Synonymous Codon Usage, Accuracy of Translation, and Gene Length in Caenorhabditis elegans

In many unicellular organisms, invertebrates, and plants, synonymous codon usage biases result from a coadaptation between codon usage and tRNAs abundance to optimize the efficiency of protein synthesis. However, it remains unclear whether natural selection acts at the level of the speed or the accuracy of mRNAs translation. Here we show that codon usage can improve the fidelity of protein synthesis in multicellular species. As predicted by the model of selection for translational accuracy, we find that the frequency of codons optimal for translation is significantly higher at codons encoding for conserved amino acids than at codons encoding for nonconserved amino acids in 548 genes compared between Caenorhabditis elegans and Homo sapiens. Although this model predicts that codon bias correlates positively with gene length, a negative correlation between codon bias and gene length has been observed in eukaryotes. This suggests that selection for fidelity of protein synthesis is not the main factor responsible for codon biases. The relationship between codon bias and gene length remains unexplained. Exploring the differences in gene expression process in eukaryotes and prokaryotes should provide new insights to understand this key question of codon usage. Received: 18 June 2000 / Accepted: 10 November 2000     

An Evaluation of Measures of Synonymous Codon Usage Bias

Synonymous codons are not generally used at equal frequencies, and this trend is observed for most genes and organisms. Several methods have been proposed and used to estimate the degree of the nonrandom use of the different synonymous codons. The estimates obtained by these methods, however, show different levels of both precision and dispersion when coding regions of a finite number of codons are under analysis. Here, we present a study, based on computer simulation, of how the different methods proposed to evaluate the nonrandom use of synonymous codons are affected by the length of the coding region analyzed. The results show that some of these methods are heavily influenced by the number of codons and that the comparison of codon usage bias between coding regions of different lengths shows a methodological bias under different conditions of nonrandom use of synonymous codons. The study of the dispersion of the estimates obtained by the different methods gives, on the other hand, an indication of the methods to be applied to compare values of codon usage bias among coding regions of equivalent length. Received: 10 September 1997 / Accepted: 23 March 1998     

Recent Selection on Synonymous Codon Usage in Drosophila

Evidence from a variety of sources indicates that selection has influenced synonymous codon usage in Drosophila. It has generally been difficult, however, to distinguish selection that acted in the distant past from ongoing selection. However, under a neutral model, polymorphisms usually reflect more recent mutations than fixed differences between species and may, therefore, be useful for inferring recent selection. If the ancestral state is preferred, selection should shift the frequency distribution of derived states/site toward lower values; if the ancestral is unpreferred, selection should increase the number of derived states/site. Polymorphisms were classified as ancestrally preferred or unpreferred for several genes of D. simulans and D. melanogaster. A computer simulation of coalescence was employed to derive the expected frequency distributions of derived states/site under various modifications of the Wright–Fisher neutral model, and distributions of test statistics (t and Mann–Whitney U) were derived by appropriate sampling. One-tailed tests were applied to transformed frequency data to assess whether the two frequency distributions deviated from neutral expectations in the direction predicted by selection on codon usage. Several genes from D. simulans appear to be subject to recent selection on synonymous codons, including one gene with low codon bias, esterase-6. Selection may also be acting in D. melanogaster. Received: 15 April 1998 / Accepted: 13 May 1999     

Codon Usage Bias and tRNA Abundance in Drosophila

Codon usage bias of 1,117 Drosophila melanogaster genes, as well as fewer D. pseudoobscura and D. virilis genes, was examined from the perspective of relative abundance of isoaccepting tRNAs and their changes during development. We found that each amino acid contributes about equally and highly significantly to overall codon usage bias, with the exception of Asp which had very low contribution to overall bias. Asp was also the only amino acid that did not show a clear preference for one of its synonymous codons. Synonymous codon usage in Drosophila was consistent with ``optimal' codons deduced from the isoaccepting tRNA availability. Interestingly, amino acids whose major isoaccepting tRNAs change during development did not show as strong bias as those with developmentally unchanged tRNA pools. Asp is the only amino acid for which the major isoaccepting tRNAs change between larval and adult stages. We conclude that synonymous codon usage in Drosophila is well explained by tRNA availability and is probably influenced by developmental changes in relative abundance. Received: 5 December 1996 / Accepted: 14 June 1997     

Codon Usage by Transposable Elements and Their Host Genes in Five Species

We compared the codon usage of sequences of transposable elements (TEs) with that of host genes from the species Drosophila melanogaster, Arabidopsis thaliana, Caenorhabditis elegans, Saccharomyces cerevisiae, and Homo sapiens. Factorial correspondence analysis showed that, regardless of the base composition of the genome, the TEs differed from the genes of their host species by their AT-richness. In all species, the percentage of A + T on the third codon position of the TEs was higher than that on the first codon position and lower than that in the noncoding DNA of the genomes. This indicates that the codon choice is not simply the outcome of mutational bias but is also subject to selection constraints. A tendency toward higher A + T on the third position than on the first position was also found in the host genes of A. thaliana, C. elegans, and S. cerevisiae but not in those of D. melanogaster and H. sapiens. This strongly suggests that the AT choice is a host-independent characteristic common to all TEs. The codon usage of TEs generally appeared to be different from the mean of the host genes. In the AT-rich genomes of Arabidopsis thaliana, Caenorhabditis elegans, and Saccharomyces cerevisiae, the codon usage bias of TEs was similar to that of weakly expressed genes. In the GC-rich genome of D. melanogaster, however, the bias in codon usage of the TEs clearly differed from that of weakly expressed genes. These findings suggest that selection acts on TEs and that TEs may display specific behavior within the host genomes. Received: 2 May 2001 / Accepted: 29 October 2001     

Trends in Codon and Amino Acid Usage in Thermotoga maritima

The usage of synonymous codons and the frequencies of amino acids were investigated in the complete genome of the bacterium Thermotoga maritima using a multivariate statistical approach. The GC3 content of each gene was the most prominent source of variation of codon usage. Surprisingly the usage of UGU and UGC (synonymous triplets coding for Cys, the least frequent amino acid in this species) was detected as the second most prominent source of variation. However, this result is probably an artifact due to the very low frequency of Cys together with the nonbiased composition of this genome. The third trend was related to the preferential usage of a subset of codons among highly expressed genes, and these triplets are presumed to be translationally optimal. Concerning the amino acid usage, the hydropathy level of each protein (and therefore the frequency of charged residues) was the main trend, while the second factor was related to the frequency of usage of the smaller residues, suggesting that the cell economy strongly influences the architecture of the proteins. The third axis of the analysis discriminated the usage of Phe, Tyr, Trp (aromatic residues) plus Cys, Met, and His. These six residues have in common the property of being the preferential targets of reactive oxygen species, and therefore the anaerobic condition of T. maritima is an important factor for the amino acid frequencies. Finally, the Cys content of each protein was the fourth trend. Received: 22 June 2001 / Accepted: 1 October 2001     

Intron–Genome Size Relationship on a Large Evolutionary Scale

The intron–genome size relationship was studied across a wide evolutionary range (from slime mold and yeast to human and maize), as well as the relationship between genome size and the ratio of intervening/coding sequence size. The average intron size is scaled to genome size with a slope of about one-fourth for the log-transformed values; i.e., on the global scale its increase in evolution is lower than the increase in genome size by four orders of magnitude. There are exceptions to the general trend. In baker's yeast introns are extraordinarily long for its genome size. Tetrapods also have longer introns than expected for their genome sizes. In teleost fish the mean intron size does not differ significantly, notwithstanding the differences in genome size. In contrast to previous reports, avian introns were not found to be significantly shorter than introns of mammals, although avian genomes are smaller than genomes of mammals on average by about a factor of 2.5. The extra-/intragenic ratio of noncoding DNA can be higher in fungi than in animals, notwithstanding the smaller fungal genomes. In vertebrates and invertebrates taken separately, this ratio is increasing as the increase in genome size. Two hypotheses are proposed to explain the variation in the extra-/intragenic ratio of noncoding DNA in organisms with similar numbers of genes: transition (dynamic) and equilibrium (static). According to the transition model, this variation arises with the rapid shift of genome size because the bulk of extragenic DNA can be changed more rapidly than the finely interspersed intron sequences. The equilibrium model assumes that this variation is a result of selective adjustment of genome size with constraints imposed on the intron size due to its putative link to chromatin structure (and constraints of the splicing machinery). Received: 23 October 1997 / Accepted: 14 April 1999     

Evolution of Base Composition and Codon Usage Bias in the Genus Flavivirus

The extent to which base composition and codon usage vary among RNA viruses, and the possible causes of this bias, is undetermined in most cases. A maximum-likelihood statistical method was used to test whether base composition and codon usage bias covary with arthropod association in the genus Flavivirus, a major source of disease in humans and animals. Flaviviruses are transmitted by mosquitoes, by ticks, or directly between vertebrate hosts. Those viruses associated with ticks were found to have a significantly lower G+C content than non-vector-borne flaviviruses and this difference was present throughout the genome at all amino acids and codon positions. In contrast, mosquito-borne viruses had an intermediate G+C content which was not significantly different from those of the other two groups. In addition, biases in dinucleotide and codon usage that were independent of base composition were detected in all flaviviruses, but these did not covary with arthropod association. However, the overall effect of these biases was slight, suggesting only weak selection at synonymous sites. A preliminary analysis of base composition, codon usage, and vector specificity in other RNA virus families also revealed a possible association between base composition and vector specificity, although with biases different from those seen in the Flavivirus genus. Received: 29 August 2000 / Accepted: 19 December 2000     

Codon Bias and Mutability in HIV Sequences

A survey of the patterns of synonymous codon preference in the HIV env gene reveals a correlation between the codon bias and the mutability requirements of different regions of the protein. At hypervariable regions in gp120 one finds a greater proportion of codons that tend to mutate nonsynonymously, but to a target that is similar in hydrophobicity and volume. We argue that this strategy results from a compromise between the selective pressure placed on the virus by the induced immune response, which favors amino acid substitutions in the complementarity determining regions, and the negative selection against missense mutations that violate structural constraints of the env protein. Received: 9 June 1997 / Accepted: 25 May 1998     

Codon Usage in Plastid Genes Is Correlated with Context, Position Within the Gene, and Amino Acid Content

Highly expressed plastid genes display codon adaptation, which is defined as a bias toward a set of codons which are complementary to abundant tRNAs. This type of adaptation is similar to what is observed in highly expressed Escherichia coli genes and is probably the result of selection to increase translation efficiency. In the current work, the codon adaptation of plastid genes is studied with regard to three specific features that have been observed in E. coli and which may influence translation efficiency. These features are (1) a relatively low codon adaptation at the 5′ end of highly expressed genes, (2) an influence of neighboring codons on codon usage at a particular site (codon context), and (3) a correlation between the level of codon adaptation of a gene and its amino acid content. All three features are found in plastid genes. First, highly expressed plastid genes have a noticeable decrease in codon adaptation over the first 10–20 codons. Second, for the twofold degenerate NNY codon groups, highly expressed genes have an overall bias toward the NNC codon, but this is not observed when the 3′ neighboring base is a G. At these sites highly expressed genes are biased toward NNT instead of NNC. Third, plastid genes that have higher codon adaptations also tend to have an increased usage of amino acids with a high G + C content at the first two codon positions and GNN codons in particular. The correlation between codon adaptation and amino acid content exists separately for both cytosolic and membrane proteins and is not related to any obvious functional property. It is suggested that at certain sites selection discriminates between nonsynonymous codons based on translational, not functional, differences, with the result that the amino acid sequence of highly expressed proteins is partially influenced by selection for increased translation efficiency. Received: 21 July 1999 / Accepted: 5 November 1999     

Heterogeneity in Codon Usage in the Flatworm Schistosoma mansoni

Synonymous codon choices vary considerably among Schistosoma mansoni genes. Principal components analysis detects a single major trend among genes, which highly correlates with GC content in third codon positions and exons, but does not discriminate among putatively highly and lowly expressed genes. The effective number of codons used in each gene, and its distribution when plotted against GC₃, suggests that codon usage is shaped mainly by mutational biases. The GC content of exons, GC₃, 5′, 3′, and flanking (5′+ 3′+ introns) regions are all correlated among them, suggesting that variations in GC content may exist among different regions of the S. mansoni genome. We propose that this genome structure might be among the most important factors shaping codon usage in this species, although the action of selection on certain sequences cannot be excluded. Received: 10 March 1997 / Accepted: 27 June 1997     

Synonymous Nucleotide Divergence and Saturation: Effects of Site-Specific Variations in Codon Bias and Mutation Rates

The synonymous divergence between Escherichia coli and Salmonella typhimurium is explained in a model where there is a large variation between mutation rates at different nucleotide sites in the genome. The model is based on the experimental observation that spontaneous mutation rates can vary over several orders of magnitude at different sites in a gene. Such site-specific variation must be taken into account when studying synonymous divergence and will result in an apparent saturation below the level expected from an assumption of uniform rates. Recently, it has been suggested that codon preference in enterobacteria has a very large site-specific variation and that the synonymous divergence between different species, e.g., E. coli and Salmonella, is saturated. In the present communication it is shown that when site-specific variation in mutation rates is introduced, there is no need to invoke assumptions of saturation and a large variability in codon preference. The same rate variation will also bring average mutation rates as estimated from synonymous sequence divergence into numerical agreement with experimental values. Received: 10 July 1998 / Accepted: 20 August 1998     

Synonymous Codon Choices in the Extremely GC-Poor Genome of Plasmodium falciparum: Compositional Constraints and Translational Selection

We have analyzed the patterns of synonymous codon preferences of the nuclear genes of Plasmodium falciparum, a unicellular parasite characterized by an extremely GC-poor genome. When all genes are considered, codon usage is strongly biased toward A and T in third codon positions, as expected, but multivariate statistical analysis detects a major trend among genes. At one end genes display codon choices determined mainly by the extreme genome composition of this parasite, and very probably their expression level is low. At the other end a few genes exhibit an increased relative usage of a particular subset of codons, many of which are C-ending. Since the majority of these few genes is putatively highly expressed, we postulate that the increased C-ending codons are translationally optimal. In conclusion, while codon usage of the majority of P. falciparum genes is determined mainly by compositional constraints, a small number of genes exhibit translational selection. Received: 10 November 1998 / Accepted: 28 January 1999     

Selection on the codon bias of chloroplast and cyanelle genes in different plant and algal lineages

In the plant chloroplast genome the codon usage of the highly expressed psbA gene is unique and is adapted to the tRNA population, probably due to selection for translation efficiency. In this study the role of selection on codon usage in each of the fully sequenced chloroplast genomes, in addition to Chlamydomonas reinhardtii, is investigated by measuring adaptation to this pattern of codon usage. A method is developed which tests selection on each gene individually by constructing sequences with the same amino acid composition as the gene and randomly assigning codons based on the nucleotide composition of noncoding regions of that genome. The codon bias of the actual gene is then compared to a distribution of random sequences. The data indicate that within the algae selection is strong in Cyanophora paradoxa, affecting a majority of genes, of intermediate intensity in Odontella sinensis, and weaker in Porphyra purpurea and Euglena gracilis. In the plants, selection is found to be quite weak in Pinus thunbergii and the angiosperms but there is evidence that an intermediate level of selection exists in the liverwort Marchantia polymorpha. The role of selection is then further investigated in two comparative studies. It is shown that average relative codon bias is correlated with expression level and that, despite saturation levels of substitution, there is a strong correlation among the algae genomes in the degree of codon bias of homologous genes. All of these data indicate that selection for translation efficiency plays a significant role in determining the codon bias of chloroplast genes but that it acts with different intensities in different lineages. In general it is stronger in the algae than the higher plants, but within the algae Euglena is found to have several unusual features which are noted. The factors that might be responsible for this variation in intensity among the various genomes are discussed. Received: 6 June 1997 / Accepted: 24 July 1997     

Gene Expression, Amino Acid Conservation, and Hydrophobicity Are the Main Factors Shaping Codon Preferences in Mycobacterium tuberculosis and Mycobacterium leprae

Mycobacterium tuberculosis and Mycobacterium leprae are the ethiological agents of tuberculosis and leprosy, respectively. After performing extensive comparisons between genes from these two GC-rich bacterial species, we were able to construct a set of 275 homologous genes. Since these two bacterial species also have a very low growth rate, translational selection could not be so determinant in their codon preferences as it is in other fast-growing bacteria. Indeed, principal-components analysis of codon usage from this set of homologous genes revealed that the codon choices in M. tuberculosis and M. leprae are correlated not only with compositional constraints and translational selection, but also with the degree of amino acid conservation and the hydrophobicity of the encoded proteins. Finally, significant correlations were found between GC₃ and synonymous distances as well as between synonymous and nonsynonymous distances. Received: 30 October 1998 / Accepted: 16 August 1999     

Evolutionary Lability of Context-Dependent Codon Bias in Bacteria

In bacteria, synonymous codon usage can be considerably affected by base composition at neighboring sites. Such context-dependent biases may be caused by either selection against specific nucleotide motifs or context-dependent mutation biases. Here we consider the evolutionary conservation of context-dependent codon bias across 11 completely sequenced bacterial genomes. In particular, we focus on two contextual biases previously identified in Escherichia coli; the avoidance of out-of-frame stop codons and AGG motifs. By identifying homologues of E. coli genes, we also investigate the effect of gene expression level in Haemophilus influenzae and Mycoplasma genitalium. We find that while context-dependent codon biases are widespread in bacteria, few are conserved across all species considered. Avoidance of out-of-frame stop codons does not apply to all stop codons or amino acids in E. coli, does not hold for different species, does not increase with gene expression level, and is not relaxed in Mycoplasma spp., in which the canonical stop codon, TGA, is recognized as tryptophan. Avoidance of AGG motifs shows some evolutionary conservation and increases with gene expression level in E. coli, suggestive of the action of selection, but the cause of the bias differs between species. These results demonstrate that strong context-dependent forces, both selective and mutational, operate on synonymous codon usage but that these differ considerably between genomes. Received: 6 May 1999 / Accepted: 29 October 1999     

Codon Usage and Selection on Proteins

Selection pressures on proteins are usually measured by comparing homologous nucleotide sequences (Zuckerkandl and Pauling 1965). Recently we introduced a novel method, termed volatility, to estimate selection pressures on proteins on the basis of their synonymous codon usage (Plotkin and Dushoff 2003; Plotkin et al. 2004). Here we provide a theoretical foundation for this approach. Under the Fisher-Wright model, we derive the expected frequencies of synonymous codons as a function of the strength of selection on amino acids, the mutation rate, and the effective population size. We analyze the conditions under which we can expect to draw inferences from biased codon usage, and we estimate the time scales required to establish and maintain such a signal. We find that synonymous codon usage can reliably distinguish between negative selection and neutrality only for organisms, such as some microbes, that experience large effective population sizes or periods of elevated mutation rates. The power of volatility to detect positive selection is also modest—requiring approximately 100 selected sites—but it depends less strongly on population size. We show that phenomena such as transient hyper-mutators can improve the power of volatility to detect selection, even when the neutral site heterozygosity is low. We also discuss several confounding factors, neglected by the Fisher-Wright model, that may limit the applicability of volatility in practice. Electronic Supplementary Material Electronic Supplementary material is available for this article at and accessible for authorised users. [Reviewing Editor: Dr. Lauren Meyers]