Aa A



Figure 8.5

Evolution of Bcd binding sites in the hunchback genes of higher Diptera

The distribution of sites with TAAT core sequences (red hexagons) and different core sequences (blue ovals) in the upstream 1 kb of the hunchback genes of five species are shown. The number and topology of sites has diverged among species, and the Bcd protein has co-evolved with these regulatory elements.

Source: McGregor AP, Shaw PJ, Hancock JM, et al. Rapid restructuring of bicoid-dependent hunchback promoters within and between Dipteran species: implications for molecular coevolution. Evol Dev2001; 3: 397-407. Reproduced with permission of Blackwell Publishing Ltd.

from the same species. For example, the Bcd protein from D. melanogaster binds with a higher affinity to the D. melanogaster hb sequences than to the Musca hb sequences, and the Musca Bcd protein binds to the Musca hb with higher affinity than to the D. melanogaster hg element. Consistent with these observations, the Drosophila Bcd protein rescued embryos from bcd mothers more efficiently than did Musca Bcd. These data are consistent with compensatory coevolution occurring between the Bcd transcription factor and the hb P2 promoter.

Functional modification of cis-regulatory DNA sequences

Many of the comparative studies presented in Chapters 5, 6, and 7 inferred a change in cis-regulatory function but, with the exception of Hoxc8 element evolution in vertebrates (see Fig. 5.7), the specific functional changes within elements have not been identified. Nevertheless, some well-analyzed cases of the evolution of tissue-specific cis-regulatory elements do illustrate the likely mechanisms underlying the functional modification of such elements of developmental regulatory genes.

Co-option or modification of cis-regulatory elements is the most likely explanation for the evolution of lens crystallins (a diverse group of water-soluble proteins that have been recruited to function in eye lenses in the refraction of light onto the retina). Various enzymes or proteins related to enzymes such as lactate dehydrogenase B, argininosuccinate lyase, a-enolase, glutathione-S-transferase, and small heat-shock proteins are expressed at very high levels in the lens in particular species. The recruitment of these proteins as crystallins has occurred independently in various lineages, suggesting that many gene co-option events have occurred during the evolution of lens crystallins.

Analyses of cis-regulatory elements of a wide variety of crystallin genes have revealed that certain transcription factors are frequently involved in high levels of protein expression in the lens. They include the Pax-6 and retinoic acid receptor proteins, which play major roles in eye development in most animal phyla. The implication of these proteins in crystallin gene regulation suggests a scenario for the molecular genetic basis of lens crystallin recruitment. Namely, the evolution of sites for Pax-6, retinoic acid receptors, and other transcription factors within extant cis-regulatory elements of these genes may lead to abundant levels of expression in developing eye tissue. In general, high levels of gene expression in a given cell type or tissue could reflect the evolution of sites for abundant transcription factors within extant cis-elements.

Tissue-specific losses of gene expression also occur. For example, blind mole rats have degenerated, nonfunctional eyes. In the superspecies Spalax ehrenbergi, the small heat shock protein/a-crystallin gene enhancer has selectively lost activity in the developing lens, while its activity in the heart and muscle have been conserved. This appears to be due to relatively few changes in the mole rat gene relative to other rodents.

A second, well-analyzed case involves the glucose dehydrogenase (gld) enzyme, which is expressed in particular reproductive tissues of many Drosophila species. In D. teissieiri, for example, gld expression does not occur in a subset of these tissues. Comparisons of the homologous cis-regulatory regions of the gld gene of seven different Drosophila species have revealed extensive sequence conservation between all seven genes, including the perfect conservation of certain motifs between all seven species. On the other hand, the D. teissieiri gld element lacks all copies of a sequence motif that occurs three times in a D. melanogaster gld regulatory element. This motif directs gene expression in D. melanogaster in the tissues from which gld expression in D. teissieiri is absent. Thus evolutionary changes in the D. teissieiri gld element correlate with tissue-specific loss of gld expression in this lineage.

The gain or loss of gene expression can, in some cases, be as specific as involving a single cell. In the two nematode worms C. elegans and C. briggsae, the expression of the lin-48 gene is shared in hindgut cells and neuronal support cells but is present in the excretory duct cell only in C. elegans. Examination of the respective lin48 cis-regulatory elements of both species suggests that this difference is due in part to the presence of binding sites for the CES-2 transcription factor in the lin48 element in C. elegans that were either gained in the C. elegans lineage or lost in the C. briggsae lineage.

Variation in regulatory DNA and gene expression

The evolutionary divergence in gene expression between species associated with morphological diversity, the turnover of binding sites in cis-regulatory DNA, and the association of some quantitative trait variation with noncoding DNA sequences all point towards the importance of variation in regulatory DNA sequences as a major component of morphological evolution. How widespread is functional variation in regulatory sequences in natural populations?

Variation in gene expression has also been examined at the level of individuals, populations, and closely related species. For example, in the fish Fundulus heteroclitus, 18% of loci surveyed differed in expression between individuals within a population. Among different strains of D. melanogaster, a range of 2-6% of genes surveyed exhibited differences in expression levels at a single developmental stage. Pairwise comparisons of D. melanogaster and two closely related species, D. simulans and D. yakuba, suggest that 10% of genes exhibit expression differences at a single stage, and that 27% of all genes differ in expression levels between at least two strains or species. Perhaps even more compeling, comparisons of gene expression profiles from the brains of humans, chimpanzees, orangutans, and macaques suggest that a substantial fraction of genes, perhaps a quarter, are expressed at different levels in different primates.

We are only beginning to directly survey nucleotide variation and its attendant effects on gene expression. Allelic variation in gene expression has been best studied thus far in mice and humans. In general, polymorphisms with affects on the order of 1.5-4-fold on the level of transcription have been found for as many as half of all genes surveyed. These findings suggest that regulatory variation is widespread within natural populations. Furthermore, small-scale nucleotide variation is sufficient to impart potentially functionally significant differences in levels of gene expression. Indeed, it may be that humans, for example, are heterozygous at more functional cis-regulatory sites than amino acid positions, underscoring the substantial potential of regulatory variation to underlie phenotypic variation.

The emerging picture, then, is that regulatory variation is abundant at the level of individuals, populations, and species. The future challenge will be to link this regulatory variation to discrete phenotypic characters.

0 0

Post a comment