Persistence or Rapid Generation of Dna of Humans Length Polymorphism at the Zeta-globin Locus

Extensive restriction mapping of 76 human genomic DNAs defines multiple sites of length and point mutation near the zeta-globin locus, which codes for an embryonic alpha-like globin chain. There are two major sites of DNA length variation: one in the intergenic region with three alleles and one in the first intron of the zeta I gene with at least four alleles. Our mapping establishes that the intronic polymorphism is associated with a tandem array of short, repeated sequences. T h e length alleles occur in each of four human populations sampled, suggesting an ancient origin with persistence of several length alleles, o r rapid regeneration of these particular variants. Four polymorphic restriction sites were also found; the frequency of polymorphic sites is comparable to that found in the human beta-globin gene region. Analysis of haplotypes indicates either that multiple recombinations have occurred near the 5' end of the zeta I gene o r that this region is prone to recurrent length mutation. HE use of recombinant DNA methods in genetic studies has led to the T discovery of a new class of polymorphisms that result from rearrangements of chromosomal DNA. The origin and behavior of DNA length variants can be studied with a combination of molecular and population genetic techniques. We have applied such an approach to the length polymorphisms associated with the zeta-globin genes, located on the short arm of human chromosome 16 (MCKUSICK 1982). The zeta genes are embryonic members in a cluster of five genes encoding alpha-like subunits of hemoglobin, shown in Figure 1 (LAUER, SHEN and MANIATIS 1980). At least one allele of each gene has been cloned and sequenced together with extensive segments of flanking DNA (PROUDFOOT, GIL and MANIATIS 1982; PROUDFOOT and MANIATIS 1980; LIEBHABER, GOOSSENS and KAN 1980; MICHELSON and ORKIN 1980; SAWADA et al. 1983; GOODBOURN et al. 1983). The two zeta genes are located approximately 10 kb apart at the 5' end of the alpha gene cluster. Their coding regions differ by three amino acid substitutions, one of which terminates translation of the zeta 1 mRNA at codon ' Present address: Chiron Corporation, 4560 Horton Street, Emeryville, California 94608. Genetics 114: 79-92 January, 1986. 80 B. S. CHAPMAN, K . A. VINCENT AND A. C. WILSON 3 ’ Zeta2 Zero1 +a a2 ai = Y/A N I I f 5’ -t I I I I 0 5 IO 15 20 25 30 35 Kb FIGURE ].-The zeta-globin genes shown in relation to the three other genes in the human alpha-globin gene family (5 ’ to 3’, left to right). Two genes, shown as hatched boxes, encode nontranslatable niKNAs and have been designated pseudogenes (LAUER, SHEN and MANIATIS 1980; PROUDFOOI‘, GI[. and MANIATIS 1982). To name the zeta genes we have adopted the convention of LAUER, SHEN arid MANIATIS ( 1 980). 6. Since there are no silent substitutions in the coding DNA, and only three substitutions in the unique sequence portions of the introns, inactivation of the zeta 1 gene would appear to have occurred about 1 million years ago (PROUDFOOT, GIL and MANIATIS 1982; WILLARD et al. 1986). Although the zeta I gene cannot produce a globin polypeptide, it remains transcriptionally active ( PROUDFOOT, RUTHERFORD and PARTINCTON 1 984). There is extensive DNA polymorphism near the embryonic zeta genes of the cluster (CHAPMAN, VINCENT and WILSON 1981; HIGCS et al. 1981; CHAPMAN and WILSON 1982, 1983; GOODBOURN et al. 1983). These polymorphisms in restriction fragments appear to be generated by variation in the length of DNA sequences, as well as by mutation in restriction sites. Length variations detectable by genomic Southern blotting have been observed near other known loci, including the human insulin gene (BELL, KARAM and RUTTER 1981) and the c-Ha-ras-I gene (GOLDFARB et al. 1982). Nucleotide sequences of polymorphic regions in the insulin (BELL, SELBY and RUTTER 1982), c-Ha-ras-I (CAPON et al. 1983) and zeta-globin (GOODBOURN et al. 1983) loci have revealed tandem arrays of related oligonucleotides similar to the “minisatellite” sequences identified by JEFFREYS, WILSON and THEIN (1 985). The zeta-globin locus offers an opportunity to examine a highly polymorphic region using both molecular and population genetic approaches. Our restriction mapping of genomic DNA defines two regions of length polymorphism and four polymorphic restriction sites in the vicinity of the zeta-globin genes. We have mapped variations in the length of the zeta 2 gene to the tandem array in intron I . To collect data on incidence in human populations, a panel of 76 human DNAs was screened by comparative Southern blotting. The stability of the length variation was examined in two small pedigrees and in cultured leukemic cells. In addition, haplotypes were constructed and used to search for recombination within the region. An unexpected outcome of this analysis is that DNA length alleles of the zeta-globin region may be ancient and stable, a possibility not previously considered. We discuss our results and those of others with respect to two models: one in which the length alleles are repeatedly generated by unequal recombination, and the other in which ancient alleles persist. MATERIALS A N D METHODS D N A was prepared from blood or placenta of 75 human individuals and from an erythroleukemic cell line (K562) that can be induced to synthesize zeta globin. This DNA POLYMORPHISM IN THE ZETA LOCUS 81 t rp p s a taf aff sevf p gf p v t s t t c e f t p b b t