• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Nov 25, 2003; 100(24): 13863–13868.
Published online Nov 13, 2003. doi:  10.1073/pnas.1936123100
PMCID: PMC283512

Coevolution of an aminoacyl-tRNA synthetase with its tRNA substrates


Glutamyl-tRNA synthetases (GluRSs) occur in two types, the discriminating and the nondiscriminating enzymes. They differ in their choice of substrates and use either tRNAGlu or both tRNAGlu and tRNAGln. Although most organisms encode only one GluRS, a number of bacteria encode two different GluRS proteins; yet, the tRNA specificity of these enzymes and the reason for such gene duplications are unknown. A database search revealed duplicated GluRS genes in >20 bacterial species, suggesting that this phenomenon is not unusual in the bacterial domain. To determine the tRNA preferences of GluRS, we chose the duplicated enzyme sets from Helicobacter pylori and Acidithiobacillus ferrooxidans. H. pylori contains one tRNAGlu and one tRNAGln species, whereas A. ferrooxidans possesses two of each. We show that the duplicated GluRS proteins are enzyme pairs with complementary tRNA specificities. The H. pylori GluRS1 acylated only tRNAGlu, whereas GluRS2 was specific solely for tRNAGln. The A. ferrooxidans GluRS2 preferentially charged equation M1. Conversely, A. ferrooxidans GluRS1 glutamylated both tRNAGlu isoacceptors and the equation M2 species. These three tRNA species have two structural elements in common, the augmented D-helix and a deletion of nucleotide 47. It appears that the discriminating or nondiscriminating natures of different GluRS enzymes have been derived by the coevolution of protein and tRNA structure. The coexistence of the two GluRS enzymes in one organism may lay the groundwork for the acquisition of the canonical glutaminyl-tRNA synthetase by lateral gene transfer from eukaryotes.

Faithful protein biosynthesis relies on the correct attachment of amino acids to their corresponding (cognate) tRNAs catalyzed by the aminoacyl-tRNA synthetases (AARSs) (1). Given the presence of 20 canonical amino acids, one would expect each organism to contain at least 20 AARSs to ensure high accuracy during aminoacylation of the complete set of tRNAs. However, this is true only for eukarya and some bacteria. Most of the other bacterial organisms (2, 3), all known archaea (4, 5), and eukaryotic organelles (6) lack the AARS specific for glutamine. In addition, a canonical asparaginyl-tRNA synthetase is missing from many archaeal and bacterial genomes (7), whereas a canonical cysteinyl-tRNA synthetase is not distinguishable in some methanogenic archaea (8).

In addition to direct acylation of the cognate tRNA by AARSs, an indirect pathway was found to be responsible for the formation of both Gln-tRNAGln and Asn-tRNAAsn. In these cases, formation of the mischarged product (Glu-tRNAGln and Asp-tRNAAsn, respectively) is a required intermediate in the synthesis of the cognate amino acid–tRNA pairs. To accomplish this, the indirect pathway to Gln-tRNAGln formation uses a nondiscriminating (ND) form of glutamyl-tRNA synthetase (GluRS) that is able to efficiently glutamylate both tRNAGlu and tRNAGln (9). In the next step, a tRNA-dependent amidotransferase, Glu-tRNAGln amidotransferase, amidates Glu-tRNAGln to the correctly charged Gln-tRNAGln, which is then used as a substrate for protein synthesis (reviewed in ref. 10).

In bacteria, the ND Bacillus subtilis GluRS has been extensively studied (9). Due to the lack of a canonical glutaminyl-tRNA synthetase (GlnRS) in this organism, the ND-GluRS is an essential enzyme in Gln-tRNA formation as it generates Glu-tRNAGln. This product is then converted to Gln-tRNAGln by Glu-tRNAGln amidotransferase (1012). However, the mechanism of how the ND-GluRS recognizes two different tRNA substrates is not known, and the tRNA identity set for such an enzyme has never been determined. In contrast, organisms with a canonical GlnRS enzyme possess a discriminating GluRS (D-GluRS), which recognizes only tRNAGlu. The tRNA identity elements of a D-GluRS enzyme were reported for E. coli (13). Major identity determinants for this enzyme are located in the augmented D-helix (indicated in Fig. 1), which is formed by nucleotides in the D-stem and D-loop, as well as the bases connecting the acceptor helix and D-stem, and some bases in the extra arm (13). In vitro mutagenesis of Thermus thermophilus D-GluRS expanded its tRNA specificity to the recognition of tRNAGlu with a glutamine-specific anticodon (14). However, this work did not afford real insight into a ND-GluRS, because the tRNA specificity was not changed.

Fig. 1.
Cloverleaf representation of A. ferrooxidans tRNAGlu and tRNAGln species. The bases that form part of the augmented D-helix (13), in both isoacceptors of tRNAGlu and in equation M20, are shown in white letters and black shadow.

In addition to the lack of some synthetases (see above), whole genome analysis also revealed several organisms containing more than one AARS for certain amino acids. In many cases, the function of such a redundancy is unknown; however, in some organisms, the roles of the duplicated AARSs have been explained (15). A second aspartyl-tRNA synthetase in Deinococcus radiodurans, for instance, is shown to be involved in tRNA-dependent asparagine biosynthesis (7). In E. coli, two genes for lysyl-tRNA synthetase (LysRS) exist; one LysRS gene (lysS) is constitutively expressed, and the other one is heat-inducible (lysU) (16, 17). Similarly, two paralogous gltX genes (encoding GluRS) exist in some proteobacteria and a scattering of other bacterial groups (18); however, the function of this duplication is not clear. There are indications that the two GluRSs in Helicobacter pylori preferentially glutamylate either tRNAGlu or tRNAGln.** To investigate further the duplicated gltX genes, we chose the Acidithiobacillus ferrooxidans and H. pylori systems and examined the substrate preferences of the two GluRS enzymes that coexist in these organisms.

Materials and Methods

Oligonucleotides, DNA Sequencing, and Radiochemicals. Oligonucleotides were synthesized, and DNAs were sequenced by the Keck Foundation Biotechnology Resource Laboratory at Yale University. Uniformly labeled [14C]Glu (254 mCi/mmol) and [14C]Gln (244 mCi/mmol) and [γ-32P]ATP (6,000 Ci/mmol) were from Amersham Pharmacia Biosciences.

Cloning of gltX Genes from H. pylori and A. ferrooxidans. The H. pylori genes (HP0476, gltX1; HP0643, gltX2) were obtained by PCR from the genomic DNA (strain 26695) by using the Expand High Fidelity PCR System (Roche Molecular Biochemicals). Oligonucleotide primers introduced an NdeI site at the 5′ end and a SapI site at the 3′ end of each gene. The PCR products, digested with NdeI and SapI, were cloned into the pTYB1 vector (New England BioLabs) digested with the same enzymes, to form the plasmids pHPgltX1 and -X2. The identity of the clones was determined by DNA sequencing.

Analysis (http://tigrblast.tigr.org/ufmg/index.cgi?database=a_ferrooxidansseq) of the incomplete A. ferrooxidans genome sequence revealed two gltX genes. They were cloned into the pCR2.1-TOPO vector (Invitrogen) after PCR amplification of genomic A. ferrooxidans DNA (ATCC 23270) by using oligonucleotides that introduced an EcoRI site on both ends of each gene. After sequence confirmation, the genes were cloned into the EcoRI site of pGEX-2T (Amersham Biosciences) to yield the plasmids pAFgltX1 and -X2.

Expression and Purification of GluRSs. The plasmids pHPgltX1 and -X2 were transformed into Escherichia coli BL21(DE3) (Invitrogen). The expression of both genes and the purification and cleavage of the intein fusion proteins were done as described (New England BioLabs). For the overexpression of A. ferrooxidans genes and purification of the GST-fusion protein, we followed the previously described protocols (19).

Preparation of Unfractionated tRNA. The A. ferrooxidans tRNAGln genes (sequence from the TIGR database), were constructed by synthesis of the corresponding oligonucleotides and subsequent ligation into the pKK223-3 vector (Amersham Biosciences) and digested with EcoRI and HindIII enzymes, generating the plasmids pJSQUUG and pJSQCUG. These were transformed into E. coli DH5α. Purification of unfractionated tRNA containing the specified tRNA gene transcript was carried out as described (20). Comparison of aminoacylation reactions by using E. coli GlnRS and unfractionated tRNA obtained from the cells, transformed with pJSQUUG, pJSQCUG, or empty vector, showed that the A. ferrooxidans equation M3 and equation M4 isoacceptors comprised between 10% and 20% of the total tRNA.

To obtain mature unfractionated tRNA from both organisms, H. pylori cells (strain 43526) were grown (21), and the unfractionated tRNA was prepared as described (20). A. ferrooxidans (strain 19859) was grown in a bioreactor in 9-K medium [0.4 g/liter (NH4)2SO4,0.4g/l MgSO4(7H2O), 0.056 g/lK2HPO4(3H2O)/5g/l FeSO4(7H2O) adjusted to pH 1.6 with concentrated H2SO4, and sterilized by filtration. The ferrous ion was regenerated by 0.5 A of current]. Total tRNA extracted from 0.5 g of A. ferrooxidans cells was purified by the Qiagen RNA/DNA kit as described by the manufacturer (Qiagen, Valencia, CA).

Purification of tRNAGln and tRNAGlu by Affinity Chromatography. The A. ferrooxidans equation M5 and equation M6 (see above) and E. coli tRNAGln and tRNAGlu were purified from unfractionated tRNA by affinity chromatography on immobilized T. thermophilus EF-Tu (22) with the following specifications. Unfractionated E. coli tRNA was aminoacylated with E. coli GlnRS or GluRS. The Gln-tRNAGln or Glu-tRNAGlu was separated from the uncharged tRNA by the formation of a ternary complex with the EF-Tu-GTP immobilized on a Ni-NTA-agarose column. After the elution of Gln-tRNAGln or Glu-tRNAGlu, the tRNA was deacylated for 15 min at 65°C in 0.1 mM borate-KOH, pH 9.0. The pure tRNAGln was not contaminated with tRNAGlu and vice versa, as judged by aminoacylation with pure E. coli GluRS or GlnRS, respectively.

Aminoacylation Assays with GluRS1 and -RS2. In vitro acylations with H. pylori and A. ferrooxidans GluRS enzymes was carried out at 37°C in an 80-μl reaction mixture containing 100 mM Hepes-KOH, pH 7.2; 30 mM KCl; 12 mM MgCl2; 5 mM ATP; 2 mM DTT; 25 μM [14C]Glu plus 75 μM unlabeled glutamate and 40–60 μg of unfractionated tRNA or 0.7–1.3 μg of pure tRNA. The reactions were started by adding the enzyme (100–150 nM). Aliquots (15 μl) were taken at different times, spotted on 3MM filter paper discs, and washed twice with 5% trichloroacetic acid. For acidic gel separations (see below), unfractionated H. pylori or A. ferrooxidans tRNA was charged in the presence of 0.1 mM unlabeled glutamate.

Acid Urea Gel Electrophoresis of tRNA and Aminoacyl-tRNA. This method (23) allows the separation of charged from uncharged tRNA due to a difference in electrophoretic mobility between the two species. Hybridization of a sequence-specific probe permits the determination of the identity of the tRNA on the gel. Unfractionated tRNA (80 μg) from H. pylori or A. ferrooxidans was glutamylated for 15 min at 37°C with the homologous synthetases (see above), and half of the reaction was deacylated for 15 min at 65°C with 0.1 mM borate-KOH, pH 9.0. After extraction with phenol (saturated with 0.3 M sodium acetate, pH 5.0/10 mM EDTA), the charged and uncharged tRNAs were recovered by ethanol precipitation. They were dissolved (at a final concentration of 5 μg/μl) in sample buffer (0.1 M sodium acetate, pH 5.0/8 M urea/0.05% bromophenol blue/0.05% xylene cyanol) The RNA samples (10 μg) were loaded on a 9.5% polyacrylamide gel (50 × 20 cm, 0.4 mm thick) containing 7 M urea and 0.1 M sodium acetate, pH 5.0 and run at 4°C, 600 V in 0.1 M sodium acetate, pH 5.0, for 40 h.

Detection of the tRNAs was performed by Northern blotting. For this purpose, the portion of the gel containing the tRNAs was electroblotted onto a Hybond-N + membrane (Amersham Bioscience) by using a Hoefer Electroblot apparatus (Amersham Bioscience) at 10 V for 10 min and then at 30 V for 90 min with 10 mM Tris acetate, pH 7.8/5 mM sodium acetate/0.5 mM Na-EDTA as transfer buffer. The membranes were then baked at 72°C for 2 h. The tRNAs were detected by hybridization with a 5′-32P-labeled oligodeoxyribonucleotide probe. For H. pylori tRNAGln, the probe complementary to nucleotides 1–21. H. pylori has two slightly different equation M7 genes; however, by Northern blot, we detected the product of one gene as the main tRNAGlu (data not shown); the probe used to detect that tRNA was complementary to nucleotides 1–21 (5′-TAACCACTAGATGAAGGAGCC-3′).

The A. ferrooxidans genome sequence reveals two tRNAGln isoacceptors, equation M8 and equation M9 (see Fig. 1). The probes were complementary to nucleotides 13–37 and 53–75 for equation M10 and to nucleotides 14–37 for equation M11. In the case of tRNAGlu, we also found two isoacceptors that were detected by using the probes complementary to nucleotides 15–39 for each tRNA.


Selection of the Experimental Systems. The presence of two paralogous gltX genes has been previously noticed in a few bacterial genomes (18). To survey the presence of duplicated GluRS genes or pieces thereof, we performed tblastn searches (24) of all currently available bacterial genomes. As a result, gltX duplications were found in the genomes of >20 organisms (Table 1). With the exception of two members of the Clostridiales and the deep-rooted Thermotoga maritima, all organisms with duplicated gltX genes are proteobacteria.

Table 1.
Organisms with two gltX genes

Apart from duplications of the entire gltX gene, many genomes were found to contain gluX (yadB in E. coli), a truncated form of gltX encoding a GluRS protein that lacks the enzyme's entire C-terminal anticodon-binding domain (≈35% of the total protein) and has no ascribed function; it was suggested to be a pseudogene (25) and shown to be not essential in E. coli (ref. 26, J.C.S., unpublished results). The relatively wide distribution (from E. coli to D. radiodurans and Corynebacterium) of the genes encoding these precisely truncated GluRS fragments suggests that these proteins may have a yet unknown function. Nonetheless, GluX was not analyzed further.

Class I lysyl-tRNA synthetase (LysRS) is structurally very related (27) and displays sequence homology to GluRS near its C terminus (COG 0008). Its rare occurrence in bacteria is confined largely to α-proteobacteria. It is interesting to note that the entire set of α-proteobacteria with two gltX genes (Table 1) is identical to that containing a class I LysRS. Whether this is a functionally significant correlation remains to be seen.

Phylogenetic analysis of GluRS sequences in a general tree showed a segregation of the duplicated GluRS proteins into two groups (Fig. 2). The first enzyme, GluRS1, branched with discriminating GluRS of E. coli and other proteobacteria. The second enzyme, GluRS2, appeared quite diverged but was in some cases closer to the ND-GluRSs, e.g., B. subtilis GluRS. Therefore, we assumed that GluRS1 is a canonical discriminating synthetase, whereas GluRS2 is the ND enzyme capable of forming Glu-tRNAGln, the tRNA substrate in the transamidation pathway of Gln-tRNA formation (reviewed in ref. 10). The key enzyme in this pathway, the Glu-tRNAGln amidotransferase (encoded by gatCAB) (1012), is present in all of the organisms with two gltX genes (Table 1). A relationship, mediated by a common partner, has been established between H. pylori gltX2 and gatA in a genomic protein–protein interaction map (29).

Fig. 2.
Phylogeny of bacterial-type GluRS proteins. Amino acid sequences were aligned by using the clustalw program (Version 1.82) (28). Initial analyses were performed with GluRS proteins from numerous bacterial organisms, and later the same was confirmed with ...

To complement the enzyme analysis, we proceeded to examine the tRNAGlu and tRNAGln species in these organisms. Our survey showed that some of the organisms possess only one isoacceptor of each tRNAGlu and tRNAGln (e.g., H. pylori and Campylobacter jejuni), whereas others have two (e.g., A. ferrooxidans and Novosphingobium aromaticivorans). Closer examination brought better insight: in most cases where a genome contains two tRNAGln isoacceptors, they are dissimilar and structurally distinguished by the presence or absence of an augmented D-helix (13). The tRNAGlu and tRNAGln species of A. ferrooxidans are depicted in Fig. 1; the equation M12 has an augmented D-helix, whereas the equation M13 lacks this element; in addition it has a different acceptor stem sequence. Thus, the A. ferrooxidans equation M14 may be a substrate for the E. coli-type GluRS (GluRS1). Together, these findings suggest that the two GluRS proteins may bind structurally different tRNA substrates. Therefore, we chose A. ferrooxidans and H. pylori, organisms with duplicated gltX genes, to examine the possible reason for the presence of two GluRS proteins by determining the tRNA substrate preferences of these enzymes.

The Two H. pylori GluRSs Possess Complementary tRNA Specificities. The gltX1 and gltX2 genes from H. pylori and A. ferrooxidans were cloned into expression vectors to encode intein or GST fusion proteins, respectively. After expression in E. coli and affinity chromatography, the native H. pylori enzymes were generated by intein cleavage, whereas the A. ferrooxidans GluRS enzymes were used as GST fusion proteins. Preliminary experiments showed that GluRS1 and -RS2 from both organisms efficiently charge total E. coli tRNA with glutamate but, as expected, not glutamine (data not shown). We then proceeded to determine the enzymes' tRNA specificities.

Because H. pylori possess only one tRNAGlu and one tRNAGln species, we analyzed this system first. Because H. pylori tRNAs are very similar in sequence to E. coli tRNAs (>80% sequence identity for both species), we proceeded to purify tRNAGlu and tRNAGln from unfractionated E. coli tRNA by EF-Tu affinity chromatography (22). The resulting pure E. coli tRNA species were then acylated with glutamate by the H. pylori GluRS1 and -RS2 enzymes. GluRS1 charged tRNAGlu well, whereas GluRS2 did not recognize this tRNA (Fig. 3A). In contrast, the result with tRNAGln was completely reversed, because this tRNA was charged exclusively by GluRS2 (Fig. 3B). To ascertain the validity of these results obtained in charging the heterologous E. coli tRNA, we also analyzed glutamylated mature H. pylori tRNA samples by Northern blot of acid/urea gels (23). Glu-tRNAs generated by H. pylori GluRS1 and -RS2 and the control Gln-tRNA formed by E. coli GlnRS were used. After running the gel, Northern blot analysis was performed by using a specific probe for each tRNA of interest. Glu-tRNAGln was formed only in the presence of GluRS2 (Fig. 4A, lane 3), whereas Glu-tRNAGlu was formed only in the presence of the GluRS1 (Fig. 4A, lane 6). These results confirmed that GluRS1 and -RS2 recognize the homologous tRNAGlu and tRNAGln, respectively (Fig. 4A, compare lanes 2 and 3 to lanes 6 and 8). As expected, Gln-tRNAGln, made by E. coli GlnRS, moved more slowly than Glu-tRNAGln (Fig. 4A, compare lanes 1 and 3), which suggests that this method could also be used to provide a clue to the amino acid identity of the reaction products.

Fig. 3.
Aminoacylation of tRNAGlu and tRNAGln from E. coli and A. ferrooxidans. The reaction conditions were as indicated in Materials and Methods by using the pure tRNAs. (A) E. coli tRNAGlu.(B) E. coli tRNAGln.(C) A. ferrooxidans equation M21. (D) A. ferrooxidans equation M22. The ...
Fig. 4.
Northern blot analysis of H. pylori and A. ferrooxidans tRNA after in vitro acylation by the different GluRS enzymes. The blots were probed with 32P-labeled oligonucleotide complementary to the anticodon stem or plus acceptor stem (see Materials and Methods ...

tRNA Specificity of GluRS Is Linked to Idiosyncratic tRNA Features. The A. ferrooxidans GluRS system is expected to be more complex than that of H. pylori, because there are two tRNAGlu and two tRNAGln isoacceptors (Fig. 1). Although these four tRNAs have different anticodons, both tRNAGlu isoacceptors are uniform in their secondary structure and similar to E. coli and H. pylori tRNAGlu, one of the tRNAGln isoacceptors shows a striking divergence in its structure and is in some aspects more similar to tRNAGlu than tRNAGln (Fig. 1). We expressed the A. ferrooxidans tRNA genes in E. coli and purified them by EF-Tu chromatography (22). The aminoacylation results show clearly that equation M15 was not charged by A. ferrooxidans GluRS1 but moderately by GluRS2 and well by E. coli GlnRS (Fig. 3C).†† However, equation M16 was charged well by GluRS1 and E. coli GlnRS but less well by GluRS2 (Fig. 3D). As expected, only A. ferrooxidans GluRS1 could charge E. coli tRNAGlu (Fig. 3A). These tRNA specificities were confirmed by glutamylation of mature unfractionated A. ferrooxidans tRNA followed by acidic gel analysis, as described above for H. pylori. The A. ferrooxidans GluRS1 substrates are both tRNAGlu isoacceptors (Fig. 4 B and C, lane 7) and equation M17 (Fig. 4C, lane 3), but not equation M18 (Fig. 4B, lane 3). On the other hand, the A. ferrooxidans GluRS2 recognized primarily equation M19 (Fig. 4B, lane 5) but also exhibited some charging of the other tRNAGln isoacceptor and tRNAGlu (Figs. 4B, lane 9, and C, lanes 5 and 9), thus retaining some characteristics of the ND-GluRS enzymes. That GluRS1 charged exclusively tRNAs with augmented D-helix structure strongly suggests this is the major identity element for this group of enzymes.

Altogether, these results imply the existence of a definite relationship between GluRS types and the specific structure of their tRNA substrates.


RNA Sequence Suggests Structural Differences in Bacterial tRNAGln Species. Inspection of the tRNAGlu and tRNAGln sequences in the Genomic tRNA Database (http://rna.wustl.edu/tRNAdb) revealed that most archaeal and eukaryal tRNAGlu and tRNAGln species (with the exception of Saccharomyces cerevisiae) possess sequences containing the augmented D-helix element. Thus, this RNA structural motif may be a relatively old element contributing to the function of these tRNAs. However, restricted to bacteria, structurally dissimilar (with respect to the augmented D-helix) tRNAGlu and tRNAGln species, are frequently observed. In addition, tRNAGln isoacceptors exist that differ in the presence or lack of augmented D-helix (e.g., H. pylori, one species with a nonaugmented D-helix; A. ferrooxidans, two tRNAGln species, one with and one without augmented D-helix; Caulobacter crescentus, one species with an augmented D-helix). Thus, it is reasonable to assume the existence of different GluRS types as being related to this unusual tRNA diversity.

Bacterial gltX Gene Duplications Occurred Several Times. Phylogenetic analyses suggest a common origin of the GluRS and GlnRS family of AARSs (25, 30, 31). It is widely accepted that GlnRS probably evolved within the eukaryal ancestor through duplication of a gltX gene and later spread by horizontal gene transfer to the bacterial domain (25). Thus, it is plausible that a similar gltX gene duplication event may have occurred in some bacterial species and gave rise to a new subgroup of genes encoding enzymes with the tRNA specificity of GlnRS. Phylogenetic analysis shows that the tRNAGln-specific GluRS (GluRS2) described here arose from the ancestral gene duplication within bacterial GluRSs (ref. 30; this work), possibly in α-proteobacteria. A standard ND-GluRS may have preceded the GluRS2-type enzyme, because A. ferrooxidans GluRS2 retains traces of tRNAGlu-binding capability. Such a gene duplication must have taken place several times in bacterial evolution, because GluRS2 enzymes in proteobacteria, in Thermoanaerobacter tengcongensis and in Thermotoga maritima, for instance, are clearly not of the same origin. This gene duplication in T. maritima is a much more recent event, and its two GluRSs group together in the phylogenetic tree (Fig. 2).

GluRS Duplication, a Strategy to Enhance Fidelity of Aminoacylation? The consistent occurrence of structurally different tRNAGlu and tRNAGln sequences in organisms with a gltX1 and a gltX2 gene suggests that the separation of tRNA specificities in the corresponding GluRS enzymes may have provided higher fidelity of aminoacylation. It is reasonable to assume that use of specialized enzymes like GluRS1, GluRS2, or GlnRS would be favorable in terms of aminoacylation accuracy, compared with the broader specificity of GluRS-ND enzymes.

GluRS2, a New Type of AARSs. The essential ND-GluRS (e.g., B. subtilis GluRS) found in many organisms is a synthetase specific for one amino acid (Glu) and two acceptor RNA families (tRNAGln and tRNAGlu). This simple view of a single type of ND-GluRS among bacteria is not universally correct. Our results uncovered at least two additional types of GluRS. The first is still a ND-GluRS that, in addition to tRNAGlu, recognizes only the subset of the tRNAGln isoacceptor having an augmented D-helix (e.g., A. ferrooxidans GluRS1). The second enzyme is unique among AARSs; it is a new “mischarging” synthetase with an amino acid specificity (Glu) different from its single tRNA specificity (tRNAGln). This GluRS (e.g., H. pylori GluRS2) is an enzyme halfway between the discriminating GluRS and GlnRS enzymes, each of which recognizes only (and correctly) their cognate amino acid/tRNA pair.

GluRS2 and GlnRS Evolved Independently. It is tempting to speculate that GluRS2 might be on its way to developing into a true bacterial-origin GlnRS that has not yet evolved a glutamine-binding capacity. However, all known bacterial GlnRS proteins show the highest similarity to eukaryal GlnRSs (25) and are apparently not descendants of the tRNAGln-specific GluRS2 proteins discussed here. Furthermore, examination of the still incomplete Desulfitobacterium hafniense genome revealed (Table 1) the coexistence of ORFs for GluRS1 and -RS2 proteins with the predicted complementary tRNA specificities and for a canonical GlnRS. This implies that the GluRS2 and GlnRS enzymes evolved independently. Possibly the coexistence of GluRS1 and -RS2 may have prepared the organism for the acquisition of a canonical GlnRS by horizontal gene transfer from eukarya and subsequent loss of the nonorthologous gltX2 gene. This might have been an easier scenario than evolving a new amino acid specificity (Glu to Gln) in gltX2. These assumptions could explain the relatively narrow distribution of organisms with duplicated gltX genes and the very rare example (D. hafniense) of an organism with two gltx genes and a canonical GlnRS.

Diversity of Bacterial GluRS Enzymes. Based on currently available information, there appear to be three groups of GluRS covering most of the examples investigated. Many bacteria [e.g., Sinorhizobium meliloti (3) or Chlamydia trachomatis (20)] have a single ND-GluRS and tRNAGln isoacceptor without the augmented D-helix. The GluRS2-type enzyme is adapted to recognize only tRNAGln lacking the augmented D-helix, a tRNA species also recognized by the canonical GlnRS. Last, there is GluRS1, an enzyme on its way to becoming a discriminating GluRS and preferentially recognizing tRNA species with the augmented D-helix. Therefore, the discriminating or ND natures of GluRS may have formed by coevolution of protein and tRNA structure.

Considering all of the data, it is clear that the GluRS/GlnRS family of enzymes shows exquisite diversity of gene arrangement and pathways, which might be linked to a relatively young evolution of canonical GlnRS activity.


We are grateful to Jesse Rinehart for assistance with the gel-shift experiment; to Klaus Melchers (ALTANA Pharma, Konstanz, Germany) for the gift of H. pylori DNA; to David Graham and Michael Ibba for critical questions; and to Milind Deshpande and Jolene Ehler for assistance in growing H. pylori cells at the Center for Biocatalysis at the University of Iowa. This work was supported by grants from the National Institute of General Medical Sciences (to L.D. and D.S.), the Department of Energy (to D.S.), the National Aeronautics and Space Administration (to D.S.), the U.S. Department of Agriculture (to L.D.), and Fondo Nacional de Desarrollo Científico y Tecnológico Chile (Grant 1020087 to O.O.).


Abbreviations: AARS, aminoacyl-tRNA synthetase; ND, nondiscriminating; GluRS, glutamyl-tRNA synthetase; GlnRS, glutaminyl-tRNA synthetase.

Note Added in Proof. During the course of publication, a similar study with comparable results was published (32).


**Hendrickson, T. L., Skouloubris, S., de Reuse, H. & Labigne, A. (2002) Abstr. Papers Am. Chem. Soc. 224, 121.

††A. ferrooxidans GluRS2 does not reach the same plateau as E. coli GlnRS (Fig. 3C). This may be due to the lack of a nucleotide modification in the heterologously produced equation M27 or to the presence of a supernumerary one introduced by E. coli. This is plausible, because mature A. ferrooxidans equation M28 is fully charged by GluRS2 (Fig. 4B, lane 5).


1. Ibba, M. & Söll, D. (2000) Annu. Rev. Biochem. 69, 617–650. [PubMed]
2. Wilcox, M. & Nirenberg, M. (1968) Proc. Natl. Acad. Sci. USA 61, 229–236. [PMC free article] [PubMed]
3. Gagnon, Y., Lacoste, L., Champagne, N. & Lapointe, J. (1996) J. Biol. Chem. 271, 14856–14863. [PubMed]
4. White, B. N. & Bayley, S. T. (1972) Can. J. Biochem. 50, 600–609. [PubMed]
5. Tumbula, D. L., Becker, H. D., Chang, W.-Z. & Söll, D. (2000) Nature 407, 106–110. [PubMed]
6. Schön, A., Kannangara, C. G., Gough, S. & Söll, D. (1988) Nature 331, 187–190. [PubMed]
7. Min, B., Pelaschier, J. T., Graham, D. E., Tumbula-Hansen, D. & Söll, D. (2002) Proc. Natl. Acad. Sci. USA 99, 2678–2683. [PMC free article] [PubMed]
8. Li, T., Graham, D., Stathopoulos, C., Haney, P., Kim, H.-S., Vothknecht, U., Kitabatake, M., Hong, K.-W., Eggertsson, G., Curnow, A. W., et al. (1999) FEBS Lett. 462, 302–306. [PubMed]
9. Lapointe, J., Duplain, L. & Proulx, M. (1986) J. Bacteriol. 165, 88–93. [PMC free article] [PubMed]
10. Feng, L., Tumbula-Hansen, D., Min, B., Namgoong, S., Salazar, J., Orellana, O. & Söll, D. (2003) in Aminoacyl-tRNA Synthetases, eds. Ibba, M., Francklyn, C. S. & Cusack, S. (Landes Bioscience, Georgetown, TX), in press.
11. Strauch, M. A., Zalkin, H. & Aronson, A. I. (1988) J. Bacteriol. 170, 916–920. [PMC free article] [PubMed]
12. Curnow, A. W., Hong, K. W., Yuan, R., Kim, S. I., Martins, O., Winkler, W., Henkin, T. M. & Söll, D. (1997) Proc. Natl. Acad. Sci. USA 94, 11819–11826. [PMC free article] [PubMed]
13. Sekine, S.-I., Nureki, O., Sakamoto, K., Niimi, T., Tateno, M., Gō, M., Kohno, T., Brisson, A., Lapointe, J. & Yokoyama, S. (1996) J. Mol. Biol. 256, 685–700. [PubMed]
14. Sekine, S.-I., Nureki, O., Shimada, A., Vassylyev, D. G. & Yokoyama, S. (2001) Nat. Struct. Biol. 8, 203–206. [PubMed]
15. Brown, J. R., Gentry, D., Becker, J. A., Ingraham, K., Holmes, D. J. & Stanhope, M. J. (2003) EMBO Rep. 4, 692–698. [PMC free article] [PubMed]
16. Hirshfield, I. N., Bloch, P. L., Bogelen, V. & Neidhardt, F. C. (1981) J. Bacteriol. 146, 345–351. [PMC free article] [PubMed]
17. Clark, R. L. & Neidhardt, F. C. (1990) J. Bacteriol. 172, 3237–3243. [PMC free article] [PubMed]
18. Woese, C. R., Olsen, G., Ibba, M. & Söll, D. (2000) Microbiol. Mol. Biol. Rev. 64, 202–236. [PMC free article] [PubMed]
19. Salazar, J. C., Zúñiga, R., Lefimil, C., Söll, D. & Orellana, O. (2001) FEBS Lett. 491, 257–260. [PubMed]
20. Raczniak, G., Becker, H. D., Min, B. & Söll, D. (2001) J. Biol. Chem. 276, 45862–45867. [PubMed]
21. Deshpande, M., Calenoff, E. & Daniels, L. (1995) Appl. Environ. Microbiol. 61, 2431–2435. [PMC free article] [PubMed]
22. Ribeiro, S., Nock, S. & Sprinzl, M. (1995) Anal. Biochem. 228, 330–335. [PubMed]
23. Varshney, U., Lee, C.-P. & RajBhandary, U. L. (1991) J. Biol. Chem. 266, 24712–24718. [PubMed]
24. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990) J. Mol. Biol. 215, 403–410. [PubMed]
25. Lamour, V., Quevillon, S., Diriong, S., N′Guyen, V. C., Lipinski, M. & Mirande, M. (1994) Proc. Natl. Acad. Sci. USA 91, 8670–8674. [PMC free article] [PubMed]
26. Gerdes, S. Y., Scholle, M. D., Campbell, J. W., Balazsi, G., Daugherty, M. D., Somera, A. L., Kyrpides, N. C., Anderson, I., Gelfand, M. S., Bhattacharya, A., et al. (2003) J. Bacteriol. 185, 5673–5684. [PMC free article] [PubMed]
27. Terada, T., Nureki, O., Ishitani, R., Ambrogelly, A., Ibba, M., Söll, D. & Yokoyama, S. (2002) Nat. Struct. Biol. 9, 257–262. [PubMed]
28. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Higgins, D. G. (1997) Nucleic Acids Res. 25, 4876–4882. [PMC free article] [PubMed]
29. Rain, J. C., Selig, L., de Reuse, H., Battaglia, V., Reverdy, C., Simon, S., Lenzen, G., Petel, F., Wojcik, J., Schachter, V., Chemama, Y., Labigne, A. & Legrain, P. (2001) Nature 409, 211–215. [PubMed]
30. Brown, J. R. & Doolittle, W. F. (1999) J. Mol. Evol. 49, 485–495. [PubMed]
31. Handy, J. & Doolittle, R. F. (1999) J. Mol. Evol. 49, 709–715. [PubMed]
32. Skouloubris, S., Ribas de Pouplana, L., de Reuse, H. & Hendrickson, T. L. (2003) Proc. Natl. Acad. Sci. USA 100, 11297–11302. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...