Unbiased estimation of the rates of synonymous and nonsynonymous substitution

Summary The current convention in estimating the number of substitutions per synonymous site (K _S ) and per nonsynonymous site (K _A ) between two protein-coding genes is to count each twofold degenerate site as one-third synonymous and two-thirds nonsynonymous because one of the three possible changes at such a site is synonymous and the other two are nonsynonymous. This counting rule can considerably overestimate theK _S value because transitional mutations tend to occur more often than transversional mutations and because most transitional mutations at twofold degenerate sites are synonymous. A new method that gives unbiased estimates is proposed. An application of the new and the old method to 14 pairs of mouse and rat genes shows that the new method gives aK _S value very close to the number of substitutions per fourfold degenerate site whereas the old method gives a value 30% higher. Both methods give aK _A value close to the number of substitutions per nondegenerate site.     

Low Nucleotide Diversity in Man

The nucleotide diversity (pi) in humans is studied by using published cDNA and genomic sequences that have been carefully checked for sequencing accuracy. This measure of genetic variability is defined as the number of nucleotide differences per site between two randomly chosen sequences from a population. A total of more than 75,000 base pairs from 49 loci are compared. The DNA regions studied are the 5' and 3' untranslated regions and the amino acid coding regions. The coding regions are divided into nondegenerate sites (i.e., sites at which all possible changes are nonsynonymous), twofold degenerate sites (i.e., sites at each of which one of the three possible changes is synonymous) and fourfold degenerate sites (i.e., sites at which all three possible changes are synonymous). The pi values estimated are, respectively, 0.03 and 0.04% for the 5' and 3' UT regions, and 0.03, 0.06 and 0.11% for nondegenerate, twofold degenerate and fourfold degenerate sites. Since the highest pi value is only 0.11%, which is about one order of magnitude lower than those in Drosophila populations, the nucleotide diversity in humans is very low. The low diversity is probably due to a relatively small long-term effective population size rather than any severe bottleneck during human evolution.     

Evolution of glucagon genes

Statistical analyses of DNA sequences of the preproglucagon genes from bovine, human, hamster, and anglerfish suggest that a gene duplication creating two anglerfish genes (AF I and II) occurred about 160 Myr ago, long after the separation of fish and mammals. The analyses further suggest that the internal duplication producing the glucagon and glucagon-like peptide II (GLP-II) regions occurred about 1.2 billion years ago, which would indicate that the GLP-II region was present in the ancestral anglerfish sequence but was silenced or deleted before the gene duplication separating AF I and II. The glucagon-like peptide I (GLP-I) was derived from a duplication of the ancestral glucagon region about 800 Myr ago. The rate of synonymous substitution in these genes is approximately 4.3 x 10(-9) substitutions per year per synonymous site. The rate of nonsynonymous substitution in the signal peptide region is about 1.1 x 10(-9) substitutions per year per nonsynonymous site, a high rate comparable to that in the C-peptide region of preproinsulin. The rate of nonsynonymous substitution in the glicentin-related pancreatic polypeptide (GRPP) region is 0.63 x 10(-9) for the comparisons between mammalian species and 1.8 x 10(-9) for the comparisons between fish and mammals; the moderate rate in mammals suggests a physiological role for GRPP. The glucagon region is extremely conservative; no nonsynonymous substitution is observed in the mammalian genes, and a nonsynonymous rate of 0.18 x 10(-9) was obtained from the comparisons between fish and mammals. In the GLP-I region, the rate of nonsynonymous substitution was estimated to be 0.08 x 10(-9) for the comparisons between mammalian species and 0.30 x 10(- 9) for the comparisons between fish and mammals. In the GLP-II region, the rate was estimated to be 0.25 x 10(-9) for the comparisons between mammalian species. Thus, GLP-I and II are also very conservative, which suggests an important physiological role for these peptides.      

Molecular evolution of mammalian lactate dehydrogenase-A genes and pseudogenes: association of a mouse processed pseudogene with a B1 repetitive sequence

A mouse genomic clone containing a lactate dehydrogenase-A (LDH-A) processed pseudogene and a B1 repetitive element was isolated, and a nucleotide sequence of approximately 3 kb was determined. The pseudogene and B1 element are flanked by perfect 13-bp repeats, and the B1 sequence starts at 14 nucleotides 3' to the presumptive polyadenylation signal of the pseudogene. The nucleotide sequences of the LDH-A genes and processed pseudogenes from mouse, rat, and human were compared, and a phylogenetic tree was constructed. The rate and pattern of nucleotide substitutions in the LDH-A pseudogenes are similar to previously reported results (Li et al. 1984). The average rate of nucleotide substitutions in the LDH-A pseudogenes is 4.3 X 10(- 9)/site/year. The substitutions of C----T and G----A are most frequent, and A----G substitutions are relatively high. The rate of synonymous substitutions in the LDH-A genes is 5.3 X 10(-9), which is not significantly higher than the average rate of 4.7 X 10(-9) for 35 mammalian genes. The rate of nonsynonymous substitutions in the LDH-A genes is 0.20 X 10(-9), which is considerably lower than the average rate of 0.88 X 10(-9) for 35 mammalian genes. Thus, the mammalian LDH-A gene appears to be highly conserved in evolution.      

Molecular evolution of the major capsid protein VP1 of enterovirus 70.

Nucleotide sequences of the genome RNA encoding capsid protein VP1 (918 nucleotides) of 18 enterovirus 70 (EV70) isolates collected from various parts of the world in 1971 to 1981 were determined, and nucleotide substitutions among them were studied. The genetic distances between isolates were calculated by the pairwise comparison of nucleotide difference. Regression analysis of the genetic distances against time of isolation of the strains showed that the synonymous substitution rate was very high at 21.53 x 10(-3) substitution per nucleotide per year, while the nonsynonymous rate was extremely low at 0.32 x 10(-3) substitution per nucleotide per year. The rate estimated by the average value of synonymous and nonsynonymous substitutions (W.-H. Li, C.-C. Wu, and C.-C. Luo, Mol. Biol. Evol. 2:150-174, 1985) was 5.00 x 10(-3) substitution per nucleotide per year. Taking the average value of synonymous and nonsynonymous substitutions as genetic distances between isolates, the phylogenetic tree was inferred by the unweighted pairwise grouping method of arithmetic average and by the neighbor-joining method. The tree indicated that the virus had evolved from one focal place, and the time of emergence was estimated to be August 1967 +/- 15 months, 2 years before first recognition of the pandemic of acute hemorrhagic conjunctivitis. By superimposing every nucleotide substitution on the branches of the phylogenetic tree, we analyzed nucleotide substitution patterns of EV70 genome RNA. In synonymous substitutions, the proportion of transitions, i.e., C<==>U and G<==>A, was found to be extremely frequent in comparison with that reported on other viruses or pseudogenes. In addition, parallel substitutions (independent substitutions at the same nucleotide position on different branches, i.e., different isolates, of the tree) were frequently found in both synonymous and nonsynonymous substitutions. These frequent parallel substitutions and the low nonsynonymous substitution rate despite the very high synonymous substitution rate described above imply a strong restriction on nonsynonymous substitution sites of VP1, probably due to the requirement for maintaining the rigid icosahedral conformation of the virus.     

Rates of nucleotide substitution and mammalian nuclear gene evolution. Approximate and maximum-likelihood methods lead to different conclusions

Rates and patterns of synonymous and nonsynonymous substitutions have important implications for the origin and maintenance of mammalian isochores and the effectiveness of selection at synonymous sites. Previous studies of mammalian nuclear genes largely employed approximate methods to estimate rates of nonsynonymous and synonymous substitutions. Because these methods did not account for major features of DNA sequence evolution such as transition/transversion rate bias and unequal codon usage, they might not have produced reliable results. To evaluate the impact of the estimation method, we analyzed a sample of 82 nuclear genes from the mammalian orders Artiodactyla, Primates, and Rodentia using both approximate and maximum-likelihood methods. Maximum-likelihood analysis indicated that synonymous substitution rates were positively correlated with GC content at the third codon positions, but independent of nonsynonymous substitution rates. Approximate methods, however, indicated that synonymous substitution rates were independent of GC content at the third codon positions, but were positively correlated with nonsynonymous rates. Failure to properly account for transition/transversion rate bias and unequal codon usage appears to have caused substantial biases in approximate estimates of substitution rates.     

Evolution of nucleic acids coding for ribonucleases: the mRNA sequence of mouse pancreatic ribonuclease

The cDNA of mouse pancreatic mRNA has been cloned. After the library was screened with a rat ribonuclease cDNA probe, the positive clones were isolated and sequenced. There were no differences from the previously determined protein sequence. The mRNA codes for a preribonuclease of 149 amino acid residues including a signal peptide of 25 amino acids. The 3' noncoding region has a length of 260 bp, and the total mRNA length is approximately 940 bp. Comparison with the rat pancreatic ribonuclease sequence showed a high rate of nucleotide substitution. Within the coding region, nonsynonymous and synonymous substitution rates are 4.3 X 10(-9) and 15 X 10(-9) nucleotide substitutions/site/year, respectively. The latter value is one of the highest rates observed in the molecular evolution of mammalian nuclear genes. In the signal sequences the synonymous substitution rate is much lower and about the same as the nonsynonymous rate. Signal sequences of other mouse and rat proteins also exhibit little difference between synonymous and nonsynonymous rates. The sequences of rat and mouse pancreatic ribonuclease messengers were compared with those of bovine pancreatic, seminal, and brain ribonuclease. While the 3' noncoding regions of rat and mouse are very similar, as are those of the three bovine messengers, there is no significant similarity between both rodent and the three bovine messengers for the greater part of these regions. There is a duplication of approximately 50 nucleotides in the 3' noncoding region of the bovine messengers, with a region rich in A and C in between. The presence of this structural feature may be correlated with recent gene duplications that have occurred in the bovine genome.     

Evolution of hominoid mitochondrial DNA with special reference to the silent substitution rate over the genome

Summary Focusing on the synonymous substitution rate, we carried out detailed sequence analyses of hominoid mitochondrial (mt) DNAs of ca. 5-kb length. Owing to the outnumbered transitions and strong biases in the base compositions, synonymous substitutions in mtDNA reach rapidly a rather low saturation level. The extent of the compositional biases differs from gene to gene. Such changes in base compositions, even if small, can bring about considerable variation in observed synonymous differences and may result in the region-dependent estimate of the synonymous substitution rate. We demonstrate that such a region dependency is due to a failure to take proper account of heterogeneous compositional biases from gene to gene but that the actual synonymous substitution rate is rather uniform. The synonymous substitution rate thus estimated is 2.37 ± 0.11 × 10^–8 per site per year and comparable to the overall rate for the noncoding region. On the other hand, the rate of nonsynonymous substitutions differs considerably from gene to gene, as expected under the neutral theory of molecular evolution. The lowest rate is 0.8 × 10^–9 per site per year forCOI and the highest rate is 4.5 × 10^–9 forATPase 8, the degree of functional constraints (measured by the ratio of the nonsynonymous to the synonymous substitution rate) being 0.03 and 0.19, respectively. Transfer RNA (tRNA) genes also show variability in the base contents and thus in the nucleotide differences. The average rate for 11 tRNAs contained in the 5-kb region is 3.9 × 10^–9 per site per year. The nucleotide substitutions in the genome suggest that the transition rate is about 17 times faster than the transversion rate.     

Molecular Evolution of the Plant R Regulatory Gene Family

Anthocyanin pigmentation patterns in different plant species are controlled in part by members of the myc-like R regulatory gene family. We have examined the molecular evolution of this gene family in seven plant species. Three regions of the R protein show sequence conservation between monocot and dicot R genes. These regions encode the basic helix-loop-helix domain, as well as conserved N-terminal and C-terminal domains; mean replacement rates for these conserved regions are 1.02 X 10(-9) nonsynonymous nucleotide substitutions per site per year. More than one-half of the protein, however, is diverging rapidly, with nonsynonymous substitution rates of 4.08 X 10(-9) substitutions per site per year. Detailed analysis of R homologs within the grasses (Poaceae) confirm that these variable regions are indeed evolving faster than the flanking conserved domains. Both nucleotide substitutions and small insertion/deletions contribute to the diversification of the variable regions within these regulatory genes. These results demonstrate that large tracts of sequence in these regulatory loci are evolving at a fairly rapid rate.     

Phylogeny and substitution rates of angiosperm actin genes

Forty-four actin genes from five angiosperm species were PCR-amplified, cloned, and sequenced. Phylogenetic analysis of 34 of these actins, along with those previously published, indicates that angiosperm actin genes are monophyletic and underwent several duplications during evolution. Orthologues have been identified between Solanaceae species, as well as between Solanaceae species and soybean. These sequences were used to calculate nucleotide substitution rates. The synonymous rate (6.96 x 10(-9) substitutions/site/year) is similar to that of other nuclear protein-coding genes, but the nonsynonymous rate (0.19 x 10(-9) substitutions/site/year) is 6-19 times higher than that of mammalian actin genes. Relative rate tests indicate that actin genes are evolving at similar rates in monocots and in dicots. Evidence is also presented that some members of the maize actin multigene family have been involved in gene conversion events, that the potato genome contains 24 +/- 12 actin genes, and that potato and tomato diverged 11.6 +/- 3.6 MYA.      

On transition bias in mitochondrial genes of pocket gophers

The relative contribution of mutation and purifying selection to transition bias has not been quantitatively assessed in mitochondrial protein genes. The observed transition/transversion (s/v) ratio is (μ _s P _s)/(μ _v P _v), where μ _s and μ _v denote mutation rate of transitions and transversions, respectively, andP _s andP _v denote fixation probabilities of transitions and transversions, respectively. Because selection against synonymous transitions can be assumed to be roughly equal to that against synonymous transversions,P _s/P_v ≈ 1 at fourfold degenerate sites, so that thes/v ratio at fourfold degenerate sites is approximately μ _s /μ _v , which is a measure of mutational contribution to transition bias. Similarly, thes/v ratio at nondegenerate sites is also an estimate of μ _s /μ _v if we assume that selection against nonsynonymous transitions is roughly equal to that against nonsynonymous transversions. In two mitochondrial genes, cytochrome oxidase subunit I (COI) and cytochromeb (cyt-b) in pocket gophers, thes/v ratio is about two at nondegenerate and fourfold degenerate sites for both the COI and the cyt-b genes. This implies that mutation contribution to transition bias is relatively small. In contrast, thes/v ratio is much greater at twofold degenerate sites, being 48 for COI and 40 for cyt-b. Given that the μ _s /μ _v ratio is about 2, theP _s/P_v ratio at twofold degenerate sites must be on the order of 20 or greater. This suggests a great effect of purifying selection on transition bias in mitochondrial protein genes because transitions are synonymous and transversions are nonsynonymous at twofold degenerate sites in mammalian mitochondrial genes. We also found that nonsynonymous mutations at twofold degenerate sites are more neutral than nonsynonymous mutations at nondegenerate sites, and that the COI gene is subject to stronger purifying selection than is the cyt-b gene. A model is presented to integrate the effect of purifying selection, codon bias, DNA repair and GC content ons/v ratio of protein-coding genes. Correspondence to: X. Xia     

Human glyceraldehyde-3-phosphate dehydrogenase pseudogenes: Molecular evolution and a possible mechanism for amplification

We screened two human genomic libraries and isolated 14 different clones, designated λG1 and EG1-EG13, homologous to human glyceraldehyde-3-phosphate dehydrogenase (GAPD) cDNA. Subcloning and sequencing these recombinant phages led us to classify them as five different pseudogenes (ψG1–ψG5). All these sequences show such features typical of processed pseudogenes as numerous mutations, insertions, and deletions. The identity of numerous mutated sites among these pseudogenes and the presence of two Alu sequences flanking both ends of ψG1 suggest that GAPD pseudogenes originated from a unique reverse transcribed mRNA followed by gene duplication. The rate of nucleotide substitutions per site per year for known GAPD functional genes is low both for the synonymous substitutions (1.87×10⁻⁹) and for the nonsynonymous substitutions (0.12￠10⁻⁹) and indicates that the GAPD cDNA sequence is well conserved not only at the amino acid level, but also at the nucleotide level. The rate of nucleotide substitutions per site per year for GAPD pseudogenes shows a higher value (5.9×10⁻⁹) and suggests that these pseudogenes do not have any functional role. This work was supported by grants from the Consiglio Nazionale delle Ricerche and the Ministero Pubblica Istruzione (Rome, Italy). Special acknowledgment is given to the “Progetto Finalizzato Ingegneria Genetica e Basi Molecolari delle Malattie Ereditarie.”     

Evolution of duplicate genes in a tetraploid animal, Xenopus laevis

To understand the evolution of duplicate genes, we compared rates of nucleotide substitution between 17 pairs of nonallelic duplicated genes in the tetraploid frog Xenopus laevis with rates between the orthologous loci of human and rodent. For all duplicated X. laevis genes, the number of synonymous substitutions per site (dS) was greater than the number of nonsynonymous substitutions per site (dN), indicating that these genes are subject to purifying selection. There was also a significant positive correlation (r = 0.915) between dN for the X. laevis genes and dN for the mammalian genes, suggesting that, at the amino acid level, the X. laevis genes and the mammalian genes are under similar constraints. Results of relative-rate tests showed nearly equal rates of nonsynonymous substitution in each copy of the X. laevis genes; apparently there are similar constraints on both copies. No correlation was found between dS for the X. laevis genes and dS for the mammalian genes. There was a significant positive correlation both between members of pairs of duplicated X. laevis genes (r = 0.951) and between human and rodent orthologues (r = 0.854) with respect to third- position G+C content but no such relationship between the X. laevis genes and either of their mammalian orthologues. The results indicate that both copies of a duplicate gene can be subject to purifying selection and thus support the hypothesis of selection against all genotypes containing a null allele at either of two duplicate loci.      

Molecular evolution of chloroplast DNA sequences

Comparative data on the evolution of chloroplast genes are reviewed. The chloroplast genome has maintained a similar structural organization over most plant taxa so far examined. Comparisons of nucleotide sequence divergence among chloroplast genes reveals marked similarity across the plant kingdom and beyond to the cyanobacteria (blue-green algae). Estimates of rates of nucleotide substitution indicate a synonymous rate of 1.1 x 10(-9) substitutions per site per year. Noncoding regions also appear to be constrained in their evolution, although addition/deletion events are common. There have also been evolutionary changes in the distribution of introns in chloroplast encoded genes. Relative to mammalian mitochondrial DNA, the chloroplast genome evolves at a conservative rate.      

Slow molecular clocks in Old World monkeys,apes, and humans

Two longstanding issues on the molecular clock hypothesis are studied in this article. First, is there a global molecular clock in mammals? Although many authors have observed unequal rates of nucleotide substitution among mammalian lineages, some authors have proposed a global clock for all eutherians, i.e., a single global rate of 2.2 x 10(-9) substitutions per nucleotide site per year. We reexamine this issue using noncoding, nonrepetitive DNA from Old World monkeys (OWMs), chimpanzee, and human. First, using the minimal date of 6 MYA for the human-chimpanzee divergence and more than 2.5 million base pairs of genomic sequences from human and chimpanzee, we estimate a maximal rate of 0.99 x 10(-9) for noncoding, nonrepetitive genomic regions for these two species. This estimate is less than half of the proposed global rate and much smaller than the commonly used rate (3.5 x 10(-9)) for eutherians. Further, using a minimal date of 23 MYA for the human-OWM divergence, we estimate a maximal rate of 1.5 x 10(-9) for both introns and fourfold degenerate sites in humans and OWMs. In addition, with the New World monkey (NWM) lineage as an outgroup, we estimate that the rate of substitution in introns is 30% higher in the OWM lineage than in the human lineage. Clearly, there is no global molecular clock in eutherians. Second, although many studies have indicated considerable variation in the mutation rate among regions of the mammalian genome, a recent study proposed a uniform rate. Using new and existing intron sequence data from higher primates, we find significant rate variation among genomic regions and a positive correlation between the rate of substitution and the GC content, refuting the claim of a uniform rate.     

Different patterns of evolution in alternative and constitutive coding regions of Drosophila alternatively spliced genes

Seven hundred and ninety Drosophila melanogaster genes, alternatively spliced in coding regions were considered together with their Drosophila pseudoobscura orthologs. It was found that nucleotide substitutions in alternative coding regions accumulate more intensively than in constitutive regions. Moreover, the evolutionary pattern of alternative regions depends on their inclusion mechanisms (use of alternative promoters, splicing sites or polyadenylation sites) significantly. The rate of synonymous substitutions varies is more dramatically than that of nonsynonymous substitutions. Nucleotide substitution patterns in different classes of alternative regions of mammalian and Drosophila genes have little in common.     

Rates of synonymous substitution in plant nuclear genes

Summary The rate of synonymous nucleotide substitution in nuclear genes of higher plants has been estimated. The rate varies among genes by a factor of up to two, in a manner that is not immediately explicable in terms of base composition or codon usage bias. The average rate, in both monocots and dicots, is about four times higher than that in chloroplast genes. This leads to an estimated absolute silent substitution rate of 6 × 10^–9 substitutions per site per year that falls within the range of average rates (2–8 × 10^–9) seen in different mammalian nuclear genomes.     

Evolution of the Sry genes

Existing DNA sequence data on the Sry gene, the mammalian sex- determining locus in the Y chromosome, were analyzed for primates, rodents, and bovids. In all three taxonomic groups, the terminal sequences evolved faster than the HMG (high mobility group) boxes, and this applies both to synonymous (Ks) and nonsynonymous (Ka) nucleotide substitutions. Similar intragenic correlation between synonymous and nonsynonymous substitution rates was not found either in other mammalian genes that contain a conservative box (Sox, Msx) or in the MADS-box genes of plants. The rate of nonsynonymous substitutions exceeds significantly that of synonymous substitutions in the terminal Sry sequences of apes. We did not find good support for the hypothesis that the high evolutionary rate of Sry would be associated with a promiscuous mating system.      

Molecular evolution of the COX7A gene family in primates.

COX VIIa is one of 10 nuclear-encoded subunits of the COX holoenzyme, and one of three that have isoforms with tissue-specific differences in expression. Analysis of nucleotide substitution rates revealed an accelerated rate of nonsynonymous substitutions relative to that of synonymous substitutions for the heart isoform gene (COX7AH) in six primate lineages. Rate accelerations have been noted for four other COX-related genes in this time period, suggesting that the COX holoenzyme has experienced an episode of adaptive evolution. A third member of the gene family, COX7AR, has recently been described. Although its function is currently unknown, low nonsynonymous substitution/synonymous substitution (N/S) ratios in mammalian evolution suggest that COX7AR is of functional importance. When the COX7A isoforms were divided into domains, examination of nucleotide substitution rates suggested that mitochondrial targeting residues experienced an accelerated nonsynonymous substitution rate in the period following gene duplication. In contrast, paralogous comparisons of the targeting residues of each isoform show they have been relatively conserved in mammalian evolution. This pattern is consistent with the evolution of tissue-specific function.