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

Fierce poison to others: the phenomenon of bacterial dependence on antibiotics

Beyond the development of resistance, the effects of antibiotics on bacteria and microbial communities are complex and far from exhaustively studied. In the context of the current global antimicrobial resistance crisis, understanding the adaptive and physiological responses of bacteria to antimicrobials is of paramount importance along with the development of new therapies. Bacterial dependence on antibiotics is a phenomenon in which antimicrobials instead of eliminating the pathogens actually provide a boost for their growth. This trait comprises an extreme example of the complexities of responses elicited by microorganisms to these drugs. This compelling evolutionary trait was readily described along with the first wave of antibiotics use and dependence to various antimicrobials has been reported. Nevertheless, current molecular characterizations have been focused on dependence on vancomycin, linezolid and colistin, three critically important antibiotics frequently used as last resource therapy for multi resistant pathogens. Outstanding advances have been made in understanding the molecular basis for the dependence to vancomycin, including specific mutations involved. Regarding linezolid and colistin, the general physiological components affected by the dependence, namely ribosomes and membrane function respectively, have been established. Nonetheless the implications of antibiotic dependence in clinically relevant features, such as virulence, epidemics, relationship with development of resistance, diagnostics and therapy effectiveness require clarification. This review presents a brief introduction of the phenomenon of bacterial dependence to antibiotics and a summary on early and current research concerning the basis for this trait. Furthermore, the available information on the effect of dependence in key clinical aspects is discussed. The studies performed so far underline the need to fully disclose the biological and clinical significance of this trait in pathogens to successfully assess its role in resistance and to design adjusted therapies.

Comparison of repressor and transcriptional attenuator systems for control of amino acid biosynthetic operons

In bacteria, expression from amino acid biosynthetic operons is transcriptionally controlled by two main mechanisms with principally different modes of action. When the supply of an amino acid is in excess over demand, its concentration will be high and when the supply is deficient the amino acid concentration will be low. In repressor control, such concentration variations in amino acid pools are used to regulate expression from the corresponding amino acid synthetic operon; a high concentration activates and a low concentration inactivates repressor binding to the operator site on DNA so that initiation of transcription is down or up-regulated, respectively. Excess or deficient supply of an amino acid also speeds or slows, respectively, the rate by which the ribosome translates mRNA base triplets encoding this amino acid. In attenuation of transcription, it is the rate by which the ribosome translates such "own" codons in the leader of an amino acid biosynthetic operon that decides whether the RNA polymerase will continue into the operon, or whether transcription will be aborted (attenuated). If the ribosome rate is fast (excess synthesis of amino acid), transcription will be terminated and if the rate is slow (deficient amino acid supply) transcription will continue and produce more messenger RNAs. Repressor and attenuation control systems have been modelled mathematically so that their behaviour in living cells can be predicted and their system properties compared. It is found that both types of control systems are unexpectedly sensitive when they operate in the cytoplasm of bacteria. In the repressor case, this is because amino acid concentrations are hypersensitive to imbalances between supply and demand. In the attenuation case, the reason is that the rate by which ribosomes translate own codons is hypersensitive to the rate by which the controlled amino acid is synthesised. Both repressor and attenuation mechanisms attain close to Boolean properties in vivo: gene expression is either fully on or fully off except in a small interval around the point where supply and demand of an amino acid are perfectly balanced.Our results suggest that repressors have significantly better intracellular performance than attenuator mechanisms. The reason for this is that repressor, but not attenuator, mechanisms can regulate expression from biosynthetic operons also when transfer RNAs are fully charged with amino acids so that the ribosomes work with maximal speed.

The accuracy of codon recognition by polypeptide release factors

The precision with which individual termination codons in mRNA are recognized by protein release factors (RFs) has been measured and compared with the decoding of sense codons by tRNA. An Escherichia coli system for protein synthesis in vitro with purified components was used to study the accuracy of termination by RF1 and RF2 in the presence or absence of RF3. The efficiency of factor-dependent termination at all sense codons differing from any of the three stop codons by a single mutation was measured and compared with the efficiency of termination at the three stop codons. RF1 and RF2 discriminate against sense codons related to stop codons by between 3 and more than 6 orders of magnitude. This high level of accuracy is obtained without energy-driven error correction (proofreading), in contrast to codon-dependent aminoacyl-tRNA recognition by ribosomes. Two codons, UAU and UGG, stand out as hotspots for RF-dependent premature termination.

Host mutations (miaA and rpsL) reduce tetracycline resistance mediated by Tet(O) and Tet(M)

The effects of mutations in host genes on tetracycline resistance mediated by the Tet(O) and Tet(M) ribosomal protection proteins, which originated in Campylobacter spp. and Streptococcus spp., respectively, were investigated by using mutants of Salmonella typhimurium and Escherichia coli. The miaA, miaB, and miaAB double mutants of S. typhimurium specify enzymes for tRNA modification at the adenosine at position 37, adjacent to the anticodon in tRNA. In S. typhimurium, this involves biosynthesis of N6-(4-hydroxyisopentenyl)-2-methylthio-adenosine (ms2io6A). The miaA mutation reduced the level of tetracycline resistance mediated by both Tet(O) and Tet(M), but the latter showed a greater effect, which was ascribed to the isopentenyl (i6) group or to a combination of the methylthioadenosine (ms2) and i6 groups but not to the ms2 group alone (specified by miaB). In addition, mutations in E. coli rpsL genes, generating both streptomycin-resistant and streptomycin-dependent strains, were also shown to reduce the level of tetracycline resistance mediated by Tet(O) and Tet(M). The single-site amino acid substitutions present in the rpsL mutations were pleiotropic in their effects on tetracycline MICs. These mutants affect translational accuracy and kinetics and suggest that Tet(O) and Tet(M) binding to the ribosome may be reduced or slowed in the E. coli rpsL mutants in which the S12 protein is altered. Data from both the miaA and rpsL mutant studies indicate a possible link between stability of the aminoacyl-tRNA in the ribosomal acceptor site and tetracycline resistance mediated by the ribosomal protection proteins.

Effect of the nucleotide-37 on the interaction of tRNA(Phe) with the P site of Escherichia coli ribosomes

The method of anticodon loop replacement has been used to make derivatives of yeast tRNA(Phe) with the substitution at position 37 (tRNA(Phe)GAAA) and at the anticodon(tRNA(Phe)GCAG). A quantitative study of the interaction of various types of deacylated yeast tRNA(Phe) (tRNA(Phe)+Y, tRNA(Phe)GAAA, tRNA(Phe)-Y) with the P site of the [70S ribosome*poly(U)]-complex was carried out at different Mg2+ concentrations and temperatures. The presence and nature of the nucleotide situated at the 3'-end of the anticodon are essential for such interaction in E coli ribosomes. Replacement of the Y base with the unmodified adenosine decreases the interaction enthalpy from 39 kcal/mol to 24 kcal/mol, whereas its removal reduces the interaction enthalpy to 16 kcal/mol. Replacement of the second anticodon nucleotide, adenosine, with cytosine further reduces the enthalpy to 6 kcal/mol, which is typical of tRNA-P site interaction in the absence of poly(U). In the absence of poly(U) the affinity of tRNA(PheY) for the P site of the 70S ribosome is five times lower than the affinity of tRNA(Phe+Y) or tRNA(Phe)GCAG. Thus, in the ribosome the modified nucleotide stabilizes the codon-anticodon interaction through its stacking interaction with the codon-anticodon base stack. In addition, this decreases the free energy of binding as a result of the interaction of the modified nucleotide itself with the hydrophobic center of the P site.

In vitro analysis of translational rate and accuracy with an unmodified tRNA

Escherichia coli tRNA(Phe) transcript lacking all the modified nucleosides was investigated in an in vitro translation system. To estimate the affinity of tRNA toward EF-Tu, Kd and K-1 were measured by the nuclease protection assay, and it was shown that the absence of modifications decreases ternary complex stability less than 2-fold. The activity of unmodified Phe-tRNA(Phe) on E. coli ribosomes was compared to modified Phe-tRNA(Phe) using the framework of the kinetic proofreading mechanism (Thompson & Dix, 1982) with both cognate and noncognate codons. Values of the individual rate constants in the elongation process showed that the modifications increased the accuracy of translation by (1) decreasing the rate of dipeptide synthesis and (2) increasing the rate of rejection with noncognate codons.

Thiolation of transfer RNA in Escherichia coli varies with growth rate

We have used an affinity electrophoresis assay which when combined with Northern hybridization techniques permits us to estimate the degree of thiolation of individual tRNA species in Escherichia coli. We observe that the levels of 4-thio 2'(3')-uridine (4-thioU) in many but not all tRNAs varies dramatically at different bacterial growth rates: Five tRNAs are completely thiolated at all growth rates, while another eight tRNAs are incompletely thiolated and the fraction of the unthiolated form of these tRNA species increases as the growth rates increase. Transfer RNA(2Glu) contains 4-thioU as well as (methylamino)methyl-2-thio uridine (mnm(5)2-thioU). The level of mnm(5)2-thioU of tRNA(2Glu) is invariant with growth rate. Surprisingly, none of the thirteen tRNA species that we have studied is completely unmodified in all growth media. In particular, at the slowest growth rates every tRNA class that we have studied contains a form that has 4-thioU residues.

Kinetic properties of Escherichia coli ribosomes with altered forms of S12

E. coli ribosomes with alterations in S12 leading to streptomycin resistance (SmR), dependence (SmD) and pseudodependence (SmP) were studied with the quench-flow technique. Kinetic changes at the various steps of the elongation cycle were identified. The rate of hydrolysis of GTP in the ternary complex in the ribosomal A-site is decreased drastically in SmD and moderately in SmP in relation to wild-type ribosomes. Addition of streptomycin restores much of the wild-type behaviour. The SmD, SmP and SmR ribosomes have an enhanced GTP-hydrolysis idling reaction on EF-Tu, which is correlated with how aggressive proofreaders these ribosomes are in steady-state assays. We use our in vitro findings to discuss the in vivo physiology of these mutants as well as mechanistic features of E. coli translation.

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).

Antifungal properties of the immunosuppressant FK-506: identification of an FK-506-responsive yeast gene distinct from FKB1

FK-506 is a novel and potent antagonist of T-cell activation and an inhibitor of fungal growth. Its immunosuppressive activity can be antagonized by the structurally related antibiotic rapamycin, and both compounds interact with cytoplasmic FK-506-binding proteins (FKBPs) in T cells and yeast cells. In this paper, we show that FK-506 and two analogs inhibit vegetative growth of Saccharomyces cerevisiae in a fashion that parallels the immunosuppressive activity of these compounds. Yeast mutants resistant to FK-506 were isolated, and at least three complementation groups (fkr1, fkr2, and fkr3) were defined. These fkr mutants show no alteration in their levels of FK-506-binding activity. Likewise, strains carrying null alleles of FKB1 (the yeast gene coding for the FKBP) remain FK-506 sensitive, indicating that depletion of yeast FKBP is not sufficient to confer an FK-506 resistance phenotype, although fkb1 null mutants are resistant to rapamycin. FKB1 does not map to the three fkr loci defined here. These results suggest that yeast FKBP mediates the inhibitory effect of rapamycin but that at least one other protein is directly involved in mediating the activity of FK-506. Interestingly, the ability of FK-506 to rescue a temperature-sensitive growth defect of the fkr3 mutant suggests that the FKR3 gene may define such a protein.

Antifungal properties of the immunosuppressant FK-506: identification of an FK-506-responsive yeast gene distinct from FKB1

FK-506 is a novel and potent antagonist of T-cell activation and an inhibitor of fungal growth. Its immunosuppressive activity can be antagonized by the structurally related antibiotic rapamycin, and both compounds interact with cytoplasmic FK-506-binding proteins (FKBPs) in T cells and yeast cells. In this paper, we show that FK-506 and two analogs inhibit vegetative growth of Saccharomyces cerevisiae in a fashion that parallels the immunosuppressive activity of these compounds. Yeast mutants resistant to FK-506 were isolated, and at least three complementation groups (fkr1, fkr2, and fkr3) were defined. These fkr mutants show no alteration in their levels of FK-506-binding activity. Likewise, strains carrying null alleles of FKB1 (the yeast gene coding for the FKBP) remain FK-506 sensitive, indicating that depletion of yeast FKBP is not sufficient to confer an FK-506 resistance phenotype, although fkb1 null mutants are resistant to rapamycin. FKB1 does not map to the three fkr loci defined here. These results suggest that yeast FKBP mediates the inhibitory effect of rapamycin but that at least one other protein is directly involved in mediating the activity of FK-506. Interestingly, the ability of FK-506 to rescue a temperature-sensitive growth defect of the fkr3 mutant suggests that the FKR3 gene may define such a protein.

Error-prone EF-Tu reduces in vivo enzyme activity and cellular growth rate

Mutations in Salmonella typhimurium encoding error-prone EF-Tu (tufA8, tufB103) enhance translational error levels and also cause a reduced growth rate. The relative changes in error level and growth rate are inversely related and dependent on the status of the two tuf genes. Possible causes of the reduced growth rate were investigated. Several important parameters with the potential to alter growth rate (the EF-Tu-ribosome interaction, the in vivo elongation rate and the processivity of translation), are all relatively unaffected by the tuf mutations. The small reduction in processivity observed in some strains is not quantitatively related to the growth rate reduction. Instead, the error-enhancing mutations are associated with a large reduction in the specific activity of a test protein, beta-galactosidase, suggesting by inference that the reduced growth rate is a consequence of the synthesis of error-containing proteins.

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.

Impaired in vitro kinetics of EF-Tu mutant Aa

The kirromycin-resistant EF-Tu mutant Aa, previously shown to be an antisuppressor for nonsense and missense suppressor tRNAs, has been characterised in a poly(U)-primed translation system in vitro. Two major defects were found in the function of the mutant. First, the dissociation constant for Aa binding to Phe-tRNA(Phe) was increased tenfold compared to wild-type EF-Tu. Second, kcat/Km for the interaction between the EF-Tu.GTP.aa-tRNA complex and the ribosome was decreased by the mutation to one third of its wild-type value. No differences were observed between mutant and wild-type factor in the regeneration of EF-Tu.GTP from EF-Tu.GDP via EF-Ts or in the mistranslation frequency by Leu-tRNA(4Leu). The relation between the in vitro results and the mutant phenotype in vivo is discussed.

How many EF-Tu molecules participate in aminoacyl-tRNA binding and peptide bond formation in Escherichia coli translation?

We have observed that two EF-Tu.GTP cycles are required to make one peptide bond during steady-state translation in an accurate and fast poly(U) translation system prepared from Escherichia coli. We have also found that there are two complexes of EF-Tu.GTP bound to one molecule of aminoacyl-tRNA under our experimental conditions. We suggest, on the basis of these data, that aminoacyl-tRNA enters the ribosomal A-site in a pentameric complex together with two EF-Tu and two GTP molecules. When the tRNA is delivered to the ribosome two GTP molecules are hydrolyzed. It is possible that the functional role of such an EF-Tu dimer is related to the function of the two L7/L12 dimers in the large ribosomal subunit.

Genetic and physiological relationships among the miaA gene, 2-methylthio-N6-(delta 2-isopentenyl)-adenosine tRNA modification, and spontaneous mutagenesis in Escherichia coli K-12

The miaA tRNA modification gene was cloned and located by insertion mutagenesis and DNA sequence analysis. The miaA gene product, tRNA delta 2-isopentenylpyrophosphate (IPP) transferase, catalyzes the first step in the biosynthesis of 2-methylthio-N6-(delta 2-isopentenyl)-adenosine (ms2i6A) adjacent to the anticodon of several tRNA species. The translation start of miaA was deduced by comparison with mod5, which encodes a homologous enzyme in yeasts. Minicell experiments showed that Escherichia coli IPP transferase has a molecular mass of 33.5 kilodaltons (kDa). Transcriptional fusions, plasmid and chromosomal cassette insertion mutations, and RNase T2 mapping of in vivo miaA transcription were used to examine the relationship between miaA and mutL, which encodes a polypeptide necessary for methyl-directed mismatch repair. The combined results showed that miaA, mutL, and a gene that encodes a 47-kDa polypeptide occur very close together, are transcribed in the same direction in the order 47-kDa polypeptide gene-mutL-miaA, and likely form a complex operon containing a weak internal promoter. Three additional relationships were demonstrated between mutagenesis and the miaA gene or ms2i6A tRNA modification. First, miaA transcription was induced by 2-aminopurine. Second, chromosomal miaA insertion mutations increased the spontaneous mutation frequency with a spectrum distinct from mutL mutations. Third, limitation of miaA+ bacteria for iron, which causes tRNA undermodification from ms2i6A to i6A, also increased spontaneous mutation frequency. These results support the notion that complex operons organize metabolically related genes whose primary functions appear to be completely different. In addition, the results are consistent with the idea that mechanisms exist to increase spontaneous mutation frequency when cells need to adapt to environmental stress.

Presence of the hypermodified nucleotide N6-(delta 2-isopentenyl)-2-methylthioadenosine prevents codon misreading by Escherichia coli phenylalanyl-transfer RNA

The overall structure of transfer RNA is optimized for its various functions by a series of unique post-transcriptional nucleotide modifications. Since many of these modifications are conserved from prokaryotes through higher eukaryotes, it has been proposed that most modified nucleotides serve to optimize the ability of the tRNA to accurately interact with other components of the protein synthesizing machinery. When a cloned synthetic Escherichia coli tRNAPhe gene was transfected into a bacterial host that carried a defective phenylalanine tRNA-synthetase gene, tRNAPhe was overexpressed by 11-fold. As a result of this overexpression, an undermodified tRNAPhe species was produced that lacked only N6-(delta 2-isopentenyl)-2-methylthioadenosine (ms2i6A), a hypermodified nucleotide found immediately 3' to the anticodon of all major E. coli tRNAs that read UNN codons. To investigate the role of ms2i6A in E. coli tRNA, we compared the aminoacylation kinetics and in vitro codon-reading properties of the ms2i6A-lacking and normal fully modified tRNAPhe species. The results of these experiments indicate that while ms2i6A is not required for normal aminoacylation of tRNAPhe, its presence stabilizes codon-anticodon interaction and thereby prevents misreading of the genetic code.

Molecular cloning of the Escherichia coli miaA gene involved in the formation of delta 2-isopentenyl adenosine in tRNA

Escherichia coli mia strains were shown to lack delta 2-isopentenylpyrophosphate transferase activity, the first step in the synthesis of the 2-methylthio derivative of 6-(delta 2-isopentenyl) adenosine (ms2i6A). A double mutant, rpsL (Smp) miaA, was streptomycin dependent. The wild-type miaA gene was cloned by selecting for lambda recombinant bacteriophage which eliminated the streptomycin-dependent phenotype and was subsequently recloned into plasmid vectors. The cloned miaA gene restored the ms2i6A modification to tRNA. The miaA gene mapped to 95 min on the E. coli map, and we propose the order mutL-miaA-hflA-purA.

Is efficiency of suppressor tRNAs controlled at the level of ribosomal proofreading in vivo?

Ribosomal rpsD mutations did not stimulate nonsense suppressor tRNAs in a general manner according to their increased ribosomal ambiguity and decreased proofreading efficiency. Streptomycin, which stimulates error production by blocking proofreading in vitro, did not increase efficiency of suppressor tRNAs in strains with normal or streptomycin-resistant (rpsL) ribosomes. It did so only in combination with one rpsL mutation which is associated with streptomycin pseudodependence.

Mutant forms of tufA and tufB independently suppress nonsense mutations

The level of nonsense suppression in Salmonella typhimurium carrying error-enhancing mutations in either or both of the genes coding for the elongation factor EF-Tu has been measured. Suppression of both UGA and UAG is observed. There is no significant suppression of any of six UAA sites tested. Nonsense suppression does not require that both genes for EF-Tu be mutant. Strains carrying one mutant and one wild-type tuf gene suppress nonsense mutations. The level of suppression increases approximately additively when both tuf genes are mutant. It is suggested that these mutant forms of EF-Tu act independently of each other to suppress nonsense mutations. Suppression is not observed at all UGA and UAG sites, but instead shows a strong site specificity.

Effects of miaA on translation and growth rates

We have measured the growth rates and elongation rates for different proteins in wild-type, miaA, rpsL, and miaA, rpsL double mutants of Escherichia coli in the presence as well as the absence of streptomycin. The data show that while miaA and rpsL mutants inhibit elongation rates to equivalent levels, miaA inhibits the growth rate twice as effectively as does rpsL. The double mutant is more effectively inhibited than either single mutant and Sm repairs in part the growth rate as well as protein elongation rates. The data suggest that the conditional streptomycin-dependent phenotype of the double mutant cannot be due simply to the depressed polypeptide elongation rates of the double mutant.

Does streptomycin cause an error catastrophe?

We have examined the interpretation that streptomycin kills a bacterial culture by initiating the so-called error catastrophe. In particular, we asked whether the increased translational error rate induced by the antibiotic gives rise to an autocatalytic loss of functional fidelity of the devices responsible for gene expression, which ultimately causes the death of the culture. We have analyzed the performance characteristics of one of these devices, namely the ribosome in streptomycin-treated bacteria. We find that, although the treated ribosomes are constructed from error-containing proteins, they are not significantly different in elongation rate and fidelity from those ribosomes taken from untreated bacteria. We conclude that the bacteriocidal effect of streptomycin is not due to the initiation of an error catastrophe.

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).

Constraints on the accuracy of messenger RNA movement

Theoretical as well as experimental studies of translational accuracy have most often been concerned with the selection of aminoacyl-tRNA by codon-programmed ribosomes. The selection of the successive codons on the mRNA has received much less attention, probably because it represents both conceptually and experimentally, a much more demanding physical problem. Nevertheless, it would seem that errors in the selection of the codon are potentially much more destructive than errors in selection of aminoacyl-tRNA species. This can be appreciated from the following.