Use of long sequence alignments to study the evolution and regulation of mammalian globin gene clusters

The determination of long segments of DNA sequences encompassing the beta-and alpha-globin gene clusters has provided an unprecedented data base foranalysis of genome evolution and regulation of gene clusters. A newlydeveloped computer tool kit generates local alignments between such longsequences in a space-efficient manner, helps the user analyze thealignments effectively, and finds consistently aligning blocks of sequencesin multiple pairwise comparisons. Such sequence analyses among thebeta-like globin gene clusters of human, galago, rabbit, and mouse haverevealed the general patterns of evolution of this gene cluster. Alignmentsin the flanking regions are very useful in assigning orthologousrelationships. Investigation of such matches between the mouse and humanbeta-like globin gene clusters has led to a reassessment of someorthologous assignments in mouse and to a revision of the proposed pathwayfor evolution of this gene cluster. In general, the interspersed repetitiveelements have inserted independently, presumably via a retrotranspositionmechanism, in the different mammalian lineages. However, some examples ofancient L1 repeats are found, including one between the epsilon- andgamma-globin genes that appears to have been in the ancestral eutheriangene cluster. Prominent matching sequences are found in a long region 5' tothe epsilon-globin gene, the locus control region (LCR) that is a positiveregulator of the entire gene cluster. Three-way alignments among the human,goat, and rabbit sequences can extend for > or = 3 kb in part of the LCR(DNase hypersensitive site 3), indicating that the cis-acting components ofthis complex regulatory region cover a long segment of DNA. In contrast tothe beta-like globin gene clusters, the alpha-like globin gene clusters ofmany mammals occur in very G+C-rich isochores and contain prominent CpGislands. The regions between the alpha-like globin genes are evolvingfaster than the intergenic regions of the beta-like globin gene clusters.The contrasts between the two gene clusters can be attributed todifferences in DNA metabolism in the isochore. The proximal controlelements of the rabbit alpha-globin gene are located both 5' to and withinthe gene. All of this region is part of a prominent CpG island that may beacting as an extended, enhancer- independent promoter. One can hypothesizethat the analogue to the LCR in the alpha-globin gene cluster may interfacewith the distinctive alpha-globin promoter in ways different from theinteraction between the beta LCR and the promoters of beta-like globingenes.(ABSTRACT TRUNCATED AT 400 WORDS)