Figure 4-10. Modular nature of the information transfer system illustrated with human Apoe. Single letter amino acid sequence codes (translation of mRNA) sit above the three letter codons which specify them. Below is the protein structure, showing the Apoe lipid binding domain only (for the "e4"allele). Based on GenBank accession number M10065.

Figure 4-10. Modular nature of the information transfer system illustrated with human Apoe. Single letter amino acid sequence codes (translation of mRNA) sit above the three letter codons which specify them. Below is the protein structure, showing the Apoe lipid binding domain only (for the "e4"allele). Based on GenBank accession number M10065.

simple beginning (de Duve 1991; Maynard Smith and Szathmary 1995; Trifonov 1999). It seems very likely that a more direct relationship between RNA and protein existed before the conversion of mRNA to an intermediary message bearer. The most likely scenarios that have been suggested have this evolving from primitive RNA-amino acid interactions, perhaps related to amino acid synthesis, and there are relationships between the similarity of codon sequences, the genealogical relationships among the tRNA sequences, the chemical properties, and biosynthetic relationships of their corresponding amino acids (Maynard Smith and Szathmary 1995; Ronneberg et al. 2000). In this sense, the coding system can be viewed in adap-tational terms as being selected to minimize the impact of mutational error on protein characteristics (Freeland and Hurst 1998; Freeland et al. 2000; Haig and Hurst 1991).

Today, the coding system is nearly universal. There are a few exceptions (for example, there are slight changes in the codons used by mitochondria, the organelles in cells that are used for energy metabolism), but its core elements seem to go back to a single beginning.

A Digression on Selection: Genetic Load

We have raised several points about the potential illusory aspects of adaptation and that selection is generally a probabilistic process (even if the common notion is more deterministic) whose characterization in nature can be elusive. We've suggested various reasons why darwinian selection by fitness competition may not need to be invoked as universally or strongly as is often done to account for evolution. But, as so often happens, there is a reverse problem as well. When we look at genetic variation across life, we may have a problem providing an explanation for too much evidence for selection. It is clear from the patterns of relative variation (shown in Figure 3-5) that functional regions of genomes vary much less within and between species than regions for which either there is no known function or strong evidence only for weak function.

Pseudogenes are the clearest case, and provide an approximate baseline for the rate of accumulation and amount of standing sequence variation to be expected when selection is not acting at all. A pseudogene is a gene, or duplicate copy of a gene, that no longer has the regulatory sequences needed for transcription; it is a sequence like a gene or a former gene that is clearly "dead" today and hence without function and whose variation is invisible to selection—a truly neutral sequence. By contrast, splice junctions, first and second positions in codons, regulatory sequences, and regulatory sequences are systematically less variable, and this constraint on variation is explained today as the result of selection. That is entirely plausible, indeed, the only good explanation until or unless some chemical error correction mechanism is found that applies specifically to such regions of DNA.

However, the order of magnitude of the amount of selection required to maintain this reduced variation is worth considering: for every 104 genes there are about 102 amino acids (and hence codons), in each and the variation in the first two nucleotides is constrained relative to the third, most redundant position—amounting to roughly 20,000,000 constrained sites per 10,000 genes, not counting enhancers, splice junctions, start signals, polyadenylation sites, TATA boxes, and so on. From today's estimates, the human genome has somewhere between 30,000 and 50,000 genes. This means something on the order of magnitude of 100,000,000 constrained variation sites. Yet, in the long term, each human has on average only one surviving child (two per couple) each generation.

Not only are codons less variable than noncoding sequences, presumably because of selection related to function, but there is sometimes a further codon bias related to the availability of the alternative tRNA genes, based on codon redundancy, that use the possible codons specifying each amino acid. The genes for alternative tRNA molecules for a given amino acid vary in their number of copies in the genome. Codon bias is a slight statistical excess of codons for the most abundant of the alternative tRNAs. Codon bias has been found in screens of many species, although at least one search in mammals found no statistically meaningful evidence for codon bias (Urrutia and Hurst 2001) and the bias is not found uniformly among genes.

The subtle and basically statistical conservation of so many parts of genomes raises the problem known as genetic load. Genetic load became a critical issue in biology when protein electrophoresis methods made it possible to characterize variation in detail. How is such variation maintained? There would seem to be simply too much variation to be sustained by the amount of overreproduction needed to screen out mutational variation in the millions of conserved sites by mechanisms of darwinian selection. It was this problem that led to the development of the neutral theory of evolution (sometimes for this reason called "nondarwinian evolution"), because selectively neutral variation places no reproductive burden on the organism. The neutral theory holds that most variation most of the time is not affected, or barely affected, by selection and proposes neutrality and genetic drift in finite populations as the baseline for evolutionary population genetics.

The genetic load problem has been largely ignored recently but was never very satisfactorily solved. Even if we accept that introns and so on are neutral, the remaining subset of the genome that we take as constrained in variation involves a huge potential amount of selective screening. If selection is illusory in some ways and very difficult to document in nature, there is also strong evidence for pervasive selection. Several explanations have been suggested to reconcile this seeming paradox.

Species with high rates of reproduction, in which each individual produces thousands or even millions of eggs or pollen grains, can sustain a lot of selective loss, since on average an individual in a species needs to only just replace itself so that only one of these eggs or pollen grains will be successful (populations are not typically growing in size). Think again of the number of acorns shed by a single oak tree over its lifetime, only one of which on average will become a new tree. Bacteria, too, can reproduce rapidly and have smaller genomes and may be able to sustain extensive but subtle selection. But the same kinds of constraints on variation across the genome are also seen in vertebrates, who are comparatively slow reproducers. This suggests but does not prove that there may be common mechanisms not wholly related to high or low reproductive potential.

One possible way to account for ubiquitous selection is that gametes even in slow reproducers are generated in profusion and may be screened even before fertilization (prezygotic selection), at least for genes related to cell division and normal metabolism, which gametes must do successfully.

Another possibility is that selection is not even approximately gradual the way Darwin suggested; the idea of selection against a variant because of a single trait is probably also too simple. Because of pleiotropy, selection on several different functions of a gene could constrain its variation so that there may be many ways to select against a given variant. Furthermore, because of epistasis, or the functional interaction among genes, any given instance of selection may simultaneously affect many genes. Individual variants might be harmless in themselves but become harmful in rare circumstances or in combinations with many other variants in the genome, so that a selective event removes many of them at a time. Theory for epistatic effects of this kind of process has been advanced (e.g., Eyre-Walker et al. 2002; Keightley and Hill 1983) but does not seem able to account adequately for the amount of constrained variation seen across genomes.

We assume that mutations occur more or less randomly with respect to function and that selection can only purge them after they occur and in comparison with contemporary alleles in the population. Unless understanding of evolution is missing something basic, there is no known way for mutations in specific regions of a gene to be prevented from happening. For that to be possible, parts of codons would have to be specifically protected from mutation; perhaps this could somehow be achieved by DNA packaging in areas to be transcribed, but it is not something generally known as yet.

No matter what, it seems clear that for most variation of this kind the average selection coefficient has to be very small, probably so small as to be immeasurable even in large samples at any given time. In practice, most of these variants are probably essentially neutral most of the time. How to test some of these notions about genetic load is unclear, and the issue this raises is whether or to what extent there is something else important and widespread going on in life besides selection that we may not yet have stumbled upon.

Part of the problem applies generally to evolution. It is difficult to understand things that happen so slowly that we can hardly perceive them. In a famous Amazonian travelogue, H. M. Tomlinson noted that a dense tropical rain forest seemed eerily silent and almost lifeless as one moved through the dark floor under the remotely high canopy. But if time were accelerated, we would have seen the movements in that "war of phantoms ... we should have seen ... the greater trees running upwards to starve the weak of light and food, and heard the continuous collapse of the failures, and have seen the lianas writhing and constricting, manifestly like serpents, throttling and eating their hosts" (Tomlinson 1912).

Turbo Metabolism

Turbo Metabolism

Forget Silly Diets-They Don't Work. Weight loss has got to be the most frustrating experience for many people, young and old alike. Eating foods that are just horrible, denying yourself foods you truly love and enjoy. Exercising, even though you absolutely hate exercising, and end up stiff as a board with no results.

Get My Free Ebook

Post a comment