Escherichia coli elongation factor G blocks stringent factor

Subinhibitory Concentrations of Bacteriostatic Antibiotics Induce relA-Dependent and relA-Independent Tolerance to β-Lactams

The nucleotide (p)ppGpp is a key regulator of bacterial metabolism, growth, stress tolerance, and virulence. During amino acid starvation, the Escherichia coli (p)ppGpp synthetase RelA is activated by deacylated tRNA in the ribosomal A-site. An increase in (p)ppGpp is believed to drive the formation of antibiotic-tolerant persister cells, prompting the development of strategies to inhibit (p)ppGpp synthesis. We show that in a biochemical system from purified E. coli components, the antibiotic thiostrepton efficiently inhibits RelA activation by the A-site tRNA. In bacterial cultures, the ribosomal inhibitors thiostrepton, chloramphenicol, and tetracycline all efficiently abolish accumulation of (p)ppGpp induced by the Ile-tRNA synthetase inhibitor mupirocin. This abolishment, however, does not reduce the persister level. In contrast, the combination of dihydrofolate reductase inhibitor trimethoprim with mupirocin, tetracycline, or chloramphenicol leads to ampicillin tolerance. The effect is independent of RelA functionality, specific to β-lactams, and not observed with the fluoroquinolone norfloxacin. These results refine our understanding of (p)ppGpp's role in antibiotic tolerance and persistence and demonstrate unexpected drug interactions that lead to tolerance to bactericidal antibiotics.

Ribosome•RelA structures reveal the mechanism of stringent response activation

Stringent response is a conserved bacterial stress response underlying virulence and antibiotic resistance. RelA/SpoT-homolog proteins synthesize transcriptional modulators (p)ppGpp, allowing bacteria to adapt to stress. RelA is activated during amino-acid starvation, when cognate deacyl-tRNA binds to the ribosomal A (aminoacyl-tRNA) site. We report four cryo-EM structures of E. coli RelA bound to the 70S ribosome, in the absence and presence of deacyl-tRNA accommodating in the 30S A site. The boomerang-shaped RelA with a wingspan of more than 100 Å wraps around the A/R (30S A-site/RelA-bound) tRNA. The CCA end of the A/R tRNA pins the central TGS domain against the 30S subunit, presenting the (p)ppGpp-synthetase domain near the 30S spur. The ribosome and A/R tRNA are captured in three conformations, revealing hitherto elusive states of tRNA engagement with the ribosomal decoding center. Decoding-center rearrangements are coupled with the step-wise 30S-subunit 'closure', providing insights into the dynamics of high-fidelity tRNA decoding.

The weird and wonderful world of bacterial ribosome regulation

In every organism, translation of the genetic information into functional proteins is performed on the ribosome. In Escherichia coli up to 40% of the cell's total energy turnover is channelled toward the ribosome and protein synthesis. Thus, elaborate networks of translation regulation pathways have evolved to modulate gene expression in response to growth rate and external factors, ranging from nutrient deprivation, to chemical (pH, ionic strength) and physical (temperature) fluctuations. Since the fundamental players involved in regulation of the different phases of translation have already been extensively reviewed elsewhere, this review focuses on lesser known and characterized factors that regulate the ribosome, ranging from processing, modification and assembly factors, unusual initiation and elongation factors, to a variety of stress response proteins.

Dissection of the mechanism for the stringent factor RelA

During conditions of nutrient deprivation, ribosomes are blocked by uncharged tRNA at the A site. The stringent factor RelA binds to blocked ribosomes and catalyzes synthesis of (p)ppGpp, a secondary messenger that induces the stringent response. We demonstrate that binding of RelA and (p)ppGpp synthesis are inversely coupled, i.e., (p)ppGpp synthesis decreases the affinity of RelA for the ribosome. RelA binding to ribosomes is governed primarily by mRNA, but independently of ribosomal protein L11, while (p)ppGpp synthesis strictly requires uncharged tRNA at the A site and the presence of L11. A model is proposed whereby RelA hops between blocked ribosomes, providing an explanation for how low intracellular concentrations of RelA (1/200 ribosomes) can synthesize (p)ppGpp at levels that accurately reflect the starved ribosome population.

Purification of elongation factors EF-Tu and EF-G from Escherichia coli by covalent chromatography on thiol-sepharose

The elongation factors EF-Tu and EF-G of Escherichia coli are involved in the transport of aminoacyl-tRNA to ribosomes and the translocation of ribosomes on mRNA, respectively. Both possess cysteine residues that are important for activity. We took advantage of this property to design a purification protocol based on thiol-Sepharose chromatography, a method involving thiol-disulfide interchange between protein thiol groups and the glutathione-2-pyridyl-disulfide conjugate of the affinity resin. Bacterial cells were lysed by a lysozyme-EDTA method, and the lysate supernatant was purified by chromatography on, first, DEAE-Sephacel and, then thiol-Sepharose. Both elongation factors were purified in a single procedure, since DEAE-Sephacel fractions containing both factors were loaded on the thiol-Sepharose column. Thiol-Sepharose chromatography efficiently separates each elongation factor from all contaminating proteins. The purified elongation factors were characterized by SDS-PAGE, protein sequencing, and biological activity. The specific reactivities of the elongation factors with thiol-Sepharose allow their efficient purification and suggest that they possess hitherto undiscovered properties connected with their reactive thiols.

ms2i6A deficiency enhances proofreading in translation

The hypermodified base 2-methylthio-N6-isopentenyladenosine (ms2i6A) at position 37 occurs frequently in tRNAs that read codons starting with uridine. Here we have studied how ms2i6A affects the accuracy of poly(U) translation in vitro. Deficiency leads to a higher rejection rate of tRNA4(Leu) by more aggressive proofreading on the wild-type ribosome, but with the initial selection step unchanged. Our data indicate that ms2i6A has no effect on codon-anticodon interactions on wild-type ribosomes as long as aminoacyl-tRNA is in ternary complex with EF-Tu and GTP. ms2i6A deficiency in the cognate poly(U) reader tRNA(Phe) leads to increased misreading when the near-cognate competitor tRNA4(Leu) is wild-type. ms2i6A deficiency in tRNA4(Leu) gives a decreased error level in competition with wild-type tRNA(Phe).

How does ppGpp affect translational accuracy in the stringent response?

With an in vitro poly(Phe) synthesis system we have tested recent models concerning translational accuracy in the stringent response during aminoacid starvation. We have found that cognate, deacylated tRNA of very high concentrations is unable to block the A-site. No influence of EF-Tu.ppGpp on ribosomal proofreading has been found. Alternative mechanisms to keep translational errors low by the stringent response are discussed.

Novel mutants of elongation factor G

A novel mutant form of elongation factor G (EF-G) in Escherichia coli is described. This variant EF-G restricts reading frame errors by a factor of 2 to 3 in vivo at two different positions in a lacIZ fusion. In addition, a conventional fusidic acid resistant (fusR) mutant of EF-G was compared with the restrictive mutant. Both mutants were characterized in vitro in a steady-state poly(U) translating system. The data indicate that the restrictive EF-G variant has an altered interaction with the ribosome both in vivo and in vitro. In contrast, the conventional fusR variant is altered in its interaction with GTP, which is evident in vitro.

Is translation inhibited by noncognate ternary complexes?

We studied the influence of an error-prone isoacceptor (tRNALeu4), as well as an intermediate (tRNALeu2) and a weak (tRNAVal) competitor of tRNAPhe on the poly(Phe) synthesis rate. Even at very high excess concentrations of these noncognate ternary complexes there was no significant effect on the translation rate. Our result argues against the assertion that in vivo translation is slowed down by noncognate tRNA and favours the hypothesis that the incorrect ternary complex concentrations are too low to saturate the ribosomes in vivo.

Antagonistic effects of mutant elongation factor Tu and ribosomal protein S12 on control of translational accuracy, suppression and cellular growth

Kirromycin-resistant mutant forms of elongation factor Tu, which are coded by tufA (Ar) or tufB (Bo) and are associated with an increased rate of translational error formation, have been analysed. In vivo, Ar was found to increase misreading as well as suppression of non-sense codons irrespective of Bo in a strain with wild type ribosomes. It is therefore not necessary to evoke both tufA (Ar) and tufB (Bo) mutations together in order to increase translational error as suggested earlier [1]. When combined with a hyperaccurate ribosomal rpsL (S12) mutation, Ar counteracts the restrictive effects on translational error formation caused by the altered protein S12, thus restoring the levels of missense error in vitro and non-sense error and suppression in vivo to near wild type values. As judged from in vitro experiments this results principally from a lowered selectivity of the Ar ternary complex at the initial discrimination step on the ribosome during translation. In vivo, this compensatory effect on the rpsL mutation on non-sense error formation and suppression is seen irrespective of the nature of tRNA or codon context. Furthermore, the tufA mutation enhances the cellular growth rate of the rpsL mutant, whereas it decreases growth of strains with normal ribosomes. Inactivation of one of the two genes coding for EF-Tu (tufB), while leaving the other gene (tufA) intact, can by itself, increase non-sense error formation and suppression.

Mutant EF-Tu increases missense error in vitro

We have studied the consequences of mutational alteration in the structure of EF-Tu on the missense errors and proofreading activity of bacterial ribosomes in vitro. Our data show that the EF-Tu Bo mutant form of EF-Tu (van der Meide et al. 1983a) is inactive in polypeptide synthesis on the ribosome, even though it binds aminoacyl-tRNA. A second mutant form, EF-Tu Ar (van der Meide et al. 1983a), is active in polypeptide synthesis but supports a much higher messense incorporation with either leucine isoacceptor 2 or leucine isoacceptor 4 in the in vitro system. Further analysis of the kinetic basis of this enhanced missense frequency revealed that the mutation responsible for the alteration in EF-Tu Ar increases the errors at both the proofreading step and the initial selection. In this respect the effect of this particular mutation is similar to the mode of action of the antibiotic kanamycin (Jelenc and Kurland 1984).

How do combinations of rpsL- and miaA- generate streptomycin dependence?

Petrullo et al. (1983) have studied the consequences of combining a mutation (rpsL-) that normally generates streptomycin resistant (Smr) ribosomes with a mutation (miaA-) that leads to loss of a tRNA hypermodification. They found surprisingly that such doubly mutant bacteria become streptomycin dependent (Smd). Here, we show in vitro that ribosomes purified from an Smr mutant behave very like Smd ribosomes when they are combined with tRNA from an miaA- mutant. Our analysis suggests that proofreading becomes excessively intense when the mutant components are combined, and that this reduces the efficiency of translation to the very low levels characteristic of Smd ribosomes. We show that Sm increases the efficiency of translation in vitro by suppressing the proofreading flows. We suggest that this will explain the growth stimulatory effect of Sm on the rpsL-, miaA- double mutants.

Kinetic impairment of restrictive streptomycin-resistant ribosomes

Comparisons in vivo and in vitro of wild-type and otherwise isogenic bacteria with five different mutant alleles of the gene (rpsL) specifying ribosomal protein S12, all resistant to high levels of streptomycin, show that the streptomycin-resistant (Smr) phenotype can be subdivided into major groups: restrictive and non-restrictive. The restrictive bacteria have a characteristically lower frequency of nonsense suppression in vivo, and are also slower than the wild type in their rate of protein synthesis. Non-restrictive Smr bacteria on the other hand do not differ significantly from the wild type either in nonsense suppression frequencies or in the rate of translation. A complementary pattern is seen in vitro, where ribosomes from the restrictive Smr bacteria translate poly(U) with a significantly lower missense error frequency than wild-type ribosomes, and also show an increased Michaelis constant (KM) with respect to their substrate, i.e. ternary complexes. Both effects are correlated with the more aggressive proofreading function that is characteristic of these restrictive ribosomes. In contrast, ribosomes isolated from the non-restrictive Smr bacteria do not show any major difference in either proofreading or missense error in vitro when compared to the wild type.

Streptomycin preferentially perturbs ribosomal proofreading

We have studied the influence of streptomycin (Sm) on the kinetics and accuracy of translation by wild-type as well as Ram-mutant ribosomes in an in vitro system that mimics the performance characteristics of ribosomes in bacteria. It can be shown in this system that the accuracy of translation is made up of an initial selection step and one or more proofreading steps. The data show that the antibiotic has only a small influence on the initial selectivity step of wild-type or mutant ribosomes. Streptomycin stimulates the missense rate primarily by suppressing the proofreading of the ribosomes. The kinetic effects of Sm and of Ram alteration are not additive, but seem to be overlapping if not identical.

Hyper-accurate ribosomes inhibit growth

We have compared both in vivo and in vitro translation by ribosomes from wild-type bacteria with those from streptomycin-resistant (SmR), streptomycin-dependent (SmD) and streptomycin-pseudo-dependent (SmP) mutants. The three mutant bacteria translate more accurately and more slowly in the absence of streptomycin (Sm) than do wild-type bacteria. In particular, the SmP bacteria grow at roughly half the rate of the wild-type in the absence of Sm. The antibiotic stimulates both the growth rate and the translation rate of SmP bacteria by approximately 2-fold, but it simultaneously increases the nonsense suppression rate quite dramatically. Kinetic experiments in vitro show that the greater accuracy and slower translation rates of mutant ribosomes compared with wild-type ribosomes are associated with much more rigorous proofreading activities of SmR, SmD and SmP ribosomes. Sm reduces the proofreading flows of the mutant ribosomes and stimulates their elongation rates. The data suggest that these excessively accurate ribosomes are kinetically less efficient than wild-type ribosomes, and that this inhibits mutant growth rates. The stimulation of the growth of the mutants by Sm results from the enhanced translational efficiency due to the loss of proofreading, which more than offsets the loss of accuracy caused by the antibiotic.

Multiple effects of kanamycin on translational accuracy

We have studied the effects of kanamycin on the accuracy of translation in vitro by wild-type and mutant ribosomes from Escherichia coli. Kanamycin stimulates the leucine missense error of poly(U) translation by wild-type, Ram, and streptomycin-resistant ribosomes in characteristic ways; in particular, the streptomycin-resistant ribosomes are significantly less error-prone than wild-type or Ram ribosomes at all concentrations of the antibiotic. Kinetic analysis of the effects of kanamycin on the translational accuracy of wild-type ribosomes reveals a different concentration dependence for the perturbation of the initial selectivity of streptomycin-resistant ribosomes is not affected by kanamycin; the drug enhances only the error of proofreading by this mutant ribosome. We suggest that the multiple effects of kanamycin on the errors of translation are due to separate effects at different ribosomal sites.

Ram ribosomes are defective proofreaders

We have studied the kinetics of poly(U) translation by three ribosomal ambiguity (Ram) mutants in an in vitro system with performance characteristics similar to those expressed in vivo. The leucine missense frequency supported by Ram ribosomes with tRNALeu2 increases between six and twelve-fold over that of wild-type ribosomes, while the corresponding increase with tRNALeu4 was between four and eight-fold, depending on the rpsD allele. We have used a steady-state assay for proofreading to identify the kinetic lesion responsible for the Ram phenotype. We were unable to detect any difference between Ram and wild-type ribosomes with respect to the initial kinetics of amino-acyl tRNA selection. All of the increased error rates could be associated with a decreased capacity of these Ram ribosomes to discard non-cognate aminoacyl-tRNA by proof reading.

Chemical cross-linking of elongation factor G to both subunits of the 70-S ribosomes from Escherichia coli

Ribosomal proteins situated at or near the binding site of elongation factor G (EF-G) on the Escherichia coli ribosome have been identified by use of the bifunctional, cleavable cross-linker, dimethyl-4,9-diaza-5,8-dioxo-6,7-dihydroxy-dodecanebisimidate. Five different bimolecular EF-G x ribosomal-protein complexes were isolated electrophoretically. The ribosomal proteins found in each of these complexes were identified as the 50-S-subunit proteins L6, L7/L12 and L14 and the 30-S-subunit proteins S12 and S19. In the presence of thiostrepton, which prevents binding of EF-G to the ribosome, there was a considerable decrease in the yield of each of these cross-linked complexes. The data suggest that EF-G is bound close to the ribosomal subunit interface.

Kinetic suppression of translational errors by (p)ppGpp

The effect of (p)ppGpp on the accuracy of translation was studied in vitro using a poly(U)-programmed poly(Phe)-synthesizing system operating at incorporation rates and missense error frequencies close to the values obtained in vivo. Simulation of a relaxed phenotype in vitro was accomplished by limitation of the cognate aminoacyl-tRNA synthetase, while the noncognate aminoacyl-tRNA synthetase was included at saturating concentrations. This protocol yielded a Phe-tRNA-starved steady state system displaying the expected decrease in Phe polymerization rates accompanied by a drastic increase in relative Leu misincorporation errors. The use of purified enzymes permitted us to assay for the effects of the individual nucleotides ppGpp and ppGpp as well as their potential targets, the elongation factors Tu and G, upon the missense error rates. Our results support the conclusion that ppGpp reduces misincorporation in a starved in vitro system by preferentially inhibiting EF-Tu. The details of the proposed mechanism and their relevance to an in vivo situation are discussed.

Rate of elongation of polyphenylalanine in vitro

Ribosomes purified from Escherichia coli were preincubated with AcPhe-tRNA and poly(U). Then purified components necessary for polypeptide synthesis were added. Incubation of the complete system led to a burst of elongation which lasted for nearly 10 s. During the initial burst approximately 10% of the ribosomes participated in the elongation of poly(Phe) at an average rate per ribosome close to eight peptide bonds/s. The missense error rate with leucine was 4 x 10(-4) during the burst. Accordingly, the preincubated elongation system functions at a rate, as well as an accuracy, close to those of protein synthesis in vivo.

Concerning the mode of action of micrococcin upon bacterial protein synthesis

The antibiotic, micrococcin, binds to complexes formed between bacterial 23-S ribosomal RNA and ribosomal protein L11 and, in doing so, inhibits of thiostrepton. In assay systems simulating partial reaction of protein synthesis, micrococcin inhibits a number of processes believed to involve the ribosomal A site while stimulating GTP hydrolysis dependent upon ribosomes and elongation factor EF-G. The latter effect is not observed upon ribosomes lacking a protein homologous with protein L11. Nor is it apparent upon those containing 23-S RNA previously subjected to the action of a specific methylase known to render ribosomes resistant to thiostrepton. It is concluded that stimulation by micrococcin of factor-dependent GTP hydrolysis results from the binding of the drug to its normal target site which involves 23-S RNA and protein L11.