Specificity of codon recognition by Escherichia coli tRNALeu isoaccepting species determined by protein synthesis in vitro directed by phage RNA

Transfer RNA Modification

Transfer RNA (tRNA) from all organisms on this planet contains modified nucleosides, which are derivatives of the four major nucleosides. tRNA from Escherichia coli/Salmonella enterica contains 31 different modified nucleosides, which are all, except for one (Queuosine[Q]), synthesized on an oligonucleotide precursor, which through specific enzymes later matures into tRNA. The corresponding structural genes for these enzymes are found in mono- and polycistronic operons, the latter of which have a complex transcription and translation pattern. The syntheses of some of them (e.g.,several methylated derivatives) are catalyzed by one enzyme, which is position and base specific, but synthesis of some have a very complex biosynthetic pathway involving several enzymes (e.g., 2-thiouridines, N6-threonyladenosine [t6A],and Q). Several of the modified nucleosides are essential for viability (e.g.,lysidin, t6A, 1-methylguanosine), whereas deficiency in others induces severe growth defects. However, some have no or only a small effect on growth at laboratory conditions. Modified nucleosides that are present in the anticodon loop or stem have a fundamental influence on the efficiency of charging the tRNA, reading cognate codons, and preventing missense and frameshift errors. Those, which are present in the body of the tRNA, have a primarily stabilizing effect on the tRNA. Thus, the ubiquitouspresence of these modified nucleosides plays a pivotal role in the function of the tRNA by their influence on the stability and activity of the tRNA.

Transfer RNA Modification: Presence, Synthesis, and Function

Transfer RNA (tRNA) from all organisms on this planet contains modified nucleosides, which are derivatives of the four major nucleosides. tRNA from Escherichia coli/Salmonella enterica serovar Typhimurium contains 33 different modified nucleosides, which are all, except one (Queuosine [Q]), synthesized on an oligonucleotide precursor, which by specific enzymes later matures into tRNA. The structural genes for these enzymes are found in mono- and polycistronic operons, the latter of which have a complex transcription and translation pattern. The synthesis of the tRNA-modifying enzymes is not regulated similarly, and it is not coordinated to that of their substrate, the tRNA. The synthesis of some of them (e.g., several methylated derivatives) is catalyzed by one enzyme, which is position and base specific, whereas synthesis of some has a very complex biosynthetic pathway involving several enzymes (e.g., 2-thiouridines, N  6-cyclicthreonyladenosine [ct6A], and Q). Several of the modified nucleosides are essential for viability (e.g., lysidin, ct6A, 1-methylguanosine), whereas the deficiency of others induces severe growth defects. However, some have no or only a small effect on growth at laboratory conditions. Modified nucleosides that are present in the anticodon loop or stem have a fundamental influence on the efficiency of charging the tRNA, reading cognate codons, and preventing missense and frameshift errors. Those that are present in the body of the tRNA primarily have a stabilizing effect on the tRNA. Thus, the ubiquitous presence of these modified nucleosides plays a pivotal role in the function of the tRNA by their influence on the stability and activity of the tRNA.

Over expression of a tRNA(Leu) isoacceptor changes charging pattern of leucine tRNAs and reveals new codon reading

During mRNA translation, synonymous codons for one amino acid are often read by different isoaccepting tRNAs. The theory of selective tRNA charging predicts greatly varying percentages of aminoacylation among isoacceptors in cells starved for their common amino acid. It also predicts major changes in tRNA charging patterns upon concentration changes of single isoacceptors, which suggests a novel type of translational control of gene expression. We therefore tested the theory by measuring with Northern blots the charging of Leu-tRNAs in Escherichia coli under Leu limitation in response to over expression of tRNA(GAG)(Leu). As predicted, the charged level of tRNA(GAG)(Leu) increased and the charged levels of four other Leu isoacceptors decreased or remained unchanged, but the charged level of tRNA(UAG)(Leu) increased unexpectedly. To remove this apparent inconsistency between theory and experiment we postulated a previously unknown common codon for tRNA(GAG)(Leu) and tRNA(UAG)(Leu). Subsequently, we demonstrated that the tRNA(GAG)(Leu) codon CUU is, in fact, read also by tRNA(UAG)(Leu), due to a uridine-5-oxyacetic acid modification.

Modified nucleosides in the first positions of the anticodons of tRNA(Leu)4 and tRNA(Leu)5 from Escherichia coli

Minor leucine tRNA species, tRNA(Leu)4 and tRNA(Leu)5, from Escherichia coli B have been reported to recognize leucine codons UUA and UUG [Goldman, E., Holmes, W. M., and Hatfield, G. W. (1979) J. Mol. Biol. 129, 567-585]. In the present study, these two tRNA(Leu) species were purified from E. coli A19, and the nucleotide sequences were determined by a post-labeling method. tRNA(Leu)5 was found to correspond to the tRNA gene reported as su degrees6 tRNA [Yoshimura, M., Inokuchi, H., and Ozeki, H. (1984) J. Mol. Biol. 177, 627-644]. The first letter of the anticodon was identified to be 2'-O-methylcytidine (Cm). tRNA(Leu)4 was identified as the minor leucine tRNA that has been sequenced previously (tRNA(Leu)UUR) [Yamaizumi, Z., Kuchino, Y., Harada, F., Nishimura, S., and McCloskey, J. A. (1980) J. Biol. Chem. 255, 2220-2225]. There was an unidentified modified nucleoside (N*) in the first position of the anticodon of tRNA(Leu)4. Nucleoside N* was isolated to homogeneity (1 A260 unit). By 1H NMR spectroscopy, nucleoside N was found to be a 2'-O-methyluridine derivative with a substituent having a -CH2NH2+CH2COO- moiety in position 5 of the uracil ring. On the basis of these NMR analyses together with mass spectrometry, the chemical structure of nucleoside N* was determined as 5-carboxymethylaminomethyl-2'-O-methyluridine (cmnm5Um). Nucleoside N* was thus found to be a novel type of naturally occurring modified uridine. Because of the conformational rigidity of Cm and cmnm5Um in the first position of the anticodon, these tRNA(Leu) species recognize the leucine codons UUA++ and UUG correctly, but never recognize the phenylalanine codons UUU and UUC.

Structure of a 16-mer RNA duplex r(GCAGACUUAAAUCUGC)2 with wobble C.A+ mismatches

The crystal structure of a 16-mer, the longest known RNA duplex, has been determined at 2.5 A resolution. The hexadecamer r(GCAGACUUAAAUCUGC) contains isolated C.A/A.C mismatches with two hydrogen bonds. The two hydrogen bonds in the mismatches suggests that N1 of A is protonated even though the crystallization was done at neutral pH. Therefore, the C.A mismatch is a C.A+ wobble similar to the G.U wobble. The two C.A+ pairs are isolated by four Watson-Crick pairs and flanked by five Watson-Crick base-pairs on either sides. Kinks/bends of 20 degrees are observed at the wobble sites. The Watson-Crick base-pair A5.U26 on the 5'-side of the first C6.A27(+) wobble has a twist angle of 27 degrees compared to the 3'-side U7.A28 pair of 36 degrees. The twist angles are reversed (37 degrees and 26 degrees) in the second A11(+).C22 wobble because of the approximate dyad in the molecule, the flanking base-pair sequences are A.U pairs. The wobbles expand the major groove to 7.1 A/7.3 A. The duplexes form helical columns and are tightly packed around the 31-screw axis. The minor grooves of adjacent columns in juxtaposition interact through the O2' atoms and the anionic phosphate oxygen atoms.

Novel temperature-sensitive mutants of Escherichia coli that are unable to grow in the absence of wild-type tRNA6Leu

Escherichia coli has only a single copy of a gene for tRNA6Leu (Y. Komine et al., J. Mol. Biol. 212:579-598, 1990). The anticodon of this tRNA is CAA (the wobble position C is modified to O2-methylcytidine), and it recognizes the codon UUG. Since UUG is also recognized by tRNA4Leu, which has UAA (the wobble position U is modified to 5-carboxymethylaminomethyl-O2-methyluridine) as its anticodon, tRNA6Leu is not essential for protein synthesis. The BT63 strain has a mutation in the anticodon of tRNA6Leu with a change from CAA to CUA, which results in the amber suppressor activity of this strain (supP, Su+6). We isolated 18 temperature-sensitive (ts) mutants of the BT63 strain whose temperature sensitivity was complemented by introduction of the wild-type gene for tRNA6Leu. These tRNA6Leu-requiring mutants were classified into two groups. The 10 group I mutants had a mutation in the miaA gene, whose product is involved in a modification of tRNAs that stabilizes codon-anticodon interactions. Overexpression of the gene for tRNA4Leu restored the growth of group I mutants at 42 degrees C. Replacement of the CUG codon with UUG reduced the efficiency of translation in group I mutants. These results suggest that unmodified tRNA4Leu poorly recognizes the UUG codon at 42 degreesC and that the wild-type tRNA6Leu is required for translation in order to maintain cell viability. The mutations in the six group II mutants were complemented by introduction of the gidA gene, which may be involved in cell division. The reduced efficiency of translation caused by replacement of the CUG codon with UUG was also observed in group II mutants. The mechanism of requirement for tRNA6Leu remains to be investigated.

Characterization of an Escherichia coli mutant, feeA, displaying resistance to the calmodulin inhibitor 48/80 and reduced expression of the rare tRNA3Leu

We previously described a mutation feeB1 conferring a temperature-sensitive filamentation phenotype and resistance to the calmodulin inhibitor 48/80 in Escherichia coli, which constitutes a single base change in the acceptor stem of the rare tRNA3Leu recognizing CUA codons. We now describe a second mutant, feeA1, unlinked to feeB, but displaying a similar phenotype, 48/80 resistance and a reduced growth rate at the permissive temperature, 30 degrees C, and temperature-sensitive, forming short filaments at 42 degrees C. In the feeA mutant, tRNA3Leu expression (but not that of tRNA1Leu) was reduced approximately fivefold relative to the wild type. We previously showed that the synthesis of beta-galactosidase, which unusually requires the translation of 6-CUA codons, was substantially reduced, particularly at 42 degrees C, in feeB mutants. The feeA mutant also shows drastically reduced synthesis of beta-galactosidase at the non-permissive temperature and reduced levels even at the permissive temperature. We also show that increased copy numbers of the abundant tRNA1Leu, which can also read CUA codons at low efficiency, suppressed the effects of feeA1 under some conditions, providing further evidence that the mutant was deficient in CUA translation. This, and the previous study, demonstrates that mutations which either reduce the activity of tRNA3Leu or the cellular amount of tRNA3Leu confer resistance to the drug 48/80, with concomitant inhibition of cell division at 42 degrees C.

Codon pair utilization biases influence translational elongation step times

Two independent assays capable of measuring the relative in vivo translational step times across a selected codon pair in a growing polypeptide in the bacterium Escherichia coli have been employed to demonstrate that codon pairs observed in protein coding sequences more frequently than predicted (over-represented codon pairs) are translated slower than pairs observed less frequently than expected (under-represented codon pairs). These results are consistent with the findings that translational step times are influenced by codon context and that these context effects are related to the compatabilities of adjacent tRNA isoacceptor molecules on the surface of a translating ribosome. These results also support our previous suggestion that the frequency of one codon next to another has co-evolved with the structure and abundance of tRNA isoacceptors in order to control the rates of translational step times without imposing additional constraints on amino acid sequences or protein structures.

Hydrophobicity, expressivity and aromaticity are the major trends of amino-acid usage in 999 Escherichia coli chromosome-encoded genes

Multivariate analysis of the amino-acid compositions of 999 chromosome-encoded proteins from Escherichia coli showed that three main factors influence the variability of amino-acid composition. The first factor was correlated with the global hydrophobicity of proteins, and it discriminated integral membrane proteins from the others. The second factor was correlated with gene expressivity, showing a bias in highly expressed genes towards amino-acids having abundant major tRNAs. Just as highly expressed genes have reduced codon diversity in protein coding sequences, so do they have a reduced diversity of amino-acid choice. This showed that translational constraints are important enough to affect the global amino-acid composition of proteins. The third factor was correlated with the aromaticity of proteins, showing that aromatic amino-acid content is highly variable.

Recognition of UUN codons by two leucine tRNA species from Escherichia coli

Codon recognition by Escherichia coli tRNA(Leu)4 and tRNA(Leu)5 was investigated by analysis of the competition between two aminoacyl-tRNA species in an in vitro protein synthesis. Both tRNA species strictly obey the wobble rule when they are in competition with other tRNA species. This is probably due to the post-transcriptional modifications at the first position of the anticodon of these tRNA(Leu) species, supporting the proposal that the conformational rigidity of post-transcriptionally modified pyrimidine nucleotides guarantees the correct codon recognition.

Increased ribosomal accuracy increases a programmed translational frameshift in Escherichia coli

We have tested the effect of increased ribosomal fidelity on a modified version of the programmed release factor 2 (RF2) translational frameshift. In the constructs tested, the original UGA codon at the site of the shift was replaced by either of two sense codons, UGG (tryptophan), which allows a frameshift of approximately 13%, or CUG (leucine), which allows a frameshift of only approximately 2%. We confirmed the results of Curran and Yarus [Curran, J. F. & Yarus, M. (1989) J. Mol. Biol. 209, 65-77] in a wild-type ribosomal host, including a reduction of the UGG shift following induction of tRNA(Trp) from a plasmid copy of the tRNA gene. But to our surprise, in a hyperaccurate streptomycin pseudo-dependent host, the UGG frameshift increased to more than 50%. When we added a tRNA(Trp) plasmid to these cells, induction of the tRNA(Trp) gene reduced the shift back to approximately 7%. Messenger RNA levels did not vary greatly under these different induced conditions. Other increased accuracy alleles also showed increased frameshifting with UGG at the frameshift site. All increased accuracy alleles led to slower translation rates, and there appeared to be a proportionality between the extent of reduction of synthesis for the in-frame reporter and the extent of UGG frameshift for the out-of-frame reporter. There were little effects of increased accuracy on the lower level CUG frameshift. However, over-production of the cognate tRNA(1Leu) dramatically reduced even this lower level of shift, despite the fact that tRNA(1Leu) is already the most abundant isoacceptor in Escherichia coli. These results can be rationalized by following the hypothesis of Curran and Yarus as follows: with wild-type ribosomes, limited availability of tRNA(Trp) (about 1% of total tRNA) facilitates a pause at the UGG codon (due to the vacant A site), allowing increased opportunity for ribosome realignment. Excess tRNA(Trp) reduces the time the A site is vacant and thus reduces the frameshift. The slower hyperaccurate ribosomes increase the pause time and thus increase the opportunity for shifting, a process again reversed by increasing the in-frame cognate tRNA(Trp). These data provide strong support for a model in which the extent of ribosome pause time at a programmed frameshift site is a major determinant in the efficiency of the frameshift and in which tRNA availability can be a major influence on this process.

Effects of tRNA(1Leu) overproduction in Escherichia coli

Strains of Escherichia coli have been produced which express very high levels of the tRNA(1Leu) isoacceptor. This was accomplished by transforming cells with plasmids containing the leuV operon which encodes three copies of the tRNA(1Leu) gene. Most transformants grew very slowly and exhibited a 15-fold increase in cellular concentrations of tRNA(1Leu). As a result, total cellular tRNA concentration was approximately doubled and 56% of the total was tRNA(1Leu). We examined a number of parameters which might be expected to be affected by imbalances in tRNA concentration: in vivo tRNA charging levels, misreading, ribosome step time, and tRNA modification. Surprisingly, no increase in intracellular ppGpp levels was detected even though only about 40% of total leucyl tRNA was found to be charged in vivo. Gross ribosomal misreading was not detected, and it was shown that ribosomal step times were reduced between two- and threefold. Analyses of leucyl tRNA isolated from these slow-growing strains showed that at least 90% of the detectable tRNA(1Leu) was hypomodified as judged by altered mobility on RPC-5 reverse-phase columns, and by specific modification assays using tRNA(m1G)-methyltransferase and pseudo-uridylate synthetase. Analysis of fast-growing revertants demonstrated that tRNA concentration per se may not explain growth inhibition because selected revertants which grew at wild-type growth rates displayed levels of tRNA comparable to that of control strains bearing the leuV operon. A synthetic tRNA(1Leu) operon under the control of the T7 promoter was prepared which, when induced, produced six- to sevenfold increases in tRNA(1Leu) levels. This level of tRNA(1Leu) titrated the modification system as judged by RPC-5 column chromatography. Overall, our results suggest that hypomodified tRNA may explain, in part, the observed effects on growth, and that the protein-synthesizing system can tolerate an enormous increase in the concentration of a single tRNA.

Temperature sensitivity caused by missense suppressor supH and amber suppressor supP in Escherichia coli

The temperature-sensitive missense suppressor supH and amber suppressor supP in Escherichia coli are mutations of the serU and leuX genes, respectively. The supH tRNA, tRNA(SerCAA), is expected to recognize UUG codons, which are normally read by tRNA(LeuCAA) and tRNA(LeuUAA), coded for by the leuX gene and the leuZ gene, respectively. We show that supP and supH are incompatible and that strains carrying both supP and a restrictive rpsL allele are temperature sensitive. It is suggested that the temperature sensitivity of both supH and supP strains is caused by deficient reading of UUG codons by tRNA(LeuUAA).

Codon recognition patterns as deduced from sequences of the complete set of transfer RNA species in Mycoplasma capricolum. Resemblance to mitochondria

The nucleotide sequences of the complete set of tRNA species in Mycoplasma capricolum, a derivative of Gram-positive eubacteria, have been determined. This bacterium represents the first genetic system in which the sequences of all the tRNA species have been determined at the RNA level. There are 29 tRNA species: three for Leu, two each for Arg, Ile, Lys, Met, Ser, Thr and Trp, and one each for the other 12 amino acids as judged from aminoacylation and the anticodon nucleotide sequences. The number of tRNA species is the smallest among all known genetic systems except for mitochondria. The tRNA anticodon sequences have revealed several features characteristic of M. capricolum. (1) There is only one tRNA species each for Ala, Gly, Leu, Pro, Ser and Val family boxes (4-codon boxes), and these tRNAs all have an unmodified U residue at the first position of the anticodon. (2) There are two tRNAThr species having anticodons UGU and AGU; the first positions of these anticodons are unmodified. (3) There is only one tRNA with anticodon ICG in the Arg family box (CGN); this tRNA can translate codons CGU, CGC and CGA. No tRNA capable of translating codon CGG has been detected, suggesting that CGG is an unassigned codon in this bacterium. (4) A tRNATrp with anticodon UCA is present, and reads codon UGA as Trp. On the basis of these and other observations, novel codon recognition patterns in M. capricolum are proposed. A comparatively small total, 13, of modified nucleosides is contained in all M. capricolum tRNAs. The 5' end nucleoside of the T psi C-loop (position 54) of all tRNAs is uridine, not modified to ribothymidine. The anticodon composition, and hence codon recognition patterns, of M. capricolum tRNAs resemble those of mitochondrial tRNAs.

Aminoacylation of synthetic DNAs corresponding to Escherichia coli phenylalanine and lysine tRNAs

Synthetic DNA oligomers (tDNAs) corresponding to Escherichia coli tRNA(Phe) or tRNA(Lys) have been synthesized with either deoxythymidine (dT) or deoxyuridine (dU) substituted in the positions occupied by ribouridine or its derivatives. The tDNAs inhibited the aminoacylation of their respective tRNAs with their cognate amino acids, but not the aminoacylation of tRNA(Leu) with Leu. In the presence of aminoacyl-tRNA synthetase, species of both a tDNA(Phe) synthesized with a 3' terminal riboadenosine and a tDNA(Lys) containing only deoxynucleotides could be aminoacylated with the appropriate amino acids, although the Michaelis constant Km and observed maximal rate Vmax values for aminoacylation were increased by three- to fourfold and decreased by two- to threefold, respectively. The aminoacylation of synthetic tDNAs demonstrates that the ribose backbone of a tRNA is not absolutely required for tRNA aminoacylation.

Leucine tRNA family of Escherichia coli: nucleotide sequence of the supP(Am) suppressor gene

We describe the cloning and the DNA sequence of an amber suppressor allele of the Escherichia coli leuX (supP) gene. The suppressor allele codes for a tRNA with anticodon CUA, presumably derived by a single base change from a CAA anticodon. The mature coding sequence of the leuX gene is preceded by a putative Pribnow box sequence (TATAAT) and followed by a termination signal. The sequence of the leuX-coded tRNA is compared with the sequences of the four remaining tRNALeu isoacceptors of E. coli and with two tRNALeu species from bacteriophage T4 and T5. The conserved nucleotides in these seven tRNAs recognized by E. coli leucyl-tRNA synthetase are located mainly in the aminoacyl stem and in the D-stem/loop region.

The nucleotide sequence of bacteriophage T5 glutamine transfer RNA

Uniformly 32P-labeled phage-specific tRNAGln has been isolated from bacteriophage T5-infected Escherichia coli cells and its nucleotide sequence has been determined using thin-layer chromatography on cellulose to fractionate the oligonucleotides. The sequence is: pUGGGGAUUAGCUUAGCUUGGCCUAAAGCUUCGGCCUUUGAAG psi CGAGAUCAUUGGT psi CAAAUCCAAUAUCCCCUGCCAOH. The main feature of this tRNA is the absence of Watson-Crick pairing between the 5'-terminal base and the fifth base from its 3'-end. The structure of tRNA was confirmed by DNA sequencing of its gene.

Molecular basis for modulated regulation of gene expression in the arginine regulon of Escherichia coli K-12

We compare the nucleotide sequences of the regulatory regions of five genes or groups of genes of the arginine regulon of Escherichia coli K-12: argF, argI, argR, the bipolar argECBH operon and the carAB operon. All these regions harbour one or two copies of a conserved 18 bp sequence which appears to constitute the basic arginine operator sequence (ARG box). We discuss the influence of ARG box copy number, degree of dyad symmetry, base composition, and position relative to the cognate promoter site on the derepression-repression ratios of the genes of the regulon. A novel hypothesis, based on structural considerations, is also put forward to account for the absence ot attenuation control.

Aminoacyl-tRNAs from Physarum polycephalum: patterns of codon recognition

Isoacceptors of Physarum polycephalum Ala-, Arg-, Glu-, Gln-, Gly-, Ile-, Leu-, Lys-, Ser-, Thr-, and Val-tRNAs were resolved by reverse-phase chromatography and isolated, and their codon recognition properties were determined in a ribosomal binding assay. Codon assignments were made to most isoacceptors, and they are summarized along with those determined in other studies from Escherichia coli, yeasts, wheat germ, hymenoptera, Xenopus, and mammals. The patterns of codon recognition by isoacceptors from P. polycephalum are more similar to those of animals than to those of plants or lower fungi.

Does quantitative tRNA adaptation to codon content in mRNA optimize the ribosomal translation efficiency? Proposal for a translation system model

Neither a dynamic nor an energetic approach of the translation process has taken into account that intracellular levels of iso-tRNA species are adapted or adjusted to the codon frequency of mRNA being decoded (Bombyx mori silk gland, rabbit reticulocyte). A critical study of available experimental data suggests that the average elongation rate of a protein is maximized in the presence of an adapted tRNA population, usually an homologous tRNA. In addition, the amount of synthesized protein parallels that of corresponding mRNA. Other evidences--including in vitro and in vivo elongation assays with fibroin mRNA--show that individual elongation rates are not uniform. Pauses occur at certain sites of the mRNA chain. The relative lifetime of these pauses depends on the tRNA pool used. Finally, it appears that translation accuracy also depends on the balanced tRNA population. We propose to explain these different effects by using a codon-anticodon recognition model, called "trial and error system" based on a stochastic processing of the ribosome. Accordingly, various acylated tRNA species which surround a ribosome randomly encounter the receptor A site. Every trapped tRNA species is tested for a proper pairing with the codon to be recognized at the level of a comparator or discriminator function. If the pairing is correct, transpeptidation becomes irreversible. If not, the aminoacyl-tRNA is rejected and another randomly trapped tRNA is processed in turn. Mathematical analysis of this model shows that the mean number of trials used for translating the whole sequence of a mRNA is minimized when the proportion of different iso-tRNA species is correlated with the square root of codon frequency. Quantitations of reticulocyte tRNA support such a parabolic relation. Our translation system model brings some light into the role of tRNA adaptation for optimizing translation efficiency, i.e. maximizing both speed and accuracy. Some consequences of the model are discussed.

The evolving tRNA molecule

The study of tRNA molecular evolution is crucial to understanding the origin and establishment of the genetic code as well as the differentiation and refinement of the machinery of protein synthesis in prokaryotes, eukaryotes, organelles, and phage systems. The small size of the molecule and its critical involvement in a multiplicity of roles distinguish its study from classical protein molecular evolution with respect to goals and methods. Here, the authors assess available and missing data, existing and needed methodology, and the impact of tRNA studies on current theories both of genetic code evolution and of the evolution of species. They analyze mutational "hot spots", the role of base modification, synthetase recognition, codon-anticodon interactions and the status of organelle tRNA.

Mistranslation in cells infected with the bacteriophage MS2: direct evidence of Lys for Asn substitution

The coat protein of the bacteriophage MS2 was found to show an increased level of charge heterogeneity when synthesized in Escherichia coli starved for Asn or Lys. No such increase was found when the host was starved for Arg, His Ile or Pro. This is the pattern predicted by "two-out-of-three" codon misreading in the coat protein gene. In the case of Asn starvation, direct measurements of the relative incorporation of Lys demonstrate that the observed charge heterogeneity is the result of mistranslation. Asn starvation increased the error frequency in coat protein to over 0.3 mistake per asparagine codon. The small amount of charge heterogeneity seen in unstarved cells seems also to be the result of misreading Asn codons.

"Two out of three" codon reading leading to mistranslation in vivo

Strains of Escherichia coli were starved for asparagine or lysine in order to increase the in vivo level of mistranslation. In a relA strain, asparagine starvation increased the error frequency in elongation factor Tu to 0.12 mistake per asparagine codon, while with lysine starvation in the same strain the error frequency per lysine codon was 0.008. The pattern of isoelectric point changes in the altered protein produced is consistent with third position misreading in the AAN codon group. This high level of mistranslation is not seen in streptomycin resistant (rpsL) strains or in most relA+ strains.

Normal tRNAs promote ribosomal frameshifting

The addition of Ser AGC AGU tRNA to an E. coli cell-free protein synthesizing system which contains the endogenous tRNA levels results in up to 100% of the ribosomes translating the MS2 coat gene shifting into the -1 reading frame. An analogous phenomenon is seen at a much lower level without the tRNA addition, where a shift into the +1 frame can also be detected. Thus translation with the endogenous tRNA levels yields proteins which have the amino terminus of the coat protein but which are substantially larger than the coat protein and comprise about 5% of the coat translation. Since the lysis gene overlaps the 3' end of the coat gene in the +1 frame, we conclude that the reading frame shift into the +1 frame yields a hybrid protein. Also, we present evidence that ribosomes translating the synthetase gene shift into the -1 frame near the distal end of the gene. This frameshifting is promoted by thrACU ACC tRNA. Specific competitor tRNAs for both Thr and Ser tRNA-promoted frameshifting have been characterized. The generality of this new mechanism for producing additional proteins is unclear, but it investigation should increase understanding of the coding mechanism and its origin.