|Home | About | Journals | Submit | Contact Us | Français|
Elongation Factor Tu (EF-Tu) binds and loads elongating aminoacyl-transfer RNAs (aa-tRNAs) onto the ribosome for protein biosynthesis. Many bacteria biosynthesize Gln-tRNAGln and Asn-tRNAAsn by an indirect, two-step pathway that relies on the misacylated tRNAs Glu-tRNAGln and Asp-tRNAAsn as essential intermediates. Previous thermodynamic and experimental analyses have demonstrated that Thermus thermophilus EF-Tu does not bind Asp-tRNAAsn and predicted a similar discriminatory response against Glu-tRNAGln [Asahara, H. and Uhlenbeck, O., (2005), Biochemistry 46, 6194–6200 and Roy, H. et al. (2007) Nucleic Acids Research 35, 3420–3430]. By discriminating against these misacylated tRNAS, EF-Tu plays a direct role in preventing misincorporation of aspartate and glutamate into proteins at asparagine and glutamine codons. Here we report the characterization of two different mesophilic EF-Tu orthologs, one from Escherichia coli, a bacterium that does not utilize either Glu-tRNAGln or Asp-tRNAAsn, and the second from Helicobacter pylori, an organism in which both misacylated tRNAs are essential. Both EF-Tu orthologs discriminate against these misacylated tRNAs, confirming the prediction that Glu-tRNAGln, like Asp-tRNAAsn, will not form a complex with EF-Tu. These results also demonstrate that the capacity of EF-Tu to discriminate against both of these aminoacyl-tRNAs is conserved even in bacteria like E. coli that do not generate either misacylated tRNA.
By necessity, translation of the genetic code proceeds with high fidelity. This accuracy is ensured by a variety of mechanisms including high specificity in transfer RNA (tRNA) aminoacylation (1), editing of misacylated tRNAs (2–4), and proofreading of codon-anticodon interactions in the A-site of the ribosome (3). However, many bacteria are missing genes for glutaminyl- and/or asparaginyl-tRNA synthetase (GlnRS and AsnRS, respectively), the enzymes that directly generate Gln-tRNAGln and Asn-tRNAAsn. In these organisms, Gln-tRNAGln and Asn-tRNAAsn are still obligate substrates for protein translation, but they are biosynthesized indirectly via parallel two-step processes that proceed through designed misacylation of these two tRNAs. First, a non-discriminating glutamyl-tRNA synthetase (ND-GluRS) (5, 6), a tRNAGln-specific glutamyl-tRNA synthetase (GluRS2) (7, 8), or a non-discriminating aspartyl-tRNA synthetase (ND-AspRS) (9–12) misacylates tRNAGln or tRNAAsn with glutamate or aspartate: the resultant products are the misacylated tRNAs Glu-tRNAGln and Asp-tRNAAsn, respectively. These misacylated tRNAs are subsequently converted to Gln-tRNAGln and Asn-tRNAAsn by a glutamine-dependent Asp-tRNAAsn/Glu-tRNAGln amidotransferase (Asp/Glu-Adt) (13, 14).
The use of misacylated Glu-tRNAGln and Asp-tRNAAsn as essential aminoacyl-tRNA precursors necessitates the existence of proofreading mechanisms to prevent their fatal misuse in ribosomal protein synthesis. Asp/Glu-Adt and elongation factor Tu (EF-Tu), the G-protein responsible for loading aminoacyl-tRNAs onto the ribosome, have recently emerged as key players in the maintenance of translational accuracy (15, 16). In Thermus thermophilus, Asp/Glu-Adt forms a “transamidosome” complex with ND-AspRS and tRNAAsn, but not with tRNAAsp, to sequester Asp-tRNAAsn and directly deliver it to Asp/Glu-Adt for repair (17, 18). The T. thermophilus ND-AspRS is an archaeal-type AspRS; the formation of a bacterial-type transamidosome, at least in a transient fashion, has been postulated and is supported by the observation that the affinity of Asp-tRNAAsn for Asp/Glu-Adt is improved upon the addition of Helicobacter pylori ND-AspRS (19).
EF-Tu is responsible for loading nearly all elongating aminoacyl-tRNAs into the A site of the ribosome at the expense of GTP (20). (The one major exception is selenocysteinyl-tRNASec, which has its own elongation factor (21)). Despite this broad specificity, EF-Tu does distinguish between different amino acids and tRNAs in its aminoacyl substrates (9, 15, 22–27). Asp-tRNAAsn shows little to no affinity to Thermus thermophilus EF-Tu˙GTP, under conditions where Asp-tRNAAsp and Asn-tRNAAsn each bind with low nanomolar Kd values (28). Binding of Glu-tRNAGln to T. thermophilus EF-Tu has not been directly measured but has been predicted to be ~2 kcal/mol less than that of Glu-tRNAGlu (24). Formation of a Pisum sativum Glu-tRNAGln˙EF-Tu˙GTP complex does not occur with this organism’s chloroplast EF-Tu, however binding of this aa-tRNA was observed with the E. coli EF-Tu ortholog (29). Taken together, these experiments demonstrate that both Asp/Glu-Adt and EF-Tu participate in preventing Asp-tRNAAsn and Glu-tRNAGln from entering the ribosome.
The role of EF-Tu as a proofreading protein has almost exclusively been studied using the thermophilic EF-Tu ortholog from T. thermophilus, an organism that utilizes GlnRS to directly generate Gln-tRNAGln but relies on indirect aminoacylation to generate Asp-tRNAAsnen route to Asn-tRNAAsn (9). Here we present the characterization of two mesophilic EF-Tu orthologs – The EF-Tu from Escherichia coli, an organism that uses GlnRS and AsnRS to directly generate both Gln-tRNAGln and Asn-tRNAAsn (30) and consequently does not require proofreading of Glu-tRNAGln and Asp-tRNAAsn, and the EF-Tu from Helicobacter pylori, a bacterium that lacks both GlnRS and AsnRS and consequently uses the Asp/Glu-Adt transamidation pathway for the biosynthesis of both Glu-tRNAGln and Asp-tRNAAsn (14, 31). Each of these EF-Tu orthologs discriminated against both misacylated tRNAs, demonstrating that EF-Tu’s role as a proofreading enzyme is conserved in both mesophilic and thermophilic bacteria, independent of the method by which Asn-tRNAAsn and Gln-tRNAGln are generated in vivo.
The E. coli trpA34 strain (provided by Professor Dieter Söll, Yale University) was transformed with either pQE-80L (empty vector) or pPTC001 (the plasmid encoding H. pylori ND-AspRS (12)). A single colony from each strain was used to inoculate a 200 mL overnight culture in M9 minimal media supplemented with 20 amino acids (20 µg/mL) and 0.5% glucose as a carbon source. The overnight culture was harvested by centrifugation at 5,000 rpm for 10 minutes at 4 °C. The cell pellets were gently washed twice with 200 mL of M9 minimal media supplemented with 0.5% glucose and each encoded amino acid (20 µg/mL) except tryptophan. The cell pellets were then gently resuspended in the same media. These cell suspensions were used to streak M9 minimal media agar plates supplemented with 0.5% glucose, 19 amino acids (20 µg/mL), in the absence and presence of tryptophan, and 0, 0.1, or 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). Plates were incubated at 37 °C and the growth of cells was assessed after 24 hours.
The plasmid encoding a six histidine tagged E. coli EF-Tu construct (32) was obtained from Professor David Draper (Johns Hopkins University). The gene encoding H. pylori EF-Tu (tufB, Hp1205) was amplified from H. pylori strain 26695 chromosomal DNA (ATCC) using the following primers: TLH-EFTU-A (5’-ggagaaggatccatggcaaaagaaaagtttaacagaactaagcc) and TLH-EFTU-B (5’-tttttgaagcttttattcaataatattgctcacaacacc); these primers introduced BamHI and HindIII sites (bold) onto the 5’ and 3’ ends of the tufB gene. The resultant product was cloned into the BamHI and HindIII sites of pQE-80L (Qiagen). This vector (pPTC033) introduces an N-terminal 6-histidine tag onto the encoded protein.
Both EF-Tu orthologs were purified using Ni-NTA spin columns (Qiagen), adjusting the manufacturer’s instructions so that all buffers contained 50 µM guanosine diphosphate (GDP). Eluted EF-Tu was concentrated using microcon spin filters (Millipore) and exchanged into buffer containing 50 mM Tris˙HCl (pH 7.5), 25 mM KCl, 15 mM MgCl2, 2 mM β–mercaptoethanol, 50 µM GDP (33). EF-Tu was stored at -20 °C in the same buffer with 50% glycerol.
Each H. pylori tRNA was overexpressed in vivo in E. coli as previously described (8, 12). Small RNAs were purified using the Nucleobond RNA/DNA Maxi Kit (Clontech), according to the manufacturer’s instructions. This procedure produces a mixture of tRNAs that is enriched with the H. pylori tRNA of interest but also contains total E. coli tRNA as well as other small RNAs. Yields of a given tRNA isoacceptor were quantified by aminoacylation by GluRS1, GluRS2, or ND-AspRS (see below). Typical yields ranged from 500–1600 pmoles per OD260. Ec AspRS was used to verify the amount of Ec tRNAAsp in samples of overexpressed Hp tRNAAsn.
Prior to aminoacylation, the five different H. pylori tRNAs (tRNAGlu1, tRNAGlu2, tRNAGln, tRNAAsp and tRNAAsn) were adjusted by the addition of batch E. coli tRNA (Sigma-Aldrich) to similar levels of purity (~500 pmoles/OD). H. pylori GluRS1, was used to aminoacylate tRNAGlu1 and tRNAGlu2; H. pylori GluRS2 was used to aminoacylate tRNAGln; and H. pylori ND-AspRS was used to aminoacylate tRNAAsp and tRNAAsn. Each tRNA (~20–100 µM) was incubated for 1 hour at 37 °C in 10 mM Hepes˙OH (pH 7.5), with 2 mM ATP, and 8 mM MgCl2, glutamic acid or aspartic acid, and 1 µM enzyme. The amino acid concentrations were varied as needed for downstream experiments and ranged from 2 to 200 µM. For assays with low levels of amino acid, commercial 3H-[3, 4]-glutamic acid (Perkin Elmer, 45.4 Ci/mmol) or 3H-[2, 3]-aspartic acid (Amersham, 34 Ci/mmol) was used without subsequent dilution. For higher amino acid concentration assays, unlabelled amino acid was used alone or supplemented with trace levels of radioactivity.
Binding of different aminoacyl-tRNAs to EF-Tu was initially analyzed using a Ni2+ affinity column, essentially as previously described (9). Prior to column preparation, tRNA aminoacylation reactions were conducted as described above using 2–3 µM 3H-amino acid (undiluted). Columns were loaded with 400 µL Ni-NTA slurry (Qiagen) and washed with 1 mL Buffer A (50 mM Tris˙Cl, 10 mM MgCl2, 50 mM NH4Cl, 50 mM KCl, 5 mM β-mercaptoethanol, 15 mM guanosine triphosphate (GTP), 15 mM phosphoenolpyruvate (PEP)). EF-Tu˙GDP (1400 pmoles in 200 µL) was added to the resin and Buffer A (200 µL) was added; columns were agitated at 4 °C for one hour. After EF-Tu binding, buffer was allowed to flow through. EF-Tu˙GDP was then converted to EF-Tu˙GTP by incubation for 30 minutes at room temperature in 200 µL Buffer B (Buffer A supplemented with 0.3 mg/mL pyruvate kinase (PK)). After activation, buffer was allowed to flow through. Columns were then treated with 1–7 pmoles 3H-aa-tRNA in Buffer B (200 µL final volume) and agitated at room temperature for 30 minutes. The total concentration of EF-Tu on these columns is 3.5 µM based on total binding of 1400 pmoles EF-Tu and the combined volume of 200 µL resin and 200 µL aa-tRNA-containing buffer; because the enzyme is resin-bound, local concentrations are likely to be higher. Initial flow through was collected for quantification of unbound aa-tRNA. Columns were washed ten times each with 200 µL Buffer A. Columns were subsequently treated five times each with 200 µL Buffer A, supplemented with 0.1 M NaCl. Bound 3H-aa-tRNA was eluted by treating columns seven times with 200 µL Buffer C (0.1 M Sodium Borate (pH 7.5), 1 M NaCl). All fractions were collected and quantified by liquid scintillation separately. Reported data (Figure 2) represents the average of experiments run in triplicate and the error bars reflect standard error.
EF-Tu˙GDP was converted to EF-Tu˙GTP according to protocols previously reported (34). Briefly, EF-Tu (5 – 15 µM as determined by Bradford Assay (Biorad)) was agitated at room temperature in 50 mM Hepes (pH 7.0), NH4Cl (150 mM for RNAse protection assay and 50 mM for deacylation assays), 20 mM MgCl2, 5 mM β-mercaptoethanol, 20 µM GTP, 3 mM PEP, 50 µg/mL PK. EF-Tu˙GTP was kept on ice and used immediately.
The concentration of active EF-Tu˙GTP (defined as protein capable of binding aa-tRNAs) was determined using an RNAse protection assay modified from that previously described (35). EF-Tu˙GTP (~5 µM as determined by Bradford assay) was incubated at 4 °C for 30 minutes (50 µL total volume) under buffer conditions described above supplemented with either H. pylori3H-Asp-tRNAAsp (34 Ci/mmol) or 3H-Glu-tRNAGlu1 (45.4 Ci/mmol); the final concentration of each aa-tRNA was 15–20 µM. RNAse A (Sigma-Aldrich) was added to a final concentration of 0.01 mg/mL and the reaction was incubated at 4 °C for 20 seconds. The reaction was quenched with 100 µL 10% trichloroacetic acid (TCA, Fisher) that had been supplemented with 0.1 mg/mL total E. coli tRNA (Sigma-Aldrich). These assays were conducted in triplicate. An aliquot of each reaction (100 µL) was loaded onto a 2.5 cm filter (Whatman) that had been saturated with 200 µL 5% TCA. Filters were washed 3 × 15 minutes in 5% TCA followed by a single 5 minute wash in EtOH. Pads were dried and soaked in 3 mL Ecolite + Scintillation cocktail (Mp Biomedicals) and the amount of intact aa-tRNA was quantified using an LS6500 Scintillation counter (Beckman). These data were corrected for the amount of intact aminoacylated-tRNA remaining in a control reaction where the EF-Tu was omitted.
The active concentration of H. pylori EF-Tu was typically ~10% of that predicted by Bradford Assay, whereas that of E. coli EF-Tu was typically ~40% of that observed by Bradford Assay. The percent active EF-Tu was independent of the nature of the cognate aminoacyl-tRNA (Asp-tRNAAsp versus Glu-tRNAGlu1). This enigmatic disparity between the level of active EF-Tu and amount of protein present is a well-documented phenomenon (27, 36, 37). Unless otherwise stated, all reported EF-Tu˙GTP concentrations are based on the results of this binding assay.
Dissociation constants were determined for each H. pylori aa-tRNA using a protocol adapted from that developed by Uhlenbeck and colleagues (27, 34). A stock of active EF-Tu˙GTP was prepared using the above procedure in the presence of 150 mM NH4Cl. Twelve concentrations of EF-Tu˙GTP (25 µL each, ranging in concentration from ~1 nM – 2 mM) were prepared by serial dilution in a 96-well plate at 4 °C. To each well, aa-tRNA was added (25 µL, ~ 20 nM final concentration in 150 mM Hepes (pH 7.0), 150 mM NH4Cl, 20 mM MgCl2, 5 mM β-mercaptoethanol, 20 µM GTP, 3 mM PEP, 50 µg/mL PK). After incubation for 30 minutes, RNAse A was added to a final concentration of 0.01 µg/mL. After a 20 second incubation on ice, aa-tRNA degradation was quenched by the addition of 100 µL 10% TCA that had been supplemented with 0.1 mg/mL unfractionated E. coli tRNA (Sigma-Aldrich). The amount of intact aa-tRNA was quantified as described above. Data was analyzed using Kaleidagraph v. 3.6.2. Kd values were determined by plotting the concentration of EF-Tu˙GTP˙aa-tRNA complex against the initial concentration of EF-Tu˙GTP and applying the formula: [Complex] = ((Kd + [EF-Tu]i + [aa-tRNA]i) – ((Kd + [EF-Tu]i + [aa-tRNA]i)2 – (4 * [aa-tRNA]i * [EF-Tu]i))0.5) / 2. All experiments were conducted in triplicate and error bars represent standard error from these replicates.
A stock of active EF-Tu˙GTP was prepared using the above procedure in the presence of 50 mM NH4Cl. Reactions containing 1 µM active EF-Tu˙GTP and 50 nM 3H-aa-tRNA in 50 mM Hepes (pH 7.0), 50 mM NH4Cl, 20 mM MgCl2, 5 mM β-mercaptoethanol, 20 µM GTP, 3 mM PEP, 50 µg/mL PK were incubated at 4 °C for 30 minutes. Solutions were transferred to 37 °C and the rate of deacylation was determined by quenching aliquots (10 µL) at various times over 90 minutes. Each time point was quenched onto pads prepared with 5% TCA and counted as described above to assess the degree of deacylation.
A ribonuclease (RNAse) protection assay (27, 34) was used to determine apparent dissociation constants (K d) between each of the two non-cognate and three cognate aa-tRNAs and EF-Tu˙GTP (Figure 1 and Table I). As expected, each cognate aa-tRNA bound both EF-Tu orthologs with high affinity. The E. coli EF-Tu revealed tighter Kd values ranging from about 35 to 95 nM, whereas those for the H. pylori EF-Tu ranged from about 90 to 180 nM. (For reasons not well understood, only about 35% of the Glu-tRNAGlu1 was protected in this RNAse assay. However, complete complex formation was observed using a deacylation assay, see Figure 2 and Figure S1.) When the two misacylated tRNAs were examined neither bound significantly to either EF-Tu ortholog. (Analysis with Ec D-AspRS confirmed that the low levels of complex formation observed with Asp-tRNAAsn (Figure 3 A and C) are due to contaminating E. coli Asp-tRNAAsp, data not shown.) In fact, neither ortholog was capable of protecting even 5% of the Asp-tRNAAsn or Glu-tRNAGln present in each experiment, even at the highest concentration examined (~1.6 – 3 µM EF-Tu). While these observations are consistent with previous in vitro experiments showing that E. coli EF-Tu can discriminate against T. thermophilus Asp-tRNAAsn, this is the first example of a quantitative in vitro assay demonstrating a similar discrimination against Glu-tRNAGln (9, 28).
Deacylation rates were examined for each of the five H. pylori aa-tRNAs being evaluated herein in the absence of protein and in the presence of either E. coli or H. pylori EF-Tu˙GTP (Figure 2 and Figure S1). Aminoacyl-tRNAs are inherently labile under aqueous conditions and this spontaneous hydrolysis is suppressed when an aa-tRNA is bound to EF-Tu (38). For this reason, deacylation assays offer a method for examining complex formation between EF-Tu and various aa-tRNAs that is orthogonal to the RNAse protection assay described above. This method offers the additional advantage that it does not rely on the action of an additional enzyme (e.g. RNAse). The results of these experiments recapitulate those from the RNAse protection assay: The three accurately aminoacylated aa-tRNAs were protected from deacylation in the presence of either EF-Tu orthologs (Figure S1), however the addition of EF-Tu did not protect either Asp-tRNAAsp or Glu-tRNAGln from hydrolysis (Figure 2).
It has previously been shown that overexpression of B. subtilis ND-AspRS leads to suppression of tryptophan auxotrophy in the E. coli strain trpA34 (39). This strain carries a GAT→AAT mutation at codon 60 of the trpA gene, introducing an Asp60Asn mutation into the α-subunit of tryptophan synthase and inactivating the enzyme; consequently, this strain requires exogenous tryptophan for viability (40). Because ND-AspRS can aminoacylate E. coli tRNAAsn with aspartate, expression of B. subtilis ND-AspRS facilitated the incorporation of aspartate into tryptophan synthase at position 60, via ribosomally accessible Asp-tRNAAsn, at sufficient levels to restore tryptophan synthase activity and viability in the absence of tryptophan (39). This observation is in apparent contradiction with the results described above which show that E. coli EF-Tu˙GTP does not bind Asp-tRNAAsn.
We similarly tested the overexpression of H. pylori ND-AspRS in trpA34 under varying conditions of induction. Tryptophan auxotrophy was suppressed specifically within a narrow range of IPTG concentrations (Figure 3). No growth was observed in the absence of tryptophan and IPTG (top left panel). When 0.1 mM IPTG was added, sufficient Asp-tRNAAsn was used in the biosynthesis of tryptophan synthase to suppress the requirement for tryptophan (middle left panel). At higher IPTG concentrations, lethality in the absence of tryptophan was restored, suggesting that levels of aspartate incorporation into proteins at asparagine codons had become too high to promote viability (left bottom panel). In the presence of tryptophan, these strains were all viable under the conditions screened (right panels). Interestingly, the strains grown in the presence of tryptophan were viable in the presence of 0.5 mM IPTG; a level high enough to introduce ND-AspRS toxicity in the -Trp counterpart. This viability is thought to be due to a higher resilience afforded by the healthier environment of the tryptophan enriched plates.
The observation that expression of H. pylori ND-AspRS can suppress tryptophan auxotrophy in the trpA34 strain demonstrates that this enzyme misacylates E. coli tRNAAsnin vivo and that E. coli EF-Tu can bind the resultant Asp-tRNAAsn at levels that are sufficient to promote errors in protein biosynthesis under selective conditions (similar to previous results with B. subtilis ND-AspRS (39)). Other complementation studies have also demonstrated in vivo incorporation of glutamate at glutamine codons via Glu-tRNAGln (41). To resolve these results with our quantitative demonstration that the E. coli EF-Tu˙GTP discriminates against Asp-tRNAAsn and Glu-tRNAGln, we sought to identify conditions where binding of Asp-tRNAAsn and/or Glu-tRNAGln to EF-Tu˙GTP could be recapitulated in vitro. The E. coli and H. pylori EF-Tu orthologs were both examined in order to determine whether differences exist in organisms that misacylate tRNAAsn and tRNAGln (H. pylori) versus organisms that accurately aminoacylate these tRNAs directly (E. coli).
Each relevant aminoacylated H. pylori tRNA (Asp-tRNAAsp, Asp-tRNAAsn, Glu-tRNAGlu1, Glu-tRNAGlu2, and Glu-tRNAGln, each with a radiolabel incorporated into the amino acid) was separately loaded onto a Ni2+ column containing either 6-His tagged E. coli EF-Tu˙GTP or H. pylori EF-Tu˙GTP at high local concentrations (≥3.5 µM). Based on a previously published approach (9), the columns were washed and then bound aa-tRNAs were eluted by treatment with high salt. Consistent with the in vivo tryptophan auxotrophy experiments, all five different H. pylori aa-tRNAs, including the two misacylated tRNAs, bound both EF-Tu orthologs under the conditions tested (Figure 4). Asp-tRNAAsp and Asp-tRNAAsn bound each EF-Tu at essentially identical levels (~80% bound); in contrast, Glu-tRNAGln did not bind either EF-Tu to the same extent as the two Glu-tRNAGlu acceptors (~70% bound versus 80–95 % bound, respectively).
We have demonstrated herein that two different mesophilic EF-Tu orthologs, one from H. pylori and one from E. coli, bind the accurately aminoacylated tRNAs Asp-tRNAAsp, Glu-tRNAGlu1, and Glu-tRNAGlu2 with low nanomolar affinities as expected. Under these same conditions, binding was not observed for either Asp-tRNAAsn or Glu-tRNAGln, the two misacylated tRNAs that are essential for H. pylori viability but not for E. coli. These two non-cognate species do exhibit a capacity to complex with EF-Tu under non-native conditions (e.g. overexpression of AspRS, Figure 3, or high local concentrations of EF-Tu, Figure 4). This last result (Figure 4) explains the discrepancy between in vivo experiments reported herein (Figure 3) and by Söll and colleagues (39, 40), which demonstrate formation of functional EF-Tu˙GTP˙Asp-tRNAAsn/Glu-tRNAGln complexes in vivo, and in vitro experiments (Figure 1–Figure 2 and references, which show that EF-Tu˙GTP discriminates against Asp-tRNAAsn and Glu-tRNAGln in favor of accurately aminoacylated tRNAs (9, 28, 29)). Clearly, endogenous levels of the aminoacyl-tRNA synthetases and EF-Tu are carefully balanced in vivo to prevent catastrophic errors. Perturbation of this system via overexpression of a non-discriminating aminoacyl-tRNA synthetase, in combination with the high in vivo concentration of EF-Tu (~170 µM (42)), ablates the proofreading capacity of EF-Tu such that Asp-tRNAAsn and Glu-tRNAGln are incorrectly taken into the ribosome at levels sufficient for complementation.
The observation that the two EF-Tu orthologs examined herein discriminate against Asp-tRNAAsn is consistent with work prior to the present study that had demonstrated that the T. thermophilus and E. coli EF-Tu orthologs both differentiate against this misaspartylated species in favor of Asp-tRNAAsp and Asn-tRNAAsn (28). With respect to Glu-tRNAGln, researchers have previously calculated that EF-Tu would be incapable of binding this misacylated species (24) (a result qualitatively observed with the chloroplast EF-Tu from P. sativum (29). Here we have quantitatively demonstrated that both the E. coli and H. pylori EF-Tu orthologs do indeed discriminate against this misacylated tRNA, with complex formation being disfavored by at least 2 kcal/mol compared to accurately aminoacylated tRNAs. In total, the results herein further support the hypothesis that EF-Tu has universally retained the ability to prevent the endogenously produced, potentially toxic, misacylated tRNAs Asp-tRNAAsn and Glu-tRNAGln from entering ribosomal translation. The observation that an organism such as E. coli, that contains no biochemical means of producing either misacylated tRNA, exhibits this vestigial proofreading capacity suggests that this discrimination mechanism existed in the last common universal ancestor along with the indirect transamidation pathway for Asn-tRNAAsn and Gln-tRNAGln biosynthesis.
The authors thank Professor Dieter Söll (Yale University) for providing a sample of the trpA34 strain, Marc Greenberg and David Draper (Johns Hopkins University) for generous access to their equipment, and David Draper (JHU), Olke Uhlenbeck, and Lee Sanderson (Northwestern University) for helpful discussions.
†This work was supported by a Grant from the National Institutes of Health (GM071480). P.C. was supported by a DPST Fellowship from the Royal Thai Government
Abreviations and Textual Footnotes: EF-Tu, non-discriminating aminoacyl-tRNA synthetase, aspartyl-tRNA synthetase, glutamyl-tRNA synthetase, proofreading