The +4G site in Kozak consensus is not related to the efficiency of translation initiation

The optimal context for translation initiation in mammalian species is GCCRCCaugG (where R = purine and "aug" is the initiation codon), with the -3R and +4G being particularly important. The presence of +4G has been interpreted as necessary for efficient translation initiation. Accumulated experimental and bioinformatic evidence has suggested an alternative explanation based on amino acid constraint on the second codon, i.e., amino acid Ala or Gly are needed as the second amino acid in the nascent peptide for the cleavage of the initiator Met, and the consequent overuse of Ala and Gly codons (GCN and GGN) leads to the +4G consensus. I performed a critical test of these alternative hypotheses on +4G based on 34169 human protein-coding genes and published gene expression data. The result shows that the prevalence of +4G is not related to translation initiation. Among the five G-starting codons, only alanine codons (GCN), and glycine codons (GGN) to a much smaller extent, are overrepresented at the second codon, whereas the other three codons are not overrepresented. While highly expressed genes have more +4G than lowly expressed genes, the difference is caused by GCN and GGN codons at the second codon. These results are inconsistent with +4G being needed for efficient translation initiation, but consistent with the proposal of amino acid constraint hypothesis.     

A covariotide model explains apparent phylogenetic structure of oxygenic photosynthetic lineages

The aims of the work were (1) to develop statistical tests to identify whether substitution takes place under a covariotide model in sequences used for phylogenetic inference and (2) to determine the influence of covariotide substitution on phylogenetic trees inferred for photosynthetic and other organisms. (Covariotide and covarion models are ones in which sites that are variable in some parts of the underlying tree are invariable in others and vice versa.) Two tests were developed. The first was a contingency test, and the second was an inequality test comparing the expected number of variable sites in two groups with the observed number. Application of these tests to 16S rDNA and tufA sequences from a range of nonphotosynthetic prokaryotes and oxygenic photosynthetic prokaryotes and eukaryotes suggests the occurrence of a covariotide mechanism. The degree of support for partitioning of taxa in reconstructed trees involving these organisms was determined in the presence or absence of sites showing particular substitution patterns. This analysis showed that the support for splits between (1) photosynthetic eukaryotes and prokaryotes and (2) photosynthetic and nonphotosynthetic organisms could be accounted for by patterns arising from covariotide substitution. We show that the additional problem of compositional bias in sequence data needs to be considered in the context of patterns of covariotide/covarion substitution. We argue that while covariotide or covarion substitution may give rise to phylogenetically informative patterns in sequence data, this may not always be so.      

Codon usage and gene expression.

The hypothesis that codon usage regulates gene expression at the level of translation is tested. Codon usage of Escherichia coli and phage lambda is compared by correspondence analysis, and the basis of this hypothesis is examined by connecting codon and tRNA distributions to polypeptide elongation kinetics. Both approaches indicate that if codon usage was random tRNA limitation would only affect the rarest tRNA species. General discrimination against their cognate codons indicates that polypeptide elongation rates are maintained constant. Thus, differences in expression of E. coli genes are not a consequence of their variable codon usage. The preference of codons recognized by the most abundant tRNAs in E. coli genes encoding abundant proteins is explained by a constraint on the cost of proof-reading.     

Nonrandom frequency patterns of synonymous substitutions in homologous mammalian genes

All 69 homologous coding sequences that are currently available in four mammalian orders were aligned and the synonymous positions of quartet and duet (fourfold and twofold degenerate) codons were divided into three classes (that will be called conserved, intermediate, and variable) according to whether they show no change, one change, or more than one change, respectively. We observed (1) that the frequencies of conserved, intermediate, and variable positions of quartet and duet codons are different in different genes; (2) that the frequencies of the three classes are significantly different from expectations based on a random substitution process in the majority of genes (especially for GC-rich genes) for quartet codons and in a minority of genes for doublet codons; and (3) that the frequencies of the three classes of positions of quartet codons are correlated with those of duet codons, the conserved positions of quartet and duet codons being, in addition, correlated with the degree of amino acid conservation. Our main conclusions are that synonymous substitution frequencies: (1) are gene-specific; (2) are not simply the result of a stochastic process in which nucleotide substitutions accumulate at random, over time; and (3) are correlated in quartet and duet codons.     

Rapid evolution through gene duplication and subfunctionalization of the testes-specific alpha4 proteasome subunits in Drosophila

Gene duplication is an important mechanism for acquiring new genes and creating genetic novelty in organisms. Evidence suggests that duplicated genes are retained at a much higher rate than originally thought and that functional divergence of gene copies is a major factor promoting their retention in the genome. We find that two Drosophila testes-specific alpha4 proteasome subunit genes (alpha4-t1 and alpha4-t2) have a higher polymorphism within species and are significantly more diverged between species than the somatic alpha4 gene. Our data suggest that following gene duplication, the alpha4-t1 gene experienced relaxed selective constraints, whereas the alpha4-t2 gene experienced positive selection acting on several codons. We report significant heterogeneity in evolutionary rates among all three paralogs at homologous codons, indicating that functional divergence has coincided with genic divergence. Reproductive subfunctionalization may allow for a more rapid evolution of reproductive traits and a greater specialization of testes function. Our data add to the increasing evidence that duplicated genes experience lower selective constraints and in some cases positive selection following duplication. Newly duplicated genes that are freer from selective constraints may provide a mechanism for developing new interactions and a pathway for the evolution of new genes.     

Synonymous codon usage and gene function are strongly related in Oryza sativa

The relationship between codon usage and gene function was investigated while considering a dataset of 2106 nuclear genes of Oryza sativa. The results of standard chi(2) test and F-statistic showed that for every 59 synonymous codons, a strongly significant association with gene functional categories existed in rice, indicating that codon usage was generally coordinated with gene function whether it was at the level of individual amino acids or at the level of nucleotides. However, it could not be directly said that the use of every codons differed significantly between any two functional categories. Notably, there existed large difference both in selection for biased codons or selection intensity among functional categories. Therefore, we identified at least two classes of genes: one group of genes, mainly belonging to the "METABOLISM" category, was tended to use G- and/or C-ending codons while the other was more biased to choose codons ending with A and/or U. The latter group contained genes of various functions, especially those genes classified into the "Nuclear Structure" category. These observations will be more important for molecular genetic engineering and genome functional annotation.     

Evidence from simulation studies for selective constraints on the codon usage of the Angiosperm psbA gene

The codon usage of the Angiosperm psbA gene is atypical for flowering plant chloroplast genes but similar to the codon usage observed in highly expressed plastid genes from some other Plantae, particularly Chlorobionta, lineages. The pattern of codon bias in these genes is suggestive of selection for a set of translationally optimal codons but the degree of bias towards these optimal codons is much weaker in the flowering plant psbA gene than in high expression plastid genes from lineages such as certain green algal groups. Two scenarios have been proposed to explain these observations. One is that the flowering plant psbA gene is currently under weak selective constraints for translation efficiency, the other is that there are no current selective constraints and we are observing the remnants of an ancestral codon adaptation that is decaying under mutational pressure. We test these two models using simulations studies that incorporate the context-dependent mutational properties of plant chloroplast DNA. We first reconstruct ancestral sequences and then simulate their evolution in the absence of selection on codon usage by using mutation dynamics estimated from intergenic regions. The results show that psbA has a significantly higher level of codon adaptation than expected while other chloroplast genes are within the range predicted by the simulations. These results suggest that there have been selective constraints on the codon usage of the flowering plant psbA gene during Angiosperm evolution.     

Same-strand overlapping genes in bacteria: compositional determinants of phase bias

Background

Same-strand overlapping genes may occur in frameshifts of one (phase 1) or two nucleotides (phase 2). In previous studies of bacterial genomes, long phase-1 overlaps were found to be more numerous than long phase-2 overlaps. This bias was explained by either genomic location or an unspecified selection advantage. Models that focused on the ability of the two genes to evolve independently did not predict this phase bias. Here, we propose that a purely compositional model explains the phase bias in a more parsimonious manner. Same-strand overlapping genes may arise through either a mutation at the termination codon of the upstream gene or a mutation at the initiation codon of the downstream gene. We hypothesized that given these two scenarios, the frequencies of initiation and termination codons in the two phases may determine the number for overlapping genes.

Results

We examined the frequencies of initiation- and termination-codons in the two phases, and found that termination codons do not significantly differ between the two phases, whereas initiation codons are more abundant in phase 1. We found that the primary factors explaining the phase inequality are the frequencies of amino acids whose codons may combine to form start codons in the two phases. We show that the frequencies of start codons in each of the two phases, and, hence, the potential for the creation of overlapping genes, are determined by a universal amino-acid frequency and species-specific codon usage, leading to a correlation between long phase-1 overlaps and genomic GC content.

Conclusion

Our model explains the phase bias in same-strand overlapping genes by compositional factors without invoking selection. Therefore, it can be used as a null model of neutral evolution to test selection hypotheses concerning the evolution of overlapping genes.

Reviewers

This article was reviewed by Bill Martin, Itai Yanai, and Mikhail Gelfand.     

"Silent" sites in Drosophila genes are not neutral: evidence of selection among synonymous codons

The patterns of synonymous codon usage in 91 Drosophila melanogaster genes have been examined. Codon usage varies strikingly among genes. This variation is associated with differences in G+C content at silent sites, but (unlike the situation in mammalian genes) these differences are not correlated with variation in intron base composition and so are not easily explicable in terms of mutational biases. Instead, those genes with high G+C content at silent sites, resulting from a strong "preference" for a particular subset of the codons that are mostly C- ending, appear to be the more highly expressed genes. This suggests that G+C content is reduced in sequences where selective constraints are weaker, as indeed seen in a pseudogene. These and other data discussed are consistent with the effects of translational selection among synonymous codons, as seen in unicellular organisms. The existence of selective constraints on silent substitutions, which may vary in strength among genes, has implications for the use of silent molecular clocks.      

Interaction of silent and replacement changes in eukaryotic coding sequences

Summary We examined the codon usages in wellconserved and less-well-conserved regions of vertebrate protein genes and found them to be similar. Despite this similarity, there is a statistically significant decrease in codon bias in the less-well-conserved regions. Our analysis suggests that although those codon changes initially fixed under amino acid replacements tend to follow the overall codon usage pattern, they also reduce the bias in codon usage. This decrease in codon bias leads one to predict that the rate of change of synonymous codons should be greater in those regions that are less well conserved at the amino acid level than in the better-conserved regions. Our analysis supports this prediction. Furthermore, we demonstrate a significantly elevated rate of change of synonymous codons among the adjacent codons 5 to amino acid replacement positions. This provides further support for the idea that there are contextual constraints on the choice of synonymous codons in eukaryotes.     

Forbidden synonymous substitutions in coding regions

In the evolution of highly conserved genes, a few "synonymous" substitutions at third bases that would not alter the protein sequence are forbidden or very rare, presumably as a result of functional requirements of the gene or the messenger RNA. Another 10% or 20% of codons are significantly less variable by synonymous substitution than are the majority of codons. The changes that occur at the majority of third bases are subject to codon usage restrictions. These usage restrictions control sequence similarities between very distant genes. For example, 70% of third bases are identical in calmodulin genes of man and trypanosome. Third-base similarities of distant genes for conserved proteins are mathematically predicted, on the basis of the G+C composition of third bases. These observations indicate the need for reexamination of methods used to calculate synonymous substitutions.      

Codon catalog usage and the genome hypothesis.

Frequencies for each of the 61 amino acid codons have been determined in every published mRNA sequence of 50 or more codons. The frequencies are shown for each kind of genome and for each individual gene. A surprising consistency of choices exists among genes of the same or similar genomes. Thus each genome, or kind of genome, appears to possess a "system" for choosing between codons. Frameshift genes, however, have widely different choice strategies from normal genes. Our work indicates that the main factors distinguishing between mRNA sequences relate to choices among degenerate bases. These systematic third base choices can therefore be used to establish a new kind of genetic distance, which reflects differences in coding strategy. The choice patterns we find seem compatible with the idea that the genome and not the individual gene is the unit of selection. Each gene in a genome tends to conform to its species' usage of the codon catalog; this is our genome hypothesis.     

Multisequence comparisons in protein coding genes. Search for functional constraints

A very powerful method for detecting functional constraints operative in biological macromolecules is presented. This method entails performing a base permanence analysis of protein coding genes at each codon position simultaneously in different species. It calculates the degree of permanence of subregions of the gene by dividing it into segments, c codons long, counting how many sites remain unchanged in each segment among all species compared. By comparing the base permanence among several sequences with the expectations based on a stochastic evolutionary process, gene regions showing different degrees of conservation can be selected. This means that wherever the permanence deviates significantly from the expected value generated by the simulation, the corresponding regions are considered "constrained" or "hypervariable". The constrained regions are of two types: alpha and beta. The alpha regions result from constraints at the amino acid level, whereas the beta regions are those probably involved in "control" processing. The method has been applied to mitochondrial genes coding for subunit 6 of the ATPase and subunit 1 of the cytochrome oxidase in four mammalian species: human, rat, mouse, and cow. In the two mitochondrial genes a few regions that are highly conserved in all codon positions have been identified. Among these regions a sequence, common to both genes, that is complementary to a strongly conserved region of 12S rRNA has been found. This method can also be of great help in studying molecular evolution mechanisms.     

Constraints on codon context in Escherichia coli genes. Their possible role in modulating the efficiency of translation

The constraints on nucleotide sequences of highly and weakly expressed genes from Escherichia coli have been analysed and compared. Differences in synonymous codon spectra in highly and weakly expressed genes lead to different frequencies of nucleotides (in the first and third codon positions) and dinucleotides in the two groups of genes. It has been found that the choice of synonymous codons in highly expressed genes depends on the nucleotides adjacent to the codon. For example, lysine is preferably encoded by the AAA codon if guanosine is 3' to the lysine codon (AAA-G, P less than 10(-9)). And, on the contrary, AAG is used more often than AAA (P less than 0.001) if cytidine is 3' adjacent to lysine. Guanosine occurs more frequently than adenosine 5' to all the lysine codons (AAR, P less than 10(-5), i.e. NNG codons are preferred over the synonymous NNA codons 5' to the positions of lysine in the genes. The context effect was observed in nonsense and missense suppression experiments. Therefore, a hypothesis has been suggested that the efficiency of translation of some codons (for which the constraints on the adjacent nucleotides were found) can be modulated by the codon context. The rules for preferable synonymous codon choice in highly expressed genes depending on the nucleotides surrounding the codon are presented. These rules can be used in the chemical synthesis of genes designed for expression in E. coli.     

Structure of a human smooth muscle actin gene (aortic type) with a unique intron site.

A recombinant phage containing an actin gene (lambda Ha201) was isolated from a human DNA library and the structure of the actin gene was determined. The amino acid sequences deduced from the nucleotide sequences of lambda Ha201 were compared with those of six actin isoforms; they matched those of bovine aortic smooth muscle actin, except for codon 309, which was valine (GTC) in lambda Ha201 and alanine (GCN) in bovine aortic smooth muscle actin. Southern blot hybridization experiments showed that the gene of normal human cells did not have the TaqI-sensitive site around position 309, whereas half of the genes of HUT14 cells did. These results indicate that one allele of the aortic smooth muscle actin gene in HUT14 cells has a transition point mutation (C----T) at codon 309 and that the amino acid sequences of normal human aorta and bovine smooth muscle actins are probably identical. In addition to the five introns interrupting exons at codons 150, 204, and 267, and between codons 41 and 42 and 327 and 328, which are common to skeletal muscle and cardiac muscle actin genes, the smooth muscle actin gene has two more intron sites between codons 84 and 85 and 121 and 122. The previously unreported intron site between codons 84 and 85 is unique to the smooth muscle actin gene. The intron site between codons 121 and 122 is common to beta-actin genes but is not found in other muscle actin genes. A hypothesis is proposed for the evolutionary pathway of the actin gene family.     

Sense codon in Escherichia coli are translated in context

The nucleotide frequencies 5' and 3' to the sense codons in highly and weakly expressed genes have been investigated by the chi-squares method. A comparison between the experimental and computer-generated random nucleotide sequences (in which each codon is substituted by a random synonymous one) was made. It was shown that the choice of a particular codon among the synonymous ones in a given position of the gene depends on the three nucleotides 3' and 5' adjacent to the codon in highly expressed genes (the triplet 3' and a single nucleotide 5' to the codons in weakly expressed genes). Concrete patterns for the preferable choice of synonymous codons depending on their contexts are presented. It is suggested that these constraints are related to the efficiency of messenger translation. The constraints on the amino acid sequences of encoded proteins also lead to statistically significant bases in nucleotide frequencies around the sense codons. The biological role of these constraints is discussed.     

The Wilhelmine E. Key 1999 Invitational lecture. Predicting the evolution of human influenza A

We studied the evolution of the HA1 domain of the H3 hemagglutinin gene from human influenza virus type A. The phylogeny of these genes showed a single dominant lineage persisting over time. We tested the hypothesis that the progenitors of this single evolutionarily successful lineage were viruses carrying mutations at codons at which prior mutations had helped the virus to avoid human immune surveillance. We found evidence that eighteen hemagglutinin codons appeared to have been under positive selection to change the amino acid they encoded in the past. Retrospective tests show that viral lineages undergoing the greatest number of mutations in the positively selected codons were the progenitors of future H3 lineages in nine of eleven recent influenza seasons. Codons under positive selection were associated with antibody combining sites A or B or the sialic acid receptor binding site. However, not all codons in these sites had predictive value. Monitoring new H3 isolates for additional changes in positively selected codons might help identify the most fit extant viral strains that arise during antigenic drift.     

Codon distribution in vertebrate genes may be used to predict gene length

I have analysed the coding regions of 96 eukaryotic genes for their use of iso-coding codons. Specific codons occur more frequently in specific positions in all members of some gene families than would be expected if codon choice was determined solely by the frequency of codon usage. In the absence of evidence a priori for selection for particular codons at particular positions, I term such co-occurring codons “coincident codons”. Coincident codons are not confined to particular regions of genes, and their occurrence is not detectably linked with the location of introns in the genomic sequence. Their presence is partly but not completely explained by the exchange of sequence between similar functional genes within a species: homologous genes from different organisms also possess the same codons at some sites with greater than expected frequencies. The relative excess of coincident codons correlates well with the overall length of the genes analysed, but not with the length of mRNA or coding regions, or with qualitative features of gene structure or expression. This, and the unusual sequence environment of coincident codons, suggests that they are a feature of the overall secondary structure of the heterogeneous nuclear RNA. Such considerations suggest approaches for optimizing the expression of exogenous genes in eukaryotic systems, and for predicting the structure of genes for which only partial sequence data is available.