Does codon composition influence ribosome function?

Escherichia coli ribosomes pre-initiated with N-acetyl-Val-tRNAVal elongate strictly alternating poly(U-G) at a rate between eight and 12 peptide bonds per second per ribosome in vitro. Comparisons with poly(U)-primed poly(Phe) synthesis show that these systems function with the same rates which are close to those of protein synthesis in vivo. This indicates that, at least in vitro, codon composition has no marked influence on the speed of elongation when the concentration of ternary complex is saturating. Furthermore, the missense frequencies for the two polymers are within the same range: the missense substitution of Trp for Cys is 10(-4) and that of Met for Val is 10(-3) in the poly(U-G)-primed system. These data argue against models that explain the codon preference of certain gene families by postulating effects of high or low GC content of codons on the performance characteristics of ribosomes.     

5' contexts of Escherichia coli and human termination codons are similar.

The nearest 5' context of 2559 human stop codons was analysed in comparison with the same context of stop-like codons (UGG, UGC, UGU, CGA for UGA; CAA, UAU, UAC for UAA; and UGG, UAU, UAC, CAG for UAG). The non-random distribution of some nucleotides upstream of the stop codons was observed. For instance, uridine is over-represented in position -3 upstream of UAG. Several codons were shown to be over-represented immediately upstream of the stop codons: UUU(Phe), AGC(Ser), and the Lys and Ala codon families before UGA; AAG(Lys), GCG(Ala), and the Ser and Leu codon families before UAA; and UCA(Ser), AUG(Met), and the Phe codon family before UAG. In contrast, the Thr and Gly codon families were under-represented before UGA, while ACC(Thr) and the Gly codon family were under-represented before UAG and UAA respectively. In an earlier study, uridine was shown to be over-represented in position -3 before UGA in Escherichia coli [Arkov,A.L., Korolev,S.V. and Kisselev,L.L. (1993) Nucleic Acids Res., 21,2891-2897]. In that study, the codons for Lys, Phe and Ser were shown to be over-represented immediately upstream of E. coli stop codons. Consequently, E. coli and human termination codons have similar 5' contexts. The present study suggests that the 5' context of stop codons may modulate the efficiency of peptide chain termination and (or) stop codon readthrough in higher eukaryotes, and that the mechanisms of such a modulation in prokaryotes and higher eukaryotes may be very similar.     

Termination of translation in bacteria may be modulated via specific interaction between peptide chain release factor 2 and the last peptidyl-tRNA(Ser/Phe).

The 5' context of 671 Escherichia coli stop codons UGA and UAA has been compared with the context of stop-like codons (UAC, UAU and CAA for UAA; UGG, UGC, UGU and CGA for UGA). We have observed highly significant deviations from the expected nucleotide distribution: adenine is over-represented whereas pyrimidines are under-represented in position -2 upstream from UAA. Uridine is over-represented in position -3 upstream from UGA. Lysine codons are preferable immediately prior to UAA. A complete set of codons for serine and the phenylalanine UUC codon are preferable immediately 5' to UGA. This non-random codon distribution before stop codons could be considered as a molecular device for modulation of translation termination. We have found that certain fragment of E. coli release factor 2 (RF2) (amino acids 93-114) is similar to the amino acid sequences of seryl-tRNA synthetase (positions 10-19 and 80-93) and of beta (small) subunit (positions 72-94) of phenylalanyl-tRNA synthetase from E. coli. Three-dimensional structure of E. coli seryl-tRNA synthetase is known [1]: Its N-terminus represents an antiparallel alpha-helical coiled-coil domain and contains a region homologous to RF2. On the basis of the above-mentioned results we assume that a specific interaction between RF2 and the last peptidyl-tRNA(Ser/Phe) occurs during polypeptide chain termination in prokaryotic ribosomes.     

Mistranslation in twelve Escherichia coli ribosomal proteins. Cysteine misincorporation at neutral amino acid residues other than tryptophan

The misincorporation of cysteine (codon: UGU/C) into twelve ribosomal proteins devoid of cysteine has been studied. Although it is generally assumed that cysteine is misincorporated at arginine and tryptophan residues (codons: CGU/U and UGG respectively), our results are consistent with the idea that cysteine is also misincorporated at phenylalanine residues (codon: UUU/C) through a second-position C:U mismatch. Cysteine was found in ribosomal proteins L29, L32/L33 and S10, under conditions where only its misincorporation at neutral residues was measured. Since these proteins contain no tryptophan, the date imply that cysteine has replaced a neutral amino acid other than tryptophan. Because there was a statistically significant correlation between the total level of cysteine in the twelve proteins under study and their content of phenylalanine and arginine residues, we conclude that there is a likelihood of cysteine misincorporation at phenylalanine residues, in addition to its misincorporation at arginine and tryptophan residues. Our measurements are consistent with the existence of a cluster of ribosomal proteins having an average mistranslation frequency of 2.5 X 10(-4)/residue and another having an average mistranslation frequency of 10(-3)/residue. There was three times less cysteine misincorporated into ribosomal protein L1 than into L7/L12, although the L1 mRNA contains eleven CGU/C codons and four UUU/C codons while the L7/L12 mRNA contains only one arginine and two phenylalanine codons (both proteins are free of tryptophan). Furthermore, the mRNAs for both L1 and L7/L12 contain a CGU codon located in the context GUA-codon-GG and there was as much cysteine incorporated at this codon in L7/L12 [Bouadloun, F., Donner, D. and Kurland, C.G. (1983) EMBO J. 2, 1351-1356] than in the whole of L1. This suggests that, relatively speaking, little cysteine is to be found at the phenylalanine and the other ten arginine positions of L1 and that the phenylalanine residues of L7/L12 are particularly error-prone.     

Role of GC-biased mutation pressure on synonymous codon choice in Micrococcus luteus, a bacterium with a high genomic GC-content.

The GC (G + C, or G or C)-contents of codon silent positions in all two-codon sets and three codons AUY/A (IIe), and in most of the family boxes of Micrococcus luteus (genomic GC-content: 74%) are 95% to 100% in both the highly and weakly expressed genes. In some family boxes, there is a decrease in NNC codons and an increase in NNG codons from the highly expressed to weakly expressed genes without apparent involvement of NNU and NNA codons. From these observations, we conclude that the selective use of synonymous codons in M. luteus may be largely determined by GC-biased mutation pressure and that in the highly expressed genes tRNAs would act as a weak selection pressure in some family boxes. Available data suggest that the effect of selection pressure by tRNAs on the synonymous codon choice becomes more apparent in the highly expressed genes in eubacteria with intermediate GC-contents such as Escherichia coli and Bacillus subtilis, and that the U/C ratio of the codon third positions in NNU/C-type two-codon sets in the weakly expressed genes would represent the approximate magnitude of directional mutation pressure throughout eubacteria.     

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.     

Selection Intensity for Codon Bias

The patterns of nonrandom usage of synonymous codons (codon bias) in enteric bacteria were analyzed. Poisson random field (PRF) theory was used to derive the expected distribution of frequencies of nucleotides differing from the ancestral state at aligned sites in a set of DNA sequences. This distribution was applied to synonymous nucleotide polymorphisms and amino acid polymorphisms in the gnd and putP genes of Escherichia coli. For the gnd gene, the average intensity of selection against disfavored synonymous codons was estimated as approximately 7.3 X 10(-9); this value is significantly smaller than the estimated selection intensity against selectively disfavored amino acids in observed polymorphisms (2.0 X 10(-8)), but it is approximately of the same order of magnitude. The selection coefficients for optimal synonymous codons estimated from PRF theory were consistent with independent estimates based on codon usage for threonine and glycine. Across 118 genes in E. coli and Salmonella typhimurium, the distribution of estimated selection coefficients, expressed as multiples of the effective population size, has a mean and standard deviation of 0.5 +/- 0.4. No significant differences were found in the degree of codon bias between conserved positions and replacement positions, suggesting that translational misincorporation is not an important selective constraint among synonymous polymorphic codons in enteric bacteria. However, across the first 100 codons of the genes, conserved amino acids with identical codons have significantly greater codon bias than of either synonymous or nonidentical codons, suggesting that there are unique selective constraints, perhaps including mRNA secondary structures, in this part of the coding region.     

Recognition of AUG and alternative initiator codons is augmented by G in position +4 but is not generally affected by the nucleotides in positions +5 and +6.

A primer extension (toeprinting) assay was used to monitor selection by ribosomes of the first versus the second AUG codon as a function of introducing mutations on the 3' side (positions +4, +5 and +6) of the first AUG codon. Six different flanking codons starting with G (GCG, GCU, GCC, GCA, GAU and GGA) strongly augmented selection of AUG#1 when compared with matched mRNAs that had A or C instead of G in position +4. Augmentation by G in position +4 failed only when it was combined with U in position +5, as in the sequence augGUA. In contrast with the usual enhancing effect of introducing G in position +4, most mutations in position +5 had no discernible effect, as shown with the series augANA (where N = C, A, G or U) and the series augCNA. AUG codon recognition was also unaffected by mutations in position +6, as shown by testing four mRNAs that had augCCN as the start site. Thus the primary sequence context that augments the recognition of AUG start codons does not appear generally to extend beyond G in position +4. When the toeprinting assay was used with mRNAs that initiate translation at CUG instead of AUG, cugGAU was not recognized better than cugGGU, contradicting the hypothesis that initiation at non-AUG codons might be favored by A instead of G in position +5.     

The secondary structure of mRNAs from Escherichia coli: its possible role in increasing the accuracy of translation.

A secondary structure model was proposed for mRNAs during translation (in a polysome) where the secondary structure is described by a set of small unbranched hairpins. Computer simulation experiments reveal that the number of hairpins is much greater (P less than 10(-6) in highly expressed mRNAs from E. coli as compared with the random sequences coding for the same amino acid sequence, i.e. certain synonymous codons are used in definite mRNA positions to increase the number of hairpins. No constraints on the amino acid sequence, which would affect the secondary structure of mRNAs, were found. The codons UGU, UGC (Cys), GCC (Ala), ACA, ACG (Thr), CCU, CCC (Pro), etc. translated by minor tRNAs were found to occur significantly more frequently in the position 5' to the hairpins than the other codons translated by major tRNAs (P less than 5.10(-6). This correlation leads to the hypothesis that the process of hairpin unfolding can increase the time of translocation from the A to P ribosome site of the codon 5' to the hairpin, thus decreasing the probability of translational error (the latter would likely occur more frequently in the codons translated by minor tRNAs).     

Cloning of murine SeGpx cDNA and synthesis of mutated GPx proteins in Escherichia coli.

Glutathione peroxidase (GPx) of mammalian cells and Escherichia coli formate dehydrogenase both contain a selenocysteine (SeCys) in their amino acid (aa) sequence. In these two enzymes, this aa is encoded by a UGA codon, which is usually a stop codon for protein synthesis. We constructed plasmids to test the synthesis of GPx in E. coli. These constructions permitted high-level production of GPx mutants, where the SeCys codon was replaced by cysteine (UGC, UGU) or serine (UCA) codons, but synthesis of selenoprotein could not be detected: our data suggest that signals used for the recognition of the UGA codon as a SeCys codon are not conserved between E. coli and mammalian cells.     

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     

The effect of context on synonymous codon usage in genes with low codon usage bias.

The effect of neighbouring bases on the usage of synonymous codons in genes with low codon usage bias in yeast and E. coli is examined. The codon adaptation index is employed to identify a group of genes in each organism with low codon usage bias, which are likely to be weakly expressed. A similar pattern is found in complementary sequences with respect to synonymous usage of A vs G or of U vs C. It is suggested that this may reflect an effect of context on mutation rates in weakly expressed genes.     

Formylmethionyl-tRNA binding to 30 S ribosomes programmed with homopolynucleotides and the effect of translational initiation factor 3

Binding of the polynucleotides poly(U), poly(X) and poly(dT) to 30 S ribosomes of Escherichia coli triggers IF2-dependent binding of initiator-tRNA (fMet-tRNA) to these particles. Poly(A) and poly(C) are inactive. A minimum chain-length of approximately 100 residues in poly(U) is required for full activity in fMet-tRNA binding, although much shorter polymers bind tightly to 30 S particles and do stimulate the binding of acPhe-tRNA. The stimulation of fMet-tRNA binding to 30 S ribosomes is strongly reduced under conditions where the polynucleotides adopt secondary structure. Complexes containing fMet-tRNA and the non-cognate codon UUU or XXX are destabilized by IF3, whereas the formation of such a complex containing an AUG codon is slightly enhanced by the factor. Consistent with previous observations, it was found that all model initiation complexes containing acPhe-tRNA are strongly destabilized by IF3, even when the cognate codon (UUU) is present. Our results suggest that IF3 counteracts 'unnatural' initiation events in vitro and suggest a regulatory role for this factor in vivo.     

Translation of chloroplast-encoded mRNA: potential initiation and termination signals.

A survey of 196 protein-coding chloroplast DNA sequences demonstrated the preference for AUG and UAA codons for initiation and termination of translation, respectively. As in prokaryotes at every nucleotide position from -25 to +25 (AUG is +1 to +3) and for 25 nucleotides 5' and 3' to the termination codon an A or U is predominant, except for C at +5 and G at +22. A Shine-Dalgarno (SD) sequence (GGAGG or tri- or tetranucleotide variant) was found within 100 bp 5' to the AUG codon in 92% of the genes. In 40% of these cases, the location of the SD sequence was similar to that of the consensus for prokaryotes (-12 to -7 5' to AUG), presumed to be optimal for translation initiation. A SD sequence could not be located in 6% of the chloroplast sequences. We propose that mRNA secondary structures may be required for the relocation of a distal SD sequences to within the optimal region (-12 to -7) for initiation of translation. We further suggest that termination at UGA codons in chloroplast genes may occur by a mechanism, involving 16S rRNA secondary structure, which has been proposed for UGA termination in E. coli.     

On the accessibility and selection of the initiator site of mRNA in protein synthesis.

The specificity of the cell-free system of Escherichia coli for mRNA was examined, and the "accessibility" of some natural and synthetic RNAs to the ribosomes was determined by measurement of AcPhe-tRNA and fMet-tRNA binding, AcPhe-puromycin and fMet-puromycin formation, and polypeptide synthesis. The E. coli system effectively initiates the translation of various synthetic RNAs with AcPhe-tRNA or fMet-tRNA under conditions optimal for the translation of viral RNA. Poly(A,G,U) is accessible to the ribosomes according to all of the above criteria. Poly(A,C,G,U), 23 S rRNA, R17 RNA, and MS2 RNA, on the other hand, show limited accessibility when tested for initiator tRNA binding, or for AcPhe-puromycin and fMet-puromycin formation. MS2 and R17 RNA, but not poly(A,C,G,U) and 23 S rRNA, show accessibility when measured by polypeptide synthesis. The results suggest that, except at initiator sites of natural mRNA, an RNA containing about equal amounts of all four bases is inaccessible to E. coli ribosomes for polypeptide synthesis. Rate constants obtained for fMet-tRNA binding with MS2 RNA, poly(A,G,U), and poly(C,G,U) indicate that the ribosomes do not have any special affinity for the viral RNA. Thus, the selection of the initiator site in protein synthesis may be critically determined more by the accessibility of the initiator codon than by ribosomal recognition of the site.     

Mutation bias is the driving force of codon usage in the Gallus gallus genome

Synonymous codons are used with different frequencies both among species and among genes within the same genome and are controlled by neutral processes (such as mutation and drift) as well as by selection. Up to now, a systematic examination of the codon usage for the chicken genome has not been performed. Here, we carried out a whole genome analysis of the chicken genome by the use of the relative synonymous codon usage (RSCU) method and identified 11 putative optimal codons, all of them ending with uracil (U), which is significantly departing from the pattern observed in other eukaryotes. Optimal codons in the chicken genome are most likely the ones corresponding to highly expressed transfer RNA (tRNAs) or tRNA gene copy numbers in the cell. Codon bias, measured as the frequency of optimal codons (Fop), is negatively correlated with the G + C content, recombination rate, but positively correlated with gene expression, protein length, gene length and intron length. The positive correlation between codon bias and protein, gene and intron length is quite different from other multi-cellular organism, as this trend has been only found in unicellular organisms. Our data displayed that regional G + C content explains a large proportion of the variance of codon bias in chicken. Stepwise selection model analyses indicate that G + C content of coding sequence is the most important factor for codon bias. It appears that variation in the G + C content of CDSs accounts for over 60% of the variation of codon bias. This study suggests that both mutation bias and selection contribute to codon bias. However, mutation bias is the driving force of the codon usage in the Gallus gallus genome. Our data also provide evidence that the negative correlation between codon bias and recombination rates in G. gallus is determined mostly by recombination-dependent mutational patterns.     

tRNA1Ser(G34) with the anticodon GGA can recognize not only UCC and UCU codons but also UCA and UCG codons

The tRNA1Ser (anticodon VGA, V=uridin-5-oxyacetic acid) is essential for translation of the UCA codon in Escherichia coli. Here, we studied the translational abilities of serine tRNA derivatives, which have different bases from wild type at the first positions of their anticodons, using synthetic mRNAs containing the UCN (N=A, G, C, or U) codon. The tRNA1Ser(G34) having the anticodon GGA was able to read not only UCC and UCU codons but also UCA and UCG codons. This means that the formation of G-A or G-G pair allowed at the wobble position and these base pairs are noncanonical. The translational efficiency of the tRNA1Ser(G34) for UCA or UCG codon depends on the 2'-O-methylation of the C32 (Cm). The 2'-O-methylation of C32 may give rise to the space necessary for G-A or G-G base pair formation between the first position of anticodon and the third position of codon.     

Decoding with the A:I wobble pair is inefficient.

tRNAs with inosine (I) in the first position read three codons ending in U, C and A. However, A-ending codons read with I are rarely used. In Escherichia coli, CGA/U/C are all read solely by tRNAICGArg. CGU and CGC are very common codons, but CGA is very rare. Three independent in vivo assays show that translation of CGA is relatively inefficient. In the first, nine tandem CGA cause a strong rho-mediated polar effect on expression of a lacZ reporter gene. The inhibition is made more extreme by a mutation in ribosomal protein S12 (rpsL), which indicates that ribosomal binding by tRNAICGArg is slow and/or unstable in the CGA cluster. The second assay, in which codons are substituted for the regulatory UGA of the RF2 frameshift, confirms that aa-tRNA selection is slow and/or unstable at CGA. In the third assay, CGA is found to be a poor 5' context for amber suppression, which suggests that an A:I base pair in the P site can interfere with translation of a codon in the A site. Two possible errors, frameshifting and premature termination by RF2, are not significant causes for inefficiency at CGA. It is concluded that the A:I pair destabilizes codon:anticodon complexes during two successive ribosomal cycles, and it is suggested that these properties contribute to the rare usage of codons read with the A:I base pair.