While early studies in molecular evolution clamied the existance of a universal
molecular clock, more recent data have shown that the rates of
evolution of
DNA sequences in not constant in all lineages. A variety of mechanisms have been suggested to account for this obersrvation, including
DNA repair efficiency,
generation time, metabolic rates, functional
constraints on proteins, and the size of the population. Britten’s (1986) proposal that efficiency of DNA repair mechanisms may vary between lineages is backed by evidence from cell cultures, but has never been demonstrated in outside these strict lab conditions (Li 1997, p.228). The other mechanisms have received more attention, and considerable data have accumulated in the past five years.
The Generation Time Hypothesis
The generation time hypothesis (Kohne 1970) is the oldest hypothesized mechanism, and has been invoked to explain all of the instances of rate heterogeneity I have detailed above. This theory postulates that organisms with shorter generation times will have higher rates of evolution relative to organisms with longer generation times. This hypothesis assumes that the majority of mutations occur through errors in DNA replication. Organisms with shorter generation times will have more rounds of germ-cell divisions per unit time, and hence more mutations than organisms with longer generation times.
Although the studies above have data which is consistent with the generation time hypothesis, none explicitly tested the assumptions of the model. Li (1997, p. 229) tested this assumption by comparing the ratio of mutation rates between sexes to the ratio of germ-cell divisions between sexes. He found that for mice, rats and higher primates, the two ratios were in rough agreement, suggesting that there is a strong correlation between the number of germ-cell divisions and the number of mutations. He interprets these data as evidence that other mutation factors are considerably less important for explaining the rates of mutation.
The Metabolic-Rate Hypothesis
There is considerable evidence, however, that metabolic rates have a major impact on rates of evolution. Martin, Naylor and Palumbi (1992) were the first to show this effect, studying the rate of mitochondrial DNA (mtDNA) evolution in sharks. The authors examined the cytochrome b and cytochrome oxidase I genes (both mitochondrial) in 13 species of sharks which are well represented in the fossil record. They examined transversions (changes of a purine to a pyrimidine or vice versa) at four fold degenerate sites (sites where the code is unaffected by the base present). Transversions at these sites were used because changes accumulate linearly with time. (Transitions do not, as they become saturated quickly (Irwin, Kocher and Wilson 1991).) When they plotted the number of transversions (corrected for multiple hits) against the divergence times as measured from the fossil record for sharks, ungulates and primates, they found that sharks have accumulated mutations seven times a slowly as ungulates or primates.
Sharks and primates have similar generation times, so the generation time hypothesis fails to explain this case. However, sharks have metabolic rates 5-10 times lower than mammals, and this seems a likely explanation of this difference. A normal product of metabolism are oxygen radicals, which are potent mutagens (Shigenaga, Gimeno and Ames 1989). The metabolic-rate hypothesis postulates that the difference in evolutionary rates is due to the increased presence of oxygen radicals in primates and ungulates due to their increased metabolic rates.
Examining a wide variety of taxa, Martin and Palumbi have (1993) shown that body size is tightly correlated with mtDNA sequence divergence. Many physiological and life history variables are correlated with body size, including both generation time and metabolic rate. They used a multiple regression in an attempt to tease apart the various factors responsible for the rate of divergence, and found that metabolic rate, but not generation time, had a significant effect on silent site substitution rate. Rates of mtDNA divergence separate into two broad classes, an observation which cannot be explained simply by the generation time hypothesis. Rather, the two groups shown in the plot are divided between poikilotherms and homeotherms, of which poikilotherms have lower metabolic rates as well as lower rates divergence.
In the same paper, Martin and Palumbi propose that rather than examining simply metabolic rate or generation time, a more effective measure of sequence evolution rates is nucleotide generation rate. A nucleotide generation is defined as the length of time in which it takes a specific nucleotide position to be copied, due either to DNA replication or repair. Nucleotide generations are likely to be short in species with fast generation times, as germ-line cells (and their nucleotides) will be copied rapidly. Species with high metabolic rates (and hence high production of oxygen radicals) will also have short nucleotide generation times as repair mechanisms will be forced to function more often in such an organism.
It may be tempting to hypothesize that generation time effects will predominate on nuclear genes, while metabolic rate effects will predominate on mitochondrial genes. The picture is not that simple, however. Oxygen radicals can cross the mitochondrial membrane, and oxidative damage to nuclear DNA has been shown to be extensive (Richter, Park and Ames 1988). Further, little is known about mitochondrial generations within germ-line cells. Mitochondria are continually degraded and replaced at a rate independent of (and much higher than) the germ-cell division rate. This likely serves to decouple the rates of mtDNA replication (and hence mtDNA nucleotide generation time) from cellular and organismal generation times (Rand 1994). Li (1993) has suggested that "It is possible that for relatively closely related species, generation time is the major factor for rate variation, whereas for divergent taxonomic groups, difference in metabolic rate may be important." The exact meaning of "closely related" is subject to individual interpretation, but Nunn and Stanley (1998) found that metabolic rate was the major source of rate heterogeneity among the order Procellariiformes (tube-nosed seabirds.) While further evidence is lacking, Li’s prediction seems too simple.
The Functional Constraint Hypothesis
The correlation between body temperature and the rate of evolution has been explained in terms of the function of proteins as well as by metabolic rates. Thomas and Bechenbach (1989) found that, much like the data on sharks by Martin, Naylor and Palumbi (1992), the cold-blooded salmonoid fish have a slower rate of evolution than mammals. They hypothesized that because the proteins of cold-blooded vertebrates must function at a wide variety of temperatures, there are fewer alternate states for which the protein is adapted. For such a protein, more amino acid substitutions will violate these stringent functional constraints, making these mutations deleterious. By contrast, the relatively constant internal environment of homeotherms may allow for a relaxation of some of these functional constraints, and an increased number of amino acid substitutions.
Adachi, Cao and Hasegawa (1993) tested this hypothesis by comparing the ratio of the amino acid substitution rate (KA) to silent substitution rate (KS) in both salmonoids and primates. Because silent substitutions are nearly neutral, KS is very close to the actual mutation rate. As a result, KA:KS is a close approximation of the number of amino acid substitutions which were not eliminated from the population, and are assumed to be selectively neutral. The ratio of KA:KS is much higher in primates than salmonoids, indicating that the faster rate of amino acid substitution in mammals is due (at least in part) to relaxed functional constraints.
This analysis is problematic however. Third codon positions are likely to have received multiple mutation events over long time periods, yet they cannot be eliminated from these analyses because these sites constitute nearly all of the possible silent mutations. A study free of these problems has been provided by Mindell et al. (1996), who studied rates evolution in birds. Non-passerine birds have metabolic rates similar to those of mammals, so the metabolic rate hypothesis predicts that the two groups should not have different evolution rates. They analyzed 13 mitochondrial protein coding genes, and used a relative rate test to show that all 13 show slower rates of change in Gallus gallus than four mammals (Bos taurus, Homo sapiens, Mus musculus, and Rattus norvegicus.) There was no significant difference in the rates of 16S or 12S mitochondrial rDNA between eight bird species and the four mammals.
The finding that all of the protein coding genes have different evolutionary rates between birds and mammals contradicts the metabolic-rate hypothesis, as nonpasserine birds and mammals have roughly similar metabolic rates. Birds have higher body temperatures, however. Increases in temperatures can have many adverse effects on both protein structure and reactions enzymes participate in (Somero 1978). At higher body temperatures in birds, proteins may be more constrained in the possible number of alternate forms. A greater proportion of mutations in these proteins may be deleterious, leading to a slower rate of evolution in birds. They argue that the data on rDNA is also consistent with the functional constraint hypothesis, claiming that higher temperature may affect the weak bonds involved in the tertiary and quaternary structure of proteins, but may not adversely affect rRNA. Unfortunately, they provide no evidence for this assertion.
To date, all studies of functional constraint have been performed on mtDNA genes. While it is tempting to extend these studies to nuclear genes, the validity of this extension is questionable (Adachi, Cao and Hasegawa 1993). For an organism operating at a wide variety of temperatures (such as the salmoniod species), the most advantageous genotype may be the heterozygote, with each copy functioning optimally at different temperatures. Such heterozygote advantage could not hold for the mitochondrial genome, as it is haploid. Adachi, Cao and Hasegawa (1993) propose that a test of this objection would be to compare heterozygosity of the nuclear genome between warm-blooded and cold-blooded vertebrates.
The Effect of Population Size
The nearly neutral theory (Ohta 1976) predicts that molecular evolution will occur at a faster rate in smaller populations. Under this model, as Ne falls, the range of mutations which are effectively neutral increases, leading to an increase in the rate of evolution. To examine this prediction, DeSalle and Templeton (1988) examined rates of evolution of two lineages of Hawaiian Drosophila. One lineage is known to have small population sizes and have experienced repeated genetic bottlenecks due to founder events, while the other lineage has large population sizes and is has likely to have arisen by subdivision of large populations. They found that each lineage has a relatively even level of mtDNA evolution within the lineage. However, the lineage with small population sizes and frequent bottlenecks was found to have a rate of evolution three times that of the lineage with large populations and no bottlenecks.
Ohta (1995) has tested the predictions of the nearly neutral theory by examining the rates of evolution in primates, artirodactyls, and rodents. She could not simply contrast the rates of evolution of the three groups, as rodents have significantly shorter generation times and would be expected to show a faster rate of evolution simply due to that fact. She compared the ratio of rates of nonsynonymous mutations to synonymous mutations between the three lineages. For large populations, a greater proportion of nonsynonymous mutations will be non-neutral than for small populations. The ratio represents the proportion of total mutations which are nearly neutral amino acid substitutions (and hence fixed in the genome.) This ratio should be lowest in large populations. Ohta found precisely this: rodents, which have a much larger population size than either primates or artirodactyls had a correspondingly lower ratio than either group.
LITERATURE CITED
ADACHI, J., Y. CAO and M. HASEGAWA. 1993. Tempo and more of mitochondrial DNA evolution in vertebrates at the amino acid level: rapid evolution in warm-blooded vertebrates. J. Mol. Evol. 36:270-281.
DESALLE, R. and A. R. TEMPLETON. 1988. Founder effects and the rate of mitochondrial DNA evolution in Hawaiian Drosophila. Evolution 42:1076-1084.
IRWIN, D. L., T. D. KOCHER and A. C. WILSON. 1991. Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 33:128-144.
KOHNE, D. E. 1970. Evolution of higher-organism DNA. Q. Rev. Biophys. 33:327-375.
LI, W.-H. 1993. So, what about the molecular clock hypothesis? Curr. Opin. Gen. Dev. 3:896-901.
LI, W.-H. 1997. Molecular Evolution. Sinauer Associates, Inc., Sunderland, MA.
MARTIN, A. P., G. J. P. NAYLOR, and S. R. PALUMBI. 1992. Rates of mitochondrial DNA evolution in sharks are slow compared with mammals. Nature. 357:153-155.
MARTIN, A. P. and S. R. PALUMBI. 1993. Body size, metabolic rate, generation time, and the molecular clock. Proc. Natl. Acad. Sci. USA 90:4087-4091.
MINDELL, D. P., A. KNIGHT., C. BAER, and C. J. HUDDELSTON. 1996. Slow rates of molecular evolution in birds in the metabolic rate and body temperature hypotheses. Mol. Biol. Evol. 13:422-426.
NUNN, G. B. and S. E. STANLEY. 1998. Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Mol. Biol. Evol. 15:1360-1371.
OHTA, T. 1976. Role of slightly deleterious mutations in molecular evolution and polymorphisms. Theoret. Popul. Biol. 10:254-275.
OHTA, T. 1995. Synonymous and nonsynonymous substitutions in mammalian genes and the nearly neutral theory. J. Mol. Evol. 40:56-63.
RAND, D. M. 1994. Thermal habit, metabolic rate and the evolution of mitochondrial DNA. Trends Ecol. Evol. 13:125-131.
RICHTER, C., J.-W. PARK, and B. N. AMES. 1988. Normal oxidative damage to mitochondrial and nuclear DNA is extensive. Proc. Natl. Acad. Sci. USA 85:6465-6467.
SARICH, V. M. and A. C. WILSON. 1973. Generation time and genomic evolution in primates. Science 179:1200-1203.
SHIGENAGA, M. K., C. J. GIMENO, and B. N. AMES. 1989. Urinary 8 hydroxy-2’-deoxyguanosine as a biological marker of in-vivo oxydative DNA damage. Proc. Natl. Acad. Sci. USA 86:9697-9701.
SOMERO, G. N. 1978. Temperature adaptation of enzymes. Annu. Rev. Ecol. Syst. 9:1-29.
THOMAS W. K. and A. T. BECKENBACH. 1989. Variation in salmonoid mitochondrial DNA: evolutionary constraints and mechanisms of substitution. J. Mol. Evol. 29:233-245.