Codon-defined ribosomal pausing in Escherichia coli detected by using the pyrE attenuator to probe the coupling between transcription and translation

Investigation of Genome Biology by Synthetic Genome Engineering

Synthetic genomes were designed based on an understanding of natural genomic information, offering an opportunity to engineer and investigate biological systems on a genome-wide scale. Currently, the designer version of the M. mycoides genome and the E. coli genome, as well as most of the S. cerevisiae genome, have been synthesized, and through the cycles of design-build-test and the following engineering of synthetic genomes, many fundamental questions of genome biology have been investigated. In this review, we summarize the use of synthetic genome engineering to explore the structure and function of genomes, and highlight the unique values of synthetic genomics.

Regulation of the pyrimidine biosynthetic pathway by lysine acetylation of E. coli OPRTase

The de novo pyrimidine biosynthesis pathway is an important route due to the relevance of its products, its implications in health and its conservation among organisms. Here, we investigated the regulation by lysine acetylation of this pathway. To this aim, intracellular and extracellular metabolites of the route were quantified, revealing a possible blockage of the pathway by acetylation of the OPRTase enzyme (orotate phosphoribosyltransferase). Chemical acetylation of OPRTase by acetyl-P involved a decrease in enzymatic activity. To test the effect of acetylation in this enzyme, K26 and K103 residues were selected to generate site-specific acetylated proteins. Several differences were observed in kinetic parameters, emphasizing that the k_cat of these mutants showed a strong decrease of 300 and 150-fold for OPRTase-103AcK and 19 and 6.3-fold for OPRTase-26AcK, for forward and reverse reactions. In vivo studies suggested acetylation of this enzyme by a nonenzymatic acetyl-P-dependent mechanism and a reversion of this process by the CobB deacetylase. A complementation assay of a deficient strain in the pyrE gene with OPRTase-26AcK and OPRTase-103AcK was performed, and curli formation, stoichiometric parameters and orotate excretion were measured. Complementation with acetylated enzymes entailed a profile very similar to that of the ∆pyrE strain, especially in the case of complementation with OPRTase-103AcK. These results suggest regulation of the de novo pyrimidine biosynthesis pathway by lysine acetylation of OPRTase in Escherichia coli. This finding is of great relevance due to the essential role of this route and the OPRTase enzyme as a target for antimicrobial, antiviral and cancer treatments.

A Code Within a Code: How Codons Fine-Tune Protein Folding in the Cell

The genetic code sets the correspondence between the sequence of a given nucleotide triplet in an mRNA molecule, called a codon, and the amino acid that is added to the growing polypeptide chain during protein synthesis. With four bases (A, G, U, and C), there are 64 possible triplet codons: 61 sense codons (encoding amino acids) and 3 nonsense codons (so-called, stop codons that define termination of translation). In most organisms, there are 20 common/standard amino acids used in protein synthesis; thus, the genetic code is redundant with most amino acids (with the exception of Met and Trp) are being encoded by more than one (synonymous) codon. Synonymous codons were initially presumed to have entirely equivalent functions, however, the finding that synonymous codons are not present at equal frequencies in mRNA suggested that the specific codon choice might have functional implications beyond coding for amino acid. Observation of nonequivalent use of codons in mRNAs implied a possibility of the existence of auxiliary information in the genetic code. Indeed, it has been found that genetic code contains several layers of such additional information and that synonymous codons are strategically placed within mRNAs to ensure a particular translation kinetics facilitating and fine-tuning co-translational protein folding in the cell via step-wise/sequential structuring of distinct regions of the polypeptide chain emerging from the ribosome at different points in time. This review summarizes key findings in the field that have identified the role of synonymous codons and their usage in protein folding in the cell.

Emergent rules for codon choice elucidated by editing rare arginine codons in Escherichia coli

The degeneracy of the genetic code allows nucleic acids to encode amino acid identity as well as noncoding information for gene regulation and genome maintenance. The rare arginine codons AGA and AGG (AGR) present a case study in codon choice, with AGRs encoding important transcriptional and translational properties distinct from the other synonymous alternatives (CGN). We created a strain of Escherichia coli with all 123 instances of AGR codons removed from all essential genes. We readily replaced 110 AGR codons with the synonymous CGU codons, but the remaining 13 "recalcitrant" AGRs required diversification to identify viable alternatives. Successful replacement codons tended to conserve local ribosomal binding site-like motifs and local mRNA secondary structure, sometimes at the expense of amino acid identity. Based on these observations, we empirically defined metrics for a multidimensional "safe replacement zone" (SRZ) within which alternative codons are more likely to be viable. To evaluate synonymous and nonsynonymous alternatives to essential AGRs further, we implemented a CRISPR/Cas9-based method to deplete a diversified population of a wild-type allele, allowing us to evaluate exhaustively the fitness impact of all 64 codon alternatives. Using this method, we confirmed the relevance of the SRZ by tracking codon fitness over time in 14 different genes, finding that codons that fall outside the SRZ are rapidly depleted from a growing population. Our unbiased and systematic strategy for identifying unpredicted design flaws in synthetic genomes and for elucidating rules governing codon choice will be crucial for designing genomes exhibiting radically altered genetic codes.

The Yin and Yang of codon usage

The genetic code is degenerate. With the exception of two amino acids (Met and Trp), all other amino acid residues are each encoded by multiple, so-called synonymous codons. Synonymous codons were initially presumed to have entirely equivalent functions, however, the finding that synonymous codons are not present at equal frequencies in genes/genomes suggested that codon choice might have functional implications beyond amino acid coding. The pattern of non-uniform codon use (known as codon usage bias) varies between organisms and represents a unique feature of an organism. Organism-specific codon choice is related to organism-specific differences in populations of cognate tRNAs. This implies that, in a given organism, frequently used codons will be translated more rapidly than infrequently used ones and vice versa A theory of codon-tRNA co-evolution (necessary to balance accurate and efficient protein production) was put forward to explain the existence of codon usage bias. This model suggests that selection favours preferred (frequent) over un-preferred (rare) codons in order to sustain efficient protein production in cells and that a given un-preferred codon will have the same effect on an organism's fitness regardless of its position within an mRNA's open reading frame. However, many recent studies refute this prediction. Un-preferred codons have been found to have important functional roles and their effects appeared to be position-dependent. Synonymous codon usage affects the efficiency/stringency of mRNA decoding, mRNA biogenesis/stability, and protein secretion and folding. This review summarizes recent developments in the field that have identified novel functions of synonymous codons and their usage.

Nucleotides, Nucleosides, and Nucleobases

We review literature on the metabolism of ribo- and deoxyribonucleotides, nucleosides, and nucleobases in Escherichia coli and Salmonella,including biosynthesis, degradation, interconversion, and transport. Emphasis is placed on enzymology and regulation of the pathways, at both the level of gene expression and the control of enzyme activity. The paper begins with an overview of the reactions that form and break the N-glycosyl bond, which binds the nucleobase to the ribosyl moiety in nucleotides and nucleosides, and the enzymes involved in the interconversion of the different phosphorylated states of the nucleotides. Next, the de novo pathways for purine and pyrimidine nucleotide biosynthesis are discussed in detail.Finally, the conversion of nucleosides and nucleobases to nucleotides, i.e.,the salvage reactions, are described. The formation of deoxyribonucleotides is discussed, with emphasis on ribonucleotidereductase and pathways involved in fomation of dUMP. At the end, we discuss transport systems for nucleosides and nucleobases and also pathways for breakdown of the nucleobases.

Mapping codon usage of the translation initiation region in porcine reproductive and respiratory syndrome virus genome

BACKGROUND

Porcine reproductive and respitatory syndrome virus (PRRSV) is a recently emerged pathogen and severely affects swine populations worldwide. The replication of PRRSV is tightly controlled by viral gene expression and the codon usage of translation initiation region within each gene could potentially regulate the translation rate. Therefore, a better understanding of the codon usage pattern of the initiation translation region would shed light on the regulation of PRRSV gene expression.

RESULTS

In this study, the codon usage in the translation initiation region and in the whole coding sequence was compared in PRRSV ORF1a and ORFs2-7. To investigate the potential role of codon usage in affecting the translation initiation rate, we established a codon usage model for PRRSV translation initiation region. We observed that some non-preferential codons are preferentially used in the translation initiation region in particular ORFs. Although some positions vary with codons, they intend to use codons with negative CUB. Furthermore, our model of codon usage showed that the conserved pattern of CUB is not directly consensus with the conserved sequence, but shaped under the translation selection.

CONCLUSIONS

The non-variation pattern with negative CUB in the PRRSV translation initiation region scanned by ribosomes is considered the rate-limiting step in the translation process.

Codon preference optimization increases heterologous PEDF expression

Pigment epithelium-derived factor (PEDF) is widely known for its neurotrophic and antiangiogenic functions. Efficacy studies of PEDF in animal models are limited because of poor heterologous protein yields. Here, we redesigned the human PEDF gene to preferentially match codon frequencies of E coli without altering the amino acid sequence. Following de novo synthesis, codon optimized PEDF (coPEDF) and the wtPEDF genes were cloned into pET32a containing a 5′ thioredoxin sequence (Trx) and the recombinant Trx-coPEDF or Trx-wtPEDF fusion constructs expressed in native and two tRNA augmented E coli hosts - BL21-CodonPlus(DE3)-RIL and BL21-CodonPlus(DE3)-RP, carrying extra copies of tRNA^arg,ile,leu and tRNA^arg,pro genes , respectively. Trx-PEDF fusion proteins were isolated using Ni-NTA metal affinity chromatography and PEDF purified after cleavage with factor Xα. Protein purity and identity were confirmed by western blot, MALDI-TOF, and UV/CD spectral analyses. Expression of the synthetic gene was ∼3.4 fold greater (212.7 mg/g; 62.1 mg/g wet cells) and purified yields ∼4 fold greater (41.1 mg/g; 11.3 mg/g wet cell) than wtPEDF in the native host. A small increase in expression of both genes was observed in hosts supplemented with rare tRNA genes compared to the native host but expression of coPEDF was ∼3 fold greater than wtPEDF in both native and codon-bias-adjusted E coli strains. ΔGs at −3 to +50 of the Trx site of both fusion genes were −3.9 kcal/mol. Functionally, coPEDF was equally as effective as wtPEDF in reducing oxidative stress, promoting neurite outgrowth, and blocking endothelial tube formation. These findings suggest that while rare tRNA augmentation and mRNA folding energies can significantly contribute to increased protein expression, preferred codon usage, in this case, is advantageous to translational efficiency of biologically active PEDF in E coli. This strategy will undoubtedly fast forward studies to validate therapeutic utility of PEDF in vivo.

Translational machinery and protein folding: evidence of conformational variants of the estrogen receptor alpha

As an approach to understand how translation may affect protein folding, we analyzed structural and functional properties of the human estrogen receptor alpha synthesized by different eukaryotic translation systems. A minimum of three conformations of the receptor were detected using limited proteolysis and a sterol ligand-binding assay. The receptor in vitro translated in rabbit reticulocyte lysate was rapidly degraded by protease, produced major bands of about 34kDa and showed a high affinity for estradiol. In a wheat germ translation system, the receptor was more slowly digested. Two soluble co-existing conformations were evident by different degradation patterns and estradiol binding. Our data show that differences in the translation machinery may result in alternative conformations of the receptor with distinct sterol binding properties. These studies suggest that components of the cellular translation machinery itself might influence the protein folding pathways and the relative abundance of different receptor conformers.

Correlations of nucleotide substitution rates and base composition of mammalian coding sequences with protein structure

We investigated the relationships between the nucleotide substitution rates and the predicted secondary structures in the three states representation (alpha-helix, beta-sheet, and coil). The analysis was carried out on 34 alignments, each of which comprised sequences belonging to at least four different mammalian orders. The rates of synonymous substitution were found to be significantly different in regions predicted to be alpha-helix, beta-sheet, or coil. Likewise, the nonsynonymous rates also differ, although expectedly at a lower extent, in the three types of secondary structure, suggesting that different selective constraints associated with the different structures are affecting in a similar way the synonymous and nonsynonymous rates. Moreover, the base composition of the third codon positions is different in coding sequence regions corresponding to different secondary structures of proteins.

How optimized is the translational machinery in Escherichia coli, Salmonella typhimurium and Saccharomyces cerevisiae?

The optimization of the translational machinery in cells requires the mutual adaptation of codon usage and tRNA concentration, and the adaptation of tRNA concentration to amino acid usage. Two predictions were derived based on a simple deterministic model of translation which assumes that elongation of the peptide chain is rate-limiting. The highest translational efficiency is achieved when the codon recognized by the most abundant tRNA reaches the maximum frequency. For each codon family, the tRNA concentration is optimally adapted to codon usage when the concentration of different tRNA species matches the square-root of the frequency of their corresponding synonymous codons. When tRNA concentration and codon usage are well adapted to each other, the optimal content of all tRNA species carrying the same amino acid should match the square-root of the frequency of the amino acid. These predictions are examined against empirical data from Escherichia coli, Salmonella typhimurium, and Saccharomyces cerevisiae.

Non-random usage of 'degenerate' codons is related to protein three-dimensional structure

We report an analysis of a novel sequence-structure database of mammalian proteins incorporating nucleotide sequences of the exon regions of their genes together with protein sequence and structural information. We find that synonymous codon families (i.e. coding the same residue) have non-random codon distribution frequencies between protein secondary structure types. Their structural preferences are related to the third, 'silent' nucleotide position in a codon. We also find that some synonymous codons show very different or even opposite structural preferences at the N- or C-termini of structure fragments, relative to those observed for their amino acid residues.

Protein secondary structural types are differentially coded on messenger RNA

Tricodon regions on messenger RNAs corresponding to a set of proteins from Escherichia coli were scrutinized for their translation speed. The fractional frequency values of the individual codons as they occur in mRNAs of highly expressed genes from Escherichia coli were taken as an indicative measure of the translation speed. The tricodons were classified by the sum of the frequency values of the constituent codons. Examination of the conformation of the encoded amino acid residues in the corresponding protein tertiary structures revealed a correlation between codon usage in mRNA and topological features of the encoded proteins. Alpha helices on proteins tend to be preferentially coded by translationally fast mRNA regions while the slow segments often code for beta strands and coil regions. Fast regions correspondingly avoid coding for beta strands and coil regions while the slow regions similarly move away from encoding alpha helices. Structural and mechanistic aspects of the ribosome peptide channel support the relevance of sequence fragment translation and subsequent conformation. A discussion is presented relating the observation to the reported kinetic data on the formation and stabilization of protein secondary structural types during protein folding. The observed absence of such strong positive selection for codons in non-highly expressed genes is compatible with existing theories that mutation pressure may well dominate codon selection in non-highly expressed genes.

Ribosome-mediated translational pause and protein domain organization

Because regions on the messenger ribonucleic acid differ in the rate at which they are translated by the ribosome and because proteins can fold cotranslationally on the ribosome, a question arises as to whether the kinetics of translation influence the folding events in the growing nascent polypeptide chain. Translationally slow regions were identified on mRNAs for a set of 37 multidomain proteins from Escherichia coli with known three-dimensional structures. The frequencies of individual codons in mRNAs of highly expressed genes from E. coli were taken as a measure of codon translation speed. Analysis of codon usage in slow regions showed a consistency with the experimentally determined translation rates of codons; abundant codons that are translated with faster speeds compared with their synonymous codons were found to be avoided; rare codons that are translated at an unexpectedly higher rate were also found to be avoided in slow regions. The statistical significance of the occurrence of such slow regions on mRNA spans corresponding to the oligopeptide domain termini and linking regions on the encoded proteins was assessed. The amino acid type and the solvent accessibility of the residues coded by such slow regions were also examined. The results indicated that protein domain boundaries that mark higher-order structural organization are largely coded by translationally slow regions on the RNA and are composed of such amino acids that are stickier to the ribosome channel through which the synthesized polypeptide chain emerges into the cytoplasm. The translationally slow nucleotide regions on mRNA possess the potential to form hairpin secondary structures and such structures could further slow the movement of ribosome. The results point to an intriguing correlation between protein synthesis machinery and in vivo protein folding. Examination of available mutagenic data indicated that the effects of some of the reported mutations were consistent with our hypothesis.

Total synthesis and expression in Escherichia coli of a gene encoding human tropoelastin

To elucidate the structural features and interactions of tropoelastin (TEL) molecules which assist in giving the elastic fibre its physical properties, a 2210-bp synthetic human TEL-encoding gene (SHEL) was constructed for expression in Escherichia coli. To this end, a model of codon adjustment was tested which better suits the polypeptide biosynthetic needs of E. coli than the human sequence, where over one-third of this natural sequence contains expression-limiting rare codons and 4 amino acids alone account for 75% of the resulting polypeptide. This large synthetic TEL gene was expressed at a high level as the recombinant counterpart of human TEL and as a C-terminal fusion with glutathione S-transferase. This demonstrates that a synthetic approach based upon matching codon usage to that of the host organism can support significant expression of recombinant sequences. The synthetic gene incorporates the facility for simple cassette replacement in future insertion, deletion and mutagenesis experiments, including the introduction and removal of exon homologues. The resulting soluble polypeptide is easily purified and displays properties expected for this protein.

The Escherichia coli K-12 "wild types" W3110 and MG1655 have an rph frameshift mutation that leads to pyrimidine starvation due to low pyrE expression levels

The widely used and closely related Escherichia coli "wild types" W3110 and MG1655, as well as their common ancestor W1485, starve for pyrimidine in minimal medium because of a suboptimal content of orotate phosphoribosyltransferase, which is encoded by the pyrE gene. This conclusion was based on the findings that (i) the strains grew 10 to 15% more slowly in pyrimidine-free medium than in medium containing uracil; (ii) their levels of aspartate transcarbamylase were highly derepressed, as is characteristic for pyrimidine starvation conditions; and (iii) their levels of orotate phosphoribosyltransferase were low. After introduction of a plasmid carrying the rph-pyrE operon from strain HfrH, the growth rates were no longer stimulated by uracil and the levels of aspartate transcarbamylase were low and similar to the levels observed for other strains of E. coli K-12, E. coli B, and Salmonella typhimurium. To identify the mutation responsible for these phenotypes, the rph-pyrE operon of W3110 was cloned in pBR322 from Kohara bacteriophage lambda 2A6. DNA sequencing revealed that a GC base pair was missing near the end of the rph gene of W3110. This one-base-pair deletion results in a frame shift of translation over the last 15 codons and reduces the size of the rph gene product by 10 amino acid residues relative to the size of RNase PH of other E. coli strains, as confirmed by analysis of protein synthesis in minicells. The truncated protein lacks RNase PH activity, and the premature translation stop in the rph cistron explains the low levels of orotate phosphoribosyltransferase in W3110, since close coupling between transcription and translation is needed to support optimal levels of transcription past the intercistronic pyrE attenuator.

Effects of consecutive AGG codons on translation in Escherichia coli, demonstrated with a versatile codon test system

A system for testing the effects of specific codons on gene expression is described. Tandem test and control genes are contained in a transcription unit for bacteriophage T7 RNA polymerase in a multicopy plasmid, and nearly identical test and control mRNAs are generated from the primary transcript by RNase III cleavages. Their coding sequences, derived from T7 gene 9, are translated efficiently and have few low-usage codons of Escherichia coli. The upstream test gene contains a site for insertion of test codons, and the downstream control gene has a 45-codon deletion that allows test and control mRNAs and proteins to be separated by gel electrophoresis. Codons can be inserted among identical flanking codons after codon 13, 223, or 307 in codon test vectors pCT1, pCT2, and pCT3, respectively, the third site being six codons from the termination codon. The insertion of two to five consecutive AGG (low-usage) arginine codons selectively reduced the production of full-length test protein to extents that depended on the number of AGG codons, the site of insertion, and the amount of test mRNA. Production of aberrant proteins was also stimulated at high levels of mRNA. The effects occurred primarily at the translational level and were not produced by CGU (high-usage) arginine codons. Our results are consistent with the idea that sufficiently high levels of the AGG mRNA can cause essentially all of the tRNA(AGG) in the cell to become sequestered in translating peptidyl-tRNA(AGG) -mRNA-ribosome complexes stalled at the first of two consecutive AGG codons and that the approach of an upstream translating ribosome stimulates a stalled ribosome of frameshift, hop, or terminate translation.

Novel in-frame two codon translational hop during synthesis of bovine placental lactogen in a recombinant strain of Escherichia coli

A recombinant Escherichia coli strain was constructed for the overexpression of bovine placental lactogen (bPL), using a bPL structural gene containing 9 of the rare arginine codons AGA and AGG. When high level bPL synthesis was induced in this strain, cell growth was inhibited and bPL accumulated to less than 10% of total cell protein. In addition, about 2% of the recombinant bPL produced from this strain exhibited an altered trypsin digestion pattern. Amino acid residues 74 through 109 normally produce 2 tryptic peptides, but the altered form of bPL lacked these two peptides and instead had a new peptide which was missing arginine residue 86 and one of the two flanking leucine residues. The codon for arginine residue 86 was AGG and the codons for the flanking leucine residues 85 and 87 were TTG. When 5 of the 9 AGA and AGG codons in the bPL structural gene were changed to more preferred arginine codons, cell growth was not inhibited and bPL accumulated to about 30% of total cell protein. When bPL was purified from this modified strain, which included changing the arginine codon at position 86 from AGG to CGT, none of the altered form of bPL was produced. These observations are consistent with a model in which translational pausing occurs at the arginine residue 86 AGG codon because the corresponding arginyl-tRNA species is reduced by the high level of bPL synthesis, and a translational hop occurs from the leucine residue 85 TTG codon to the leucine residue 87 TTG codon. This observation represents the first report of an error in protein synthesis due to an in-frame translational hop within an open reading frame.

Identification of the rph (RNase PH) gene of Bacillus subtilis: evidence for suppression of cold-sensitive mutations in Escherichia coli

A shotgun cloning of Bacillus subtilis DNA into pBR322 yielded a 2-kb fragment that suppresses the cold-sensitive defect of the nusA10(Cs) Escherichia coli mutant. The responsible gene encodes an open reading frame that is greater than 50% identical at the amino acid level to the E. coli rph gene, which was formerly called orfE. This B. subtilis gene is located at 251 degrees adjacent to the gerM gene on the B. subtilis genetic map. It has been named rph because, like its E. coli analog, it encodes a phosphate-dependent exoribonuclease activity, RNase PH, that removes the 3' nucleotides from precursor tRNAs. The cloned B. subtilis rph gene also suppresses the cold-sensitive phenotype of other unrelated cold-sensitive mutants of E. coli, but not the temperature-sensitive phenotype of three temperature-sensitive mutants, including the nusA11(Ts) mutant, that were tested.

Attenuation in the rph-pyrE operon of Escherichia coli and processing of the dicistronic mRNA

We have substituted on a plasmid the native promoter of the Escherichia coli rph-pyrE operon with an inducible transcription-initiation signal. The plasmid was used to study the mRNA chains derived from the operon at different intracellular concentrations of UTP and as a function of time following induction of transcription. The results showed that dicistronic rph-pyrE mRNA was formed when the UTP pool was low, and that a monocistronic rph mRNa was the major transcription product in high-UTP pools, thus supporting an UTP-controlled attenuation mechanism for regulation of pyrE gene expression. However, the dicistronic rph-pyrE transcript was rapidly processed into two monocistronic mRNA units, and a cleavage site was mapped near the attenuator in the intercistronic region, close to the site where transcription was terminated in high-UTP pools. Furthermore, the major 3' end of the pyrE mRNA was mapped near a palindromic structure of similarity to the family of repetitive extragenic palindromic sequences, 35 nucleotide residues after stop codon of the pryE gene.

Importance of mRNA folding and start codon accessibility in the expression of genes in a ribosomal protein operon of Escherichia coli

The trmD operon of Escherichia coli consists of the genes for the ribosomal protein (r-protein) S16, a 21 kilodalton protein (21K) of unknown function, the tRNA(m1G37)methyltransferase (TrmD), and r-protein L19, in that order. The synthesis of the 21K and TrmD proteins is 12 and 40-fold lower, respectively, than that of the two r-proteins, although the corresponding parts of the mRNA are equally abundant. This translational control of expression of at least the 21K protein gene (21K), is mediated by a negative control element located between codons 18 and 50 of 21K. Here, we present evidence for a model in which mRNA sequences up to around 100 nucleotides downstream from the start codon of 21K fold back and base-pair to the 21K translation initiation region, thereby decreasing the translation initiation frequency. Mutations in the internal negative control element of 21K that would prevent the formation of the proposed mRNA secondary structure over both the Shine-Dalgarno (SD) sequence and the start codon increased expression up to about 20-fold, whereas mutations that would disrupt the base-pairing with the SD-sequence had only relatively small effects on expression. In addition, the expression increased 12-fold when the stop codon of the preceding gene, rpsP, was moved next to the SD-sequence of 21K allowing the ribosomes to unfold the postulated mRNA secondary structure. The expression increased up to 150-fold when that stop codon change was combined with the internal negative control element base-substitutions that derepressed translation about 20-fold. The negative control element of 21K does not seem to be responsible for the low expression of the trmD gene located downstream. However, a similar negative control element native to trmD can explain at least partly the low expression of trmD. Possibly, the two mRNA secondary structures function to decouple translation of 21K and trmD from that of the respective upstream cistron in order to achieve their independent regulation.

Escherichia coli orfE (upstream of pyrE) encodes RNase PH

RNase PH from extracts of Escherichia coli was purified to homogeneity and subjected to NH2-terminal sequencing. Comparison of this sequence with all open reading frames in the GenBank data base revealed at least 95% identity to an unidentified open reading frame (orfE) upstream of pyrE at 81.7 min on the E. coli chromosome. Clones of orfE overexpress RNase PH activity, verifying that orfE encodes this ribonuclease. We suggest that orfE be renamed rph.

Low-usage codons in Escherichia coli, yeast, fruit fly and primates

Codon usage is compared between four classes of species, with an emphasis on characterization of low-usage codons. The classes of species analyzed include the bacterium Escherichia coli (ECO), the yeast Saccharomyces cerevisiae (YSC), the fruit fly Drosophila melanogaster (DRO), and several species of primates (PRI) (taken as a group; includes eleven species for which nucleotide sequence data have been reported to GenBank, however, greater than 90% of the sequences were from Homo sapiens). The number of protein-coding sequences analyzed were 968 for ECO, 484 for YSC, 244 for DRO, and 1518 for PRI. Three methods have been used to determine low-usage codons in these species. The first and most common way of assessing codon usage is by summing the number of time codons appear in reading frames of the genome in question. The second way is to examine the distribution of usage in different genes by scoring the number of protein reading frames in which a particular codon does not appear. The third way starts with a similar notion, but instead considers combinations of codons that are missing from the maximum number of genes. These three methods give very similar results. Each species has a unique combination of eight least-used codons, but all species contain the arginine codons, CGA and CGG. The agreement between YSC and PRI is particularly striking as they share six low-usage codons. All six carry the dinucleotide sequence, CG. The eight least-used codons in PRI include all codons that contain the CG dinucleotide sequence. Low-usage codons are clearly avoided in genes encoding abundant proteins for ECO, YSC DRO. In all species, proteins containing a high percentage of low-usage codons could be characterized as cases where an excess of the protein could be detrimental. Low codon usage is relatively insensitive to gross base composition. However, dinucleotide usage can sometimes influence codon usage. This is particularly notable in the case of CG dinucleotides in PRI.

Codon bias and gene expression

The frequencies with which individual synonymous codons are used to code their cognate amino acids is quite variable from genome to genome and within genomes, from gene to gene. One particularly well documented codon bias is that associated with highly expressed genes in bacteria as well as in yeast; this is the so-called major codon bias. Here, it is suggested that the major codon bias is not an arrangement for regulating individual gene expression. Instead, the data suggest that this codon bias, which is correlated with a corresponding bias of tRNA abundance, is a global arrangement for optimizing the growth efficiency of cells. On the practical side, it is suggested that heterologous gene expression is not as sensitive to codon bias as previously thought, but that it is quite sensitive to other characteristics of the heterologous gene.

Unusual codon bias occurring within insertion sequences in Escherichia coli

The large open reading frames of insertion sequences from Escherichia coli were examined for their spatial pattern of codon usage bias and distribution of rarely used codons. There is a bias in codon usage that is generally lower toward the terminal ends of the coding regions, which is reflected in the occurrence of an excess of nonpreferred codons in the 3' portions of the coding regions as compared with the 5' portions. In contrast, typical chromosomal genes have a lower codon usage bias toward the 5' ends of the coding regions. These results imply that the selective forces reflected in codon usage bias may differ according to position within the coding sequence. In addition, these constraints apparently differ in important ways between genes contained in insertion sequences and those in the chromosome.

The effects of leader peptide sequence and length on attenuation control of the trp operon of E.coli

We have examined the effects of changing the length and codon content of the trp leader peptide coding region on expression of the trp operon of Escherichia coli, it had previously been shown that coupling of transcription and translation in the trp leader region is essential for both basal level control and tryptophan starvation control of transcription attenuation in this operon. We have found that increasing the length of the leader peptide coding region by 55 codons allowed normal basal level control and normal tryptophan starvation control. As expected, the presence of a nonsense codon early in the leader peptide coding region decreased basal expression and eliminated starvation control. Introducing tandem rare codons had no effect on basal level expression, but eliminated the tryptophan starvation response. Frameshifting at tandem rare codons was tested as the most likely explanation for loss of the tryptophan starvation response, but the results were inconclusive.

Three genes preceding pyrE on the Escherichia coli chromosome are essential for survival and normal cell morphology in stationary culture and at high temperature

Previous studies of the upstream region of the pyrE gene in Escherichia coli revealed three genes of unknown function. Inactivation of these genes (designated orfE, orfX and orfY) by crossing the KmR-cassette-disrupted orf into the chromosome indicated that they were not required during exponential growth (Poulsen et al., Mol., Microbiol., 1989 b). Here we report that the three genes are of importance in the stationary phase. Thus, cultures of the mutants grown to a stationary state in rich media contained bacterial filaments of abnormal morphology. In addition, flow cytometric analyses showed that outgrown cultures of the orf mutants have anomalous size distribution and DNA content, and that rifampicin treatment of exponentially growing mutants results in cell populations with chromosome numbers in the range from about 1 to 10, compared with wild type strains that end up with 4 and 8 full chromosomes. Finally, it appeared that the three orf's are indispensable at high temperatures since the insertion mutants were unable to form colonies above 45 degrees C and since cultures of exponentially growing mutants lysed upon a temperature shift from 37 degrees C to 45 degrees C.

Growth rate dependence of global amino acid composition

The global amino acid composition of bacteria growing in different media has been studied. The data reveal significant changes in the amino acid composition in the growth rate range between 0.5 and 2.1 doublings per hour at 37 degrees C. The changes are consistent with a progressive simplification of the protein population and mRNA pools as the growth rates increase.

Codon preferences in free-living microorganisms

A popular interpretation of the major codon preference is that it reflects the operation of a regulatory device that controls the expression of individual proteins. In this popular model, rapidly translated codons are thought to promote the accumulation of the highly expressed proteins and slowly translated codons are thought to retard the expression of poorly expressed proteins. However, this widely accepted model is not supported by kinetic theory or by experimental results. A less fashionable model in which the major codon preference has nothing to do with the expression level of the individual proteins is forwarded. In this model, the major codon preference is viewed as a global strategy to support the efficient function of the translation system and thereby to maximize the growth rates of cells under favorable conditions.

Antisuppression by mutations in elongation factor Tu

Two slow-growing kirromycin-resistant Escherichia coli mutants with altered EF-Tu (Ap and Aa) were studied in vivo in strains with an inactive tufB gene. Mutant form Aa was isolated as an antisuppressor of the tyrT(Su3) nonsense suppressor, as described here. Ap, the tufA gene product of strain D2216 (from A. Parmeggiani), has previously been shown to give an increased GTPase activity. The slow cellular growth rates of both EF-Tu mutants are correlated with decreased translational elongation rates. Ap and Aa significantly decrease suppression levels of both nonsense and missense suppressor tRNAs [tyrT(Su3), trpT(Su9), glyT(SuAGA/G)], but have only little or no effect on misreading by wild-type tRNAs. A particular missense suppressor, lysT(SuAAA/G), which acts by virtue of partial mischarging as the result of an alteration in the amino acid stem, is not significantly affected by the EF-Tu mutations. The combination of tufA(Aa) and a rpsD12 ribosomal mutation is lethal at room temperature and the double-mutant strain has an elevated temperature optimum (42 degrees C) for growth rate, translation rate and nonsense suppression. Our data indicate an alterated interaction between Aa and the ribosome, consistent with our in vitro results.

Expression of tetanus toxin fragment C in E. coli: high level expression by removing rare codons

Tetanus toxin fragment C had been previously expressed in Escherichia coli at 3-4% cell protein. The codon bias for tetanus toxin in Clostridium tetani is very different from that of highly expressed homologous genes in E. coli, resulting in the presence of many rare E. coli codons in the sequence encoding fragment C. We have replaced the coding sequence by sequence optimized for codon usage in E. coli, and show that the expression of fragment C is increased. Although the level of mRNA also increased this appeared to be a secondary consequence of more efficient translation. Complete sequence replacement increased expression to approximately 11-14% cell protein but only after the promoter strength had been improved.

A regulatory element within a gene of a ribosomal protein operon of Escherichia coli negatively controls expression by decreasing the translational efficiency

The trmD operon of Escherichia coli consists of the genes for the ribosomal protein (r-protein) S16, a 21 kDa protein (21K) of unknown function, the tRNA(m1G37)methyltransferase (TrmD), and r-protein L19, in this order. Previously we have shown that the steady-state amount of the two r-proteins exceeds that of the 21K and TrmD proteins 12- and 40-fold, respectively, and that this differential expression is solely explained by translational regulation. Here we have constructed translational gene fusions of the trmD operon and lacZ. The expression of a lacZ fusion containing the first 18 codons of the 21K protein gene is 15-fold higher than the expression of fusions containing 49 or 72 codons of the gene. This suggests that sequences between the 18th and the 49th codon may act as a negative element controlling the expression of the 21K protein gene. Evidence is presented which demonstrates that this regulation is achieved by reducing the efficiency of translation.

Translation rates of individual codons are not correlated with tRNA abundances or with frequencies of utilization in Escherichia coli

We analyzed 12 individual codons, which differed widely with respect to the frequency of use in Escherichia coli and the abundance of the corresponding tRNAs, for their influence on the coupling between transcription and translation. This was probed by determining the effects of codon substitutions in the leader peptide gene on transcription past the pyrE attenuator, as described previously by Bonekamp et al. (F. Bonekamp, H. D. Andersen, T. Christensen, and K. F. Jensen, Nucleic Acids Res. 13:4113-4123, 1985). In principle, the results revealed that either RNA polymerase or the (leading) ribosomes pass the different codon strings at different rates. However, under the assumption that the rate of transcription elongation is unaffected by the sequence changes, the results may be interpreted as indicating that different codons are translated at different rates and that these rates do not generally reflect the concentrations of the corresponding tRNAs or the frequencies with which the codons are used in E. coli. Moreover, it seems that codon synonyms that are served by the same isoaccepting tRNA species can deviate as much from each other in translational behavior as synonymous codons that are served by isoacceptors present in the cell in widely different amounts can.

Rates of aminoacyl-tRNA selection at 29 sense codons in vivo

We have placed aminoacyl-tRNA selection at individual codons in competition with a frameshift that is assumed to have a uniform rate. By assaying a reporter in the shifted frame, relative rates for association of the 29 YNN codons and their cognate aminoacyl-tRNAs were obtained during logarithmic growth in Escherichia coli. For five codons, three beginning with C and two with U, these relative rates agree with relative in vitro rates for elongation factor Tu-mediated aminoacyl-tRNA binding to ribosomes and subsequent GTP hydrolysis. Therefore, the frameshift assay probably measures this process in vivo. Observed rates for aminoacyl-tRNA selection span a 25-fold range. Therefore, the time required to transit different codons in vivo probably differs substantially. Codons very frequently used in highly expressed genes generally select aminoacyl-tRNAs more quickly than do rarely used codons. This suggests that speed of aminoacyl-tRNA selection is a significant factor determining biased use of synonymous codons. However, the preferential use of codons appears to be marked only for codons with the highest rates of aminoacyl-tRNA selection. Rapid selection in vivo is usually effected by elevation of the tRNA concentration for codons with moderate intrinsic speed (rate constant), not by choosing intrinsically fast codons. Despite a preference for high rate, there are quickly translated codons that are not commonly used, and common codons that are translated relatively slowly. Other factors are therefore more important than speed for some codons. Strong preference for rapid aminoacyl-tRNA selection is not observed in weakly expressed genes. Instead, there is a slight preference for slower aminoacyl-tRNA selection. The rate of aminoacyl-tRNA selection by a YNC codon is always greater than the rate of the corresponding YNU codon even though in many YNC/U pairs both codons react with the same elongation factor Tu/GTP/aminoacyl-tRNA complex. Thus, for these tRNAs, the differences between in vivo rate constants of tRNAs are dependent on the nature of anticodon base-pairing. However, no more general relationship is evident between codon/anticodon composition and rate of aminoacyl-tRNA selection. The frameshift method can be extended to all codons.

Codon usage determines translation rate in Escherichia coli

We wish to determine whether differences in translation rate are correlated with differences in codon usage or with differences in mRNA secondary structure. We therefore inserted a small DNA fragment in the lacZ gene either directly or flanked by a few frame-shifting bases, leaving the reading frame of the lacZ gene unchanged. The fragment was chosen to have "infrequent" codons in one reading frame and "common" codons in the other. The insert in these constructs does not seem to give mRNAs that are able to form extensive secondary structures. The translation time for these modified lacZ mRNAs was measured with a reproducibility better than plus or minus one second. We found that the mRNA with infrequent codons inserted has an approximately three-seconds longer translation time than the one with common codons. In another set of experiments we constructed two almost identical lacZ genes in which the lacZ mRNAs have the potential to generate stem structures with stabilities of about -75 kcal/mol. In this way we could investigate the influence of mRNA structure on translation rate. This type of modified gene was generated in two reading frames with either common or infrequent codons similar to our first experiments. We find that the yield of protein from these mRNAs is reduced, probably due to the action in vivo of an RNase. Nevertheless, the data do not indicate that there is any effect of mRNA secondary structure on translation rate. In contrast, our data persuade us that there is a difference in translation rate between infrequent codons and common codons that is of the order of sixfold.

Molecular and mutational analysis of three genes preceding pyrE on the Escherichia coli chromosome

The nucleotide sequence of two kilobase pairs (kb) 5' to the orfE-pyrE operon has been determined. The sequence revealed two open reading frames, orfX and orfY, consisting of 286 and 274 codons, respectively, and having a transcriptional orientation opposite that of the orfE-pyrE operon. Analysis of transcription initiations showed that the promoters of orfE and orfX constitute a pair of divergent promoters with overlapping -35 regions and that orfY is transcribed from an independent promoter. Translational analysis indicated that the orfs are expressed in Escherichia coli. The orfE, orfX, and orfY genes were inactivated on the bacterial chromosome by deletion-insertion mutagenesis using a kanamycin resistance cassette. The mutants were all viable. However, the orfE deletion caused a dramatic reduction in the level of pyrE expression and a partial pyrimidine requirement, because this mutation prevented transcription of pyrE. the orfE protein seemed without significance for pyr-gene expression in E. coli, and the mutations in orfX and orfY were without detectable phenotypes.

Nucleotide sequence determination of the DNA region coding for Bacillus stearothermophilus glyceraldehyde-3-phosphate dehydrogenase and of the flanking DNA regions required for its expression in Escherichia coli

The complete nucleotide sequence of a 3541-base pairs (bp) DNA fragment from Bacillus stearothermophilus able to complement an Escherichia coli glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) mutant (gapD-) has been determined. The B. stearothermophilus gap gene consists of a 1005-bp open reading frame commencing with an ATG start codon and ending with a TAA stop codon. Upstream from the start codon is a strong Shine-Dalgarno sequence typical of Gram-positive bacteria. Only one putative RNA polymerase recognition signal (-35 and -10 regions) is found 1153 bp upstream from the ATG start codon. In vivo utilization of this signal is in agreement with the study of gene expression from different subclones of the original fragment. 57 bp downstream from the TAA stop codon is a structure resembling Rho-independent transcription termination signals. Although B. stearothermophilus GAPDH-coding gene is highly expressed in E. coli, it contains several rare codons for E. coli. The predicted amino acid sequence of the GAPDH enzyme presents several differences with the amino acid sequence previously determined from the protein and is in better agreement with published crystallographic data.

Hyper-regulation of pyr gene expression in Escherichia coli cells with slow ribosomes. Evidence for RNA polymerase pausing in vivo?

UTP-modulated attenuation of transcription is involved in regulating the synthesis of pyrimidine nucleotides in Escherichia coli. Thus, expression of two genes, pyrBI and pyrE, was shown to be under this type of control. The genes encode the two subunits of aspartate transcarbamylase and orotate phosphoribosyltransferase respectively. The levels of these enzymes are inversely correlated with the intracellular concentration of UTP. Modulation of attenuation seems to be a consequence of the effect of UTP concentration on the mRNA chain growth rate. Reducing the UTP pool retards RNA polymerase movement. Mechanistically this will couple the ribosomes translating a leader peptide gene more tightly to the elongating RNA polymerase. The ribosomes will then be more prone to prevent the folding of the mRNA chains into terminating hairpin structures when RNA polymerase is at the attenuator and has to decide whether transcription should terminate or continue into the structural genes. This paper described a study of pyrBI and pyrE gene regulation in cells where the ribosomes move slowly as a result of mutation in rpsL. It appears that expression of the two genes is hyper-regulated by the UTP pool in this type of cells. Furthermore, the attenuator model can only account for the results if it is assumed that UTP-concentration-dependent pausing of transcription occurs in vivo in the two pyr gene leaders such that RNA polymerase waits for the coupled ribosomes before transcribing into the attenuator regions.

Translation initiation controls the relative rates of expression of the bacteriophage lambda late genes

The late operon of bacteriophage lambda contains the genes encoding the morphogenetic proteins of the phage. These genes are transcribed equally from the single late promoter. Although the functional half-lives of the mRNA for the various genes of this operon vary less than 2-fold, their relative rates of expression have been shown to vary by nearly 1000-fold. This variation could result from differing rates of translation initiation, from overlapping upstream translation, or from differential elongation rates due to the presence of codons for which the corresponding tRNAs are rare. To distinguish between these possibilities, we have cloned sequences surrounding the initiator codons of several of these genes and measured their ability to drive synthesis of hybrid lambda-beta-galactosidase proteins. The rates of expression of the hybrid genes thus produced correlate very well with the natural rates of expression of the corresponding phage genes, suggesting that the rate of initiation of translation controls the relative expression rates of these genes.

The AGG codon is translated slowly in E. coli even at very low expression levels

Data are presented which indicate that AGG codons for arginine are translated significantly more slowly than the CGU codons for the same amino acid even when their expression level from the probe is very low. The two types of codons were inserted (three in tandem) on a multicopy plasmid in an artificial leader peptide gene in front of the pyrE attenuator where the frequency of transcription termination is regulated by the degree of coupling between transcription and translation. Transcription of the operon is initiated from the lac-promoter dependent on the concentration of the lac-operon inducer IPTG. At all induction levels it was found that the frequency of transcription past the pyrE attenuator was approximately nine times lower when the AGG codons were present in the leader than with CGT codons present. This shows that AGG codons decouple translation from transcription in the pyrE attenuator region even when the concentration of this codon is not increased significantly relative to that in the unperturbed wild type strain. Thus the results indicate that AGG codons are always slowly translated in Escherichia coli.

Synonymous codon usage in Bacillus subtilis reflects both translational selection and mutational biases

Codon usage data for 56 Bacillus subtilis genes show that synonymous codon usage in B. subtilis is less biased than in Escherichia coli, or in Saccharomyces cerevisiae. Nevertheless, certain genes with a high codon bias can be identified by correspondence analysis, and also by various indices of codon bias. These genes are very highly expressed, and a general trend (a decrease) in codon bias across genes seems to correspond to decreasing expression level. This, then, may be a general phenomenon in unicellular organisms. The unusually small effect of translational selection on the pattern of codon usage in lowly expressed genes in B. subtilis yields similar dinucleotide frequencies among different codon positions, and on complementary strands. These patterns could arise through selection on DNA structure, but more probably are largely determined by mutation. This prevalence of mutational bias could lead to difficulties in assessing whether open reading frames encode proteins.

Translational control of phage f1 gene expression by differential activities of the gene V, VII, IX and VIII initiation sites

Phage-specific transcription and subsequent RNA processing in Escherichia coli infected with the filamentous phage (f1, M13, fd) generate a pool of abundant and relatively long-lived phage mRNA species encoding the four adjacent genes V, VII, IX and VIII. Yet the products of gene V and gene VIII are synthesized at much higher levels than the gene VII and gene IX proteins. To ask if the translational initiation sites heading these genes show corresponding differences in activity and/or functional properties, we have purified a number of the phage mRNAs from cells infected with f1 and examined them in in vitro initiation reactions. The ribosome binding patterns obtained for the phage mRNA species and for smaller defined RNA fragments containing selected initiator regions reveal a large range in apparent ribosome binding strengths. The gene V and gene VIII sites are recognized efficiently in each mRNA species in which they are present. Gene IX site activity appears to be limited by local mRNA structure: the site has undetectable or low ribosome binding activity in all of the phage mRNA species, but is at least tenfold more active if the RNA sequences required to form a potential hairpin stem-and-loop 15 nucleotides upstream from the initiator AUG have been removed. The gene VII site shows no evidence of interaction with ribosomes in any phage mRNA or RNA fragment tested. The same striking differences in initiation activity were observed in vivo by cloning small f1 DNA fragments containing gene V or gene VII initiation site sequences to drive beta-galactosidase synthesis. High levels of a gene V-beta-galactosidase fusion protein are initiated at the V site, but no detectable synthesis occurs from the VII site. If the VII site is preceded by all of the information encoding the upstream gene V, however, modest amounts of a fusion protein initiated at the VII site are produced. The overall results, in accord with the observed yields of proteins in the phage-infected cell, provide strong evidence that the properties of these translational initiation sites determine in a significant way the differential expression of phage f1 genes V, VII, IX and VIII.

The functional stability of the lacZ transcript is sensitive towards sequence alterations immediately downstream of the ribosome binding site

Various synthetic DNA sequences were inserted downstream of the fourth codon of the Escherichia coli lacZ gene on plasmids containing a hybrid lacZ-galK operon. Several different sequences, one as short as 10 bp, reduced the functional stability of the lacZ message three- to fourfold, whereas others had little or no effect. Introduction of synthetic sequences into a plasmid containing the intact lac operon resulted in similar reductions of mRNA stability. The sequence alterations also reduced the translational efficiency and transcription through lacZ as monitored by measurements of galactokinase synthesis from the downstream galK gene. There was no correlation between the average translational frequency and the stability of the lacZ message indicating that some of the inserted sequences reduced mRNA stability directly and not as a consequence of their effect on translation. The reduction of transcription through the lacZ gene correlated with the reduction of translation in agreement with current models of transcriptional polarity.

Lysis gene of bacteriophage MS2 is activated by translation termination at the overlapping coat gene

The 3' boundary of the coat gene of the RNA bacteriophage MS2 lies 46 nucleotides downstream from the beginning of the lysis (L) cistron. The translation of both reading frames is coupled; the synthesis of the lysis protein does not occur unless translation of the overlapping coat gene takes place. In the preceding paper we showed that de novo initiation at the L gene is prevented by a hairpin structure that sequesters the ribosomal binding site. Here we examine how translation of the coat gene activates the L gene start site. The experiments show that the movement of ribosomes through the hairpin is in itself not sufficient to expose the lysis gene. Rather, the endpoint of translation is important. Termination at the natural end of the coat gene triggers the lysis response, but further downstream terminations do not. Activation of the L gene is suppressed when the stability of the lysis initiator hairpin is increased by mutations that create additional base-pairs. We assume that the ribosome, terminating at the coat reading frame, covers part of the lysis hairpin, thereby destabilizing the secondary structure. This may be sufficient to promote the binding of a vacant ribosome to the L gene start. Alternatively, the terminated but not yet released ribosome may reach the L gene start by random lateral movements along the mRNA and reinitiate there. The present findings are also discussed in relation to an earlier proposal for L gene activation.

Effect of UTP and GTP pools on attenuation at the pyrE gene of Escherichia coli

We have used the galK gene, minus its promoter, to quantitate transcription of the orfE--pyrE operon of Escherichia coli in front of and after the intercistronic attenuator. Expression of the hybrid genes was studied in a bacterium with mutations that permit changes in the UTP and GTP pools during exponential growth. It was found that the greater part of pyrE gene regulation by the nucleotides takes place at the intercistronic attenuator and that promoter control contributes only little, ca. twofold. When pools of both UTP and GTP were high only 5%-6% of the mRNA chains were continued into the pyrE gene. However, when the UTP pool was reduced (from 1.3 to 0.2 mumol/g dry weight) nearly 100% of transcription passed the attenuator. Likewise, a reduction in the GTP pool (from 3.2 to 0.8 mumol/g dry weight) resulted in 25%-30% escape of attenuation. Regulation by attenuation disappeared when a premature stop-codon was introduced near the end of orfE such that translational coupling to transcription was prevented in the attenuator area. Therefore, we attribute the modulation of attenuation to nucleotide-induced variations in the kinetics of mRNA chain elongation. In support for this it was found that an RNA polymerase mutant with reduced RNA chain growth rate transcribed past the pyrE attenuator at a high frequency in the presence of a high UTP pool, but only when coupling of translation to transcription was allowed at the end of orfE.

Coevolution of codon usage and transfer RNA abundance

The use of synonymous codons is strongly biased in the bacterium Escherichia coli and yeast, comprising both bias between codons recognized by the same transfer RNA and bias between groups of codons recognized by different synonymous tRNAs. A major determinant of the second sort of bias is tRNA content, codons recognized by abundant tRNAs being used more often than those recognised by rare tRNAs, particularly in highly expressed genes, probably owing to selection at the level of translation against codons recognized by rare tRNAs. Conversely, codon usage is likely to exert selection pressure on tRNA abundance. Here I develop a model for the coevolution of codon usage and tRNA abundance which explains why there are unequal abundances of synonymous tRNAs leading to biased usage between groups of codons recognized by them in unicellular organisms.