosphate at 68' C tor 1 hr. 3 limes .n ·a simila r vo lume lsolattOn ot Structural Genes from Libraries 699 of 3X SET. 0 .1% SOS , 0 .1% sodium pyrophosphate at 68°C lor 1530 min; twice in 1X SET, 0.1% SOS. 0 .1% sodium pyropnosphote at 68°C tor 15-30 min : and once in 4X SET at room temperature . In some cases , more str ingent washing conditions (0 .5X SET) were used as a final step . The filters were blotted dry , mounted on cardboard and exposed to preflashed Kodak XR5 X-ray tdm with Dupont Cronex :11R Xtra Ute Lightning-plus intensify1ng screens at - 10"'C for 1-2 days . Plaqua Purification or R•comblnant Phage Plaques from the region of a plate correspond ing to I pos itive on the auto radiogram were picked and suspended in 0 .5 ml SM . The phage suspension was t itered and the plate containing about 100 plaques was rescreened . The process of picking positives and rescreening was repeated until - 90% of the plaques on a plate gave positive signals attar screening Plate lysates were prepared using 1 phage from an 1ndivldua1 plaque on a 10 cm plate , and the lysates were harvested as described in Library Amplification. 10 10-10 11 phage were usually recovered . o• Hybridization Probe• cONA plasmids were grown . purified and labeled in vitro by nick translation as previously described {Maniatis et al. , 1976). Complementary ONA to globin and chorion mANAs was synthesized as described by Efstratiadis et al. (1975) and Friedman and Rosbash (1977). C ■CI S.dlmentetlon Equlllbrlum Anelyal1 or Recombinant Ph ■ g• Phage from the amplified rabb it library (4 x 10' pfu ) were mixed with 3 x 10' pfu of Charon 14 phage (Blattner el al., 1977) 1n a solution of CsCI. density = 1 .514 g / ml , 10 mM Tm,-Ct (pH 7 5) and 1.0 mM MgSO • . The phage were banded by cent ri fuga1 1on to equilibrium in a Ti50 rotor , 5.38 x 10' g ", hr at 4•c . One drop fract ions were collected from the bottom of the gradient into 0 .5 ml SM for a total of 128 fractions . The tractions conta,n ing phage were titered on OPSOSupF, a Su II+ Su Ill stra in , to determme the distribution of densities of phage m the library and also on Cle (SupO) to determine the position of the marker Charon 14, a nonamber phage . The position of a second marker , Charon 4A (representing the background phage m the library), was determined by counting the number of blue plaques termed on KH802, a Su II lac - strain , using Xgal plates From the known sizes ol Charon 14 and Charon 4A, a calibrattOn curve was constructed relating traction number, or buoyant density , to tength of insert ONA . The average density of the recombinant phage provides a minimum estimate for the size ol rabbit ONA inserts, since the empirical relationship between phage density and ONA molecular weight was established using A ONA (50% G+C) (Davidson and Szybatski, 1971), while lhe base composition of rabb it DNA ,s «.2% G+C (Kritskii, Arends and Nikolaev , 1967). Roteomblnanl ONA lately Experiments involving the cloning or propagation of bactenophage A carry ing eucaryotic ONA were performed in accordance with the NIH Guidelines for recomb inant ONA reaearch . Drosophila and silkmoth cloning experiments were performed using EK1 vectors in P2 fac ilities located at the Cali fornia Institute cl Technology and Harvard University. Rabbit ONA clonlng experiments *ere perlormed using the EK2 vector-host system A Charon 4A/ OPSOSupF in a P31acihty st the California Institute of Technology . The rete'4'ant genotypes ot EK2 Vftctors were verifled prior to their use in recombinant ONA. •J.pe.nmenfs. Acknowledgment1 The procedurn tor making and screening llbrarie1, were estabhahed ·c oncurrantty in the laboratodes of E . Oavidaon, N . Oav•d· son , J.. St,nner and L. Hood er Caltech We benefited from the free a.x.chang-e of ff\formation and ideas among these labs We are gro.l~ful to D . Goldberg , B . S.Od . 0 Engel, A . Scheller and 0 . 25 Anderson ior suggestions and d1scuss,ons . and to N . Davidson for critical readings or the manuscript . We thank F Blattner tor providing Charon phage strains . N Sternberg and L . Enqu1st for prov1dmg in vitro packaging s1ra1ns , C . K llakura and R Scheller for prov1d1ng Eco Rt linkers . J Rosenberg tor prov1d1ng Eco RI methylase . R . Rob inson tor providing labeled single-copy Oro sophtla ONA !racer . and A Cortenbach for preparing med ia and materials . We thank W . Bender and 0 . Hogness for d1scuss,ons and tor commun1cat1ng unpublished data on the Orosophtla library cheractenza11on, and F C Kafatos tor the use al facd1t1es end his support We thank J . Man1alls for help w1lh the l1gures Th1s work was supported by an NSF grant 10 F C Katalos and T M ., an NtH grant to F. C Kalatos , an American Cancer Society granl to T . M . and funds from an NIH 81omed1cal Research Support grant to the Cal1lorn1a Institute of Technology R . C. H was supported by a fellowship from the Jane Collin Childs Memona l Fund for Medical Research T M 1s the rec ipient ot a Rita Allen Foundation career development award The costs of publ1cal1on of this article were defrayed m part by the payment ol page charges This article must therefore by hereby marked " advertisement " 1n accordance w ith 18 U.S .C Section 1734 solely to ,nd1cate this fact Received June 16 , 1978 Reference, Bahl. C. P., Menans . K J . Wu . R ., Stawinsky . J and Narang , S (1977 ). A genera! method lor inserting spec1t1c ONA sequences into cloning vehicles Gene 1, 81 - 92 . Ballet , A .. Busse . H .and Baldwin . FL (1971) Tandem duplications ma der 1va1tve of phage lambda In The Bacte nophage Lambda . A . J . Hershey , ed (New York Co ld Spring Harbor Laboratory). pp . 501-513 . Benton , W 0 . and Davis , A. W . (1977) . Screening A.QI recombinant clones by hybr1d1zat1on to smgle plaques 1n situ Science 196 , 160-182 Blattner , F. R ., Wtlllams . 8 . G , Blee hi . A E . Thompson , K O , Faber . H . E .. Furlong . L A .. Grunwald . D J . Kieler , D . 0 . Moore . D . D .. Schumm . J W , Sheldon . E L. and Sm,th,es . 0 . (1977) Charon phages safer der,vat1ves of bacteriophage lambda tor ONA cloning . Science 196, 161-169 Blin . N and Stafford , D W. (1976) A general method for 1solat1on of high molecular weigh! DNA lrom eukaryotes Nucl. Ac ids Ras 3 , 2303-2308 . Breathnach . R., Mande l. J L and Chambon . P. (1977) . Dvalbu· min gene is split 1n cn,cken DNA . Nature 270 . 314-319 Brutlag , D .. Appals . R .. Denn,s . E. S and Peacock . W J (1977 ) Highly repeated ONA m Orosoph,ta melanogaster J . Mel 810 1 172, 31-47 Carbon . J ., Clarke . L .. llgen . C . and Ratzk 1n . B (1977 ). The conslruct,on and use ot a hybrid plasmid gene bank in E coli In Recombinant Molecules Impact on Science and Society . R F. Beers ano E. G. Bassali , eds (New York Raven Press). pp 335378 Clarke . L. and Carbon , J (1976) A colony bank con1a1n 1ng synthetic Col El hyOr1d plasmids representative ol the entire E coli genome Cell 9 , 91-99 . Clegg . J . B . and Weatherall . 0 . J . (1976 ) Molecular basis ot Bull . 32 , 262-269 thalass■em,a . Br . Med Clisaold . P. M .. Arnstein . H R. V and Chestenon. C J. (1977) . Ouant1tat1on of glob1n mRNA levels during erythro,d development in the rabbit and discovery of a new µ-relaled species m immature ery1hroblasts Cell 71. 353-361 Davidson , N . and Szybalsk1 . W (1971) . Physical and chemrcal character1st1cs ot lambda ONA . In The Bacteriophage Lambda , A . 0 Hershey , ed {New York Cold Spring Harbor Laboratory). pp 45-82 . Cell 700 26 Oeisseroth, A., Nlenhu11 , A., Turner, P., Velez , R., Ando,-on , W . F., Ruddlt , F., Ltwronct , J ., Croagen , R. and Kuchtrlapat, , R. (1977) . LocalizaUon of the human o-globm 11ructur1I o•n• to chrom osome 16 ,n aomalic cell hyb rids b y molecular hybrid1Z1· tion auay . Cell 12, 20!>-218. Oeisuro,h. A. ., Ntenhu1I , A .. Lawrence . J ., 0 1111. R., Turne r, P. and Ruddle , F. (1978). Ch romo■ oma l locall za tlon of human JJglob,n gene on human chromosome 11 in 10mat1c cell hybrid ■ Proc . Nat. Aced . Sci . USA 75, 1•56-1460 . Oenhard t, 0 . T . (1966). A membr1n•t1lt1 r techn ique for the detect10n of complementary DNA B,ochem . Biop hya. Rea. Com mun . 23 , 6•1-646 . Kritakil , G .. Arend, . I. and N1kOIHV , Y. (1967) . Nucloot,dt com po11t1on of ONA fraction, of bone marrow in normal and X· irred 11ttd an,mels B,ochemlatry (USSR ) 32 , 372 - 376 . Lai. E . C .. Woo , S. L. C., Duga,czyk. A., C1lttr11 . J . F. end O ' M ■ l ley , 8 . W. (1978) The ovatbumm gene struc tu ral NQuenc•s in native ch ick ONA are not cont1guouI . Proc Nat. Acad . Sci USA 75 , 2205-2209 . Leder . P.. Tteme1er. 0 and Enq u1 I t , L (1 977) EK2 den vI1tves of bacteriophage lambda uHful ,n the cloning o t ONA from hig her o rganisms · the A. gtWES sy stem Sc ,ence 196 , 175-1 77 . Lehmann , H (1970) 0 1tf1rent types o f alpha thalassem ,a and I 1gn1t1cInce of humog lobm Bart·s 1n neonates Lancet 2. 78- 80 Efstratiedis , A., Man iatis . T., Kafatos , F . C .. Jeffrey , A. tnd Vournakis, J . N . (1975). Full length and discrete part ial r■ verae tran scripts of glob1n and chorion mRNAs . C.114 , 367 - 378. L,u . A . Y., Paddock . G. V., He,ndel l, H C and Selser , W (1977) Nucl•ot 1de sequences from I rabbit alpha globin gene inserted ,n • chimeric plasm id Science 196 , 192- 197. Efstrat11d is, A ., Crain , W. A .. Br itten , R. J ., Oav1dson , E. H. and Katatos , F. {1976 ). ONA sequence organ ization in the lepidopteran Antherea pernyl. Proc . Not. Acad . Sci . USA 13, 2289- 2293 . Mandel . M . and Hoga , /\ (1970). Calcium-dependent b1ctoroo ph1ge DNA 1ntor1ct1on . J . Mol . Biol. 53 , 159-162 Friedman, E. Y. and Rosbash , M . (1977) . The ayn thH IS of high yields ot full-le ngth reverse transcripts of glob,n mANA . Nucl. Acids Res 4 , 345~3471 . Galau , G. A., Kte ,n , W. H., Davia, M . M ., Wold , B . J ., Bri tton , R J . end Davidson , E. H. (1976). St ructural gene set, active in embryos and adult t1s ■ ueI of the sea urchin . Cell 7, 487-505 . Gilman , J . G. and Sm ith1t1s , 0 . (1968). Feta l hamoglobln variants ,n m,ce Science 160 , 885-886 . Glover , D. M ., Wh ,ta , R. L .. Finnegan, D. J . and HognHS . D s (1975) Characterization of I 1x cloned ONAs from Orosoph1l1 melanogaste r. includ ing one that conta i n ■ the genea for rANA . Ca 11 5, U~l57 . Greene , P. J ., Poonien , M . S., Nussbaum , A. L. , Tob ias , L ., Guion , D. E., Boyer . H. W. and Goodman , H. M . (1975 ) Rest ri ction and mod ificat ion of a self complementary octanucleot,de conta ining tht EcoRI substrate . J. Mo l. Biol. gg , 237 - 261 . Maniatis . T., Jeftroy . A and van dt Sande . H (1975) Chain length determination of ■ ma ll double· and single •stranded ONA mole• cu._, by polyIcrylam 1cie ge t electrophores is B1ochem1stry 14 , 3787-3794 Man iatis, T ., s,m , G K., Etstr1t,ad 1s, A. and K1f1tos . F. C (1978) Ampllfteation and cnaractenzation of a #-g lob1n gene synthesized in vitro . Cell 8, 183 - 182 Man.11t11, T .. Stm . G K . KIfIto1 . F C , v ,111-Komaro tf . L and Elslr111ad 11. A (1977) An approa ch to the study o l de velo p· menta lly reg ulated genea In Molec ular Clon ing ol Recomb inan t DNA (New York Acadom ,c Pron). pp 173- 203 McOonnel . M .. Simon , M N and Stud 11 r . W F (1977) Ana ty11s ol rest nction fragments o l T7 ONA ■ nd determin ation o f molecul ar we ights by ele ct,ophoreI11 in neutral and alka lme gels. J Mo l. Bio l. 110. 119- 146 Meldens . H., Steinhe•der , G and Ostertag . W (1974) Evidence to r a unique kind of o type glob1n chain m early mamm1hIn tmbryoa . Nature 250 , 774 -776 . Heyneker . H. L., Shine , J ., Ge10dm1n . H. J ., Boyer , H., Rosanbtrg , J ., D1cktr10n , R. E ., Narang . D. A . and Riggs . A . D (1976) Synthetic lac operator ONA is functiona l In vivo . N ■ h.He 283, 748752 . Nea l, M . W and Fiorini , J . A . (1972) Effect al Iucro1e gradient compos ition on re1o lut1on of RNA 1pec11I. Ana l B1ochem 45 . 271-276 . Hohn , B. and Murray , K. (1 977). Packag ing rocomb1n1nt DNA molecutes into bactenophage part icles In vitro . Proc . Nat. Acad . Sci . USA 74 , 3259- 3263 . Pytntz . R., Sc hlegel , R and Thomu . C , Jr (1972) Hydrody namic shear breakage of ONA may prod uce 11ng le·ch11ned t1rm 1• n11I . B10ch,m . B10phys Acta 272, 504 - 509 Hollaende r, A. (1955). Red ,atoon Biology , II (Now York · McGr1wH1II) Robtr1s . R. J . (1976) Reatnc t10n endonucleaaas CRC Crot Rt v Biochem . 4, 123- 164 Hollan , S. R., Sztltny i,, J . G .. Bri mhall, B .. Duera t. M .. Jonea , R. T., Ko le,, R. D and Stock len , I. (1972) Multiple alpha Ch1 1n loc , for human hMmoglobins : Hb J-Buda and Hb G-P9st . Nature 235 , 47-50 . • Rober1a , R. J ., Myers. P /\ , Mom aon , A . and Mu rr ay, K. (1976) A 1pec1hc endonucle111 from Arthobac ter luteua . J Mal. 8 101. 102 . 157- 165. Huisman . T. H. J ., Wrlghtstone , R. M , WIiton , J . B .. Schroeder . W. A. end Kendall , A. G. (1972 ). Hemoglobin Kenya , tho product of a fus ion of 'Y and fJ polypeptide cha1n1 . Arch . e,ochem . B1ophys . 153, 650-853 . Rudk in , G. T. (1972) Rephca 11on in polytene ch romoaomes In Ae1ult1 .and Problems m Cell Oifterent 1at1on , 4 , W Beerman . ~ (Now York . Spnnger - Verlag ). pp 59 - 85 Je ff reys, A. J . and Flavell , R. A. (19771). A phy11c1I map o f the ONA regions flank ing the rabb it /j-g lob,n gene Coll 12, 4211 - 439 Salsar . W , Bowen . S , Brown , D . Adl 1, F E .. Ftderoft . N . Fry , K .. He1nde ll. H . Paddock , G ., Poor. A ., Walla ce . 8 and Wh,1· come . P. (1976) lnveIt1ga t1 on ot the organ izat ion of mImmI 11In chromosomes at the ONA sequence le vel Fed Proc 35 , 23- 35 Jelfnrya, ,-. J . and fl• vall , R. A. (1977b) Th• rabbit /j-glob,n gtnt conte ina I larg,a ,nun rn t,.. cod•fl{I aequonct Cell 12, 1097\108 Scheller, R H .. D1cke r10n, R E . Boyer . H W . Riggs . A D and 1takurI. K. (1 9771) Chem ica l 1ynth1s1I o f reat t1 ction enzyme recogn it ion s1te1 uN ful to r clon ing sc ..nce 1ge , 177-180 K1t1tos. F C ., Etstr1tied i1, A., Goldsmith, r.l R., Jones, C W., Oan iatia, T., Re iger, J . C., Rod1k i1, G .. Rosentllal, N., Sim, G. K., Thireos, G. and Villt-Komarott , L. (1 978) . The - topmentally ,.gulated multigene tam 1hes encoding chonon p rotein• ,n allk· moth In Structure and Organ izatio n o t O...-topmentally Reg u• fated Genes, J . Schultz and F. Ahmed , eda . (New York · Acadom,c Presa) , in prHs . Scheller , R. H., Thomu . T L . LM , A. S , Kle on. W H .. Nolts , W D ., Britton , R J . and Dav,daon , E . H. (1877b). Clonea of 1nd 1vodua l ,.petihve aequencea from NI urch in ONA con1truct1d with 1ynthel 1c eco RI 111eI. sc ..nce 196. 197- 200 . Kim , J . S. and D1vld10n , N. (1974). Electron mlcroacope hetoroduple x study of sequence relat ions of T2 , T4 and T6 btct•rlophage DNAa. Virology 57, 93-111. S..burg , P. H., Shone , J . M1r111I. J. A., Baxter. J . D. and Goodman , H M . (19 77) Nuc}eohde aequence and amphhcat ,on In bacteria of 1tructural oene to r rat growth hormone . Nature 270 , •ee-•e•. Sharp , P. A., Sugden , B and Sambrook , J . (1973) . Oetect10n of two ,..trict1on endonucleaae 1cttvitt11 1n HaempophiluI para,n - Isolation of Structural Genes from Libraries 701 27 fluenzae using analytical agarose ethid1um bromide electrophoresis . Biochemistry 12, 3055-3063 . saying e1ttracts without the addition or e1to0enous >. ONA are no longer required (tor current rules . consult OH1ce of Recombinant Shine . J. , Seeburg, P. H., Martial . J . A .. Butor , J. O. and Goodman , H. M . (1977) . Construction and analyaia of recomb1 · nant ONA for human chonon 1c somatomammotropin . Nature 270 , 494-499 . ONA Activities) . Sim , G. K., Efstr1ti1d1s . A .. Jones , C W ., Kafatos , F. C., Koehler. M ., Kronenberg , H. M ., Maniatis . T .. Regier. J . C .. Roberts . 8 . F. and Rosenthal , N . (1978) . Studies on the structure of genes expressed during development. Cold Spring Harbor Symp . Quant. Biol. 42 . 933-945 . Southern , E. M . (1975). Detection of specific sequences among ONA fragments separated by gel electrophoresis J . Mol. 8101. QB , 503-517 . Steinheider, G., Melderis, H. and Ostertag , W . (1975) Embryonic ,: chains of mice and rabbits. Nature 257, 7i~716 . Ste1nheider , G., Melderis. H. and Ostertag , W. (1977) . In Intern&· tional Symposium on the Synthesis. Structure and Function of Hemoglobin , H. Mart in and L. Nov,cke , eds . (Munich : Lehmans), pp. 222-235 . Sternberg , N., Tiemeier . 0 . and Enqu ist , L. (1977) . In vitro pack aging of a A Dem vector containing Eco Al ONA fragments of E . and phage Pl . Gene 1, 255-280 . Thomas . M ., Cameron , J . R. and Davis , R. W. (1974) . Viable molecular hybrids of bacteriophage lambda and eukaryot1c ONA . Proc . Nat. Acad . Sci. USA 71 , 4579-4583 . Tilghman , s. M ., Tiemeier . 0 . C ., Polsky , F., Edgell , M . H., Seidman , J . G., Leder . A., Enquist . L. W., Norman , 8 . and Leder , P. (1977). Cloning specific segments of the mammalian genome bacteriophage A conta ining mouse glob tn and surrounding gene sequences . Proc . Nat . Aced Sc i USA 74 , 4406-4410 . Tilghman , S. ~ -. Tiemeter , 0 . C., Seidman , J . G., Peterlm , B M .. Sullivan . M .. Maizel, J . and Leder , P. (1978 ). Intervening sequence of ONA identified in the structural portion ot a mouse ,8-globm gene . Proc . Nat. Acad Sc i. USA 75 , 725-729 . Tonegawa. S., Brack . C ., Hozum1 , N. and Scholler. R. (1977) . Cloning of an ,mmunoglobul in vanable region gene from mouse embryo . Proc . Nat. Acad . Sci. USA 74, 3518-3522 . Ullrich , A., Shine , J ., Chirgw in . J .. O,ctet. A . T11cher , E .. Roltes , W. J . and Goodman, H. J . (1977 ). Rat insul in genes : construction of plasmtds conta ining the coding sequences . Science 1Q6, 13131319. Weigle , J ., Meselson , M . and Pa 1gen , K. (1959) . Density altera • t ions associated with transduc ing ability in bacteriophage lambda. J . Mol. Biol. 1, 37~386 . Weinstock , R., Sweet . R., Weiss. M .. Cedar , H. and A1tel , A . (1977). lntragenic spacers int erru pt t'he ovalbum ln gene . Proc . Nat . Acad . Sci. USA 75 , 1299-1303. Wens ink . P. C ., Finnegan , D. J ., Donelaon , J . E . and Hogneaa , D. S. (1974). A system for mapping ONA oequences in the chromosomes of Drosophila mel ■ nogaster. Cell 3 , 315-325 . Yamamoto , K. A ., Alberts . 8 . M .. Benzmger. A .. Lawhorne , L. and Treiber , C. (1970). Rapid bacteriophage sedimentation in the presence of polyethylene glycol and its application to large acale virus puri ficattOn Virology 40. 734-744 . Vovng , M . and Holiness , D. (1977). A new approach tor identifying and mapping structural genes in Orosophit. cne\u 10Quter ,n Eucaroytic Genet,c Systems , ICN-UCLA Symposia on Mollcu~ r and Cellular Biology, VIJ/ /Hew Vorl<: Academic Press) , pp J l f>.331 . The NIH has ,_,tly revised the rules for in vitro packaging in EK2 systems using phage ;. vectors . In particular . ultraviolet irradiation of extracts to • dooe of 40 phag&-lethal hits and as- 28 Chapter 2 The chromosomal arrangement of human a-like globin genes: Sequence homology and a-globin gene deletions 29 Cati, Vol. 20, 119-130, May 1980, Copyright C 1980 l)y MIT The Chromosomal Arrangement of .Human a-Like Globin Genes: Sequence Homology and a-Globin Gene Deletions Joyce Lauer, Che-Kun James Shen and Tom Maniatis Division of Biology California Institute of Technology Pasadena, California 91125 Summary We report the Isolation of a cluster of four a-like globin genes from a bacteriophage .\ library of human ONA (Lawn et al., 1978). Analysis of the cloned ONA confirms the linkage arrangement of the two adult a-globin genes (a1 and a2) previously derived from genomic blotting experiments (Orkin, 1978) and identifies two additional closely linked a-like genes. The nucleotide sequence of a portion of each of these a-like genes was determined. One of these sequences is tentatively Identified as an embryonic f-globin gene (r,) by comparison with structural data derived from purified f-globln pr~ tein (J. Clegg, personal communication), while the other sequence cannot be matched with any known a-like polypeptide sequence (we designate this . . quence lj'(Jl ). Localization of the four a-like . . quences on a restriction map of the gene cluster indicates that the genes have the same transcriptional orientation and are arranged in the order 5'r,-lj'(J1-a2-a1-3' . Genomic blotting experiments identified a second, nonallelic f-like globin gene (t2) located 10-12 kb 5 ' to the cloned f-globin gene. Comparison of the locations of restriction sites within a1 and a2 and heteroduplex studies reveal extensive sequence homology within and flanking the two genes. The homologous . . quences, which are interrupted by two blocks of nonhomology, span a region of approximately 4 kb. This extensive sequence homology between two genes which are thought to be the products of an ancient duplication event suggests the existence of a mechanism for sequence matching during evolution. One consequence of this arrangement of h~ mologous sequences is the occurrence of two types of deletions In recombinant phage DNA during propagation in E. coli . The locations and sizes of the two types of deletions are indistinguishable from those of the two types of deletions associated with athalassemia 2 (Embury et al., 1979; Orkin et al., 1979; S. Embury et al., manuscript submitted). This information strongly suggests that the genetic dlaea&e is a consequenc• of unequal crossing over between homologous sequences within and / or sur• rounding the two adult a-globin genes. Introduction The a-like and ,8-llke subunits of hemoglobin are encoded by a small group of genes which are expressed sequentially during development. The earliest embryonic hemoglobin consists of ( (.8-like) and (a-like) polypeptide chains. The ( polypeptide is gradually replaced by y (fetal /!) and this, in tum, is followed by the adult globins 8 and ,8. In contrast , the f chain is succeeded immediately by a , which functions during both fetal and adult life (Weatherall and Clegg, 1979). Genetic studies have demonstrated linkage between the human fetal and adult ,8-like genes (Weatherall and Clegg, 1979), but the adult ,8-like and a-like genes are located on separate chromosomes (Deisseroth et al. . 1977, 1978). Physical linkage between the ,8-like genes was demonstrated by genomic blotting experiments (Flavell et al.. 1978; Mears et al. . 1978; Fritsch. Lawn and Maniatis. 1979; Little et al .. 1979a; Tuan et al.. 1979) and examination of cloned ONA (Lawn et al. , 1978; Fritsch et al. . 1979; Fritsch , Lawn and Maniatis. 1980). All the P-like genes have the same transcriptional orientation and are arranged in the order 5'-f-Gy-•y-6-/3-3' . The 5 ' to 3' organization of the genes reflects the order of their expression during development. It is not known whether this developmentally correlated gene arrangement plays a role in the mechanism of sequential activation of globin genes or is simply a consequence of the duplication events which gave rise to the gene cluster. Genetic and structural studies of a-chain variants suggest that most individuals carry at least two linked a-globin genes per haploid genome (see Bunn . Forget and Ranney, 1977; Weatherall and Clegg, 1979). It appears that both of these genes are expressed since individuals heterozygous for two different a-globin variants produce normal a-globin in addition to both of the variant polypeptides. Since only one a-globin amino acid sequence is found in most individuals (Dayhoff, 1972). the two linked genes encode identical polypeptide chains . Physical linkage between the two adult a-globin genes was demonstrated by genomic blotting experiments. These experiments also revealed that both genes are transcribed from the same ONA strand and are separated by approximately 3 . 7 kb (Orkin, 1978; Embury et al. . 1979). The location of the embryonic f-globin gene with respect to the a genes was previously unknown . In this study we used molecular cloning procedures to demonstrate physical linkage between the two a- globin genes and two additional alike sequences, one of which appears to be We also used genomic blotting procedures to identify a second f-like gene located 5 ' to the cloned r-globin gene. Finally, we present an analysis of homologous regions within the coding , intervening and flanking sequences of the two adult a-globin genes, and characterize two types of deletions which occur in recombinant phage DNA during propagation In E. coli. r r. 30 Catt 120 Results Localization of a-Like Sequences in Cloned DNA Two clones denoted .\.HaG1 and .\.HaG2 containing a-globin genes were isolated from a library of human DNA (Lawn et al.. 1978) by screening with a 32 Plabeled human a-globin c0NA plasmid, JW101 (Wilson et al. , 1978). a-Globin seQuences were located on the restriction map of each clone using the blot hybridization procedure (Southern, 1975). The adult a-globin genes were identified by comparison with the restriction map derived from genomic blotting experiments (Orkin, 1978; Embury et al.. 1979). The two genes have the same transcriptional orientation. For convenience we will refer to the 5' and 3' members of the a-globin gene pair as a2 and a1. respectively . One intensely hybridizing band and two fainter bands are detected when .\.HaG1 ONA is digested with Hpa I plus Barn HI and hybridized to 32 P-labeled a-globin c0NA plasmid (Figure 1, lanes a and b). The intense band at 4 .3 kb corresponds in size to a Hpa I fragment containing the a2 gene. The identification of the two fainter bands of Figure 1b will be discussed below. The clone .\.HaG2 contains both adult a-globin genes . .\.HaG2 ONA digested with Hind Ill plus Bgl II displays four strongly hybridizing bands (Figure 1, lanes c and d). Hind Ill cuts within each a-globin gene to generate a 3.8 kb intergene fragment (Orkin, 1978) which is cleaved by Bgl II to produce 1.8 and 2.0 kb fragments. Correlation of the blot of Figure 1d with the restriction map of .\.HaG2 (Figure 1 l indicates that the 1.8 and 2.0 kb fragments contain, respectively , the 3' half of a2 and the 5' half of al . Similarly, the 10.2 kb Bgl II plus Hind Ill hybridizing fragment contains the 5' half of a2 (Figure 1). The 3' half of a1 is located within the 3.8 kb hybridizing fragment, which includes part of the lambda vector (Figure 1). These data confirm the arrangement of the adult genes derived from genomic blotting experiments (ontin , 1978; Embury et aL, 1979). A de1ai\ed restriction map of A.HaG1 and .\.HaG2 DNA is presented in Figure 2. The precise localization of the four a -ike ~nces on this map will be CMCU~ t>elo>N. Comparison of the Positions of R"trlctlon Sltff In Adult a-Globln Genes To determine the extent _ o f similarity between the two adult a-globin genes and to establish whether these genes contain intervening seQuences, the distances between many restriction enzyme sites within and near al and a2 were compared with the distances between corresponding sites predicted by the nucleotide sequence of a-globin mRNA (Forget et al., 1979). This comparison was accomplished by labeling the single Hind Ill site within a1, a2 and the cDNA clone JW101 (Wilson et al ., 1978) with 32 P, digesting with a second restriction enzyme and purifying the end-labeled fragments containing the 5' and 3' halves of the 0 b d C --:-9.0 . . . -10.2 tlP"!la -;- 4 . 3 ..... -3.6 - - 3. 8 - - - - 2.0 --- AHaGI 3.6 9.0 4.3 I II I Hpo Born Hpo AHaG2 =:::J, Bgl 10.2 1. 8 ., r:;::: 3.4 Hpo Hpo 1.8 2.0 3.8 " "' I Hind Bc;il Hind I Bc;il Figure 1. L.ocallzation of a-Uk• Globin 5-Quenc;• In >.HaG 1 and .l.HaG2 ONAa Purified .l.HaG 1 and l<HoG2 DNA• - • diQ . .led with restriction endonucleasea as indicated below . fractionated on 0 .7% agaroae geta. tranaferr..:I to nitrocellulose fitters (Southern. 1975) and hybri<Slzed to the "P-labeled a-globin cDNA plasmid JW101 (Wlloon at al .. 1978). (a) Ethidivm bromic»-etained gel of l<HoG 1 ONA digested wi1h Hpa I plus 8am HI. (b) Autoradi<>Qram of the blot of the get shown in (a). The band labeled 4.3 corresponds to a Hpa I fragment containing the a2 g - (see Af(aGt map below autoradioorem). The fainter bands at 3 .15 and 9 .0 kb COff-,d to Sam HI plus Hoa I fragments containing a-Uke eequenc... The, _,, faint band marl<ed with an aatenak does not align wttn a flt.ooreacent band In (al and therefore CO<T8Sl)0nds to a minor eontammant. !he nal\Jre of wnich hu not been determined . (cl Ethidium ~ get of >JiaG2 ONA diQested with Bgl H plusHind tll. (d) Autoradlogram of the blot of !he get _ , In (c). The four bands co,respond "'-"ttvely to a 10.2 kb Bgl K plut Hind llt fragment containing two a-like MQuen<:n plut the 5' haH of the 5' adutt or-globin · a 3 .8 kb Hind Ill plus Bgl H !raQment containing the 3' hall of the 3' e-globin · a 2.0 kb Bgl H plus Hind llt frag""!"t containing the 5' hall of the 3' a-globin and a 1.8 kb Hind Ht plus Bgl I fragment containing the 3' haH of the 5' .-globin gene . TMM hybridizing fragments are indicated on the mac, of .\HaG2 ONA ~ !he auto<aclloQram. three a-like sequences. These fragments were then digested with a variety of restriction enzymes and the sizes of the products were determined by polyacrytamide gel electrophoresis to give the distance be- 31 Human a-Glol)jn Gene Cluster 121 ,...,,, ..., m \M,1GZ ( IIDI •• ~ SKI .... .... S-1 ~ :u ..... -· -·-· - Q,O •• "" 1, . 11 I Z, LU - ,. Flgure 2. "481) of Restriction Endonucleue Cleavage SitH in ;1.HaG 1 and ;1.HaG2 DNAs .., ••2 IIDI 11 01 I., 0 ~ Restriction sites are indicated by -1ical lines. and the we of each fragment ia given on kllobaM pan (kb). The poaltlona Of at , a2 , ,io1 and fare lhown. Blacl< box• r_...,,, mRNA coding regiona and white boxes repre_,, noncoding inten,ening sequencea. The region of human DNA included the inaert ct .\HaG 1 o, ;1.HaG2 is indicated by Ille horizontal liMe 11 the IOC> ot the figure. 0 I . IZO 1,4 021 2.0 05 I. ()40 01, CWI °'' 0~ •• ' 23 21 0 12 l 4 !1 04!1023 SI 0 12 10 tween the Hind Ill site and the nearest recognition site tor each enzyme. The results of this analysis (Figure 3) indicate that both of these genes contain two intervening sequences (IVS). The 5' half of each gene digested with Mbo II or Hint I generates fragments 95 bp longer than the corresponding fragments of the cDNA, indicating the presence of a 95 bp IVS (IVS I). Because there is no Sma I Of Hpa II site in the 5' coding region. those sites in each gene must lie within IVS I. Hph I cleaves the 3' half of each gene to yield a fragment 120 (a2) or 1 28 bp (a1) longer than the corresponding fragment of the cDNA, indicating the presence of a second intervening sequence (IVS II) in each gene. A number of sites (Hae Ill , Taq I, Hha I, Tha I, Bst NI, Mbo 11 , Alu I) are found in the 3' half of both cloned adult genes but are not present in the corresponding positions in the cDNA clone , suggesting that these sites are within IVS II . Two IVSs have been located in the mouse aglobin gene between codons 31 and 32 and 99 and 100 (Nishioka and Leder, 1979). These locations are coosistent with the approximate positions of the human a-globin IVSs I and II (Figure 3). It seems likely that the human and mouse a-globin IVS positions are identical, since the locatioos of two IVSs in all mammalian p-globin genes examined thus far are identical (for references see Hardison et al. , 1979) and are . analogous to the positions of the IVSs in the mouse a-globin gene (Nishioka and Leder, 1979). The results presented in Figure 3 indicate that the coding , intervening and flanking sequences of a 1 and a2 share considerable homolog,'{. The locaitions o1 restriction sites in the two genes can be aligned by assuming that IVS II of a1 contains a block of approxima1ely' seven ms<:leotides which is not found in a2. This unshared region can be located precisely as an insertion within the few nucleotides separating the Bst NI and Mbo II sites of a2 (Of a deletioo from a1 ). The homology between the coding sequences of a1 and a2 is COflsistent with the tact that only one a-globin protein (Dayhoff, 1972) and mRNA (for references 04!10.2] see Forget et al., 1979) sequence have been reported . However, we were surprised lo observe the homology within intervening and Hanking sequences which is indicated by the similar positions in al and a2 of restriction sites in those regions , since other globin gene pairs show little homology in those regions (Konkel , Maizels and Leder, 1979; Hardison et al., 1 979). Heteroduplex Analysis of a1 and a2 The homology indicated by the results of Figure 3 was demonstrated directly by examining a heteroduplex between 4 .3 and 5.7 kb Hpa I fragments of >.HaG2 which contain a2 and a1 , respectively (Figure 4A, 1 3). Alignment of the restriction map of Figure 2 with length measurements of a1 -a2 heteroduplexes indicates that the region of homology begins 100 bp 3' to the mRNA coding sequence of each gene and continues through the genes into the 5' flanking sequence for a total of 1 .8 kb (Figure 4B, 1 and 3). Beyond this, we observe a small bubble of nonhomologous DNA. a short duplex region , a second , larger nonhomology bubble and finally a 1 .0 kb stretch of homologous sequence which extends to the 5' end of the Hpa I fragments. The large nonhomology bubble contains at least one inverted repeat sequence which is evident in the alternative structures shown in Figure 4A, 2 and 3. Since homoduplexes formed by the a2 Hpa I fragment also contain the inverted repeats (Figure 4A, 4 and 5) and no inverted repeats are observed in a1 homoduplexes, we conclude that the inverted repeats in the a1 -a2 heteroduplex (Figure 4A, 2 and 3) are contributed by the a2 strand. Formation of this hairpin loop . results in a reduction in the length of the duplex region 5' to the large nonhomology bubble and a concomitant increase in the length of the shorter strand of the bubble. This suggests that part of the sequence in the hairpin stem is also present in the 5. 7 kb Hpa I fragment. To determine whether there are additional regions of sequence homology 3' to a 1 and a2. a heteroduplex 32 c.tt 122 ,,,, es, Hot Toq HIia ThoBstMloAii H!lfl I I I I "' \ I Alu I / I Alu / I MtooAlu Hcl/l / \ a2 at lilbO Alu I Smo 1100 tOOO I I 900 100 1'00 I I!H,,ri Smo 5"'° 600 500 •oo I ,., I es, 5lftC :100 lilbO Hfl cONA 200 / HiNI \ Hoelflo8st Alu Alu , I 1 'I 1Toq1111 II1 •I Tholb) Pw Psi O tOO 200 300 -00 500 Soc! I Hool I l I Hp!\ tOO 8SI / 600 1'00 100 900 t000 1100 bp Hind A,tu Hp!\ Alu 1 11·----••-•-• Figure 3 . Comparison of the Location• of Restriction Sites in a 1, a2 and ct-Gtobin cDNA Positions of restriction endonucleue cleavage artes wrtflin al and a2 were determined aa deacribed in Experimental Procedures, excaot for a s,tes wniclt _,.e de termined by partial 0< double digestion expenments. Sites wrtf\111 tne cDNA - e determined by Forget et al. (1979). Slacl< bO.a:e.s reQfe&ent sequences present in mRNA. and tne unshaded reok>ns between tt\esa boxes represent interven1nQ sequences. Abbreviations Sltown in the liQure cottesQOnd I<> tile following enzymes: (Alu) Alu I: (Bst) Bat NI: (Hae) Hae Ill; (Hha) Hha I: (Hind) Hind Ill; (Hint) Hint I: (Hpa) Hpa II: (Hph) Hplt I: (Mbo) Mbo II: (Pst) Pst I; (Pvu) Pvu II: (Sma) Sma I: (TaQ) TAQ I: (Thal Tha I. In cues wl>ere the cleavage lite diflen from the recognition s.ite tor an enzyme (fOf exampte, Mbo ti) the poSJtion of the cleavage site is indicated. The number of b&N paira between each site and tile Hind Ill site within eeclt g - was determined USlt1ll labeled fragments -1¥9d fr0<n tile cONA clone aa standards. between DNA of subclones containing the 3' half of a 1 or a2 plus adjacent 3' sequences was examined in the electron microscope. This experiment confirms that the a1 -a2 homology extends only 100 bp beyond the sequences encoding the 3' ends of the two mRNAs (Figure 4A, 6 and Figure 48, 2 and 3). No other sequence homology is detected except for a 260 bp duplex region located about t kb 3' to each gene. This 260 bp region may correspend to a sequence which is highly repeated in the human genome and is found at several locations within the ,8-like globin gene cluster (Fritsch et at., 1980). This is suggested by the fact that a hybridization probe containing this repeat derived from the ,8-like gene cluster hybridizes to four different restriction fragments within the a-like globin gene cluster (E . Fritsch and R. Lawn, personal communication). One of these fragments contains only 400 bp of human DNA and maps to the same position as the 260 bp region of homoJogy which is located 3' to a1 . The locations of homologous seque,,ces determined by heteroduple:,: analysis are aligned with a restriction map of the gene cluster in Figure 48, 3. This alignment reveals that a1 and a2 are each located within an approximately 4 kb homology unit interrupted by two regions of nonhomology. Under the conditions used in the formation of the heteroduplexes described above, a sequence with as much as 2530% base mismatch would be scored as duplex (Davis, Simon and Davidson, 1971 ). The actual degree of sequence homology between the a1 and a2 regions is therefore not known. Specific Deletions In >.HaG1 and >.HaG2 The restriction map of Figure 2 is consistent with data obtained from genomic blotting experiments (Orkin, 1978; Embury et al., 1 979; S. Embury et at. , manuscript submitted; our unpublished data), indicating that the a-globin gene region can be cloned without detectable sequence rearrangements . However, both >.HaG1 and >.HaG2 give rise to deletions at high frequency. When these phage are purified by CsCI equilibrium sedimentation, two bands are observed, with the majority of the phage in the upper band. The Figure ◄ . He!<KO<l~x Anatyaia o! the Adult a-Glooin 0 - Region1 (Al Electron mierograi:,lts and interpretive drawings of hetero- and homodul)lex• formed by rea1riCtion fragments or plasmid ONAa containing a1 and a2 . The arrow in each tracing indicates tile l)08ition o! tile mRNA coding f9Q1on and the direction of tnonacriC>tlon. (1-3) Heteroduc,iexes fo<med between tile ◄ . 3 and 5. 7 kb Hpa I fragments of >.HaG2 DNA (tltese fragments contain a2 and a1 . reec,ectiwfy). 80% of tile molecules are o! the type I/town in (1 ) . Comparison o! the summed tengtlts of ainQte- and ~stranded regions with the known len<;llhs of the two Hpa I fragments indicates tllal the 1.03 kb llrand of the larger nonhomology bubble ia contrtbuted by the ◄ .3 kb Hpa I ~ t . F,__tty lhia lingleatranded DNA t0<ms • hairpin loo!> (2 and 3). (◄ and 5) Hoffiodul)lexes formed by clenaturatlon and reaaaociation o! the ◄ . 3 kb Hpa I fragment . Tl>e lengttt of Clu!>iex ONA 5' to the hairpin looc>I ia tit• same u that 5' to tile large nonltomo4ogy bubble in tile heteroduplexes of (2) and (3). (6) Heterodu()lexes formed between tile plaamtd pRHa2 CU1 wi1f\ Hind Ill (tttia molecule contains tile 3' half or a2 r:,lua 3' ftanking NQUenCH) and ttte plasmid P8Ra 1 cut witlt Hind UI plus SaJ I (tltia molecule contains the 3' half ol al r:,lua 3' flanking aeQuences), Comparisons o! the summed lengtlts Of single- and double-stranded regions wi1f\ tile tengtha of tile human DNA _ , . in pRH end pRB indicate tllat tile 1, 17 and 1.01 kb strands sre from tile regiona 3' to a2 and a1 , respectively, (Bl Schematic drawings showing tile locations o! homologoua and noo •"tomologoua uq...,,.,.. between tile a 1 and a2 regions. ( 1) A heteroduplex exemplified by the molecule of (A 1). The double- and single-stranded DNA ragiona are rep-nented by thick and tNn lines, r ~ , and the lenQtha are gr,en in kitobase pairs and kilonucleotides, respectMlly . (2) A heteroduplex exeml)ltfled by the molecule of (A6). (3) A map I/towing tile PoSitlons o! restriction enzyme lites in relation to tile tltree regions of uquence homology ,_._,led by tile croasha1clted boxes, tile emaJf OQert boxes and tile stll)l)led arrowa. The amaH arrows reprnent tile 260 t,p region of llomoloQy located on the 3 ' o! al and a2. The OQert area a between tile boxes and arrows corrffl)Ond to reg,ona ot nonhemology. 33 Human a-Globin 0 - Cluster 123 A B 1 f-03\ i. e ◄ _a_2_ _ _ 0_ .9 _ 8 _ - ~ 0.18 0 .06 0 al 0 .12 -------,lo ,1 - ■ 2 c2 Pw Hi>oSoc Smo Pw □ Smo Hmd Pw- Soc al llql ,...________;...> r n s- Hind P.., o ,_ .. --·..•·--·...-·-----> 34 Cell 1 24 restriction map of the human DNA insert in lower band phage DNA corresponds to the map derived by genomic blotting experiments, whereas two types of deletions are detected in upper band phage DNA of both AHaG1 and AHaG2 . The breakpoints of these deletions were mapped as described in the Appendix and are shown in Figure 5. The breakpoints of the leftward type of deletion which removes 4.3 kb of DNA map within regions of sequence homology located approximately 3-4 kb 5' to a1 and a2 (Figure 5A). The distance between corresponding restriction sites within these regions (for example. Hpa I to Hpa I or Sac I to Sac I; Figure 2) is also 4 .3 kb, strongly suggesting that the leftward type of deletion is generated by unequal crossing over between homologous sequences. Similarly, the breakpoints of the rightward type of deletion which removes 3 .8 kb of DNA map within the regions of sequence homology within and immediately flanking a1 and a2 . The corresponding restrict ion sites in these homologous regions are separated by 3.8 kb (for example , Hind Ill to Hind Ill or Pvu II to Pvu II; Figure 2). suggesting that the rightward type of deletion is generated by unequal crossing over between these homologous regions . Thus the occurrence of both types of deletions must be a consequence of the distribution of direct repeats around the a1 and a2 genes. We do not know whether this recombination is intra- or intermolecular. Since recombination can occur anywhere within the repeats , the breakpoints of the deletions cannot be mapped precisely. In fact , the leftward and rightward types of deletions might each consist of a number of different deletions generated by recombination at sites anywhere within the region of homol- ··- 2 ....... , - - a : J -- ---t11DD>------.1no- --•-ca--c== --,,---i .. _, t - - - - 11 - - - - j ·· - 1 111 ~ ~ " __,... z .... , -tD-----.-D>------.IDD>----G-<a--C== . .. F'IQUfe 5. Locations of Oetetiona In AHaGi and .\HaG2 ONAa Restriction enzyme clnveoe lites are lndicallld by VW1lc# ..._, and the m• of Mell lraqmen/ le giv«I in kilobaae pairs. Fragments dltteetad onty i n - • be.nd pt,age ONA are rltl)rnentlld by horizontal .,es interrupted by a bracketed reoion which lndlcateo the approximate location of a deletion. The lines labeled ··..-ll'lal 2" indicate the poaition8 of deletions In tne ONA of lndMduals wilt! ..-thal.(Embury et al., 1979; Orkin et al., 1979: S. Embury et al ., manuacnpt IUbmined). The dashed arroww below tne brackets r l t l ) r - the limits OYtt< which the deletion breakl)Oints can occur In elll1er the cloned ONA o, the a-lhal 2 deletion&. The length of each arrow equals the length of the region of hOmoiooY (Figure 48). (See Appendix tor a discussion of lhis point and a deacription of the mai,p;ng data). (A) Leftward-type defetlon: (B) ngl!twarcl-lype deletion. ogy. This uncertainty is represented by the dashed arrows in Figure 5. Characterization of Linked a-Uke Globln Sequences The restriction fragments which hybridize weakly to the a-globin cDNA plasmid (Figure 1b) were located on the map of Figure 2 by blot hybridization experiments and DNA sequence analysis. Table 1 shows a portion of the nucleotide sequence of the two nonadult a-like regions which can be aligned with codons 73-96 of the a-globin mRNA sequence. This sequence information precisely locates each a-like sequence on the map of Figure 2 and establishes the relative transcriptional orientation of each gene. In both non-adult a-like regions the mRNA sequence is read from left to right (5' to 3') in the map of Figure 2. a1 and a2 have the same transcriptional orientation . An embryonic a-like polypeptide, t has been identified (Huehns et al., 1961 ; Capp, Rigas and Jones, 1967, 1970; Weatherall , Clegg and Wong , 1970; Todd et al ., 1970; Huehns and Farooqui , 1975; Kamuzora and Lehmann, 1975; Gale, Clegg and Huehns, 1979). The chromosomal location of the r globin gene has not been established. To determine whether either of the a-like sequences linked to a1 and a2 encodes an embryonic f-globin, the amino acid sequence predicted by the nucleotide sequence corresponding to codons 73-96 of each gene was compared with amino acid sequence data for the r protein. Neither sequence could be aligned with a preliminary sequence derived by matching the amino acid composition of rand a chain peptides (Kamuzora and Lehmann, 1975). However, recent peptide composition data derived from a more purified preparation of r chain (J . Clegg, personal communication) provide a sequence for codons 73-89 which is identical to that predicted by the nucleotide sequence of the 5'most gene. We therefore tentatively identify this sequence as a f-globin gene. The a-like sequence between rand a2 does not encode any known globin polypeptide (for convenience, we refer to this sequence as ,?a1 ). The complete nucleotide sequence of this gene has been determined and will be reported elsewhere (N . Proudfoot and T. Maniatis, manuscript submitted). In addition to the four a-like sequences contained in >.HaG1 and AHaG2, we have detected one other a-like sequence in the human genome . DNA from four individuals was digested with the restriction enzyme Bgl II and probed with a labeled restriction fragment containing the 5' half of the cloned f-globin gene. Comparison of the map of Figure 2 with the autoradiogram of Figure 6A indicates that each of the four DNAs contains a hybridizing fragment of 1 2 kb which corresponds In size to the AHaG 1 Bgl II fragmen t containing a f-globin gene. In addition, at least one other Bgl II fragment not found in the cloned DNA is 35 - a-Glol>in Gene Cluster 125 (.) (.) "'"' 0 0 0 :::, "'"' (.) :::, 0 0 0 < (.) (.) (.) (.) (.) (.) (.) (.) (.) ii > 5 'i> A iE .. ... 2 ,: £ ct ct 0 0 < < < i i i "'• 0 0 0 < < < "' > 0 0 (.) 0 0 :::, 0 !? < ,:= :, :, (.) (.) (.) ..,j ..,j 0 0 (.) > 0 g << < 0 (.) (.) :::, < .,,. .ii ..,j 0 12.0- --a::s 10.9 _ _ .. 'i • ..,j (.) (.) (.) (.) (.) (.) (.) C> ~ U C (.) (.) ~ (.) (.) :, :, :, ..,j ..,j ..,j 0 " o" I C ~ <O ., < << 0< <i .,,. a "1?,: :l1.. ,n u C, C, :, C, Cl., - ..,j u .., < < < "' < "' ~..~ ; 0 .,.., .. .. .. ..8 ".,.. ., ., u < < .. a ., ~1 < < .i:, (.) C, (.) 0 0 0 ,ii C> ~ ,ii C, :::, 0 :::, :, :, :, (.) (.) ..,j ..,j ..,j I,) 0 C, :, C, 0 ,. (.) (.) :::, ~ Q, ,( 0 N C 0 o " ~ .!i D 2 ~ :::, u :::, ..,j (.) u :::, " :2 u ~ z :, "' l .., ~ ..,j ..,j j ! c " J "'2 .5 .. g ., 16 ::, :::,0 0:::, .. 'i> il .!-~ .!!. C, :::, ., i:: "' C, 0< 0 <• <i < ..., !g 's.., ... .g ., u :::, g ; il.E ,: 5i z ... < < 0 c3 C, C, C, (.) (.) (.) (.) (.) ,!ii ..,j I C O' .. = 0 ... .... ... g < < (.) (.) (.) (.) (.) (.) (.) C, C, < < 0 0 ii i a ..i1,.., C> ct ct - .,... 0:::, :::,0 g :I;; :I;; .. i ~ < < < - & ~ .S..?!,, ~ 0 J! (.) < ~ 0 g (.) < 0 (.) < C, j < i < - :, . -- i i & < ~ 0 il .. ...... 0< 0< 0<" <i <i <i = .E '::;) z 0 ..,... 0:::,0 5< :::, 'i i ! jJ < > !i a ~ 1! • i. • i. < E 0 (.) .§ ..,j Q, - C ...i ! c -6 .8 -5 .2 F",gure 6. Identification of a Second, Linked i-Globin Gene (A) ONA from four individuals (lanes 1-4) was diQested WIiii 89111 and the products were trachonated on a 0 .7% aoarose get . transferred to nitroceUuloM (Soutl>am . 1975) and hybridized witt, a "P4abeled P,.,u 114-iinc II fragment containing Ille 5' haH of p . (8) ONA from individual 1 of (A) was digested wi1li 891 II plu1 Hp.a I (lane 1), 891 II plus Eco RI (Jane 2) <>< 891 II (lane 3), fractionated on a 0 . 7'11, agaroae gel, tranaf8ff9d to nltrocellulose and probed wi1h a "P-labeled Eco RI-Sac I fragment contain,r,g the tenninal 0 .4i kb of 1he human DNA lnaen of >.HaG 1 . (See Figura 7 for the location of tliis fragment .) "1? C, (.) :::, : l: -- - ~1 ,:O' ;: (.) ~ (.) (.) <D -- mv z .., ... < :::,< < i.. i.. :f <'° 0 u .. "' 0 0 0 :! .5~ ., :::, :::, u:::, ., .. (.) • - : ;;G !.; .. i5 . "'"' < < :::,< i I ...>- ..... -~ " .., C, .,"' C, C,:::, 0 <.. >ii <!! !;= (.) 3 --10.9 u <"' .. "' :::, 0 0 ;; ;;; :::, :::, :::, ., ., il ...a,~ N 2 ~ E C ~ ~ 4 1:. .., 00:::, g0 00:::, 'i 'i 'i .2.., > B 3 i ii§~ detected in DNA from each individual. The size of this latter fragment varies between individuals, suggesting the presence of restriction site polymorphisms in the regions flanking the f-like sequence. Individual 4 appears to be heterozygous for polymorphisms in the Bgl II sites surrounding both f-globin sequences . since four different Bgl II fragments are detected (Figure 6A) . To determine whether the second f-like seouence is a nonallelic f-globin gene, we established a physical map of the restriction sites surrounding this sequence (Figure 7). The sizes of hybridizing fragments resulting from single or double digestion experiments are presented in Table 2. Analysis of these data made it possible to construct a restriction map distinct from that of the cloned f-globin gene, indicating that the hybridizing seouence does in fact correspond to a aecond f-globin gene. The data in Table 2 also suggest that the two sequences are linked, as indicated in the map .of Figure 7. Linkage was definitively established by genomic blotting experiments -using a terminal fragment of ,\HaG1 as a hybridization probe. As expected from the map of Figure 7, this probe detected the 10.9 kb Bol II fragment containing the f-like sequence, as well as fragments of 6.8 and 5.2 kb resulting from double digests using Bgl II plus Eco RI and Bgl II plus Hpa I, respectively (Figure 68). The sizes of these latter r 36 Cell 126 {2 •• ':" ':' -II '! ' ~ ' ~ 13 ' ~ ' ~ ' '! ' ' '1 ' ~ ' 1' ~ ~ Ml , .2 &o RI 22., 120 10.9 IQI II II 7 Hpo I ...... 1 '? 117 10.3 '-9 10.6 •on I Table 2 . S.zea of Fragments Detecled by p Proo. on Genomic Blots Hind 1H 16, 13 Hpal 11 . 7 doublet Eco RI 22.5, 5 .2 BamHI 10.3 , 5 .9 Bolll 12.0, 10.9 Kon I >'0. 10.6 Bgl N/ Hind Ill 10.2, 5.1 8am RI HI / Eco 10, 2 .8 Bol n/ Eco RI 12.0 . 4 .2 8am HI / Bol 11 3 .0 , 1.9 Hpa I/ Hind Ill 11 .7 , 11 . 1 Kon I/ Hind Ill 10.2 , 15.5 8am HI / Hind Ill 9 .0, 5.8 Pvu II 2.0 , 1.5 Sall Xho I >'0 >'0 ONA lrom individual 1 of Figure 6A was dioested with the indicated enzymes. lraChOnllted on O. 7'11, aQarose gels. lransferred to nitrocellutoae . and probed with a up~a~ Pvu II- Hine ti fragment contain. ing the 5' hall of p . The sizes of hybridizing lr"9ments resul ting from each dioast are g,ven in kdobase pairs. fragments are those predic ted by the linkage map of Figure 7 . We therefore conclude that the !-like sequence corresponds to a second !-globin gene located 10-12 kb 5' to the cloned {-globin gene. For convenience, we will refer to the cloned {-globin gene and the {-globin gene identified by genomic blotting as t1 and s2, respectively. !1 and s2 appear to share considerable sequence homology. Identical strips of Bgl U or Hind Ill blots were hybridized with f1 f)(Obe and washed in 0.33, 0 .1 6 or 0.08 x SSC at 68°C. Both the r1 and ri bands are still visible and are eciually in.tense atrer washing in 0 .08 x SSC (data n.o t shown}. 011scuulon Llnug4l Arranqement of Globin Genes We have used molecular cloning and genomic blotting procedures to establish the linkage arrangement of the entire human or-like globin gene cluster. The cluster consists of two adult or-globin genes (or1 and or2), an apparently nonfunctional or-like gene (,j,ar1) and two embryonic f-globin genes (f1 and s2l. An exact correspondence between the amino acid sequence encoded by a short region of r, and the sequence of a2 al ~ ' l ' Flgunt 7. Unkaoe Arr81'Q9ffl9nt of Emb,yonic and AduH a-Like Globin Genes The lizN of fragments detected by a P probe (Table 2) were used 10 construct a map of restriction sites surrounding the i2 sequence. The solid line connecting p , ,j,a 1 . a2 and a 1 Indicates the ONA region which has been cloned. The locations of the p prooe, and of Iha I.HaG 1 terminal fragment probe used in the ex""""""'t of P,gure 6B , are indicated by bars the kb scale. The asterialc marks a Barn HI Site p r - in the cloned ONA but notintheONAof indiYidual 1. the corresponding region of the {-globin polypeptide suggests that t1 is a functional globin gene. This suggestion is strengthened by comparison of additional amino acid (J. Clegg, personal communication) and DNA sequence (C. O'Connell, personal communication) information. r2, on the other hand, has been demonstrated to be a functional f-globin gene by analysis of a deletion associated with a case of hydrops fetalis (L. Pressley, B. R. Higgs, J. B. Clegg and D. J. Weatherall , manuscript submitted). The r, sequence information and the fact that this gene is indistinguishable from the functional s2 gene in hybridization-melting experiments suggest that there are two functional f-globin genes. The chromosomal organization of the human or-like and ,!I-like globin gene clusters is very similar. Within each cluster all the genes have the same transcriptional orientation. The or-like genes are arranged in the order 5'-embryonic-adult-3' (this paper), while the ,!I-like genes are arranged in the order 5' -embryonic-fetal-adult-3' (Fritsch et al., 1980). A similar organization has been reported for the ,!I-like globin gene cluster of rabbit (Hardison et al. . 1979; Lacy et al., 1979) and the or-like gene cluster of chicken (Engel and Dodgson , 1980). An apparent exception to this pattern is the chicken ,!I-like globin gene cluster. whece a se~uence which preferentially hybridizes embryonic globin mRNA is located 3' to an adult globin gene. This sequence. however, has not been definitivety identified as a funclionaf embryonic gene (Dodgson. Strommer and Engel, 1979; D. Engel, personal communication) . The possible significance of globin gene linkage has been discussed elsewhere (Fritsch et al., 1979). Another feature common to globin gene clusters is the presence of globin-like sequences such as ,i,ar1 which cannot be identified with known polypeptides. The complete nucleotide sequence of ,j,ar1 indicates that it cannot encode a functional globin polypeptide (N . Proudfoot and T. Maniatis, manuscript submitted). Two globin-like sequences which have not been identified with polypeptides are found within the human ,8-like gene cluster. One of these sequences (i/,/I2) is located 5' to the E-globin gene and the other (i/,/11) is located between the •-, and cS-globin genes (Fritsch 37 Human a-Globin Gene Cluster 127 et al., 1980). Similarly, an apparently unexpressed ,8-like sequence (JJ2) is found 5' to the rabbit adult ,8-globin gene (Hardison et al., 1979; Lacy et al., 1979). These sequences may represent vestigial genes which resulted from gene duplication followed by sequence divergence. In many cases. globin genes immediately adjacent to one another are expressed at the same stage of development. Examples of this pairwise arrangement are the human 8 and ,8 (Flavell et al., 1978; Lawn et al. , 1978; Mears et at., 1978), 0 y and •y (Fritsch et al.. 1979. 1980; Little et al., 1979; Tuan et al., 1979) and a1 and a2 genes (Orkin. 1978), the rabbit ,83 and ,84 genes (Hardison et al. , 1979), the mouse ,8...., and p,_, genes (Konkel et al., 1979; M. Edgell, personal communication) and the chicken a. and a0 genes (Engel and Dodgson. 1980). The single rabbit adult p gene (Hardison el al., 1979) and the human E gene (Proudfoot and Baralle, 1979; Fritsch et at., 1980) are exceptions . However, each of these genes is immediately adjacent to a ,8-like sequence of unknown function . In many cases. comparison of the members of a gene pair indicates that the intervening and flanking sequences are more divergent than the coding sequences (Hardison et al., 1979; Konkel et al. , 1979). The human °y and •y genes, however, are essentially identical throughout their coding, intervening and flanking sequences (J . Slightam, A. Blechl and 0 . Smithies, manuscript in preparation), as is the case for the human a-globin genes. Maintenance of Sequence Homology between Adult a-Globin Genes during Evolution Genetic studies and structural analyses of a-gfobin polypeptides from a variety of vertebrate species indicate that duplication of adult a-globin genes is not limited to humans (Clegg , 1974; Kitchen, 1974; Nute, 1974 ; Russell and McFarland. 1974; Dresler et al., 1974; Chapman , Tobin and Hood . 1980). Comparison of a-globin amino acid sequences between various primate species reveals differences of 1-7%, which are consistent with !he expected rate of divergence of glob1ns over evolutionary time (Dayhoff, 1972). However, the two adult a-glot:>m proteins within a species show considerably less dive,-gence. F<:Jf example, human and gonlla ..-gtobins differ by 6%, yet no differences are obset!Ved when the intrasoecies comparisons are made . The observed similarity between the a-globin proteins within a species could result from recent independent duplication events or from a mechanism for gene correction . Comparison of aglobin genes from a number of primate species has revealed a remarkable similarity in the distribution of restriction sites surrounding the genes (Zimmer et al. , 1980), suggesting that the a-gfobin gene duplication occurred prior to the time of primate divergence. This observation, in conjunction with the evidence for a- gfobin gene duplication in most vertebrates, indicates that the a-globin gene duplication is ancient, and therefore favors the existence of a process which leads to sequence matching between a-globin genes within a species. Maintenance of homology among a family of evolving genes within a species has been termed coincidental evolution (Hood, Campbell and Elgin, 1975). Coincidental evolution can be imposed by natural selection, or by a mechanism for gene correction . Two such mechanisms are gene conversion and expansion and contraction of gene number by homologous but unequal crossing over (Smith, 1973; Tartof, 1973). Comparison of the complete nucleotide sequences of the 0 y- and •y-globin genes has led to a specific proposal of intrachromosomal gene conversion (J. Slightam, A. Blechf and 0. Smithies. manuscript in preparation). Similarly, intrachromosomal recombination has been proposed as an explanation for the presence of an identical restriction enzyme. site polymorphism within the intervening sequences of the two linked human y-globin genes (Jeffreys, 1979). Evidence for unequal crossing over between human a-globin genes is provided by analysis of deletiontype a-thafassemias, which are caused by deletion of one (or, less frequently, both) a-gfobin genes (Dozy et al., 1979; Embury et al. , 1979; Orkin et al. , 1979; S. Embury et al., manuscript submitted). Moreover. chromosomes containing three a-gfobin genes have recently been observed (Goossens et al., 1980; D. Higgs and J. Clegg, personal communication) . Unequal crossing over has been proposed to explain the sequence homology between the human °y- and •ygfobin genes (Little et al. , 1979b). Coincidental evolution of duplicated a-globin genes resulting from gene conversion or unequal crossing over has been proposed independently on the basis of a comparison of the positions of restriction enzyme sites surrounding the a-gfobin genes in various primate species (Zimmer et al., 1980). ft is not known why intra- or interchromosomal recombination mechanisms would operate on the aand y-gfobin gene pairs and not on the other more divergent gfobin gene pairs mentioned above. It is possible that the a and --, genes lie adjacent to special sites which promote recombination (Stahl, 1979). Another possibility is that the homology units of the more divergent gene pairs were interrupted at some time after the duplication event by a random insertion of ONA into the noncoding region within or flanking one of the two genes. This region of nonhomofogy would be expected to reduce the frequency of recombination, as is the case in bacteriophage lambda (Sodergren and Fox, 1979) and in yeast (G . Fink, personal communication), and thus allow the flanking and intervening sequences of the two genes to diverge . Perhaps the large and small nonhomotogy bubbles 38 Cell 128 which we have detected 5' to a 1 and a2 are the results of such insertion events and the genes are in the ear1y stages of divergence. Globin Gene Deletions and a-Thalauemla The occurrence of deletions in clones containing aglobin genes and their correlation with the locations of extensive regions of homology surrounding those genes strongly suggests that intra- or interchromosomal unequal crossing over occurs during propagation of these recombinant phage in E. coli. II is remarkable that the locations of the breakpaints of deletions which occur in cloned DNA are indistinguishable from those associated with a form of a-thalassemia designated a-thalassemia 2. a- Thalassemia 2 has been shown to result from a deletion of one of the two a-globin genes (Orkin et al. , 1979). Two types of a-thalassemia 2 deletions have been characterized. One type correspands to the leftward-type deletion described above (Figure 5A) and is found primarily among Asians, whereas the other type corresponds to the rightward-type deletion in cloned DNA (Figure 5B) and occurs frequently in Asian, black and Mediterranean papulations ($. Embury et al., manuscript submitted). Thus the precise localization of regions of homology surrounding the ar-globin genes in cloned ONA and the occurrence of deletions in E. coli which appear to involve these sequences provide an explanation for the occurrence of specific deletions associated with ar-thalassemia 2. The structural studies reparted here in conjunction with studies of the molecular basis of ar-thalassemia portray a dynamic gene family undergoing rapid evolutionary change. This family includes genes which have specialized to function at different developmental stages Ci and a), a gene for which no globin polypeptide has been identified and which may therefore no longer be functional (l/,<r1 ), and genes which are presently undergoing deletion and duplication in human popylations ( ar1 and ar2). ~HaG2 ONA <& prOVlded by ~ Sol II &igeW ol band DNAs . Ol<,)est,on 01 tl«r!<l ONA _ , Sgt If ;r,oduoeea • '2 .0 lb fragment which contains ri1. ar,d a S .1 kb traoment ...,\Ch contain, a 1 and ••lends Into Ille l a - ..e\0< \fiQUNI BA: Figura 5). Tne 12 .0 kb !!QI H fragment ii mlUing fTom upper band ONA and the relative amount of Ille 5.7 kb fragment la reduced (Figure 88). This SUQQHIS tllll one type of deletion · - · part of Ille 12.0 and 5. 7 kb !!QI II fragments (rightward deletion) while another type of delehOn (lellw1td) is contained entirely wtlnin the 12.0 kb fragment. OQaring the 5.7 kb fragment . Tne miuing 12.0 and 5.7 kb !!QIN fragments in upper bend DNA are rel)4aced by fragmenu of 13.11 and 7.7 kb (F"tgure 88). The 13.II kb fragment cannot be I deleted torn, of Ille 12 .0 kb fragmanl and must lllerefore corrffl)()nd to tile ngntward type of deiet,on whic:11 removes part Of bolll the 12.0 and 5.7 kb fragments . The deletion event ,...,.,.,,.. the !!QI N aile whic:h ordinarily - " " llleM two fragments, reeu111ng in a fuaion of the rema,nong parts of thes4! fragments to y;ald the 13.9 kb fragment (f",gur1 5). In intact DNA these two frlQtMnts total 17.7 kb (12 .0 ptua 5.7 kb). 3.8 Sma I C 8 D Figure 8 . Comparison of R"tnctlon Endonucleue Digests ot Intact and Deleted .\HoG2 DNA The recomolnanf " " - AHaG2 was grown in liQuid cultur• and purified by CaCl - t l l t , o n aQUitlb<1um u de9cribed in Ex-.. !Mntal ProcedurH. The two " " - t>andl _.. coli.cted aeparately and rerun on aeparate ~ m gradients to obtain c,urffied upper and l>encl ONAs. These DNAs - · digested wtlll BQI " 0, Sma I and Ille products - • fractionated by eiectrOl)hO,H11 on a o. 7'lb agarou gel. (A) L"""" bend DNA d!Qflted "''"' !!QI II; (8) upper bend DNA digflted with !!QI It; (Cl bend ONA digHted with Sma I; (D) upper band DNA digested with Sm• I. Tl>e taint. high molecular we,oht tr1oment tn lanes C and O ia due to reauoc1alion of vector tragmanta II tile e<» tit• of lambda. kb (17 .7 minua 13.9 kb) haw been remoYed by Ille rightward type of deletion. The 7. 7 kb fragment muat result from Ille delehOn wnich is contained entirely within Ille 12.0 kg fragment (Figure 5). 4.3 kb (12.0 mlnua 7.7 kbl • • - by the - r d type of datehOn. Sm. f WU IIMd to locatll further the breli<Pofflll of " - two - Lo,ow,- ONA dlQe.- with Sma I Cliap4a)'I fragments of 4.5 ■nd , .45 ;.to lie to \el! of a2. ltnd 1 3.0 kb 1nog,_,, whic:11 conlaina moat of a2 and moat of ... region e2attdo t (Agur.6<:; Figure 5). In - b a n d ONAbothllle4 .5 llld , .45 t.t> Sir.a I 1nf;ments are reduced In amounl , _ tc !flat ot non-ts unaffected by the d■letlona , and Ille 3 .0 kb fraQment is - • (Figure 80). Botti the ten..ard end righ...,lfd types of delet,ons must lllera1ore ,.,,,.,.. part o, •• of Ille 3 .0 kb freg..-t . which i5 conaistent with tile tact Illa! no a2 coding uquenc., can be detected (data n o t _ )_Because tile 4.5 kb fragment is reduced., amount. Ille leftward type of deletion must axtltnd fr0<n a2 to include Ille Sma I lite which lies b e - tile 1.45 and 4.5 kg tn,gments (Figure 5). The 5.7 kb fragment r:,rewnt In uPl)er bend ONA dioested with Sma I (Figure 80) la most llkely • fusion fragment ,-,tt,ng from tile leftward-type deletion (Figure 5). A small amount of me 1.45 and 4.5 kb tragm«1ts remaina, · 10 mos1 o, al ot me rlQhtward-type deletion& must lie to the right of the 1.45 kb fragment, beginning within a revion which inctudeS the 0.45 kb Sma I fragment ltnd tile a2 gene (f",gure 5). Tne p r _ . of an -,,ximately 4 kb deletion ha& been confirmed by •umtnation of hetarodul)lexH belwffn and bend DNAa (data noc lhown). !ypeS o< - lntonnation • -ding the for_._.,,.,. lenQIII& "' _..., reg'°"" in Bgl II A 39 Human o-Globin Gene Cluat• 129 leolatlon and Charac1erizatlon ol Human a-Uk• G-.. Ge,,A bacteriophage A library or human letaJ IMlr ONA <uwn et al., 1978) waa s c r - witn "'P-labeled a-Qlobln cONA plumod (JW101 : eon et al., 1978) using procedurn described by Frit9ch et al. (1980). The isolation al recombinant bacteriopnage ONA. tne recover, o1 -ifie r,.aoments from agarose a, aerytamide gela, tne mapping o1 restrietion endonuelease cleavage sitn arid blotting arid hybridization e,q,erimenta were earned out using pubilsned proeedunn (Maniatis et al., 1978: ucy et al .. 19711: Frit9eh et al., 1980). unpu!>lianed ,-,,11a. We are grateful to L I-food, N. Proudfoot arid B. s.ec:t for diaeuuiona. We thank A. Cortenbac/1 fa, preparing media and materiala. J. L arid CA<. J. S. - . . supported t>y NIH graduate and poeldoctoraf training granl9, rffl)8CtlW/y. to tne C&lifomia 1n,1... tut• o1 TeehnoioQy. arid T. M. supported t)y me Rita Allen Foundation. Thia research was funded by a grant from tne NIH . The eoatI of i:,ubileatlon of tt'lia artlele were defrayed in part by the peymenf of peoe eharQes. Thia artlele must therefore be hereby mwl<ed ·-- , -· 1n aecordance wttll 18 u.s.c. Section t 7~ - t y to Indicate thie tact. Received January 22, 1980; ra,riMd February 25, 1980 Construc:tion or Plaamld ~ Restriction endonueleaae fraomeni. of human ONA - • ieolated from AHaG1 a, AHaG2 arid eloned in pBR322 (Rodriguez et al., 1978) dlQeoted witn the apprapriate enzymes, esaentlally u described by uey et al . (1979). pRBa t extends from tne right-hand AHaG2 amfieial Eeo RI site (lhia site waa ereated by tne addition of Eeo RI linkers during eonatruction of tne human library) to tne Bol II site b e .,, arid o2 . 1>RH02 exterodl from me AHaG1 riQht-harld artificial Eeo RI srt• to Ille Hind Ill site witnin o2. pHBa2 ..,, exterodl from me Hind Al site within o2 to the right-hand Bem HI site of me pair of sitea located at about 5 kb In me map of Figure 2 . 1>BAi axtends from the lelt-hand Bem HI site <Flour• 2) to tne arflfk:lal Eeo RI site whietl eonatitutea me lelt-hand boundary of AHaG 1. - s ~ Rntric:tlon MaPIN"liJ In those eases - • single a, cloubie reetrietlon endonuelease ~ lion yielded nurneroua small ONA fregments. the a , - of theae fragments was determined ua&no the doutMe-strandeO exonucieaae of Alteromonaa espe1iana BAL31 , wNeh was t)rOYKled by H. Gray. A plasmid subclone waa ltnearized by eleavage at one of me sitn bounding the human ONA inaert. and then exonuclease and rNtrictton erodonueteaae digestlona were eanied out aa deeenbed (LeQerski. Hodnen and Gray, t 978). The locations of restriction aites in o t . o2 arid me a-Qlobin eONA -.-e com pared by digesting pHBo2,j,at , pRHa2, pRBot and JWt Ot witn Hind Mt arid e n ~ with »p _ l)HBo2,/ot and pRHo2 were then rediQH1ed with Pvu II, pRBot wu rediQeated with Ps1 I and JW10t was redigested wrtl'I Hint I, and end-labeled gene-conta,nlnQ fregments were isolated by c;iel purification. These tragment1 - e then digested with I variety of enzymes and tne lizes of rnultlno traQmenta were determined by MetroPho,eala in parallel wttt, labeled markera in an aer,iamode get. Bunn. H.F .. Fo,gat. e. G. and Ranney, H. M. (197n. Human Hemo-globina (PttH-pNa: w. e. Saundenl. Cal>!>. G., RiQaa, 0 . a r i d -· R. (111&n_ Science 157, fl!S-66 . Caoo. G.. RIQU. 0 . a n d -· R. (1970). Natura 228. 278-290 . Chai)man, B. . Tobin . A. and Hood. L (1980). 0.,,. Biol., in preea. Clegg, J. B. (1974). Ann. NY Aead . Sel. 241 , fl1-69 . a..... R. w.. Slmon, M. and O.vidaO<l. N. (1971). Meth. Enzymol. 210, 413 ◄ 28 . 0.yt,olf, M . 0 . (1972). AtlU o1 Protein SeQuene• arid Struetun, . 5 (WUhinQton. 0 .C.: National Biomedieal R-areh Foundation). Oel.-oth. " ·· Nlenhuia. A. , Turner. P .. Velez. R .. Anderson. W. F .. Ruddle . F.. u........ee. J .. Cntaoan. R. arid Kueherl&l>ati, R. (1977). Cell 12, 205-218 . Oei.-oth, A., Nienhuis. A.. Lawrence, J .. GIies. R. , Turner. P. and Ruddle , F. H. (1978). Pnx:. Nat. Aead. Sci. USA 75. t 456-1460 . Oodgaon, J. 8 ., Stromrner, J. and Engel, J. 0 . (1979). Cell I 7. 879887 . Oozv. A., l<Ml. Y. w.. Embury. s .. Mentzer. w.. Wano . w .. Lubin. a.. Devis. J. arid Koenig, H. (1979). Nature 280. 605-607. er... s. L . Runkel , o.. Strenzel, P .. (1974). Ann. NY Acad. Sci. 241 , 411 ◄ 1 5 . - e. and Jones, R. T. Embury, s.. LAl>o. R .. Oozv. A. and Kan . Y. w. (1979). J. Clin. lnveat. fl3, 1307-1310. E"lile\, J. 0. and Oodgaon, J . B. (1980). Pnx:." Nat. Aead. Sci. USA, lnpreas. Flavell. R. A., Kooter. J. M., Ce Boer. E.. Uttle. P. F. R. arid Williamson . R. (1978). Celt 15, 25-41 . DNA Sequence ANlyel• o8R{ ClffA waa diges1ed witn Hine 11, end-labeled wttt, »p, '9diQnted wrth Sae I, eluted r,.om an aerytamode get arid aequeneed by the procedure of Maxam and G~bert (1977). pHS..2,/ot ONA was dignted with Pvu ti , - rediQNted with Sae I a n d ~ asabove. , Fo,Qet, B. G .. 0.vallnc:o. C.. de Riel, J. K .. Spntz _ R. A .. Choudary, P. v .. wuaon. J . T .. Wilaon. LB .. Red<Jy, V . B. and Wetsaman . S. M . (1979). In EuUryone Gene ReouLat,on, R. Axel , T. Maniatis and c. F. Fox.-.. ~ YO<k: -.n.e Praaa). Et.ctron M i a _ , To examine the extent of homoloQy b e - _ . . within and 5' to c,I and o2 , >.HaG2 ONA was dlQeeted witn Hpe I and the 0 containing fragments (5. 7 and 4.3 kb, raspeetlvely) _ . purified by aoarose gef eleetroC)llocwa. He\erodu1)1e•"" - e formed in O. t 8 M NaPO, at &e• c. u clHeribed previouaty ~ and Maniatia, tll$l). To examine the homok>Qy b e - the 3' flanjcJng MQuencH c,,I the two genes. the p1. - pRHa2 and pRBo t • d!Qested -1-ilnd Ill and I-find Ill plull Salt. • -- ¥- Tl'Nt ~A - ~ t n o et ed. - •xlnK:fed and ethano1-9<eciQ,tated. and lleteroduQle•• were formed H described at,o,,e _ The DNA &aml)les -.-a -Nd In 50'!b ~ - 0 .1 M Tria-1-fCI. 0 .01 M Na EOTA <Oevia et al., 1970. SinQI&- and clouble-straroded .X ONA - • used •• lenQth markers. The grid9 - • shadowed and examined in a Phllipo 300 e1ec1ron Frft9ct1. E. F., La-. R. M. and Manialla, T. (1980). Cell 19, 959972. mictOeCo!)e. Frttach. E. F., ~ - R. M. and Maniatla, T. (1979). Nature 279, 598- 803. Gale. R., Clegg, J. and Hue/Ina, E. (11179). Nature 280, 1fl2-164. Gooaena. M .. Oozy, A.. Embury, S .. Zac:Nridft , Z .. Hadjiminia. M .. Stamatoyannol)OUloa, M. and Kan . w. Y. (1980). Proc. Nat. Aead . &:J. USA 77, 518-521 . Haldlaon . R. C., Butler, E. T., 11. uey, E .. - I l a . T .. ~ a l. N. andEtatraliedi&, A. (1979). Celt 18, 1285-1297. l-4ood, L . c:amc,c,etl. J. H. and Elgin, S. C. R. (1975). Ann. Rev. Genet II. 305-353. Huenna, E. and FarooQui. A. (1976). Nature 254. 335-337. Huehna. E., Flynn, F., Butw, E. and S . -, G. (196 1). Nature 189, 4116-497 . Jefhya, A. J. (1979). Celt 18, 1-10. Kamuzora. H. and Lehmann. H. (1975). Nature 256. 5 11. Kite'-. H. (1974). NV\. NY Aead. Sci. 241 , 12-24 . We thank J . Clegg, s . Embury, E. Frttaeh. Y. w. Kan , R. Lawn, N. Proudfoot, 0 . Smithies and A. WNaon for permiUion lo clle their Konltet. 0 . A., Malzeta, J. V., JI. a n d ~. P. (1979). Catt 18. 865 1173. 40 Cell 130 Lawn. R. M., Frttsch, E. F .. Parker. R. C.. Blake. G. end Manietla. T. (1978). Ceff 15, 1157-1174. Laey, E .. Hardison. R. C.. Ouon. D. end Maniatia. T. (1979). Cell 18, 1273-1283. Legerll<i. R .. Hodnett. J. and Gray. H.B .. Jf. (1978). Nucl. Aclde RH. 5. 1445-1464. Uttte, P. F. R .. Flavell. R. A.. Kooter. J. M., Annioon, G. end 10n, R. (1979a). Natura 278. 227-231 . Uttie, P. F. R .. Williemson. R .. AMilOn , G .. Fla..tl. R. A. . De Boer, E .. Bemoni. L. F .. Ottolenghi, S.. Saglio. G. and Maua U. (1979b). Nature 282, 316-318. Maniatis, T .. Hardison . R. C .. Lacy, E.. Lauer. J.. O' Connell , C.. Quon, 0 ., Sim. G. K. and Efstratiadis. A. (1978). Cell 15. 687-701 . Maxam . A. M. and G~t>ert. W. (1977). Proc. Nat. Ac:ad. Sci. USA 74. 560-564 . Mears. J. G., Ramirez. F .. Leibowitz, D. end Bani<, A. (1978). Cell 15, 15-23. NisniOka. Y. and L-.. P. (1979). Cell 18. 875-882. Nute. P. (1974). Ann. NY Ac:ad. Sci. 241 . 39-70. Orkin , S. H. (1978). Proc . Nat. Acad. Sci. USA 75, 5950-595'1 . Orkin. S. H .. Old. J .. Lazarus. H .. Allay, C.. Gurgey, A.. Weatherall . 0 . J. and Nathan. D. G. (1979). Cell f 7. 3342. PrOUdfool. N. J. and Baralie , F. E. (1979). Proc. Nat. Acad . Sci. USA 76. 5435-5439 . Rodriguez. R. L.. Bolivar. F.. Goodman. H. M ., Boye, , H. W . and Bellach , M. (1976). In Mo'-cular Mechanosms in Iha Control of Gene Expresa,on, 0 . P. Nterlich, W. J. Rutter and C. F. Fox. eds. (New York : Academic Presa). go. 4 71-4 78. Russell, E. S. and McFarland , E. C. (1974). Ann . NY Ac:ad . Sci . 241 . 25-38 . Shen. C.-K. J. and Maniatis. T. (1980). Cell 19. 379-391. Sm,11'1 . G. P. (19731. Cold Spnng Harbor Symp. Quant. Biol. 38. 507514 . Sodergren , E. and Fox. M . (1979). J. Mol. Biol. 130. 357-377 . Sou1hem . E. M. (1975). J. Mol. Biol. 98. 503-517. Stahl . F'. ( 1979). Ann. Rev. G-1. 13, 7-24. Tanof. K. 0 . (1973). Cold Spring Hart>or Symp. Quant. Biol. 38, 491500. • Todd . D .. Lai, M . C. S., Beaven , G. and Huehns, E. (1970). Brit. J. Haematol. f 9, 27-31 . Tuan. D.. Biro. P. A.. De Riel. J. K .. Lazarus. H. and Forget, B. G. (1979). Nucl. Acids Res. 6. 2519-2544 . Weatnera«. D. J. and Clegg , J. B. (1979). Cell re. 467479 . Weatherall. 0 .. Cle9g, J . and Wong , H. (1970). Brit. J. Haemalol . 18. 357-367 . Wilson . J. T .. Wilson, L. B .. de Rief. J. K .. VIHa-Komaroff , L .. Ets1r• 1iedis. A.. Forget. B. G. and Weiaaman, S. M. (1978). Nuc:J . Acid& Res. 5. 563-581 Zimmer, E. A.. Martin. S. L.. SeverMy, S. M .. Kan. Y. W. and WIIIOn. A. C. (1980). Proc. Nat. Ac:ad. Sci . USA. in press. Nola Added In Proof The ex111ence of the embryonic p-globin gene and tts Nnl<aoe 10 1he other a-~ke was recen~y C C < \ - by anafysls o( a beclenophage recombirulnl bav1n11 a 10.9 kb Bgl II fnlQm.tnt wtlich - P • .I.HoG1 · and contains.,,. r;z~ CJ. Lauer. uni><lbllahed r9SUl1S) . .,_,., !llis manuscripl - • IUbmitted, learned ot a ltudy of clon4td mouse ribosomal ONA wnich revealed tha1 deiellons which occur in the ribosomal gene inaens during pr-lion In E. coli ara aimilar 10 thoae wtrich occur in mice (N . Amhelm end M . Kuehn, 1979. J. Mol. Biol. r 34 . 7 43-765). Aa in the cue of the human a-91obin gene deletions reported he<e. the defe1iona in mouae .-naJ DNA can be correlated with the preMnee 01 dif'ect ,_at ~ •- 41 Appendix: Isolation of the z;;2 globin gene The linkage of the z; 2 globin gene to the other a-like genes was confirmed by analysis of a bacteriophage recombinant (AHi; G 1) bearing a 10.9 kb Bgl II fragment which overlaps AHaG 1 and contains the z; 2 gene. The human fetal liver DNA used in isolation of this clone was from the same individual as the DNA that was used in constructing the human DNA library (Lawn et al., 1978) from which AHaGl and AHaG2 were isolated. Human DNA was digested with Bgl II and DNA fragments of 9-12 kb were isolated by preparative agarose gel electrophoresis. The bacteriophage lambda vector, Charon 28 (F. Blattner, personal communication), was digested with BamH I and the annealed left and right arms were isolated by preparative agarose gel electrophoresis. The human DNA Bgl II fragments and the Charon 28 vector arms were ligated and packaged in vitro. screened using a 32 -- Approximately 350,000 recombinant phage were / P-labeled Eco RI-Sac I fragment containing the terminal 0.49 kb of the human DNA insert of AHaGl (see Figure 7 of Lauer et al., 1980 for the location of this fragment). Ten positive phage were isolated. The restriction map of one of these phage (AHi; G 1) is shown in Figure la. AHl;Gl DNA was digested with various restriction enzymes and the products were fractionated on an agarose gel, transferred to nitrocellulose, and hybridized with a 32 P-labeled Pvu II-Hine Il fragment containing part of z; 1 (Figure le). The approxi- mate location of the z; 2 gene as determined in this manner agrees with that previously determined by genomic blotting. The precise position of the middle exon of z; 2 was determined by alignmen t of the restriction maps of z; 2 (Figure lb) and z; 1 (Figure le). 42 References Lawn, R. M., Fritsch, E. F., Parker, R. c., Blake, G. and Maniatis, T. (1978). The isolation and characterization of linked c- and S-globin genes from a cloned library of human DNA. Cell 15, 1157-1174. Lauer, J., Shen, c.-K. J. and Maniatis, T. (1980). The chromosomal arrangement of human a-like globin genes: sequence homology and a-globin gene deletions. Cell 20, 119-130. 43 Figure 1. Map of Restriction Endonuclease Cleavage Sites in >..H l;;Gl DNA. a) The locations of cleavage sites in >..H l;;Gl DNA for the enzymes Bgl II (Bg), BamH I (Ba), Eco RI (RI), Hind III (H) and Sac I (S) are indicated. Sac I cleaves at additional sites in the vector which are not shown on this map. Charon 28 vector sequences are represented by the heavy bars, and the 10.9 kb Bgl II fragment of human DNA which contains l;;2 is represented by the thin line. b) and c) Selected restriction endonuclease cleavage sites within and adjacent to the coding region of 1:2 (b) and l;;l (c) are indicated. sequences (exons). Black boxes represent coding The approximate position of the first exon of l;;2 and 1',;l has been determined by in vitro transcription experiments (N. Proudfoot and M. Shander, personal communciation). The positions of the second and third exons of r,;1 have been determined by DNA sequence analysis (Lauer et al., 1980; C. O'Connell and N. / Proudfoot, personal communication). The position of the second exon of r,;2 is assumed by alignment of the Hine II, Sac I, and Pvu II sites of 1;:l and 1;;2. The position of the thir d exon of 7;; 2 is unknown and it is therefore tentatively indicated by an open rather than solid box. The hatched box indicates the location of the Pvu II-Hine II z;;1 probe fragment. 44 ..... 0 <t I - E C 0 (I) H 0 0 0 E I{) N Cl) H 0 H 0 E I r<) Cl) E 0 (I) / / / / 0 / a:: Cl) / / Cl) iE (I) C I I E Cl) Cl) u ..... 0 0 0 Cl) (/) Cl) Cl) l=I l=I c,, ::::, 0 0 I{) u H '\ c,, E / >-4 (I) 0 H Cl) E (I) ..... 0 0 0 u C ..... / / 0 l=I ( .) / I Cl) l=1 / Cl) E 0 E / Cl) N 0 H / N 0 0 0 ..... ::::, > > a.. a.. (I) '\ '\ '\ 0 0 0 ·"- '\ '\ r') '\ 0 '\. " t -1.() '\. C) v I ~ '\ '\ 0 0 '\ I{) 1=1'\ c,, (I) c,, 0 (I) 0 .D u Figure 1 45 Chapter 3 The molecular genetics of human hemoglobins 46 THE MOLECULAR GENETICS OF HUMAN HEMOGLOBINS Tom Maniatis, Edward F. Fritsch*, Joyce Lauer+, and Richard M. Lawn* Division of Biology California Institute of Technology Pasadena, California 91125 Present Addresses: +Department of Biochemistry, Michigan State University, Ea.st Lansing, Michigan 48824. / +Department of Molecular Biology, University of Edinburgh, Edinburgh, Scotland. *Genentech, Inc., 460 Pt. San Bruno Blvd., South San Francisco, California 94080. Shor tened Tit le: MOLECULAR GENETICS OF HUMAN HEMOGLOBINS Send Proofs to: Tom Maniatis, Division of Biology, California Institute of Technology, Pasadena, California 91125. 47 CONTENTS PAGE INTRODUCTION The Ontogeny of Globin Gene Expression Globin Gene Mapping and Molecular Cloning THE HUMAN '3-LIKE GLOBIN GENES '3-Like Globin Gene Fine Structure '3-Like Globin Gene Linkage Repetitive Sequence Elements within Globin Gene Clusters '3-Globin Variants DNA Sequence Polymorphisms within the '3-Like Globin Gene Cluster GENETIC DISORDERS IN '3-GLOBIN GENE EXPRESSION '3 +-Thalassemia '3 °-Thalassemia '3-Globin Gene Deletions THE HUMAN a-LIKE GLOBIN GENES a- Like Globin Gene Fine Structure a- Like Globin Gene Linkage Globin Gene Duplication GENETIC DISORD ERS IN a-GLOBIN GENE EXPRESSION a-Globin Gene Deletions (l- AND '3-GLOBIN PSEUDOGENES CONCLUDING R EM ARKS 48 INTRODUCTION The human glob in gene family is a paradigm for studying differential gene activity during development and the molecular basis of genetic disorders in gene expression (see Bunn et al 1977; Weatherall and Clegg 1979a; Weatherall et al 1979 for recent reviews). Structural characterization of globin polypeptides and messenger RNAs and extensive clinical investigations of inherited disorders in hemoglobin expression have provided information which is not available for any other eukaryotic gene system. The primary objective of this review is to discuss our current understanding of the structure, chromosomal arrangement, and expression of normal and abnormal globin genes, emphasizing new information derived from gene mapping and molecular cloning studies. The Ontogeny of Globin Gene Expression The a-like and f3-like subunits of hemoglobin (Hb) are encoded by a small group of genes which are expressed sequentially during d velopment. The earliest embryonic hemoglobin tetramer, Gower 1, consists of e: ( f3-like) and z; (a-like) polypeptide chains (Huehns and Farooqui 1975; Gale et al 1979). Beginning at approximately eight weeks of gestation the embryonic chains are gradually replaced by the adult a-globin chain and two different fetal f3-like chains, designated G y and Ay. The y chains differ only in the presence of glycine or alanine at position 136, respectively (Schroeder et al 1968). During the transition period between embryonic and fetal development, Hb Gower 2 fo e: ) and Hb Portland (z; 2 y 2) are detected. 2 2 Hb F (a y ) eventually becomes the predominant Hb tetra.mer throughout the 2 2 remainder of fetal life. Beginning just prior to birth, the y-globin chains are gradually replaced by the adult f3- and c-globin polypeptides. At six m<:>nths after birth 97-98% of the hemoglobin is Hb A (a2 f3 2), while Hb A2 (a2 c 2) accounts for approximately 2%. Small amounts of Hb F (1 %) are also found in adult peripheral blood. The site of erythropoiesis changes from the yolk sac in the early embryo, 49 to the developing liver, spleen and bone marrow in the fetus, and finally to the bone marrow in adults. Because the ratio of Hb A to Hb Fis the same in all fetal erythroid tissues, the switch from fetal to adult globin production is not correlated with the site of erythropoiesis (Wood and Weatherall 1973). In addition to the gene switching described above, globin gene expression is regulated within a particular developmental stage. For example, the patterns of expression of the o and f3 genes during adult red cell maturation are quite different (Roberts et al 1972). o-globin mRNA can be detected in nucleated erythroid precursor cells but not in mature reticulocytes (Wood et al 1978). Thus, the small amount of o-globin polypeptide found in circulating reticulocytes was synthesized in more immature cells in the bone marrow. Furthermore, bone marrow nuclei contain tenfold less o-globin mRNA precursor than f3-globin precursor (Kantor et al 1980), suggesting that the difference in expression of the two genes is at the level of transcription or RNA processing. The bi logical significance of the restriction of o gene expression to immature cells is unknown. The information summarized above indicates that the a-like and f3-like globin gene families have coordinated programs for differential gene expression. The primary difference between the two gene families is that two switches in gene expression (embryonic to fetal to adult) are observed for the f3-like genes, while a single switch results in activation of adult a-globin production early in fetal life. Globin Gene Mapping and Molecular Cloning Initial advances in globin gene mapping and isolation were made possible by the development of procedures for synthesizing and cloning double-stranded DNA copies of poly(A)-containing mRNAs (see Efstratiadis and Villa-Komaroff 1979; Maniatis 1980, for reviews). The successful introduction of double-stranded mouse (Rougeon et al 1975) and rabbit (Maniatis et al 1976; Higuchi et al 1976) 50 8-globin cDNA into bacterial plasmids provided homogeneous hybridization probes for gene mapping experiments. More recently cDNA plasmids carrying human a-, 8- and y-globin mRNA sequences were cloned and characterized (Wilson et al 1978; Little et al 1978). The availability of cloned hybridization probes and the successful application of the Southern blotting procedure (Southern 1975) to mapping single copy sequences in mammalian genomes (genomic blotting; Botchan et al 1976) made it possible to construct a physical map of restriction endonuclease cleavage sites within and flanking the rabbit adult 8-globin gene in chromosomal DNA (Jeffreys and Flavell 1977a). The interrupted colinearity between the restriction map of a rabbit 8-globin cDNA plasmid (Maniatis et al 1976) and the corresponding map of the rabbit 8-globin gene (Jeffreys and Flavell 1977a,b) provided evidence for the presence of noncoding intervening sequences (introns) in cellular genes (Jeffreys and Flavell 1977b). An intron within the mouse 8-globin ene was independently discovered using molecular cloning techniques (Tilghman et al 1978a). The presence of similar sequences in the human globin genes will be discussed below. Primarily two procedures have been used for cloning mammalian globin genes (see Maniatis 1980 for review). The mouse adult 8-globin genes were isolated by gene enrichment followed by molecular cloning in a bacteriophage A vector (Tilghman et al 1977). Globin genes were also isolated without a pre-enrichment step (Maniatis et al 1978; Blattner et al 1978) by constructing and screening collections (libraries) of recombinant phage containing the products of a limit restriction endonuclease digestion of mammalian DNA (Blattner et al 1978) or containing random, high molecular weight fragments of mammalian DNA (Maniatis et al 1978). The advantage of libraries of high molecular weight DNA is that the isolation of a set of overlapping clones, all of which contain a given gene, permits study of the sequences extending many kilobases (kb) from the gene. 51 Moreover, even more distant regions along the chromosome can be obtained by rescreening a library using terminal fragments of the initially selected clones as hybridization probes. Thus, this approach to gene isolation makes it possible to study the organization of closely linked genes. The entire human S-like (Lawn et al 1978; Fritsch et al 1979, 1980; Ramirez et al 1979) and a-like (Lauer et al 1980; J. Lauer, unpublished results) globin gene clusters including the intergene sequences have been isolated by this approach. In the following sections, the human a-like and S-like globin gene clusters will be described separately in terms of the structure, chromosomal arrangement and expression of normal and abnormal genes within each cluster. Some of the information presented has been previously reviewed (Bunn et al 1977.; Weatherall and Clegg 1979a) and is included for background. This review will focus on the application of molecular cloning and gene mapping procedures to the study of human globin genes. THE HUMAN a-LIKE GLOBIN GENES S- Like Globin Gene Fine Structure The structure and organization of the human S-like globin genes have been studied by genomic blotting using globin cDNA or cDNA plasmids as hybridization probes, and by cloning segments of the genome containing globin genes. Restriction sites within and surrounding t he o and S genes were mapped using the genomic blotting proeedu.re (Flavell et al 1978; Mears et al 1978). The existence of a large intron within each gene was demonstrated by comparing the restriction endonuclease cleavage map of the S-globin gene in genomic DNA with the map of a S-globin cDNA plasmid (Flavell et al 1978; Mears et al 1978). These conclusions were independently reached by analysis of bacteriophage recombinants which carry both the o and S genes (Lawn et al 1978). The identities of the two genes 52 and the precise location of the large intron within each gene were established by DNA sequence analysis. In addition, a second smaller intron similar to that found in the mouse 8 gene (Konkel et al 1978) was identified by comparing the fine structure maps of a human 8-globin cDNA plasmid (Wilson et al 1978) and of the chromosomal gene (Lawn et al 1978). All five of the known human 8-like globin genes have been obtained from recombinant clones of genomic DNA: 8 and o (Lawn et al 1978), G y and Ay (Smithies et al 1978; Fritsch et al 1979, 1980; Ramirez et al 1979) and e: (Proudfoot and Baralle 1979; Fritsch et al 1980). Identification of the e: gene was made possible by the recent determination of the partial amino acid sequence of e-globin polypeptide (Gale et al 1979). All five genes are interrupted by two introns at identical locations: the first, of 125-150 base pairs (bp) in length is located between codons 30 and 31, and the second, of 700-900 bp is between codons 104 and 105 (Figure 1). Since the human a-globin genes (Figure 1; see below) as we as the mouse aand 8-globin genes, the chicken 8-like globin genes and the four known rabbit 8-like globin genes also contain two introns at homologous locations, it is reasonable to assume that these interruptions antedate the emergence of separate a- and 8-globin genes about 500 million years ago (Tilghman et al 1977; Jeffreys and Flavell 1977b; Van den Berg et al 1978; Dodgson et al 1979; Hardison et al 1979; Konkel et al 1979; Nishioka and Leder 1979; Lauer et al 1980; Baralle et al 1980). As was shown for the mouse (Tilghman et al 1978b; Kinniburgh et al 1978; Kinniburgh and Ross 1979) and rabbit (Flavell et al 1979a; Hardison et al 1979) 8-globin genes, the introns of the human 8-like globin genes are transcribed as part of a nuclear mRNA precursor and are removed in steps by splicing to produce the mature globin mRNA (Maquat et al 1980; Kantor et al 1980). RNA-DNA hybridization mapping experiments have shown that the 5' ends of the rabbit (Flavell et al 1979a) and mouse (Weaver and Weissman 1979) 8-globin nuclear mRNA precursors are coterminal 53 with their respective cytoplasmic mRNAs. Although the function of introns is presently unknown, it has been proposed that they play a role in evolution by joining different combinations of DNA sequences encoding protein structural domains (Gilbert 1978, 1979; see Wahli et al 1979; Crick 1979 for discussion). Evidence for this hypothesis is provided by the observation that introns separate the three constant region domains within heavy chain immunoglobulin genes (Sakano et al 1979; Early et al 1979). In addition, the distribution of functional amino acid residues in the a- and S-globin polypeptides can be correlated with the arrangement of coding (exon) and noncoding (intron) sequences in these genes (Gilbert 1978, 1979; Blake 1979; Eaton 1980). For example, the middle exon encodes the region of the protein containing most of the heme contacts, while the a - S protein-protein contacts are located predominantly in the region of the globin 1 1 chain encoded by the third exon. Further evidence that the middle exon encodes a functional domain is provided by the demonstration thay the polypeptide fragment encoded by the middle exon can bind heme tightly and specifically (Craik et al 1980). The nucleotide sequences of human S- and y-globin mRNAs have been determined (see Forget et al 1979 for discussion and references). The complete nucleotide sequences of the five human S-like globin genes and flanking regions have been determined (Lawn et al 1980; Spritz et al 1980; Slightom et al 1980; Baralle et al 1980) and a detailed comparison of these and other mammalian globin gene sequences has been presented (Efstratiadis et al 1980). These comparisons made it possible t o establish accurate phylogenetic relationships among the human S-like globin genes and revealed interesting sequence homologies in regions which are potentially involved in globin gene transcription and splicing (Efstratiadis et al 1980). Alignment of the sequences 5' to the human S-like globin genes provides an example of these homologies. The most notable feature of this alignment is the presence of two blocks of sequence homology which are present in analogous positions adjacent 54 to most eukaryotic genes. The first homology block (designated the ATAA box) is found 29-30 bp 5' to each gene and the second block (designated the CCAAT box) is located 70-78 bp 5' to each gene. A possible role of these sequences in transcriptional initia tion and/or RNA processing has been discussed (Efstratiadis et al 1980). In addition to the ATAA and CCAAT box sequences which are found in many eukaryotic genes, other regions of homology which are shared among the 13-like globin genes are scattered throughout the 5' flanking sequence. The conservation of these sequences during the 200 million years since the time of e:, y-o, 13 divergence (Efstratiadis et al 1980) suggests that they are functionally significant. The identification by sequence comparisons of regions of possible functional importance may be useful in the search for structural differences between normal and mutant glob in genes. Comparison of the complete nucleotide sequences of the G y and A y genes has revealed t hat the coding, intervening and flanking seq ences of these two genes are virtually identical (Slightom et al 1980). The distribution of the infrequent sequence differences has led to formulation of a model for gene correction (see below). 13-Like Globin Gene Linkage Cell fusion studies have shown that the human adult 13-globin gene is located on chromosome 11 (DeisseI"oth et al 1978). More recent experiments have localized the 13-like globin gene cluster to the distal portion of the short arm of chromosome 11 (Jeffreys et al 1979; Gusella et al 1979; Lebo et al 1979). Genetic studies (Weatherall and Clegg 1979b) and structural analysis of hemoglobin fusion proteins (Baglioni 1962; Huisman et al 1972) predicted linkage of the y, o and 13 genes. These predictions were recently confirmed by genomic blotting and molecular cloning experiments (Figure 2). Physical linkage between the o and 13 genes was demonstrated by genomic blotting experiments (Flavell et al 55 1978) and by analysis of a recombinant bacteriophage containing both the o and f3 genes (Lawn et al 1978). Linkage of the G y and Ay genes was determined by genomic blotting experiments (Little et al 1979a) and was confirmed by analysis of clones containing both of these genes (Fritsch et al 1979, 1980; Ramirez et al 1979). Linkage between the Ay and o genes was established using a cloned hybridization probe from the Ay-o intergenic region and a well-defined deletion located in this region (Fritsch et al 1980; Tuan et al 1979), by mapping DNA fragments from the Ay-o intergenic region generated by non-limit restriction enzyme digests of genomic DNA (Bernards et al 1979a), and ultimately by cloning overlapping DNA fragments which span the Ay-o intergenic region (Fritsch et al 1980). Analysis of overlapping cloned DNA fragments also made it possible to demonstrate physical linkage between the e: gene, the isolation of which was initially reported by Proudfoot and Baralle (1979), and the remainder of the gene cluster (Fritsch et al 1980). The linkage arrangement of the human 8-like ~obin gene cluster is shown in Figure 2. All of the genes are transcribed from the same DNA strand and are arranged on the chromosome in the order of their expression during development: 5'-e:- 0 y-Ay-o-f3-3'. A sim~lar pattern of gene organization has been found in the human a-like (Lauer et al 1980; see below) and rabbit (Lacy et al 1979; Hardison et al 1979) and mouse (Jahn et al 1980) 8-like globin gene clusters. Linkage between four chicken f3 - like globin genes has also been demonstrated (Dodgson et al 1979), but the identities of the g enes. have not been unambiguously determined (J. D. Engel, personal communication). Two 8- like sequences, \jJ 81 and ¢82, which cannot be identified with known polypeptides, are also shown on the map of Figure 2. These sequences probably correspond to pseudogenes, which have been studied_in detail in other glob in gene clusters (see below). Repetitive Sequence Elements within Globin Gene Clusters The universal occurrence of repetitive sequences in eukaryotic DNA and 56 their interspersion with single-copy sequences have led to the proposal that repetitive sequences play a role in the regulation of gene expression. Previous studies of repetitive sequence elements have described their genomic distribution and their differential transcription during development (see Davidson and Britten 1979 for review). The isolation of globin gene clusters by molecular cloning provides an opportunity to study repeat sequences in relation to well-defined sets of developmentally regulated genes. A detailed study of the rabbit 8-like globin gene cluster revealed a complex array of sequences which are repeated within the gene cluster as well as throughout the genome (Shen and Maniatis 1980). A sequence which is repeated within the human 8-like globin gene cluster (Fritsch et al 1980) is also interspersed with the human a-like globin genes (E. Fritsch, J. Shen, R. Lawn and T. Maniatis, unpublished results), is highly repeated elsewhere in the genome (Fritsch et al 1980), and is homologous to a sequence which is found in the rabbit 8-like globin ~ e cluster (Shen and Maniatis 1980; J. Shen, unpublished results). Nucleotide sequence analysis of the copy of this repeat located 5' to the G y gene indicates that the repeat is a member of a family of sequences which is reiterated approximately 300,000 times in the human genome (Houck et al 1979; Jelinek et al 1980). This family of repeated sequences is characterized by the following diverse set of properties: 1) the repeats are transcribed by RNA polymerase III in vitro (Duncan et al 1980; Fritsch et al in preparation); 2) the repeats share sequence homology with an abundant small nuclear RNA molecule (Jelinek and Leinwand 1978; Jelinek et al 1980) and with double-stranded heterogeneous nuclear RNA (Jelinek et al 1980; Fritsch et al 1980); 3) the repeats contain a sequence which is homologous to a sequence found near the replication origins of SV40, polyoma and BK DNA tumor viruses (Jelinek et al 1980). Whether such repeat sequences are involved in differential gene expression or play a more general role in DNA replication or chromosome structure remains to be determined. 57 8-Globin Variants The most common structural variants of the 8-globin polypeptide are simple amino acid substitutions, presumably resulting from a single base change in the nucleotide sequence (see Bunn et al 1977 for review). The majority of such substitutions have no demonstrable effect on globin chain activity or stability. However, some amino acid substitutions result in severe clinical effects such as those associated with sickle cell anemia (Weatherall and Clegg 1979b). The widespread occurrence of a variant, Hb FSa. rd inia (Ty; Ricco et al 1976) and differences in the Ay/G y ratio during fetal development (Huisman et al 1970; Schroeder et al 1973) led to the proposal that there are more than two y-globin genes per haploid genome. However, only two y genes are detected by genomic blotting in both normal individuals and individuals with Hb FSa. rd inia (Little et al 1979a,b). Moreover, structural analysis of the Ty chain of Hb FSa. rd inia indicates that Ty is a variant of the A'Y chain (Saglio et al 1979). / Some 8-globin structural variants lack one to eight amino acids (Bunn et al 1977). Since the amino acid sequence is normal on either side of the deleted residues the genes which encode these variants must contain small deletions which do not alter the translational reading frame. Examination of the nucleotide sequences in normal DNA within and adjacent to the codons for the amino acids which are deleted in the variants reveals the presence of small (2-8 bp) direct repeat sequences (Marotta et al 1977). Small r epeats are also associated with deletions in the lac i gene of E. coli (Farabaugh and Miller 1978). These deletions are thought to result from base mispa.iriug and slippage during DNA replication (Streisinger et al 1966; Farabaugh and Miller 1978). As noted by Marotta et al (1977) and Efstratiadis et al (1980), a similar mechanism may have generated the deletions which are associated with the 8-globin variants described above. 58 Another type of hemoglobin variant is the 6-globin fusion proteins. Hb Lepore is a fused N-terminal o- and C-terminal 6- globin protein (Baglioni 1962) which is the result of a fusion between the o and 6 genes (Flavell et al 1978; Mears et al 1978) presumably caused by unequal crossing over during meiosis (Baglioni 1962). Another well-characterized fusion protein is Hb Kenya which must result from fusion of the 5' portion of the Ay- and the 3' portion of the 6-globin genes (Huisman et al 1972). As mentioned above, the existence of Hb Lepore and Hb Kenya provided the first strong evidence for linkage of the y, cS and 6 genes (Baglioni 1962; Huisman et al 1972). DNA Sequence Polymorphisms within the 6-Like Globin Gene Cluster Genetic polymorphisms of restriction endonuclease cleavage sites have been detected in t he human 6-like globin gene cluster. The first example of such a linked genetic polymorphism was an alteration in an Hpa I recognition site located 5 kb 3' to the 6 gene (Kan and Dozy 1978). The frequen association of the absence of this Hpa I site with the allele for sickle cell anemia was exploited by Kan and Dozy (1978) as a tool for prenatal diagnosis. Rarely, absence of this Hpa I site is associated with the normal allele, leading to uncertainties in diagnosis (Kan et al 1980). 6°-thalassemia in Sardinia may be diagnosed by the presence of a polymorphic BamH I site located 9.3 kb 3' to the 6 gene (Kan et al 1980). Although the frequent occurrence of this BamH I site in normal in dividuals leads to considerable uncertainty in diagnosis, genomic blotting information is a valuable complement to existing procedures for prenatal diagnosis. Sequence polymorphisms have also been detected within the large introns of the cS (Lawn et al 1978; Jeffreys 1979) and y (Jeffreys 1979; Tuan et al 1979) genes. These studies indicate that restriction site polymorphisms are frequent within the 6-like globin gene cluster. Analysis of linked polymorphisms associated with other genetic disorders may provide an important tool for studying the genetics of human disease (Jeffreys 1979; Kan et al 1980; Little et al 1980). 59 GENETIC DISORDERS IN 6-GLOBIN GENE EXPRESSION + f3 °- and f3 -thalassemia are the most frequently occurring mutations in S-globin gene expression. f3 +-thalassemia is char~cterized by reduced levels of S-globin production while in f3°-thalassemia there is no detectable a-chain synthesis (Weatherall and Clegg 1979b). S-globin gene deletions, which affect the expression of other globin genes in addition to f3, constitute a third class. In the discussion which follows we will briefly summarize the information currently available regarding the molecular basis of various types of S-thalassemia. f3 +_Thalassemia In f3 +-thalassemia a small amount of normal S-globin protein is made sug- gesting that the gene and mRNA are intact. Globin cDNA/DNA solution hybridization (Old et al 1978; Benz et al 1978), in vitro translation (Benz et al 1978), and genomic blotting (Flavell et al 1979b; Orkin et al 1979c) experiments have confirmed these conclusions. Thus, the primary defect in f3 +-thalassemiaJAppears to be a reduced level of transcription or inefficient processing of mRNA precursors. Comparison of the ratios of globin RNA sequences in bone marrow nuclear (a/S = 1) and cytoplasmic (a/S = 15-20) RNA .obtained from individuals homozygous for f3 +-thalassemia suggested that the molecular defect in these individuals is abnormal processing or decreased stability of S-globin mRNA precursor (Nienhuis et al 1977). Evidence for the former possibility was recently provided by analysis of pulse-labeled a- and '3,-globin mRNA precursors from nucleated bone marrow cells (Maquat et al 1980}. Following pulse-labeling with tritiated nucleosides a series of S-globin mRNA precursors is observed. In normal individuals these precursors are quantitatively processed to mature mRNA during a chase period with unlabeled nucleosides. By contrast, in f3 +-thalassemic individuals discrete intermediates accumulate and only a fraction of the labeled RNA is chased into S-globin mRNA. Processing of a-globin mRNA precursors is identical in normal and f3 +-thalassemic 60 individuals. Similar conclusions were obtained by Kantor et al (1980) who used a hybridization probe specific for the large intron of the e gene to quantitate the level of e-globin RNA and to distinguish between o- and e-globin mRNA precursors. Furthermore, using an RNA blotting procedure (Alwine et al 1977), an abnormal 650 nucleotide (nt) intermediate was observed in one patient while a normal 1300 nt intermediate was found to accumulate in another (Kantor et al 1980). These studies strongly suggest that some e +-thalassemias result from mutations which alter RNA processing. f3 °-Thalassemia f3°-thalassemia is characterized by complete absence of e-globin polypeptide synthesis (Weatherall and Clegg 1979b). The molecular defects in f3°-thalassemia appear to be quite heterogeneous. In the majority of cases no deletions or alterations of the e gene can be detected by genomic blotting (Flavell et al 1979b; Orkin et al 1979c). Rarely, this phenotype is associated with a 60J}'6p deletion which removes part of the large intron, the third exon, and 150 bp beyond the 3' end of the normal gene (Figure 2; Orkin et al 1979c, 1980). Three classes of e 0 -thalassemia have been defined on the basis of cDNA/mRNA hybridization experiments (Old et al 1978; Benz et al 1978). In the first class, no e-globin mRNA can be detected although in at least one case, e-globin sequences were detected in the nuclear RN A (Comi et al 1977). This class of e 0 -thalassemia probably results from a block in RNA transcription or processing. In the second class of 6°-thalassemia up to 30% of the normal level of e-globin mRNA is present although no e-globin protein is synthesized in vivo or in vitro (Benz et al 1978; see Conconi et al 1972; Conconi and Del Senno 1974 for a possible exception). In one case of this type Temple et al (1977) showed by RNA fingerprinting that an apparently normal e-globin mRNA was present. Sequence analysis of this RNA revealed, however, the presence of a single base change at 61 codon position 17 resulting in a nonsense mutation (Chang and Kan 1979). The fact that this nonsense mutation can be suppressed in vitro demonstrates that it is the primary defect (Chang et al 1979). In another case of B0 -thalassemia which appears to contain intact B-globin mRNA the initiation codon may be defective (Old et al 1978). In a third class of 8°-thalassemia, a partial RNA transcript is produced (Old et al 1978). B-Globin Gene Deletions Most cases of B0 - and B+_thalassemia result in moderate to severe anemia and are not associated with detectable deletions in the B-like globin gene cluster. In these cases, the level of fetal globin expression in adults is normal or only slightly increased, suggesting that the switch from fetal to adult globin synthesis is operative. Surprisingly, in other classes of genetic disorder known as o B-thalassemia and Hereditary Persistence of Fetal Hemoglobin (HPFH), the absence of B-globin chains is compensated for by continued expression of y-globin <;!)ains in the adult. The degree of fetal globin compensation differs among these disorders (see Wood et al 1979 for review). As classically defined, HPFH and o B-thalassemia are distinguished on the basis of three criteria: 1) the mean level of Hb F is higher in HPFH than in o B-thalassemia; 2) Hb Fas measured by the acid-elution technique is uniformly distributed throughout the red cell population in HPFH and non-uniformly distributed in oS-thalassem ia (more recently, however, the grea ter sensitivity of immunofluorescent techniques indie.ates t hat the cellular distribution is uniform in both HPFH and o8 - thalassernia; see Wood et al 1979 for critical discussion); 3) HPFH is not associated with clinical symptoms whereas o B-thalassemia is characterized by red cell abnormalities and mild to severe anemia. Studies of individuals heterozygous for either of these disorders indicated that these mutations which abolish B-globin production act in cis to affect y gene expression (Weatherall and Clegg 1979b; Huisman 62 et al 1974). On the basis of this information, Huisman et al (1974) predicted that both HPFH and o 8-thalassemia are associated with 8 gene deletions and that the HPFH deletion, but not the o 8-thalassemia deletion, removes regulatory sequences involved in the normal suppression of y gene expression in adults. This model has received considerable attention because it offered a genetic approach to understanding the mechanism of differential globin gene expression. 8-globin cDNA/DNA solution hybridization experiments confirmed that both HPFH and o 8-thalassemia are associated with deletions which remove the 8 and possibly the o genes (Kan et al 1975b; Forget et al 1976; Ottolenghi et al # 1976; Ramirez et al 1976). With the introduction of the genomic blotting procedure and the availability of specific hybridization probes, it became possible to map the endpoints of these and other deletions within the 8 -like globin gene cluster. The locations of the deletions which have been mapped to date, and the level of Hb F associated with each syndrome, are presented in Figure 1/A unified molecular interpretation of this information is not possible at present, due primarily to the heterogeneity of these genetic disorders. As discussed by Wood et al (1979), the wide spectrum of phenotypes observed in HPFH and o 8 -thalassemia leads to overlap between these syndromes. In addit ion, types of HPFH have been described which show genuinely non-uniform distribution of Hb F in the red cell population (British or Swiss-type HPFH; Weatherall et al 1979) or which express different relative amounts of G y and Ay chains (Wood et al 1979). Nonetheless, genomic blotting mapping studies have been very important in demonstrating the variety of molecular rearrangements which can alter y-globin gene expression, and have led to the formulation of new models for t he mechanism of hemoglobin switching. For example, the locations of some of these deletions can be interpreted in the context of the Huisman et al (1974) deletion model. Comparison of G y-Ay-o 8thalassemia (see legend to Figure 2 for an explanation of the nomenclature) with 63 four cases of HPFH is consistent with the possibility that a sequence 5' to the o gene exerts some influence on the level of Hb F production (Fritsch et al 1979; Bernards et al 1979b; Tuan et al 1979; Ottolenghi et al 1979). Similarly, comparison of Hb Lepore with G y-Ay-o S -thalassemia implicates a second region located 3' to the S gene (Bernards et al 1979b). Other deletions are apparently inconsistent with these interpretations. For example, the G y-o S-thalassemia deletion removes the regions 5' to the o gene and 3' to the S gene but does not result in an HPFH phenotype (Fritsch et al 1979; Orkin et al 1979a). However, because a large portion of the y-globin gene region is removed in G y-o S-thalassemia, the significance of this syndrome with respect to the deletion model cannot be assessed (Fritsch et al 1979). An alternative and equally plausible explanation of the deletion data is that the human S-like globin gene cluster consists of one or more functional chromosomal domains and that deletions within the cluster alter the chromosome structure and thereby affect the normal pattern of differential glo3,ill gene expression (see Fritsch et al 1979; Bernards et al 1979b; and Van der Ploeg et al 1980 for discussion). The concept that higher order chromosome structure is involved in regulating gene activity is central to certain hypotheses of eukaryotic differentiation (for example, see Cook 1973; Weintraub et al 1978). A correlation between alterations in chromosome structure and changes in gene activity is suggested by the observation that actively-transcribed genes are more sensitive to DNase I digestion than non-transcribed genes (Weintraub and Groudine 1976). Genomic blotting analysis of DNase I-digested chromatin has shown that moderate DNase I sensitivity extends for many kilobases on either side of an active gene (Wu et al 1979a,b; Stalder et al 1980), suggesting tha t differential gene activity involves a structural alteration of a large region of the chromosome. A third explanation of the data is that a second mutation, unrelated to the S gene deletion, might be responsible for the altered y gene regulation which 64 is observed in HPFH and/or c 8-thalassemia. The occurrence of different deletion events in independent cases of HPFH and the absence of the HPFH phenotype in non-deletion 8-thalassemias ( 8°- or 8 +-thalassemia) argues against the general occurrence of a second-site compensating mutation. A rare type of HPFH (Ay-HPFH, Greek type) appears not to be associated with a detectable deletion. In Ay-HPFH, 10-20% Hb Fis produced (9:1 ratio of Ay to G y; Clegg et al 1980), and the o and 8 genes are expressed at low levels in some individuals. No alterations of the yo-8-globin gene region have been detected (Tuan et al 1980). In the syndromes described above, deletions in the o-8-globin gene region are associated with changes in the pattern of y gene expression. In y-8-thalassemia, suppression of y and 8 gene expression occurs, resulting in severe anemia in newborns and thalassemia in adults. No homozygous case has been reported. A heterozygous case of this syndrome is characterized by a deletion of at least 40 kb removing both y genes and the o gene (Van der Ploeg et al 1980). However, the 8 gene and approxi- / mately 2.5 kb of 5' flanking sequence are intact. The lack of expression of the 8 gene on the mutant chromosome is not understood. Because this is the only example of y-S- thalassemia which has been characterized at the DNA level, the possibility that a second mutation has inactivated the S gene (similar to S0 - or S +-thalassemia) must be considered seriously. Alternatively, Van der Ploeg et al (1980) have suggested that within the 8-like globin gene cluster the DNA is organized into G y-Ay and c-S domains and that the y-S-thalassemia deletion removes sequences necessary for activation of the o-8 domain. In conclusion, analysis of deletion types of thalassemia has suggested that deletions may act in cis over considerable distances to influence differential gene expression within the S-like globin gene cluster. A similar association of deletions with alterations in globin gene switching has been described in the human a-like globin gene cluster (see below). 65 THE HUMAN ~LIKE GI.OBIN GENFS a-Like Globin Gene Fine Structure The existence of two a-globin genes per haploid genome was established on the basis of genetic studies (see Weatherall and Clegg 1979a for discussion). Although only one a-globin polypeptide sequence has been identified (Dayhoff 1972), both genes are expressed. A partial amino acid sequence of ~-globin polypeptide has been determined (J. Clegg, personal communication). The nucleotide sequence of human a-globin mRNA has been determined by analysis of a-globin cDNA and of a cDNA plasmid (Wilson et al 1978; see Forget et al 1979 for discussion and references). The structure and organization of the human a-like globin genes have been examined by genomic blotting experiments (Orkin 1978; Embury et al 1979) and by analysis of cloned DNA segments (Lauer et al 1980). For convenience, the two a genes are referred to as al and a2 (Figure 3). Restriction endonuclease analysis demonstrated that th~ l and a2 genes are each interrupted by two introns of approximately 95 and 125 bp (Lauer et al 1980; see above for a discussion of intron function). Determination of the complete nucleotide sequence of an independently isolated a2 gene indicated that the introns are located between the codons for amino acids 31 and 32 and 99 and 100 (S. Liebhaber, personal communication). The locations of introns in the human and mouse (Nishioka and Leder 1979) a-globin genes are identical, and are e.nalogous to the positions of introns in $-like globin genes. Fine structure mapping of restriction sites in the coding, intervening and flanking sequences o,f the al and a2 genes indicated that the two genes are vir tually identical (Lauer et al 1980). Restriction sites could be aligned by assuming the insertion of a block of approximately seven nucleotides into the large intron of the al gene (or a deletion from the a2 gene). The al-a2 homology is discussed further below. 66 A cloned z;;-globin gene (z;; 1) was identified by the exact correspondence between the DNA sequence of this gene (Lauer et al 1980; C. O'Connell, N. Proudfoot and T. Mania tis, unpublished results) and the partial amino acid sequence of z;; -glob in polypeptide (J. Clegg, personal communication). The z;; 1 gene contains two introns at locations identical to the positions of introns in the adult a-globin genes. a-Like Globin Gene Linkage Cell fusion studies have shown that the human adult a-globin genes are located on chromosome 16 (Deisseroth et al 1977). Genomic blotting experiments demonstrated that the linked adult a-globin genes are transcribed from the same DNA strand and are 3. 7 kb apart (Figure 3; Orkin 1978; Embury et al 1979). a-globin gene linkage was confirmed by isolation of recombinant bacteriophage clones (Lauer et al 1980) containing the two adult a-globin genes, the embryonic z;; 1 gene and an a-globin pseudogene (ival; see below). Nucleotide sequence analysis of a portion of the \jlal and z;; 1 genes indicated that they are orienteg.,.ffi the same direction as the adult a-globin genes (Lauer et al 1980). A second z;;-globin gene, z;; 2, was identified by genomic blotting experiments using a labeled restriction fragment containing part of the z;; 1 gene as a hybridization probe (Lauer et al 1980). Physical linkage of the z;; 2 gene to the other a-like globin genes was demonstrated by genomic blotting using a cloned restriction fragment located 5' to the z;; 1 gene as a hybridization probe. The z;; 2 gene is located approximately 12 kb 5' to the z;; 1 gene (Figure 3). The z;; 1-z;; 2 gene linkage was confirmed by analysis of overlapping clones which contain the z;; 1 and z;; 2 genes (J. Lauer, unpublished results). Evidence suggests that both z;;-globin genes are functional. Detection of z;;-globin protein in an infant with a homozygous deletion which removes the z;; 1 gene but spares the z;; 2 gene (see below) indicates that z;; 2 is a functional gene. Furthermore, since the nucleotide sequence of the z;; 1 gene (C. O'Connell, N. Proudfoot 67 and T. Maniatis, unpublished results) agrees with the amino acid sequence of z;-globin polypeptide (J. Clegg, personal communication), the z; 1 gene is probably also functional. Genomic blotting-melting experiments suggest that the two genes are very homologous (Lauer et al 1980). Since only one z;-globin polypeptide sequence has been identified, the z; 1 and z; 2 genes may encode identical polypeptide chains. Alternatively, the two genes may be expressed at different times or at different levels during embryonic development. Globin Gene Duplication A common feature of globin gene clusters is the occurrence of two immediatelyadjacent genes which are coordinately expressed during a given developmental stage. Examples of this are the human o-S, G y-Ay, cd-Cl2, and z; 1-z; 2 globin gene pairs. The o and S genes are highly homologous in the coding region, but the noncoding sequences within and surrounding the two genes have diverged considerably (Lawn et al 1978; Efstratiadis et al 1980). Extensive divergen of noncoding regions has also been observed in some other closely-linked, coordinately-expressed globin gene pairs (Konkel et al 1979; Hardison et al 1979; Jahn 1980). In contrast, the two members of the Gy-Ay gene pair are-virtually identical to one another throughout their coding, intervening and flanking sequences (Slightom et al 1980). Although the nucleotide sequences of linked human C1-globin genes have not yet been determined, restriction mapping and heteroduplex analysis of the ell and Cl2 genes indicate that the sequences within and flanking these two genes are virtually identical. Each Cl-globin gene is located within an approximately 4 kb r egion of homology interrupted by two small regions of non-homology (Figure 4; Lauer et al 1980). The virtual identity of sequences within the G y-Ay and C1l-C12 gene pairs appears to be the product of a mechanism for gene matching during evolution. Based on the nearly identical distribution of restriction sites surrounding the Cl-globin genes in a number of primate species, it has been suggested that the C1-globin gene duplication 68 occurred prior to the time of primate divergence (Zimmer et al 1980). Differences between the a-globin amino acid sequences of various primate species are consistent with sequence drift following primate divergence (Dayhoff 1972). However, intraspecies comparisons show much less divergence, indicating that the a-globin genes within a species have been corrected against one another. Maintenance of homology among a family of evolving genes within a species has been termed "horizontal" (Brown et al 1972) or "coincidental" (Hood et al 1975) or "concerted" (Zimmer et al 1980) evolution. Gene conversion and expansion/contraction of gene number by homologous but unequal crossing-over have been proposed as mechanisms for concerted evolution (see Hood et al 1975 for review). Analysis of the complete nucleotide sequences of cloned G y and Ay genes has led to formulation of a specific intrachromosomal gene conversion model to explain sequence matching between linked genes (Slightom et al 1980). The G y and Ay genes on one chromosome are identical in the r~on 5' to the center of the large intron, yet show greater divergence 3' to that position. Examination of the boundary between the conserved and divergent regions revealed a block of "simple sequence" DNA ([TG] ) (Slightom et al 1980). Slightom et al (1980) have proposed n . that this simple sequence is a "hot spot 11 for initiation of recombination events which lead to unidirectional gene conversion. As described above, an active mechanism for gene matching must also be invoked to explain the observed homology between the human a-globin genes. Although <l-globin gene matching could occur by gene conversion, sequence data which would address this possibility are not available. Evidence that a-globin gene sequence matching could occur by expansion and contraction of gene number by unequal crossing-over is provided by the frequent occurrence of one-gene (Orkin et al 1979b; Embury et al 1979) and three-gene (Goosens et al 1980; Higgs et al 1980) chromosomes in some human populations (see Zimmer et al 1980 and Lauer et al 1980 for discussion). 69 GENETIC DISORDERS IN a-GLOBIN GENE EXPRESSION The a-thalassemias are characterized by a reduced rate of a-globin synthesis (Weatherall and Clegg 1979b) resulting in the presence of excess f3 -like chains which associate to form the abnormal tetramers Hb Bart's ( y4) and Hb H ( f3 4). Four athalassemia syndromes of increasing clinical severity occur in the Asian population: 1) the silent carrier state (a-thalassemia 2) which is asymptomatic; 1-2% Hb Bart's at birth; 2) a-thalassemia trait (a-thalassemia 1) which is associated with red cell abnormalities but little or no anemia; 5-6% Hb Bart's at birth; 3) Hb H disease which is associated with anemia; Hb Bart's at birth and Hb Hin adulthood; 4) hydrops fetalis which is fatal at or before birth; 80-90% Hb Bart's, 10-20% Hb Portland (1; y ). 2 2 As will be discussed below, the occurrence and frequency of these syndromes differ in non-Asian populations. Homozygosity for the "a-thalassemia trait", leading to complete absence of a.-globin synthesis, was postulated to be the cause of h~ rops fetalis (Lie-lnjo et al 1962). Subsequently, genetic analysis of Hb H disease led Wasi et al (1964) to propose the existence of two a-thalassemia alleles: a "severe" allele corresponding to the previously-identified a-thalassemia trait, and a "mild" allele recognizable only when present along with the severe allele thereby causing Hb H disease. When it became apparent that human a.-globin genes are duplicated, it was proposed that the mild and severe alleles correspond not to differing degrees of impairment of a single a.-globin gene, but rather to a lack of function of one or both genes, respectively, on a chromosome, whether due to deletion or to other gene defects (Lehmann 1970). The four syndromes described above might thus result from involvement of one, t wo, three, or four a.-globin structural genes, respectively (Lehmann 1970). cDNA/DNA solution hybridization analysis indicated that persons with Hb H disease have only one-quarter of the normal number of a.-globin genes, supporting this 70 hypothesis (Kan et al 1975a). In addition, solution hybridization analysis of hydrops f etalis (which is characterized by complete absence of a-globin synthesis) indicated that the a-globin genes are deleted in this syndrome (Ottolenghi et al 1974; Taylor et al 1974). Although the precision of solution hybridization experiments is limited, these studies provided the first direct physical evidence for the association of gene deletion with a-thalassemia. The genotypes resulting in each a-thalassemia syndrome are diagrammed in Figure 5, where the presence of a functional a-globin gene is indicated by a black rectangle. Although the majority of a-thalassemias appear to be due to gene deletion, some non-deletion types of defects were recently reported (Kan et al 1977, 1979; Orkin et al 197 9b). An a-thalassemia phenotype also results from the unstable a-globin variant, Hb Constant Spring, which is synthesized at a normal level but is rapidly degraded (Weatherall and Clegg 1975). In the discussion which follows we will present the information currently available regarding the molec~ r basis of various forms of a-thalassemia. a-Globin Gene Deletions The association of -gene deletions with a-thalassemia was confirmed and extended by genomic blotting. In normal DNA the duplicated a-globin genes on each chromosome are detected in a 23 kb Eco RI fragment (Orkin 1978; Embury et al 1979). Chromosomes containing a single functional a gene can arise by inactivation of one gene (23 kb Eco RI fragmen t; Orkin et al 1979b; Kan et al 1979) or by deletion of one gene (19 kb Eco R1 fragment; Embury et al 1979; Orkin et al 1979b). Chromosomes without any functional a-globin genes are associated with a 2.6 kb Eco RI fragment containing a partial a gene (Orkin and Michelson 1980) or with deletion of both genes. Different combinations of these chromosomes are associated with each of the a-thalassemia syndromes, as summarized in Table 1. a-thalassemia 1 is common in Asians, blacks and Mediterraneans, yet hydrops 71 fetalis is extremely rare in the latter two populations. The rarity of hydrops fetalis could be due to rarity of the zero-gene chromosome in these populations (Lehmann 1970). Genomic blotting studies of a-thalassemia 1 in blacks confirmed that this syndrome arises from homozygosity for the single gene chromosome (Dozy et al 1979; Table 1). Restriction endonuclease mapping analysis indicated that chromosomes containing a single a-globin gene could have resulted from deletion of the a2 gene or from unequal crossing over between the a2 and al genes to form a single hybrid gene (Orkin et al 1979b; Embury et al 1979). More detailed mapping of single-gene chromosomes from Asians, blacks and Mediterraneans revealed that these chromosomes are of two types (Embury et al 1980). One type of deletion (termed leftward) removes 4.2 kb of DNA containing the a2 gene, while the other type of deletion (rightward) removes 3. 7 kb of DNA, apparently by unequal crossing over between the al and a2 genes (Figure 3). The rightward type of deletion predominates in all three populations. The leftward type has been found onzy-1n several Asian cases. Deletions associated with a-thalassemia 2 (Embury et al 1980) are indistinguishable in size and position from two types of deletions which occur during propagation of bacteriophage ).. clones containing the a-globin genes (Lauer et al 1980). The breakpoints of these deletions (Figure 3) are located within the blocks of al-a2 homology described above (Figure 4). The precise lengths of the leftward (4.3 kb) and rightward (3.8 kb) types of deletions indicate that both types of deletions in cloned DNA occur by homologous but unequal crossing-over between corresponding sequences in the al and a2 gene regions. Homologous but unequal crossing-over at the a-globin locus should produce a chromosome with three a-globin genes in addition to the chromosome with a single a-globin gene. An analogous crossing-over at the 13 -globin locus produces Hb Lepore (o-13 fusion) on one chromosome and anti-Lepore (13-o fusion; e.g., Hb P Congo or Miyada) along with the 13 and o genes on the other chromosome (Weatherall and 72 Clegg 1979a). Three-gene chromosomes have been found in black and Greek Cypriot populations where Cl-thalassemia 2 is common (Goosens et al 1980; Higgs et al 1980). Recently, the breakpoints of deletions associated with two cases of hydrops fetalis have been mapped (Pressley et al 1980). In a Greek case, the deletion removes the al, a2 and z; 1 genes but leaves the z; 2 gene intact (Figure 3). The presence of Hb Portland (z; 2 y ) in the Greek infant establishes that z; 2 is a functional gene. 2 In a Thai case the deletion removes the al and a2 genes but spares both the z; 1 and z; 2 genes (Figure 3). It is noteworthy that deletions associated with hydrops fetalis may be similar to those associated with HPFH in their effect on differential globin gene expression (Weatherall et al 1970; Pressley et al 1980). During normal development Hb Portland is found in significant amount only until about ten weeks gestation (Gale et al 1979), whereas in infants with hydrops fetalis 10< ) Hb Portland is found at birth (Weatherall et al 1970; Todd et al 1970). Thus, in hydrops fetalis, z;-~bin gene expression continues beyond the time at which it is normally switched off. In both hydrops fetalis and HPFH, a deletion in one region of a gene cluster affects the expression of a distant gene within the cluster. ~ AND 6-GLOBIN PSEUDOGENES a- and S-globin sequences which cannot be identified with known globin polypeptide chains (peuedogenes) have been detected in several mammalian species (Hardison et al 1979; Fritsch et al 1980; Lauer et al 1980; Jahn et al 1980; Nishioka and Leder 1980; Vannin and Smithies 1980). Nucleotide sequence analysis of a rabbit S pseudogene ( S 2; Hardison et al 1979; L. Lacy and T. Maniatis, manuscript submitted), a human a pseudogene (\)Jal; Figure 3; Lauer et al 1980; N. Proudfoot and T. Maniatis, manuscript submitted), a mouse S pseudogene (waw-a; Jahn et al 1980), and a mouse a pseudogene (Cl-3 as designated by Nishioka and Leder 1980 or a-30.5 as designated 73 by Vannin and Smithies 1980) has demonstrated a variety of structural differences between each gene and its functional counterpart. The human S-like sequences (ijl Sl and ijJ S 2; Figure 2; Fritsch et al 1980) have not yet been extensively characterized. Each of the pseudogenes which has been analyzed exhibits 75-80% sequence homology when compared with its corresponding normal gene. None of these pseudogenes can encode a functional globin polypeptide, due to the presence of small deletions or insertions which result in alterations of the translational reading frame. In addition, one or more of the intron/exon junctions of S 2, waw-a and ijlal are different from the "consensus" sequence common to splicing junctions in globin genes and all expressed genes studied to date (Breathnach et al 1978; Lerner et al 1980). The mouse a pseudogene differs from the other pseudogenes in that both introns of the mouse a pseudogene have been precisely removed, leaving an uninterrupted coding sequence. The location of the mouse a pseudogene with respect to the functional mouse a-globin genes is unknown. It is interesting to note that in all of the mammalian globin gene clusters thus far characterized, a pseudogene is found between the embryonic (or fetal) genes and the adult genes (for example, see Figures 2 and 3). It is possible that pseudogenes have some as yet unidentified function in globin gene clusters. Alternatively, pseudogenes may be the products of gene duplication and subsequent sequence divergence (Ohno 1970). The variation in human a-globin gene number observed in present day populations and the location of ijlal within the a-like globin gene cluster are consistent with the latter possibility. As shown in Figure 3, ijlal, a2 and al are separated from each other by approximately 4 kb, which is the size of the al-a2 duplication unit noted above (Figure 4). The nucleotide sequence of ijJ al indicates that it is a-like rather than r;;-like (N. Proudfoot and T. Maniatis, manuscript submitted). It therefore seems possible that ijlal was once part of a set of three functional a-globin genes. 74 CONCLUDING REMARKS The application of gene mapping and molecular cloning procedures to the study of human globin genes has led to significant advances in our understanding of their structure and chromosomal organization. The linkage arrangement of the known a- and f3 -like genes has been established and the nucleotide sequence determination for each of these genes has been completed or is in progress. This structural information and the information provided by clinical investigations of inherited disorders in globin gene expression can now be combined to study the molecular genetics of globin gene regulation. Recent improvements in molecular cloning techniques make it possible to isolate individual globin genes and their flanking sequences from the DNA which can be obtained from a small quantity of blood. Structural comparisons between mutant globin genes and their normal counterparts may lead to the identification of sequences involved in globin gene expression. However, in vivo or in vitro assays for globin gene expressiop/4rill be required to discriminate between functionally significant sequence differences and random sequence polymorphisms. An in vitro approach to studying globin gene expression is now possible because of the recent development of cell-free extracts for RNA polymerase IIdependent transcription of cloned eukaryotic genes. Two in vitro transcription systems have been described. One system consists of a cytoplasmic extract which requires the addition of purified RNA polymerase II for activity (Weil et al 1979), while the other system consists of a concentrated whole cell extract with endogenous RNA polymerase II activity (Manley et al 1980; C. Parker, personal communication). In both systems, specific transcription of adenovirus genes (Weil et al 1979; Manley et al 1980) was shown by the fact that the capped 5' terminus of the in vitro transcript is indistinguishable from that found in vivo. The general applicability of these in vitro 75 transcription systems was recently demonstrated by specific transcription of mouse (Luse and Roeder 1980), human (Manley et al 1980; N. Proudfoot, M. Shander and T. Maniatis, unpublished results), and rabbit (N. Proudfoot, M. Shander and T. Maniatis, unpublished results) globin genes. These in vitro transcription systems may provide a rapid assay for transcriptional defects in globin genes. Moreover, comparison of extracts prepared from a variety of erythroid and non-erythroid cells may lead to the identification of factors necessary for tissue-specific globin gene expression. In vivo assays for globin gene expression utilize DNA-mediated gene transfer procedures ("convection" procedures) and SV40 tumor virus vectors. Transient introduction of globin genes into mammalian cells can be achieved by replacing the genes for SV40 viral capsid proteins with foreign DNA and growing the recombinants in the presence of an SV40 helper virus (Mulligen et al 1979; Hamer and Leder 1979a,b). The recombinant virus is amplified many thousandfold, facilitating the study of globin gene expression. The disadvantage of this approac_!yi's that only small DNA fragments ('S._2 kb) can be introduced into the vector because of size constraints on viral DNA packaging. Furthermore, transcriptional initiation from a globin gene promoter has not yet been r~ported. Large DNA fragments containing rabbit or human globin genes have been stably introduced into mouse L cells in culture using the calcium phosphate DNA transformation procedure and the herpes virus thymidine kinase gene as a selectable marker (Wigler et al 1977, 1978; Mantei et al 1979). Usually, 1-10 copies of the gene are integrated into high molecular weight DNA and one or more of the copies are expressed at a relatively low level (5-2000 copies of globin mRNA per cell compared with 40,000-50,000 globin mRNA molecules in an erythroid cell). Analysis of the human and rabbit S-globin gene transcripts in mouse L cells indicated that both introns are removed by splicing. However, in at least one cell line, the mature globin mRNA lacks 46 nt at the 5' end (Wold et al 1979). Thus, transcriptional 76 initiation or RNA processing must be abnormal in this cell line. Direct microinjection of DNA into nuclei may provide an alternative and more efficient means of introducing DNA into cells (W. F. Anderson, personal communication; C. Lo and A. Efstratiadis, personal communication). Both convection procedures (calcium phosphate DNA transformation and microinjection) suffer from difficulties in detecting small amounts of globin transcripts because of the low levels of expression of integrated genes. In addition, there is significant heterogeneity in the level of expression of the foreign gene within a cloned cell population (Hanahan et al 1980). In spite of the difficulties inherent in the SV40 and convection procedures, it should be possible to study the basic mechanisms of globin gene transcription and processing. In addition to using these in vitro and in vivo assays for studying the expression of normal and mutant globin genes (e.g., 8°- and S+-thalassemias), it will also be possible to study genes which have been altered by site-directed in vitro mutagenesis. Procedures are available for introducing point mutations_yreletions, or rearrangements at specific sites in DNA (Carbon et al 1975; Domingo et al 1976; Shortle and Nathans 1978; Weissman et al 1979; Shortle et al 1979). A study of the consequences of alterations within globin ger:i,e clusters may lead to the identification of regions important for globin gene expression. It may also be possible to identify sequences necessary for tissue-specific gene expression by introducing globin genes into erythroid cells in culture. The feasibility of this approach is suggested by cell fusion studies which demonstrated that mouse erythroleukemia cells carrying human chromosome 16 or 11 can be induced by DMSO to express high levels of human a- or S-globin mRNA and protein (Deisseroth et al 1978). By introducing normal and in vitro altered globin genes into these cells, it may be possible to define the minimal genetic unit necessary for developmentally regulated gene expression. This unit may be an individual globin gene and its flanking sequences or the entire gene cluster. 77 Acknowledgements We thank P. F. R. Little for his critical reading of the manuscript. The work cited from the authors' laboratory was supported by grants from the National Science Foundation and the National Institutes of Health. 78 REFERENCES Alwine, J.C., Kemp, D. J., Stark, G. R. 1977. Method for detection of specific RN As in agarose gels by transfer to diazobenzyloxymethyl-paper and hybridization with DNA probes. Proc. Natl. Acad. Sci. USA 74:5350-5354 Baglioni, C. 1962. The fusion of two polypeptide chains in hemoglobin Lepore and its interpretation as a genetic deletion. Proc. Natl. Acad. Sci. USA 48:1880-1885 Baralle, F. E.i Shoulders, C., Proudfoot, N. J. 1980. The primary structure of the human epsilon gene. Cell, submitted for publication. Benz, E. J., Jr., Forget, B. G., Hillman, D. G., Cohen-Solal, M., Pritchard, J., Cavallesco, C., Prensky, W., Housman, D. 1978. Variability in the amount of S-globin mRNA in $ 0 thalassemia. Cell 14L299-312 Bernards, R., Little, P. F. R., Annison, G., Williamson, R., Flavell, R. A. 1979a. Structure of the human Gy-Ay-6-S-globin gene locus. Proc. Natl. Acad. Sci. USA 76:4827-4831 Bernards, R., Rooter, J. M., Flavell, R. A. 1979b. Physical mapping of the globin gene deletion in (o S )0-thalassemia. Gene 6:265-280 Blake, C. C. F. 1980. Exons encode protein functional units. Nature 277:598 Blattner, F. R., Blechl, A. E., Denniston-Thompson, K., Faber, H. E., Richards, J. E., Slightom , J. L., Tucker, P. W., Smithies, o. 1978. Cloning human fetal y-globin and mouse a-type globin DNA: Preparation and screening of shotgun collections. Science 202:1279-1283 Botchan, M., Topp, W., Sambrook, J. 1976. The arrangement of SV40 sequences in the DNA of transformed cells. Cell 9:269-287 Breathnach, R., Benoist, C., O'Hare, K., Gannon, F., Chambon, P. 1978. Ovalbumin gene: evidence for a leader sequence in mRNA and DNA sequences at the exon-intron boundaries. Proc. Natl. Acad. Sci. USA 75:4853-4857 79 Brown, D., Wensink, P., Jordan, E. 1972. A comparison of the ribosomal DNAs of Xenopus laevis and Xenopus mulleri: the evolution of tandem genes. !!· Mol. Biol. 63:57-74 Bunn, H. F ., Forget, B. G., Ranney, H. M. 1977. Human Hemoglobins. Philadelphia: W. B. Saunders Co. Carbon, J., Shenk, T., Berg, P. 1975. Biochemical procedure for production of small deletions in SV40 DNA. Proc. Natl. Acad. Sci. USA 72:1392-1396 Chang, J. c., Kan, Y. w. 1979. 13° Thalassemia, a nonsense mutation in man. Proc. Natl. Acad. Sci. USA 76:2886-2889 Chang, J. c., Temple, G. F., Trecartin, R. F., Kan, Y. w. 1979. Suppression of the nonsense mutation in homozygous 13 ° thalassemia. Nature 281:602 Clegg, J. B., Metaxatou-Mavromati, A., Kattamis, c., Sofroniadou, K., Wood, w. G., Weatherall, D. J. 1979. Occurrence of G y Hb Fin Greek HPFH: analysis of heterozygotes and compound heterozygotes with 7 thalassemia. Brit.!!· Hematology 43:521-536 Comi, P., Giglioni, B., Barbarano, L., Ottolenghi, S., Williamson, R., Novakova, M., Masera, G. 1977. Transcriptional and post-transcriptional defects in 13 ° thalassemia. Eur. J. Biochem. 79:617-622 Conconi, F., Rowley, P. T., Del Senno, L., Pontremoli, S. 1972. Induction of 8globin synthesis in the 13-thalassemia of Ferrara. Nature New Biol. 238:83-85 Conconi, F., Del Senno, 1974. L. The molecular defect of Ferrara 8-thalassemia. Ann. N.Y. Acad. Sci. 232:54 Cook, P. R. 1973. Hypothesis on differentiation and the inheritance of gene superstructure. Nature 245:23-25 Craik, C. S., Buchman, S. R., Beychok, S. 1980. Characterization of globin domains: Heme binding to the central exon product. Proc. Natl. Acad. Sci. USA 77: 1384-1388 80 Crick, F. 1979. Split genes and RNA splicing. Science 204:264-271 Davidson, E., Britten, R. 1979. Regulation of gene expression: possible role of repetitive sequences. Science 204:1052-1059 Dawid, I., Wahli, W. 1979. Application of recombinant DNA technology to questions of developmental biology: a review. Dev. Biol. 69:305-328 Dayhoff, M. o. 1972. Atlas of Protein Sequence and Structure. Washington, D.C.: National Biomedical Research Foundation. Deisseroth, A., Hendrick, D. 1978. Human a-globin gene expression following chromosomal dependent gene transfer into mouse erythroleukemia cells. Cell 15:55-63 Deisseroth, A., Nienhuis, A., Turner, P., Velez, R., Anderson, w. F., Lawrence, J., Creagan, R., Kucherlapati, R. 1977. Localization of the human a-globin structural gene to chromosome 16 in somatic cell hybrids by molecular hybridization assay. Cell 12:205-218 Deisseroth, A., Nienhuis, A., Lawrence, J., Giles, R., Turner, P., Ruddle, F. H. 1978. Chromosomal localization of human 8 -globin gene on human chromosome 11 in somatic cell hybrids. Proc. Natl. Acad. Sci. USA 75:1456-1460 Dodgson, J. B., Strommer, J., Engel, J. D. 1979. Isolation of the chicken 8-globin gene and a linked embryonic 8-like globin gene from a chicken DNA recombinant library. Cell 17:879-887 Domingo, E., Flavell, R. A., Weissman, C. 1976. In vitro site-directed mutagenesis: generation and properties of an infectious extracistronic mutant of bacteriophage Q 8. Gene 1:3-25 Dozy, A., Kan, Y. W., Embury, S., Mentzer, w., Wang, w., Lubin, B., Davis, J., Koenig, H. 1979. a-Globin gene organization in blacks precludes the severe form of a-thalassemia. Nature 280:605-607 81 Duncan, c., Biro, P. A., Chowdary, P. V., Elder, J. T., Wang, R. R. c., Forget, B. G., De Riel, J. K., Weissman, S. M. 1980. RNA polymerase III transcriptional units are interspersed among human non-a-globin genes. Proc. Natl. Acad. Sci. USA 76:5095-5099 Early, P. E., Davis, M., Kaback, D., Davidson, N., Hood,L. 1979. Immunoglobulin heavy chain gene organization in mice: Analysis of a myeloma genomic clone containing variable and a constant regions. Proc. Natl. Acad. Sci. USA 76:857-861 Eaton, W. A. 1980. The relationship between coding sequences and function in hemoglobin. Nature 284:183-185 Efstratiadis, A., Posakony, J., Maniatis, T., Baralle, F., Blechl, A., De Riel, J., Forget, B., Lawn, R., O'Connell, C., Proudfoot, N., Shoulders, C., Slightom, J., Smithies, O., Spritz, R., Weissman, S. 1980. DNA sequence comparison of the human 8-like globin genes. Cell, submitted for ublication. Efstratiadis, A., Villa-Komaroff, L. 1979. Cloning of double-stranded cDNA. In Genetic Engineering: Principles and Methods, ed. J. Setlow. New York: Plenum Press. Embury, S., Lebo, R., Dozy, A., Kan, Y. w. 1979. Organization of the a-globin genes in the Chinese a-thalassemia syndromes. !I_. Clin. Invest. 63:1307-1310 Embury, S. H., Miller, J. A., Dozy, A. M., Kan, Y. W., Chan, V., Todd, D. 1980. Two different molecular organizations account for the single a-globin gene of the a-thalassemia 2 genotype. Submitted for publication. Farabaugh, P. J., Miller, J. H. 1978. Genetic studies of the lac repressor. VII. On the molecular nature of spontaneous hotspots in the lac i gene of Escherichia coli. J. Mol. Biol. 126:847-863 Flavell, R. A., Kooter, J. M., De Boer, E., Little, P. F. R., Williamson, R. 1978. Analysis of the human B-o-globin gene loci in normal and Hb Lepore DNA: direct determination of gene linkage and intergene distance. Cell 15:25-41 82 Flavell, R. A., Bernards, R., Grosveld, G. C., Hooijmakers-VanDommelen, H. A. M., Kooter, J. M., De Boer, E. 1979a. The structure and expression of globin genes in rabbit and man. In Eukaryotic Gene Regulation, ICN-UCLA Symposium on Molecular and Cellular Biology, XIV, ed. Axel, R., Maniatis, T., Fox, C.F. New York: Academic Press. Flavell, R. A., Bernards, R., Kooter, J. M., De Boer, E. 1979b. The structure of the human S-globin gene in f3-thalassemia. Nucl. Acids Res. 6:2749-2760. Forget, B. G., Hillman, D. G., Lazarus, H., Barell, E. F., Benz, E. J., Jr., Caskey, C. T., Huisman, T. H.J., Schroeder, w. A., Housman, D. 1976. Absence of messenger RNA and gene DNA for f3 globin chains in hereditary persistence of fetal hemoglobin. Cell 7:323-329. Forget, B. G., Cavallesco, c., de Riel, J. K., Spritz, R. A., Choudary, P. V., Wilson, J. T., Wilson, L.B., Reddy, V. B., Weissman, S. M. 1979. Structure of the human globin genes. In Eukaryotic Gene Regulatioy, ICN-UCLA Symposium on Molecular and Cellular Biology, XIV ed. Axel, R., Maniatis, T., Fox, C.F. New York: Academic Press. Fritsch, E. F., Lawn, R. M., Maniatis, T. 1979. Characterization of deletions which affect the expression of fetal globin genes in man. Nature 279:598-603. Fritsch, E. F., Lawn, R. M., Maniatis, T. 1980. Molecular cloning and characterization of the human f3-like globin gene cluster. Cell 19:959-972. Gale, R., Clegg, J., Huehns, E. 1979. Human embryonic hemoglobins Gower 1 and Gower 2. Nature 280:162-164. Gilbert, w. 1978. Why genes in pieces? Nature 271:501. Gilbert, W. 1979. Introns and exons: playgrounds of evolution. In Eukaryotic Gene Regulation, ICN-UCLA Symposium on Molecular and Cellular Biology, XIV, ed. Axel, R., Maniatis, T., Fox, C.F. New York: Academic Press. 83 Goosens, M., Dozy, A., Embury, S., Zacharides, Z., Hadjiminas, M., Stamatoyannopoulos, G., Kan, Y. W. 1980. Triplicated a-globin loci in humans. Proc. Nat. Acad. Sci. USA 77:518-521 Gusella, J., Varsanyi-Brelner, A., Kao, F. T., Jones, c., Puck, T., Keys, c., Orkin, S., Housman, D. 1979. Precise localization of human 8-globin gene complex on chromosome 11. Proc. Natl. Acad. Sci. USA 76:5239-5243 Hardison, R. c., Butler, E. T., Lacy, E., Maniatis, T., Rosenthal, N., Efstratiadis, A. 1979. The structure and transcription of four linked rabbit 8-like globin genes. Cell 18:1285-1297 Hamer, D. H., Leder, P. 1979. Expression of the chromosomal mouse 8 maj_globin gene cloned in SV40. Nature 281:35-40 Hamer, D. H., Smith, K. D., Boyer, S. H., Leder, P. 1979. SV40 recombinants carrying rabbit 8-globin gene coding sequences. Cell 17:725-735 Hanahan, D., Lane, D., Lipsich, L., Wigler, M., Botchan, ~ 1980. Characteristics of an SV40-plasmid recombinant and its movement into and out of the genome of a murine cell. Submitted for publication. Higgs, D. R., Old, J. M., Pressley, L., Clegg, J. B., Weatherall, D. J. 1980. A novel a-globin gene arrangement in man. Nature 284:632-635 Higuchi, R., Paddock, G. v., Wall, R., Salser, W. 1976. A general method for cloning eukaryotic structural gene sequences. Proc. Natl. Acad. Sci. USA 7 3:3146-3150 Hood, L., Campbell, J. H., Elgin, S. C. R. 1975. The organization, expression and evolution of antibody genes and other multigene families. Ann. Rev. Genet. 9:305-353 Houck, c. M., Rinehart, F. P., Schmid, C. W. 1979. A ubiquitous family of repeated DNA sequences in the human genome. i!_. Mol. Biol. 132:289-306 84 Huehns, E., Farooqui, A. 1975. Oxygen dissociation properties of human embryonic red cells. Nature 254:335-337 Huisman, T. H.J., Schroeder, w. A., Bannister, w. H., Grech, J. L. 1972. Evidence for four nonallelic structural genes for the y chain of human fetal hemoglobin. Biochem. Genet. 7:131-139 Huisman, T. H.J., Wrightstone, R. N., Wilson, J. B., Schroeder, w. A., Kendall, A. G. 1972. Hemoglobin Kenya, the product of fusion of y and S polypeptide chains. Arch. Biochem. Biophys. 153:850-853 Huisman, T. H.J., Schroeder, W. A., Efremov, G.D., Duma, H., Meadenovski, B., Hyman, C. B., Rachmilewitz, E. A., Bouver, N., Miller, A., Brodie, A. R., Shelton, J. R., Appel, G. 1974. The present status of the heterogeneity of fetal hemoglobinin S-thalassemia: an attempt to unify some observations in thalassemia and related conditions. Ann. N. Y. Acad. Sci. 232:107-124 Jahn, C. L., Hutchinson, c. A. III, Phillips, S. J., Weaver, S , Haigwood, N. L., Voliva, c. F., Edgell, M. H. 1980. DNA sequence organization of the Sglobin complex in the BALB/c mouse. Submitted for publication. Jeffreys, A. J. 1979. DNA sequence variants in the G y-, Ay-, o- and S-globin genes of man. Cell 18:1-10 Jeffreys, A., Craig, I., Francke, u. 1979. Localization of the G y-, Ay-, o- and S-globin genes on the short arm of human chromosome 11. Nature 281:606-608 Jeffreys, A. J., Flavell, R. A. 1977a. A physical map of the DNA regions flanking the rabbit S-globin gene. Cell 12:429-439 Jeffreys, A. J., Flavell, R. A. 1977b. The rabbit S-globin gene contains a large insert in the coding sequence. Cell 12:1097-1108 Jelinek, w. G., Leinwand, L. 1978. Low molecular weight RNAs hydrogen-bonded to nuclear and cytoplasmic poly(A)-terminated RNA from cultured Chinese hamster ovary cells. Cell 15:205-214 85 Jelinek, w. R., Toomey, T. P., Leinwand, L., Duncan, C. H., Biro, P. A., Choudary, P. N., Weissman, S. M., Rubin, C. M., Houck, C. M., Deininger, P. L., Schmid, c. w. 1980. Ubiquitous, interspersed repeated sequences in mammalian genomes. Proc. Natl. Acad. Sci. USA 77:1398-1402 Kan, Y. w., Dozy, A., Varmus, H., Taylor, J., Holland, J., Lie-Injo, L., Ganesan, J., Todd, D. 1975a. Deletion of a-globin genes in hemoglobin H disease demonstrates multiple a-globin structural loci. Nature 255:255-256 Kan, Y. w., Holland, J.P., Dozy, A. M., Charache, S., Kazazian, H. H. 1975b. Deletion of the 8 globin structural gene in hereditary persistence of fetal hemoglobin. Nature 258:162-163 Kan, Y. W., Dozy, A., Trecartin, R., Todd, D. 1977. Identification of a non deletion defect in a-thalassemia. N. Engl.!!· Med. 297:1081-1084 Kan, Y. w., Dozy, A. M. 1978. Polymorphism of DNA sequence adjacent to human 8 -globin structural gene: relationship to sickle m~tion. Proc. Natl. Acad. Sci. USA 75:5631-5635 Kan, Y. w., Dozy, A., Stamatoyannopoulos, G., Hadjiminas, M., Zacharides, Z., Furbetta, M., Cao, A. 1979. Molecular basis of hemoglobin H disease in the Mediterranean population. Blood 54:1434-1438 Kan, Y. W., Lee, K. Y., Furbetta, M., Angius, A., Cao, A. 1980. Polymorphism of DNA sequence in the 8 -globin gene region. New Engl.!!· Med. 302:185-188 Kantor, J. A., Turner, P. H., Nienhuis, A. W. 1980. Beta+ thalassemia: mutations which affect processing of the 8-globin mRN A precursor. Cell, submitted. Kinniburgh, A. J., Mertz, J. E., Ross, J. 1978. The precursor of mouse 8-globin messenger RNA contains two intervening RNA sequences. Cell 14:681-693 Kinniburgh, A. J., Ross, J. 1979. Processing of the mouse 8-globin mRNA precursor: at least two cleavage-ligation reactions are necessary to excise the large intervening sequence. Cell 17:915-921 86 Konkel, D. A., Maizel, J. V., Leder, P. 1979. The evolution and sequence comparison of two recently diverged mouse chromosomal S-globin genes. Cell 18:865-873 Konkel, D. A., Tilghman, S. M., Leder, P. 1978. The sequence of the chromosomal mouse S-globin major gene: homologies in capping, splicing and poly(A) sites. Cell 15:1125-1132 Lacy, E., Hardison, R. c., Quon, D., Maniatis, T. 1979. The linkage arrangement of four rabbit S-like globin genes. Cell 18:1273-1283 Lauer, J., Shen, c.-K. J., Maniatis, T. 1980. The chromosomal arrangement of human a-like globin genes: sequence homology and a-globin gene deletions. Cell 20:119-130 Lawn, R. M., Efstratiadis, A., O'Connell, C., Maniatis, T. 1980. Cell, submitted for publication. Lawn, R. M., Fritsch, E. F., Parker, R. C., Blake, G., Maniatis, T. 1978. The isolation and characterization of linked c- and S-gJ91:5in genes from a cloned library of human DNA. Cell 15:1157-1174 Lebo, R., Carrano, A., Burkhart-Schultz, K., Dozy, A., Yu, L. C., Kan, Y. W. 1979. Assignment of human S-, y-, and c-globin genes to the short arm of chromosome 11 by chromosome sorting and DNA restriction enzyme analysis. Proc. Natl. Acad. Sci. USA 76:5804-5808 Lehmann, H. 1970. Different types of a-thalassemia and significance of hemoglobin Barts in neonates. Lancet ii:73 Lerner, M., Boyle, J., Mount, S., Wolin, S., Steitz, J. 1980. Are snRNPs involved in splicing? Nature 283:220-224 Lie-Injo Luan Eng, Lie Hong Gie, Ager, J. A. M., Lehmann, H. 1962. a-thalassemia as a cause of hydrops fetalis. Brit.:!_. Haematol. 8:1 Little, P. F. R., Annison, G., Darling, S., Williamson, R., Camba, L., Modell, B. 1980. A model for antenatal diagnosis of S-thalassemia and other monogenic disorders using molecular analysis of linked DNA polymorphisms. Nature, in press. 87 Little, P., Curtis, P., Coutelle, c., Van Den Berg, J., Dalgleish, R., Malcolm, S., Courtney, M., Weatway, D., Williamson, R. 1978. Isolation and partial sequence of recombinant plasmids containing human a-, 8-, and y-globin cDNA fragments. Nature 273:640-643 Little, P. F. R., Flavell, R. A., Kooter, J. M., Annison, G., Williamson, R. 1979a. Structure of the human fetal globin gene locus. Nature 278:227-231 Little, P. F. R., Williamson, R., Annison, G., Flavell, R. A., De Boer, E., Bernini, L. G., Ottolenghi, S., Saglio, G., Mazza, U. 1979b. Polymorphisms of human y-globin genes in Mediterranean populations. Nature 282:316-318 Luse, D. S., Roeder, R. G. 1980. Accurate transcription initiation on a purified mouse B-globin DNA fragment in a cell-free system. Cell, submitted for publication. Maniatis, T. 1980. Recombinant DNA procedures in the study of eukaryotic genes. In Comprehensive Cell Biology, Vol. 3, ed. y,"Goldstein, D. Prescott. New York: Academic Press. Maniatis, T., Hardison, R. c., Lacy, E., Lauer, J., O'Connell, c., Quon, D., Sim, G. K., Efstratiadis, A. 1978. The isolation of structural genes from libraries of eucaryotic DNA. Cell 15:687-701 Maniatis, T., Sim, G.-K., Efstratiadis, A., Kafatos, F. 1976. Amplification and characterization of a 8-globin gene synthesized in vitro. Cell 8:163-182 Manley, J. L., Fire, A., Cano, A., Sharp, P. A., Gefter, M. L. 1980. DNA-dependent transcription of adenovirus genes in a soluble whole-cell extract. Proc. Natl. Acad. Sci. USA, in press. Mantei, N., Boll, W., Weissman, C. 1979. Rabbit B-globin mRNA production in mouse L cells transformed with cloned rabbit B-globin chromosomal DNA. Nature 281:40-46. 88 Maquat, L. E., Kinniburgh, A. J., Beach, L. R., Honig, G. R., Lazerson, J., Ershler, W. B., Ross, J. 1980. Processing of the human 13-globin mRNA precursor to mRNA is defective in three patients with 13 + thalassemia. Proc. Natl. Acad. Sci. USA, submitted for publication. Marotta, C. A., Wilson, J. T., Forget, B. G., Weissman, S. M. 1977. Human 13-globin messenger RNA. III. Nucleotide sequences derived from complementary DNA. J. Biol. Chem. 252:5040-5053. Mears, J. G., Ramirez, F., Leibowitz, D., Bank, A. 1978. Organization of human o- and 13-globin genes in cellular DNA and the presence of intragenic inserts. Cell 15:15-23 Mulligan, R. c., Howard, B. H., Berg, P. 1979. Synthesis of rabbit 13-globin in cultured monkey kidney cells following infection with a SV40 13-globin recombinant clone. Nature 277:108-114 Nienhuis, A. W., Turner, P., Benz, E. J., Jr. 1977. Rela~,e stability of a- and 13-globin messenger RN As in homozygous 13 + thalassemia. Proc. Natl. Acad. Sci. USA 74:3960-3964. Nishioka, Y., Leder, P. 1979. The complete sequence of a chromosomal mouse a-globin gene reveals elements conserved throughout vertebrate evolution. Cell 18:875-882. Ohno, S. 1970. Evolution~ Gene Duplication, pp. 76-77. New York: SpringerVerlag. Old, J. M., Proudfoot, N. J., Wood, W. G., Longley, J. I., Clegg, J. B., Weatherall, D. J. 1978. Characterization of 13-globin mRNA in 13° thalassemias. Cell 14:289-298 Orkin, S. H. 1978. The duplicated human a globin genes lie close together in cellular DNA. Proc. Natl. Acad. Sci. USA 75:5950-5954 Orkin, S. H., Alter, B. P., Altay, c. 1979a. Deletion of the Ay-globin gene in G y-o 13-thalassemia. i!· Clin. Invest. 64:866-869 89 Orkin, S. H., Kolodner, R., Michelson, A., Husson, R. 1980. Cloning and direct examination of a structurally abnormal human beta0-thalassemia globin gene. Proc. Natl. Acad. Sci. USA, submitted for publication. Orkin, S., Michelson, A. 1980. Submitted for publication. Orkin, S., Old, J., Lazarus, H., Altay, c., Gurgey, A., Weatherall, D., Nathan, D. 1979b. The molecular basis of a-thalassemias: frequent occurrence of dysfunctional a loci among non-Asians with Hb H disease. Cell 17:33-42 Orkin, S. H., Old, J. M., Weatherall, D. J., Nathan, D. G. 1979c. Partial deletion of S-globin gene DNA in certain patients with 13 °-thalassemia. Proc. Natl. Acad. Sci. USA 76:2400-2404 Ottolenghi, S., Lanyon, W. G., Paul, J., Williamson, R., Weatherall, D. J., Clegg, J. B., Pritchard, J., Pootrakul, S., Wong, H. B. 1974. The severe form of a thalassemia is caused by a hemoglobin gene deletion. Nature 251:389-392 Ottolenghi, S., Comi, P., Giglioni, B., Tolstoshev, P., 1ay n, W. G., Mitchell, G. J., Williamson, R., Russo, G., Musumeci, S., Schiliro, G., Tsistrakis, G. A., Charache, S., Wood, W. G., Clegg, J. B., Weatherall, D. J. 1976. o13-thalassemia is due to a gene deletion. Cell 9:71-80 Ottolenghi, S., Giglions, B., Come, P., Gianni, A. M., Polli, E., Acquaye, C. T. A., Oldhom, J. H., Masera, G. 1979. Globin gene deletion in HPFH, 8°13 ° thalassemia and Hb Lepore disease. Nature 278:654-659 Pressley, L., Higgs, D., Clegg, J., Weatherall, D. 1980. Gene deletions in a-thalassemia prove that the 5' z; locus is functional. Proc. Natl. Acad. Sci. USA, in press. Proudfoot, N., Baralle, F. 1979. Molecular cloning of the human e:-globin gene. Proc. Natl. Acad. Sci. USA 76:5435-5439 Ramirez, F., O'Donnell, J. V., Marks, P.A., Bank, A., Musumeci, S., Schiliro, G., Pizzarelli, G., Russo, G., Luppis, B., Gambino, R. 1976. Abnormal or absent beta mRNA in beta° Ferrara and gene deletion in delta-beta thalassemia. Nature 263:471-475 90 Ramirez, F., Burns, A. L., Mears, J. G., Spence, S., Starkman, D., Bank, A. 1979. Isolation and characterization of cloned human fetal globin genes. Nucl. Acid Res. 7:1147-1162 Ricco, G., Muzza, u., Turi, R. M., Pich, P. G., Camashella, c., Saglio, G., Bernini, L. F. 1976. Significance of a new type of human fetal hemoglobin carrying a replacement isoleucine~ threonine at position 75 (E19) of the y chain. Hum. Genet. 32:305-313 Roberts, A. V., Weatherall, D. J., Clegg, J. B. 1972. The synthesis of human hemoglobin A2 during erythroid maturation. Biochem. Biophys. Res. Comm. 47:81-87 Rougeon, F., Kourilsky, P., Mach, B. 1975. Insertion of a rabbit 8-globin gene sequence into an _g. coli plasmid. Nucl. Acids Res. 2:2365-2378 Saglio, G., Ricco, G., Mazza, U., Camaschella, c., Pich, P. G., Gianni, A. M., Gianazza, E., Righette, P. G., Giglione, B., Comi, :e , Gusmerole, M., Ottolenghi, S. 1979. Human Ty globin chain is a variant of Ay chain (Ay sardinia). Proc. Natl. Acad. Sci. USA 76:3420-3424 Sakano, H., Rogers, J., Huppi, K., Brack, c., Traunecker, A., Maki, R., Wall, R., Tonegawa, S. 1979. Domains and the hinge region of an immunoglobulin heavy chain are encoded in separate DNA segments. Nature 277:627-633 Schroeder, w. A., Bannister, w. H., Grech, J., Brown, A., Weightstone, R., Huisman, T. 1973. Non-synchronized suppression of postnatal activity in non-allelic genes which synthesize the G y chain in human fetal hemoglobin. Nature New Biol. 244: 89-90 Shen, c.-K. J., Maniatis, T. 1980. The organization of repetitive sequences in a cluster of rabbit 8-like globin genes. Cell 19:379-391 Shortle, D., Nathans, D. 1978. Local mutagenesis: a method for generating viral mutants with base substitutions in preselected regions of the viral genome. Proc. Natl. Acad. Sci. USA 75:2170-2174 91 Shortle, D., Pipas, J., Lazarowitz, S., DiMaio, D., Nathans, D. 1979. Constructed mutants of simian virus 40. In Genetic Engineering, ed. J. K. Setlow, A. Hollaender, 1:73-92. New York: Plenum. Slightom, J., Blechl, A., Smithies, o. 1980. Human fetal G y and Ay globin genes: complete nucleotide sequences suggest that DNA can be exchanged between these duplicated genes. Cell, in press. Smithies, O., Blechl, A., Denniston-Thompson, K., Newell, N., Richards, J., Slightom, J., Tucker, D., Blattner, F. Cloning human fetal y-globin and mouse a-type globin DNA: characterization and partial sequencing. Science 202:1284-1289. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. !!· Mol. Biol. 98:503-517 Spritz, R., DeRiel, J., Forget, B., Weissman, S. 1980. Nucleotide sequence of the human 6 globin gene. Cell. Submitted for publication Stalder, J., Larsen, A., Groudine, M., Engel, J. D., Dodg~, J., Weintraub, H. 1980. Tissue-specific cuts in the globin chromatin domain introduced by DNase I. Submitted for publication. Streisinger, G., Okada, Y., Emrich, J., Newton, J., Tsugita, A., Terzaghi, E., Inouye, M. 1966. Frameshift mutations and the genetic code. Cold Spring Harbor ~ - Quant. Biol. 31:77-84 Taylor, J. M., Dozy, A., Kan, Y. w., Varmus, H. E., Lie-Injo, L. E., Ganesan, J., Todd, D. 197 4. Genetic lesion in homozygous a. thalassemia (hydrops fetalis). Nature 251:392-393 Temple, G. F., Chang, J.C., Kan, Y. W. 1977. Authentic S-globin mRNA sequences in homozygous S 0 -thalassemia. Proc. Natl. Acad. Sci. USA 74:3047-3051. Tilghman, S. M., Curtis, P. J., Tiemeier, D. c., Leder, P., Weissman, C. 1978b. The intervening sequence of a mouse S globin gene is transcribed within the 15 S S-globin mRNA precursor. Proc. Natl. Acad. Sci. USA 75:1309-1313 92 Tilghman, S. M., Tiemeier, D. C., Polsky, F., Edgell, M. H., Seidman, J. G., Leder, A., Enquist, L. W., Norman, B., Leder, P. 1977. Cloning specific segments of the mammalian genome: bacteriophage A containing mouse globin and surrounding sequences. Proc. Natl. Acad. Sci. USA 74:4406-4410 Tilghman, S. M., Tiemeier, D. c., Seidman, J. G., Peterlin, B. M., Sullivan, M., Maizel, J., Leder, P. 1978a. Intervening sequence of DNA identified in the structural portion of a mouse 8-globin gene. Proc. Natl. Acad. Sci. USA 75:1309-1313 Todd, D., Lai, M. C. s., Beaven, G., Huehns, E. 1970. The abnormal haemoglobins in homozygous a-thalassemia. Brit!!· Haematol. 19:27-31 Tuan, D., Biro, P. A., de Riel, J. K., Lazarus, H., Forget, B. G. 1979. Restriction endonuclease mapping of the human y globin gene loci. Nucl. Acids. Res. 6:2519-2544 Tuan, D., Murnane, M. J., de Riel, J. K., Forget, B. G. !J80, Heterogenity in the molecular basis of hereditary persistence of fetal hemoglobin. Nature, in press. van den Berg, J., van Ooyen, A., Mantei, N., Shambock, A., Grosveld, G., Flavell, R. A., Weissmann, C. 1978. Comparison of cloned rabbit and mouse 8-globin genes showing strong evolutionary divergence of two homologous pairs of introns. Nature 276:37-44 van der Ploeg, L. H. T., Konings, A., Oort, M., Roos, D., Bernini, L., Flavell, R. A. 1980. y-8-thalassemia studies showing that deletion of the y- and c-genes influences 8-globin gene expression in man. Nature 283:637-642 Vannin, E., Smithies, o. 1980. Submitted for publication. Wasi, P., Na-Nakorn, S., Suingdumrong, A. 1964. Hemoglobin H disease in Thailand: a genetical study. Nature 204:907-908 93 Weatherall, D. J., Clegg, J. B. 1975. The a-chain-termination mutants and their relationship to a-thalassemia. Philos. Trans. R. Soc. Lond. 271:411-455. Weatherall, D. J., Clegg, J. B. 1979a. Recent developments in the molecular genetics of human hemoglobin. Cell 16:467-479 Weatherall, D. J., Clegg, J. B. 1979b. The Thalassemia Syndromes, Blackwell Scientific Publications, 3rd ed. Weatherall, D., Clegg, J., Wong, H. 1970. The haemoglobin constitution of infants with the haemoglobin Bart's hydrops fetalis syndrome. Brit. !!_. Haematol. 18:357-367 Weatherall, D. J., Clegg, J. B., Wood, W. G., Pasvol, G. 1979. Human haemoglobin genetics. In Human Genetics: Possibilities and Realities. Ciba Foundation Symp. 66 (New Series), pp. 147-186. New York: Exerptea Medica. Weaver, R. F., Weissman, C. 1979. Mapping of RNA by a modification of the Berk-Sharp procedure: The 5' termini of 15 S S-globin mRNA precursor / and mature 10 S S-globin mRNA have identical map coordinates. Nucl. Acids Res. 7:1175-1193 Weil, P. A., Luse, D. S., Segall, J., Roeder, R. G. 1979. Selective and accurate initiation of transcription at the Ad2 major late promotor in a soluble system dependent on purified RNA polymerase II and DNA. Cell 18:469-484 Weintraub, H., Flint, s. J., Leffak, I. M., Groudine, M., Grainger, R. M. 1978. The generation and propagation of variegated chromosome structures. Cold Spring Harbor~- Quant. Biol. 42:401-407 Weintraub, H., Groudine, M. 1976. Chromosomal subunits in active genes have an altered conformation. Science 193:848-853 Weissman, c., Nagota, S., Taniguchi, T., Weber, H., Meyer, F. 1979. The use of site directed mutagenesis in reversed genetics. In Genetic Engineering, ed. J. Setlow, A. Hollaender, 1:133-150. New York: Plenum. 94 Wigler, M., Silverstein, s., Lee, L. s., Pellicer, A., Cheng, Y. C., Axel, R. 1977. Transfer of purified herpes virus thymidine kinase gene to cultured mouse cells. Cell 11:223-232 Wigler, M., Sweet, R., Sim, G. K., Wold, B., Pellicer, A., Lacy, E., Maniatis, T., Silverstein, s., Axel, R. 1979. Transformation of mammalian cells with genes from procaryotes and eucaryotes. Cell 16:777-785. Wilson, J. T., Wilson, L. B., de Riel, J. K., Villa-Komaroff, L., Efstratiadis, A., Forget, B. G., Weissman, S. M. 1978. Insertion of synthetic copies of human globin genes into bacterial plasmids. Nucl. Acids Res. 5:563-581 Wold, B., Wigler, M., Lacy, E., Maniatis, T., Silverstein, S., Axel, R. 1979. Expression of an adult rabbit S-globin gene stably inserted into the genome of mouse L cells. Proc. Natl. Acad. Sci. USA 76:5684-5688 Wood, W. G., Weatherall, D. J. 1973. Haemoglobin synthesis during human fetal development. Nature 244:162-165 Wood, W. G., Old, J. M., Roberts, A. V. S., Clegg, J. B., Weatherall, D. J. 1978. Human globin gene expression: control of S, o, and oS chain production. Cell 15:437-446 Wood, W. B., Clegg, J. B., Weatherall, D. J. 1979. Hereditary persistence of fetal hemoglobin (HPFH) and oS thalassemia. Brit. !!· Hematol. 43:509-520 Wu, C., Bingham, P. M., Livak, K. J., Holmgren, R., Elgin, S. C. R. 1979. The chromatin structure of specific genes: I. Evidence for higher order domains of defined DNA sequence. Cell 16:797-806 Wu, c., Wong, Y.-c., Elgin, S. C. R. 1979. The chromatin structure of specific genes: II. Disruption of chromatin structure during gene activity. Cell 16: 807-814 Zimmer, E. A., Martin, S. L., Beverley, S. M., Kan, Y. w., Wilson, A. C. 1980. Rapid duplication and loss of genes coding for the a chains of hemoglobin. Proc. Natl. Acad. Sci. USA 77 (in press). 95 Table 1 a-specific Eco RI fragments associated with a-thalasssemia syndromes Chromosome A Chromosome B Gene Fragment Gene Fragment number size number size Normal 2 23 2 23 a-thalassemia 2 (A,B,M) 2 23 1 19 a-thalassemia 2 (M) 2 23 1 (1) 23 a-thalasse m ia 1 (A) 0 2 23 a-thalassemia 1 (A,B,M) 1 19 1 19 a-thalassemia 1 (M) 1 19 1 (1) 23 Hb H (A,B,M) 0 1 19 Hb H (A,M) 0 1 ( 1) 23 Hb H (M) 1 19 0 ( 1) 2.6 Hb H (M) 1 (1) 23 0 ( 1) 2.6 hydrops fetalis (A) 0 0 The various a-thalassemia syndromes are listed in the left column. The size of the Eco RI fragment containing the a-globin genes is given in kilobase pairs. The number of functional a-globin genes on each chromosome is indicated, followed by a number in parentheses to indicate the number of nonfunctional genes on the same chromosome. A, B, and M in parentheses following the name of a syndrome signify that the indicated combination of Eco RI fragments has been observed in Asian, black, or Mediterranean populations, respectively. 96 Figure 1 Structure of human globin genes. The canonical structures for the human a-like and 8-like globin genes are drawn to approximate scale. Solid and open boxes represent coding (exon) and noncoding (intron) sequences, respectively. The a-like globin genes contain introns of approximately 95 and 125 bp, located between codons 31 and 32 and 99 and 100, respectively. The 8-like globin genes contain introns of approximately 125-150 and 800-900 bp, located between codons 30 and 31 and 104 and 105, respectively. 31 104 (1) i'.... 0( 0 31 32 200 99 400 f3 ~ - - - - - - - 30 100 600 · 141 - 800 1000 1200 1400 1600 iilllm?A 105 146 -.::i (0 98 Figure 2 Linkage arrangement of the human S-like globin genes and locations of deletions within the S-like gene cluster. The positi<?ns of the embryonic (e:), fetal (G y, Ay) and adult (o, S) S-like globin genes and the two S-like pseudogenes (ijJ S 1, 1jJ S 2) are shown. For each gene the black and white boxes represent the coding (exon) and noncoding (intron) sequences, respectively. The distribution of coding and noncoding sequences within 1jJ S 1 and 1jJ S 2 is not known. The locations of various deletions within the gene cluster are presented below the map. Open boxes represent areas known to be deleted; dashed lines indicate that the endpoint of the deletion has not been determined; and stippled boxes represent uncertainty in the extent of the deletions. For o S -thalassemia and HPFH, the type of fetal glob in chain which is produced (G y and/or Ay) is indicated in the name of each syndrome (for example, in G y-Ay-o S-thalassemia, the G y- and Ay-globin chains are produced). The percentage of Hb F observed in heterozygotes is given to the right of each deletion. An asterisk (*) indicates that the Hb F is entirely of the G y-Jm)e. See the text for references and discussion. to., ti) i' 1 I y-,8-thal 1 60 5 ~ 3 I I I I I 40 I I I G I I A G 30 I I A I 1·,1 I ,8°-thal I I □ lI:1 ca I f3 s 20 H b Lepore I y- y- S,8 - thal I D \Jr 131 1- y-HPFH G A y- y-HPFH Hb Kenya I 1 -c3,B- thal G Gy-HPFH 50 I (D - - D r - - - - ( D r - - - - - - - - - -(D Kb A)' E' Gy l/1132 I o 10 I 0 20-30 20-30 5-10* 10-13* I- 5 5 -15 c.o c.o 100 Figure 3 Linkage arrangement of the human a-like globin genes and locations of deletions within the a-like gene cluster. The positions of the adult (al, a2) and embryonic (i;; 1, i;; 2) a-like globin genes and the a-like pseudogene ( iµal) are shown. For each gene the black and white boxes represent coding (exon) and noncoding (intron) sequences. The introns in I;; 2 are assumed to exist by analogy with the other a-like genes. The locations of deletions associated with the leftward and rightward types of a-thalassemia 2 are indicated by the rectangles labeled a-thal 2 Land a-thal 2 R. The cross-hatched boxes at the ends of these rectangles indicate regions of sequence homology. The breakpoints of each type of a-thalassemia 2 deletion can occur anywhere within the regions of homology. The locations of deletions associated with two cases of a-thalassemia 1 (a-thal 1 Thai and a-thal 1 Greek) are shown below the linkage map. The stippled boxes indicate uncertainty in the extent of each deletion. See the text for discussion and references. w ~ f Kb 5 ~3 1 I 30 1 I I I ~-thal I Greek I llll s'2 I ['- :: .:• 20 I .. :. _) --,~'':- ;-... :. ~-thal I Thai I I ~I 1£1 I ~I ~2 m "'ex I m I 0 I 10 _ __ _ J m VZ/////41 ~"$1 ~-thal 2 R w/41/A ~-thal 2 L l:\01 ,_. 0 ,_. 102 Figure 4 The distribution of sequence homologies within a region of the human a-like globin gene cluster. Regions of sequence homology are indicated by crosshatched boxes, white boxes or stippled arrows. These homologies were detected by heteroduplex analysis of the cloned a-globin gene cluster (Lauer et al 1980). Recent nucleotide sequencing data indicate that the cross-hatched boxes extend to the left as far as the Pvu II sites (N. Proudfoot, unpublished results). 103 a2 Pvu Hpo Soc a:s:a l I Smo I I Pvu Smo 11 Hind Pvu Hpo Soc ~Ill QI Bol I L 1ea:aµI_ _ __ 1; a,,, s Hind Pvu vvz;i Figure 4 104 Figure 5 Schematic representation of genotypes associated with a-thalassemia syndromes. The pair of horizontal lines for each syndrome represents the chromosome 16 homologues. A black rectangle indicates the presence of a functional a-globin gene. Absence of a black rectangle signifies either gene deletion or a nondeletion defect leading to a nonfunctional a-glob in gene. a-thal 1 or 2 signifies the a-thalassemia 1 or 2 syndromes. U1 (I) i -- -- - cx-thal 2 normal oc-thal I - -- cx-thal I hydrops feta Ii s - Hb H ~ 0 U1