Regulation of E.coli phenylalanyl-tRNA synthetase operon in vivo.

The phenylalanyl-tRNA synthetase operon is composed of two adjacent, cotranscribed genes, pheS and pheT, corresponding respectively to the small and large subunit of phenylalanyl-tRNA synthetase. A fusion between the regulatory regions of phenylalanyl-tRNA synthetase operon and the lac structural genes has been constructed to study the regulation of the operon. The pheS,T operon was shown, using the fusion, to be derepressed when phenylalanine concentrations were limiting in a leaky auxotroph mutated in the phenylalanine biosynthetic pathway. Furthermore, a mutational alteration in the phenylalanyl-tRNA synthetase gene, bradytrophic for phenylalanine, was also found to be derepressed under phenylalanine starvation. These results indicate that the pheS,T operon is derepressed when the level of tRNAPhe aminoacylation is lowered. By analogy with other well-studied amino acid biosynthetic operons known to be controlled by attenuation, these in vivo results indicate that phenylalanyl-tRNA synthetase levels are controlled by an attenuation-like mechanism.

Translational control of ribosomal protein S15.

The expression of ribosomal protein S15 is shown to be translationally and negatively autocontrolled using a fusion within a reporter gene. Isolation and characterization of several deregulated mutants indicate that the regulatory site (the translational operator site) overlaps the ribosome loading site of the S15 messenger. In this region, three domains, each exhibiting a stem-loop structure, were determined using chemical and enzymatic probes. The most downstream hairpin carries the Shine-Dalgarno sequence and the initiation codon. Genetic and structural data derived from mutants constructed by site-directed mutagenesis show that the operator is a dynamic structure, two domains of which can form a pseudoknot. Binding of S15 to these two domains suggests that the pseudoknot could be stabilized by S15. A model is presented in which two alternative structures would explain the molecular basis of the S15 autocontrol.

The translational regulation of threonyl-tRNA synthetase. Functional relationship between the enzyme, the cognate tRNA and the ribosome.

The E. coli threonyl-tRNA synthetase gene is negatively autoregulated at the translational level by a direct binding of the enzyme to the leader region of the thrS mRNA. This region folds in four well-defined domains. The enzyme binds to the leader at two major sites: the first is a stem-loop structure located in domain II upstream of the translational initiation site (domain I) which shares structural analogies with the anticodon arm of several tRNA(Thr) isoacceptors. The second site corresponds to a stable stem-loop structure located in domain IV. Both sites are separated by a large unpaired region (domain III). In vivo and in vitro experiments show that the structural integrity of both sites is required for the regulatory process. The binding of the enzyme to its mRNA target site represses its translation by preventing the ribosome from binding to its attachment site. tRNA(Thr) suppresses this inhibitory effect by displacing the mRNA from the enzyme at both the upstream stem-loop structure and the tRNA-like anticodon arm.

The conformation of the initiator tRNA and of the 16S rRNA from Escherichia coli during the formation of the 30S initiation complex.

The conformation of the E. coli initiator tRNA and of the 16S rRNA at different steps leading to the 30S.IF2.fMet-ARN(fMet).AUG.GTP complex has been investigated using several structure-specific probes. As compared to elongator tRNA, the initiator tRNA exhibits specific structural features in the anticodon arm, the T and D loops and the acceptor arm. Initiation factor 2 (IF2) interacts with the T-loop and the minor groove of the T stem of the RNA, and induces an increased flexibility in the anticodon arm. In the 30S initiation complex, additional protection is observed in the acceptor stem and in the anticodon arm of the tRNA. Within the 30S subunit, IF2 does not significantly shield defined portions of 16S rRNA, but induces both reduction and enhancement of reactivity scattered in the entire molecule. Most are constrained in a region corresponding to the cleft, the lateral protrusion and the part of the head facing the protrusion. All the reactivity changes induced by the binding of IF2 are still observed in the presence of the initiator tRNA and AUG message. The additional changes induced by the tRNA are mostly centered around the cleft-head-lateral protrusion region, near positions affected by IF2 binding.

Cleavage by RNase III in the transcripts of the met Y-nus-A-infB operon of Escherichia coli releases the tRNA and initiates the decay of the downstream mRNA.

The metY gene coding for a minor form of the initiator tRNA is the first gene of a complex polycistronic operon also encoding the transcription termination factor NusA and the translation initiation factor IF2. The mixed tRNA-mRNA polycistronic transcript is cleaved by RNase III in a hairpin structure downstream from the tRNA. This cleavage separates the tRNA from the mRNA and initiates the rapid degradation of the 5' extremity of the downstream mRNA. Dissociation of the structural (tRNA) and informational (mRNA) RNAs from this operon is also achieved by independent transcription in vivo. The presence of two transcription terminators located downstream from metY produces a small tRNAMetf2 precursor transcript, whereas an internal promoter situated between metY and the first open reading frame directs the transcription of only the protein-coding part of the operon.

Translational autocontrol of the Escherichia coli ribosomal protein S15.

When rpsO, the gene encoding the ribosomal protein S15 in Escherichia coli, is carried by a multicopy plasmid, the mRNA synthesis rate of S15 increases with the gene dosage but the rate of synthesis of S15 does not rise. A translational fusion between S15 and beta-galactosidase was introduced on the chromosome in a delta lac strain and the expression of beta-galactosidase studied under different conditions. The presence of S15 in trans represses the beta-galactosidase level five- to sixfold, while the synthesis rate of the S15-beta-galactosidase mRNA decreases by only 30 to 50%. These data indicate that S15 is subject to autogenous translational control. Derepressed mutants were isolated and sequenced. All the point mutations map in the second codon of S15, suggesting a location for the operator site that is very near to the translation initiation codon. However, the creation of deletion mutations shows that the operator extends into the 5' non-coding part of the message, thus overlapping the ribosome loading site.

A severely truncated form of translational initiation factor 2 supports growth of Escherichia coli.

We have constructed strains carrying null mutations in the chromosomal copy of the gene for translational initiation factor (IF) 2 (infB). A functional copy of the infB gene is supplied in trans by a thermosensitive lysogenic lambda phage integrated at att lambda. These strains enabled us to test in vivo the importance of different structural elements of IF2 expressed from genetically engineered plasmid constructs. We found that, as expected, the gene for IF2 is essential. However, a protein consisting of the C-terminal 55,000 Mr fragment of the wild-type IF2 protein is sufficient to allow growth when supplied in excess. This result suggests that the catalytic properties are localized in the C-terminal half of the protein, which includes the G-domain, and that this fragment is sufficient to complement the IF2 deficiency in the infB deletion strain.

Processing of the Bacillus subtilis thrS leader mRNA is RNase E-dependent in Escherichia coli.

We have recently reported that processing occurs in the untranslated leader region of several members of a family of Gram-positive genes regulated by tRNA-mediated antitermination. We showed that cleavage at this site plays an important role in the induction of Bacillus subtilis thrS gene expression, following threonine starvation, by stabilising the downstream mRNA. Here we show that, when transferred on a plasmid, processing of the B. subtilis thrS leader can occur at the same site in Escherichia coli. Cleavage at this site is dependent on the E. coli endoribonuclease E, both in vivo and in vitro, suggesting that a functional homologue of RNase E is responsible for thrS processing in B. subtilis.

Escherichia coli protein synthesis initiation factor IF3 controls its own gene expression at the translational level in vivo.

Measurements of the relative synthesis rates of mRNAs transcribed from the gene (thrS) for threonyl-tRNA synthetase and the adjacent gene (infC) for initiation factor IF3 show four- to fivefold more infC mRNA than thrS mRNA in vivo, suggesting that infC expression can be controlled independently of thrS expression. S1 mapping experiments reveal the existence of two transcription initiation sites for infC mRNAs internal to the thrS structural gene. Both the mRNA measurements and the S1 mapping experiments indicate that the majority of infC transcription initiates at the infC proximal promoter. In agreement with these results, the deletion of the infC distal promoter from infC-lacZ gene fusions does not affect the expression of these gene fusions in vivo. Measurements of the relative synthesis rate of infC mRNA in vivo in infC- strains overproducing IF3 shows that infC mRNA levels are normal in these strains, thus suggesting that IF3 regulates the translation of infC mRNAs in vivo. Extension of these experiments using infC-lacZ gene fusions carried on lambda bacteriophage and integrated at the lambda att site on the Escherichia coli chromosome shows that the expression of infC-lacZ protein fusions, but not infC-lacZ operon fusions, is derepressed in two infC- strains. A cellular excess of IF3 represses the expression of an infC-lacZ protein fusion but not an infC-lacZ operon fusion. Measurements of the relative mRNA synthesis rates of hybrid infC-lacZ mRNA synthesized from an infC-lacZ protein fusion under conditions of a fourfold derepression or a threefold repression of hybrid IF3-beta-galactosidase expression shows that the hybrid infC-lacZ mRNA levels remain unchanged. These results indicate that the cellular levels of IF3 negatively regulate the expression of its own gene, infC, at the translational level in vivo.

Escherichia coli phenylalanyl-tRNA synthetase operon region. Evidence for an attenuation mechanism. Identification of the gene for the ribosomal protein L20.

The nucleotide sequences of pheS and of the beginning of pheT have been determined. The genes pheS and pheT code, respectively, for the small and large subunits of phenylalanyl-tRNA synthetase, an alpha 2 beta 2 enzyme. Upstream from pheS the sequence shows another open reading frame of 354 nucleotides (rplT), which accounts for a protein of Mr 13,400. The product of this gene, previously named "P12", is identified as the ribosomal protein L20. The promoter for the pheS, T operon was located 368 nucleotides in front of pheS by transcription experiments in vitro. The promoter site is followed by a short open reading frame, which codes for a 14-residue peptide containing five phenylalanine residues. Immediately downstream from the stop codon of this open reading frame, the DNA sequence indicates that the transcript can be folded into three alternative secondary structures, one of which is a site of transcription termination. In vitro, 90% of transcription products initiated at the pheS, T promoter terminate at this site. However, long run-off transcripts proceeding through the terminator and covering the pheS structural gene are observed. No other transcription initiation could be detected between the terminator and the pheS structural gene. All these results are consistent with a mechanism by which phenylalanine-mediated attenuation controls the expression of phenylalanyl-tRNA synthetase. Further evidence is provided for this model by the features of pheS, T regulation in vivo (see the accompanying paper).

Control of phenylalanyl-tRNA synthetase genetic expression. Site-directed mutagenesis of the pheS, T operon regulatory region in vitro.

Previous studies of phenylalanyl-tRNA synthetase expression in Escherichia coli strongly suggested that the pheS, T operon was regulated by a phenylalanine-mediated attenuation mechanism. To investigate the functions of the different segments composing the pheS, T attenuator site, a series of insertion, deletion and point mutations in the pheS, T leader region have been constructed in vitro on a recombinant M13 phage. The effects of these alterations on the regulation of the operon were measured after transferring each mutation onto a lambda phage carrying a pheS, T-lacZ fusion. The behaviours of the various mutants agree with the predictions of the attenuation model. The role of the antiterminator (2-3 pairing) as competitor of the terminator (3-4 pairing) is demonstrated by several mutations affecting the stability of the 2-3 base-pairing. The existence of deletions and point mutations in the 3-4 base-pairing shows that the terminator is essential for both expression level and regulation of the operon. Mutations in the translation initiation site of the leader peptide show that the expression of the leader peptide is essential for attenuation control. However, alteration of the translation initiation rate of the leader peptide derepresses the pheS, T operon, which is the opposite of what is observed with the trp operon. This difference is explained in terms of different translation initiation efficiencies of the leader peptides. Finally, insertion mutations, increasing gradually the distance between the leader peptide stop codon and the first strand of the antiterminator, derepress the pheS, T operon and show that formation of the antiterminator structure is under the control of the translation of the leader peptide.

Attenuation control of the Escherichia coli phenylalanyl-tRNA synthetase operon.

The pheST operon codes for the two subunits of the essential enzyme phenylalanyl-tRNA synthetase. The nucleotide sequence of the regulatory regions of the operon, in vitro transcription data and in vivo experiments indicate that the operon is controlled by attenuation in a way similar to many amino acid biosynthetic operons. In this work the control of the pheST operon was studied in vivo by measuring the effect of deletions in the regulatory regions on downstream expression. The presence of a strong promoter followed by an approximately 90% efficient terminator in front of the structural parts of the operon is demonstrated. An open reading frame coding for a 14 amino acid long leader peptide containing five phenylalanine residues is located between the promoter and the terminator. The presence of the transcription terminator is shown to be essential to the operon's regulation. The localization of the promoter and the terminator agrees with the results of previous in vitro experiments. It is also shown that about 30% of the transcripts covering the pheST operon come from the upstream gene, rplT, which codes for the ribosomal protein L20. Although cotranscription exists between rplT and pheST, these genes are not systematically coregulated since reducing the translation of rplT about tenfold, does not change pheST expression. The pheST operon is also shown to be derepressed by a cellular excess of phenylalanyl-tRNA synthetase. This derepression is shown to be due to the pheST attenuator.

Both forms of translational initiation factor IF2 (alpha and beta) are required for maximal growth of Escherichia coli. Evidence for two translational initiation codons for IF2 beta.

The gene infB codes for two forms of translational initiation factor IF2; IF2 alpha (97,300 Da) and IF2 beta (79,700 Da). IF2 beta arises from an independent translational event on a GUG codon located 471 bases downstream from IF2 alpha start codon. By site-directed mutagenesis we constructed six different mutations of this GUG codon. In all cases, IF2 beta synthesis was variably affected by the mutations but not abolished. We show that the residual expression of IF2 beta results from translational initiation on an AUG codon located 21 bases downstream from the mutated GUG. Furthermore, two forms of IF2 beta have been separated by fast protein liquid chromatography and the determination of their N-terminal sequences indicated that they resulted from two internal initiation events, one occurring on the previously identified GUG start codon, the other on the AUG codon immediately downstream. We conclude that two forms of IF2 beta exist in the cell, which differ by seven aminoacid residues at their N terminus. Only by mutating both IF2 beta start codons could we construct plasmids that express only IF2 alpha. A plasmid expressing only IF2 beta was obtained by deletion of the proximal region of the infB gene. Using a strain that carries a null mutation in the chromosomal copy of infB and a functional copy of the same gene on a thermosensitive lysogenic lambda phage, we could cure the lambda phage when the plasmids expressing only one form of IF2 were supplied in trans. We found that each one of the two forms of IF2, at near physiological levels, can support growth of Escherichia coli, but that growth is retarded at 37 degrees C. This result shows that both forms of IF2 are required for maximal growth of the cell and suggests that they have acquired some specialized but not essential function.

Target site of Escherichia coli ribosomal protein S15 on its messenger RNA. Conformation and interaction with the protein.

The regulatory site of ribosomal protein S15 has been located in the 5' non-coding region of the messenger, overlapping with the ribosome loading site. The conformation of an in vitro synthesized mRNA fragment, covering the 105 nucleotides upstream from the initiation codon and the four first codons of protein S15, has been monitored using chemical probes and RNase V1. Our results show that the RNA is organized into three domains. Domains I and II, located in the 5' part of the mRNA transcript, are folded into stable stem-loop structures. The 3'-terminal domain (III), which contains the Shine-Dalgarno sequence and the AUG initiation codon, appears to adopt alternative conformations. One of them corresponds to a rather unstable stem-loop structure in which the Shine-Dalgarno sequence is paired. An alternative potential structure involves a "pseudo-knot" interaction between bases of this domain and bases in the loop of domain II. The conformation of several RNA variants has also been investigated. The deletion of the 5'-proximal stem-loop structure (domain I), which has no effect on the regulation, does not perturb the conformation of the two other domains. The deletion of domain II, leading to a loss of regulatory control, prevents the formation of the potential helix involved in the pseudo-knot structure and results in a stabilization of the alternative stem-loop structure in domain III. The replacement of another base in domain III involved in pairing in the two alternative structures mentioned above should induce a destabilization of both structures and results in a loss of the translational control. However, the replacement of another base in domain III, which does not abolish the control, results in the loss of the conformational heterogeneity in this domain and yields a stable conformation corresponding to the pseudo-knot structure. Thus, it appears that any mutation that disrupts or alters the formation of the pseudo-knot impairs the regulatory mechanism. Footprinting experiments show that protein S15 is able to bind to the synthesized fragment and provide evidence that the protein triggers the formation of the pseudo-knot conformation. A mechanism can be postulated in which the regulatory protein stabilizes this particular structure, thus impeding ribosome initiation.

Escherichia coli threonyl-tRNA synthetase and tRNA(Thr) modulate the binding of the ribosome to the translational initiation site of the thrS mRNA.

Escherichia coli threonyl-tRNA synthetase binds to the leader region of its own mRNA at two major sites: the first shares some analogy with the anticodon arm of several tRNA(Thr) isoacceptors and the second corresponds to a stable stem-loop structure upstream from the first one. The binding of the enzyme to its mRNA target site represses its translation by preventing the ribosome from binding to its attachment site. The enzyme is still able to bind to derepressed mRNA mutants resulting from single substitutions in the anticodon-like arm. This binding is restricted to the stem-loop structure of the second site. However, the interaction of the enzyme with this site fails to occlude ribosome binding. tRNA(Thr) is able to displace the wild-type mRNA from the enzyme at both sites and suppresses the inhibitory effect of the synthetase on the formation of the translational initiation complex. Our results show that tRNA(Thr) acts as an antirepressor on the synthesis of its cognate aminoacyl-tRNA synthetase. This repression/derepression double control allows precise adjustment of the rate of synthesis of threonyl-tRNA synthetase to the tRNA level in the cell.

Feedback regulation of rRNA synthesis in Escherichia coli. Requirement for initiation factor IF2.

It has been shown that the transcription of rRNA in Escherichia coli is feedback-regulated by its own transcription products through a negative feedback loop which appears to require the assembly of rRNA into complete ribosomes. In order to examine whether the feedback loop involves the ribosomes' main function, translation, we have constructed a strain in which the chromosomal copy of infB, encoding IF2, was placed under lac promoter/operator control, and the effects of limitation of translation initiation factor IF2 on the regulation were examined. By varying the concentration of a lac operon inducer, isopropyl thiogalactoside (IPTG), it was possible to vary the cellular concentration of IF2. Under the growth conditions used, decreasing the concentration of IF2 about twofold affected the growth rate only slightly, but further deprivation of IF2 resulted in a significant decrease in growth rate, an increase in RNA content and a large accumulation of non-translating ribosomes. These accumulated ribosomes were apparently unable to cause feedback regulation of rRNA synthesis in the absence of sufficient IF2. When a higher concentration of IPTG was added to these IF2-deficient cells, a rapid increase in the IF2 level and a significant decrease in the rate of RNA accumulation were observed before the new steady-state growth was attained. These results indicate that IF2 apparently is necessary for feedback regulation of stable RNA and imply that ribosomes must enter translation for feedback regulation to occur.

Escherichia coli phenylalanyl-tRNA synthetase operon is controlled by attenuation in vivo.

The two subunits of phenylalanyl-tRNA synthetase are made from two adjacent, cotranscribed genes that constitute the pheS,T operon. Three different fusions between pheS,T and lac genes were constructed in order to study the regulation of the pheS,T operon in vivo. We show, using these fusions, that phenylalanyl-tRNA synthetase transcription is derepressed when the level of aminoacylated tRNAPhe is lowered by mutational alteration of the synthetase. The pheS,T operon is also derepressed in strains carrying a trpX mutation. The gene trpX codes for an enzyme that modifies both tRNATrp and tRNAPhe and a mutation in that gene causes derepression of the trp and pheA operons, both of which are controlled by attenuation. The in vivo features of the regulation of pheS,T expression described here in correlation with the DNA sequence and in vitro transcription results described in the accompanying paper by Fayat et al. indicate that phenylalanyl-tRNA synthetase is controlled by attenuation in a way analogous to several amino acid biosynthetic operons.

Autogenous control of Escherichia coli threonyl-tRNA synthetase expression in vivo.

The regulation of the expression of thrS, the structural gene for threonyl-tRNA synthetase, was studied using several thrS-lac fusions cloned in lambda and integrated as single copies at att lambda. It is first shown that the level of beta-galactosidase synthesized from a thrS-lac protein fusion is increased when the chromosomal copy of thrS is mutated. It is also shown that the level of beta-galactosidase synthesized from the same protein fusion is decreased if wild-type threonyl-tRNA synthetase is overproduced from a thrS-carrying plasmid. These results strongly indicate that threonyl-tRNA synthetase controls the expression of its own gene. Consistent with this hypothesis it is shown that some thrS mutants overproduce a modified form of threonyl-tRNA synthetase. When the thrS-lac protein fusion is replaced by several types of thrS-lac operon fusions no effect of the chromosomal thrS allele on beta-galactosidase synthesis is observed. It is also shown that beta-galactosidase synthesis from a promoter-proximal thrS-lac operon fusion is not repressed by threonyl-tRNA synthetase overproduction. The fact that regulation is seen with a thrS-lac protein fusion and not with operon fusions indicates that thrS expression is autoregulated at the translational level. This is confirmed by hybridization experiments which show that under conditions where beta-galactosidase synthesis from a thrS-lac protein fusion is derepressed three- to fivefold, lac messenger RNA is only slightly increased.

Domains of the Escherichia coli threonyl-tRNA synthetase translational operator and their relation to threonine tRNA isoacceptors.

The expression of the gene for threonyl-tRNA synthetase (thrS) is negatively autoregulated at the translational level in Escherichia coli. The synthetase binds to a region of the thrS leader mRNA upstream from the ribosomal binding site inhibiting subsequent translation. The leader mRNA consists of four structural domains. The present work shows that mutations in these four domains affect expression and/or regulation in different ways. Domain 1, the 3' end of the leader, contains the ribosomal binding site, which appears not to be essential for synthetase binding. Mutations in this domain probably affect regulation by changing the competition between the ribosome and the synthetase for binding to the leader. Domain 2, 3' from the ribosomal binding site, is a stem and loop with structural similarities to the tRNA(Thr) anticodon arm. In tRNAs the anticodon loop is seven nucleotides long, mutations that increase or decrease the length of the anticodon-like loop of domain 2 from seven nucleotides abolish control. The nucleotides in the second and third positions of the anticodon-like sequence are essential for recognition and the nucleotide in the wobble position is not, again like tRNA(Thr). The effect of mutations in domain 3 indicate that it acts as an articulation between domains 2 and 4. Domain 4 is a stable arm that has similarities to the acceptor arm of tRNA(Thr) and is shown to be necessary for regulation. Based on this mutational analysis and previous footprinting experiments, it appears that domains 2 and 4, those analogous to tRNA(Thr), are involved in binding the synthetase which inhibits translation probably by interfering with ribosome loading at the nearby translation initiation site.

Evidence for autoregulation of the nusA-infB operon of Escherichia coli.

Analysis of three different nusA mutant strains suggests that the expression of the nusA-infB operon of Escherichia coli is regulated autogenously by the nusA gene product, a protein known to mediate transcription termination and antitermination. The cellular amounts of NusA and IF2 (infB) proteins are enhanced by a nusAts mutation which causes reduced transcription-termination activity. A nusAam mutant carrying the am ts suppressor, supFts6, overproduces the IF2 protein when the amount of NusA protein is reduced by the thermal inactivation of the supFts6. A modified form of NusA with the cat protein of Mr of 24 000 attached to the C terminus of NusA is overproduced compared to the wild-type NusA and causes the overproduction of IF2.