Chironomus Part 2 (Molecular Biology)

4. Evolution of Chironomus Globins

Phylogenetic analyses support an early origin of monomeric hemoglobins about 500 million years ago and the appearance of homodimeric hemoglobins about 250 to 300 million years ago, corresponding roughly with the origin of Diptera. Recently, a search of the Drosophila genome database using a chironomid globin gene query found a distant relative, designated a neuroglobin because it was expressed in brain tissue in D. melanogaster; and mouse and human neuroglobins have since been characterized (10-13). The chironomid globin gene family has experienced many of the same evolutionary events as their multigene families, such as point-mutational inactivation, translocations, gene loss, gene conversion, etc. They also support two somewhat unorthodox hypotheses. One that has received much attention in recent years, is that eukaryotic genes can acquire introns, complete with correctly placed intron/exon junctions and splicing signals. The other hypothesis is that much of the diversity of hemoglobin structure in Chironomus has arisen by neutral evolution rather than by functional adaptations of the proteins. There is resistance to ascribe to neutral evolution those mutations causing even minor differences in amino acid sequence between the products of homologous genes in different species or paralogous genes in the same organism. Such differences between globin genes in Chironomus, and between the products of multigene family members in general, are widely thought to be the result of adaptive selection. However, evidence from the different chironomid globins and globin genes suggest that many were not selected for their adaptive value, but exist instead as a neutral consequence of positive selection of the expression of high levels of hemoglobin from multiple genes.

4.1. Origin of A Central Intron

By the time the first chironomid globin genes were cloned in 1984, the sequencing of vertebrate globin genes had become a cottage industry. Each new report presented a globin gene with the same mosaic structure: three exons and two introns, the latter always interrupting the coding DNA at the same location. In a prophetic para (14), Go analyzed X-ray diffraction data for a vertebrate globin and concluded that in addition to the outer introns, the ancestral globin gene had a third intron in the center of the gene such that the entire gene was separated into four exons of similar length, each coding for a separate structural domain. Go’s prediction was fulfilled shortly thereafter with the report of the first plant (legume) globin gene, which had a centrally located intron in addition to two introns at the same position as those in the vertebrate globin genes. Apparently the structure of plant globin genes represents the ancestral state. Many of the first chironomid globin genes to be reported were intronless, a condition presumably due to the reverse transcription of a completely processed transcript and reinsertion of the complementary DNA into the genome. The discovery of the C. thummi ct-2b and ct-9.1 genes (15), and of several other invertebrate globin genes with a central intron, however, raised questions about the origin of chironomid globin gene introns. Unlike the conservation of location of the outer vertebrate globin gene introns, the location of central introns is not conserved across kingdoms or phyla, and none is at the position predicted by Go. Could they, like the outer introns in globin genes, still be descendants of an ancestral intron? Gilbert suggested that introns at nonconserved locations have moved by "intron sliding." Others have argued that the discordant location of nitrons is more easily explained if introns can be acquired by previously uninterrupted genes. Hankeln et al. (8) reported that the globin genes of some species have a "central" intron at a slightly different location that those of other species. A chironomid mitochondrial gene phylogeny groups flies with an intron at one location ("type 1") on the "pseudothummi" branch, and groups those with an intron at the other location ("type 2") on a separate "plumosus" evolutionary branch (8, 16, 17). The phylogeny supports the separate acquisition of an intron by the homologous genes after divergence of the two lineages (Fig. 1). Several plausible mechanisms for intron acquisition have been suggested. In considering these, Kao et al. (15) offered a mechanism whereby a chironomid globin gene containing a centrally located HAGGTH tetramer (to provide donor and acceptor splice sites) could have acquired a central intron by a simple gene duplication, accompanied (or followed) by deletion of all but the DNA presently found in the intron. A mechanism proposed by Hankeln et al. (8), where introns are acquired by a reversal of the splicing reaction, might explain intron insertion in globin genes lacking these AGGT splice site precursors. An intron once acquired by a single gene might spread to nearby related genes by gene conversion (8); see Fig. 2).

Figure 2. Evolutionary history of orthologous gene clusters in C. thummi and its sibling subspecies, C. piger. ct-, and ctp save space. Arrows above the genes indicate direction of transcription. Ancestral genes are designated in italics. Shaded recent duplications, losses, and fusions.

4.2. Evolution of a Globin Multigene Family

The DNA sequences of orthologous globin gene clusters in different chironomid species support the adaptive selection of a high globin gene copy number and the essentially neutral evolution of individual globin genes. The ctp-Y, ctp-W, ctp-V, and ctp-Z genes, and five ctp-7B genes in C. piger (18), and the ctp-Y, ctp-W, ctp-V, and ctp-Z genes, and seven ctp-7B genes of sibling species C. thummi (1) are a case in point. The phylogenetic study summarized in Figure 2 traces the evolutionary history of the two clusters, showing the seven genes that are clearly homologous, and a series of gene duplications and deletions necessary to explain the current sequence and organization of the clusters. Pairwise comparisons show that at the nucleotide level, homologous globin genes or logical gene pairs (ie, genes of indeterminate orthology, but sharing immediate common ancestry) in the two species share from 96.7% to 99.2% identity. In three cases, homologous genes in each species code for identical polypeptides, despite several millions of years of species divergence. Within C. thummi, the ct-7B5, ct-7B6, ct-7B9, and ct-7B10 paralogous genes are also very similar, as shown by the amino acid differences in the pairwise comparisons shown below:

In addition to identical paralogues (ct-7B5, ct-7B9), sequencing of an independent genomic clone revealed a gene identical ct-7B5 (19). Designated ct-7B9/5 because its 5′ and 3′ flanking DNA are similar to the 5′ and 3′ regions of ct-7B9 and ct-7B5, respectively, it probably arose by intrachromosomal crossing over between oppositely oriented ct-7B9 and ct-7B5 in a looped (bent) chromosome structure (with deletion of most or all of ct-7b10). Since the recombination point is in fact external to the transcribed region of the gene, 7B9/5 must simply be another ct-7B5 gene or allele. The recombined region (without ct-7B10) may be an independent gene locus or an allelic haplotype. Clearly the 7B globin gene subfamily predates the thummi/piger divergence. The ability of C. thummi and C. piger to lose different descendants of an ancestral cluster by gene fusion, and the overall similarity of paralogous genes, implies that any small differences that accumulate in the genes before or after fusions and corrections are evolutionarily inconsequential. The 7B gene cluster seems to have undergone periodic expansion and contraction (gene duplication, deletion), while maintaining the ability to encode identical or nearly identical globins for millions of years. The best explanation for this phenomenon is that the few differences that accumulate between 7B genes are the result of random drift (neutral evolution), but that the cluster itself has experienced positive selection of a high number of copies of globin genes as a means to ensure the synthesis of large amounts of hemoglobin.

Whereas the orthologous globin gene clusters in C. thummi and C. piger illustrate early evolutionary events in the life of a multigene family, globin gene clusters in C. thummi and C. tentans are separated by 60 million years of evolution (16, 20). The extent paralogous and homologous genes sequences are much less similar to each other (except for a region of the ct-2b and ct-9A genes that underwent partial gene correction, homogenizing intronic and exonic DNA in the middle of the two genes). The divergence of the C. thummi and C. tentans gene clusters offers a wider window through which to test the concept of gene copy number selection. Figure 1 summarizes evolution of these genes after acquisition and spread of a "type 1" intron in one of the clusters. The ancestral C. thummi gene cluster must have had five intron-bearing genes before a gene duplication created ct-11 and ct-12. ct-12 has since lost its intron, and both ct-11 and ct-12 have been diverging for some time (16). In contrast, the C. tentans cluster contains only intronless homologues, originally ascribed to multiple intron losses from intron-bearing ancestors (16, 20). Figure 1 reflects more recent phylogenetic evidence that some or all of the genes in these clusters arose early in the genus, and that the C. tentans genes never acquired an intron in the first place (9). A difference in the organization of the orthologous genes between C. thummi and C. tentans is explained if a translocation is assumed between ctn-A and ctn-B (or vice-versa in C. thummi). Of special note, the recent inactivation of a different (non homologous) gene in each cluster suggests that not all globin genes in a cluster are indispensable, even after 60 million years of evolution. In C. thummi. ct-13RT suffered the insertion of a retroposon into coding DNA in exon I. This occurred very recently, because the rest of the coding region, the intron, and the flanking DNA containing promoter and polyadenylation motifs are all still intact. Recent retroposition is supported by the observation that ct-13RT is actually an allele coexisting with a normal, uninterrupted ct-13 allele (17). In C. tentans, the ctn-ORFB gene was truncated at both the 5′ and 3′ ends. Again, this event must have occurred very recently, because the gene retains all other attributes of a viable, transcribable gene. The phylogenetic analyses of the orthologous C. thummi and C. tentans gene clusters supports the recent inactivation of the ct-13RT allele and the ctn-ORFB genes, and the likelihood that accumulated amino acid differences between many (if not all) homologous and paralogous genes are the result of neutral evolution. If the greater diversity of the tentans and thummi genes cannot be completely explained as the positive adaptation of structurally diverse hemoglobins, then their maintenance, like that of the large number of 7B genes in thummi and piger, must be the result of positive gene copy number selection favoring high levels of hemoglobin synthesis.

In sum, the evolution of Chironomus globin genes spans more than 250 million years, in which time individual sequences within and across species have accumulated many amino acid substitutions. Some especially diverged hemoglobins may have evolved to serve unique functions, for example in environments that undergo cyclic changes in oxygenation. The specific expression of some globins during development might reflect habitant (and therefore, oxygenation) differences as larvae first descend from the waters’ surface to the benthos, and later ascend to the surface at the time of eclosion. On the other hand, differences between many of the globin genes may be neutral, selection favoring the proliferation of a large family of genes encoding proteins of physiologically similar function. Kimura (21) suggested that duplicated genes accumulating neutral change are preadapted, later becoming substrates for positive adaptation after speciation. A consequence of high gene copy number selection is that some of the many functionally redundant globin genes can serve as raw materials for Darwinian selection, which could explain the evolution and spread of more than 5000 species of Chironomous to diverse habitats.