Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2006 Nov 28; 103(48): 18095–18100.
Published online 2006 Nov 16. doi:  10.1073/pnas.0608762103
PMCID: PMC1838712

Emergence of the universal genetic code imprinted in an RNA record


The molecular basis of the genetic code manifests itself in the interaction of the aminoacyl-tRNA synthetases and their cognate tRNAs. The fundamental biological question regarding these enzymes' role in the evolution of the genetic code remains open. Here we probe this question in a system in which the same tRNA species is aminoacylated by two unrelated synthetases. Should this tRNA possess major identity elements common to both enzymes, this would favor a scenario where the aminoacyl-tRNA synthetases evolved in the context of preestablished tRNA identity, i.e., after the universal genetic code emerged. An experimental system is provided by the recently discovered O-phosphoseryl-tRNA synthetase (SepRS), which acylates tRNACys with phosphoserine (Sep), and the well known cysteinyl-tRNA synthetase, which charges the same tRNA with cysteine. We determined the identity elements of Methanocaldococcus jannaschii tRNACys in the aminoacylation reaction for the two Methanococcus maripaludis synthetases SepRS (forming Sep-tRNACys) and cysteinyl-tRNA synthetase (forming Cys-tRNACys). The major elements, the discriminator base and the three anticodon bases, are shared by both tRNA synthetases. An evolutionary analysis of archaeal, bacterial, and eukaryotic tRNACys sequences predicted additional SepRS-specific minor identity elements (G37, A47, and A59) and suggested the dominance of vertical inheritance for tRNACys from a single common ancestor. Transplantation of the identified identity elements into the Escherichia coli tRNAGly scaffold endowed facile phosphoserylation activity on the resulting chimera. Thus, tRNACys identity is an ancient RNA record that depicts the emergence of the universal genetic code before the evolution of the modern aminoacylation systems.

Keywords: aminoacyl-tRNA synthetase, evolution, O-phosphoseryl-tRNA synthetase, tRNA identity

The aminoacyl-tRNA synthetases (aaRSs) (1) are stewards of the genetic code in all forms of life on Earth. It is the specificity of these enzymes in catalyzing the attachment of amino acids to their cognate tRNA species which realizes the codon–amino acid relationships spelled out in the genetic code. Thus, the aaRSs provide the molecular basis of the genotype–phenotype correspondence. The tRNA synthetases exist in two unrelated classes (2, 3) based on the topology of their ATP binding site. Until the discovery of a class I lysyl-tRNA synthetase (LysRS1) (4), it was believed that each amino acid was exclusively served by a single aaRS of either class I or class II. Although the aaRSs are of interest from many perspectives, including their mechanism of catalysis and proofreading (1), the intricacies of protein–RNA interactions (5), their involvement in cellular signaling functions (6), their practical applicability as potential antimicrobial targets (7), their probable role in genetic diseases (8, 9), or the ability to engineer aaRSs to cotranslationally introduce unnatural amino acids into proteins (10), there remains the fundamental question about the role of these proteins in the evolution of the genetic code. The central dilemma is whether the evolution of the aaRSs is the evolution of the genetic code itself, or whether the tRNA synthetases emerged in the context of a previously established code. Already four decades ago Woese (11) argued for the latter possibility, predicting that protein enzymes with enough specificity to interpret the genetic code could have been produced only by a translation apparatus accurate enough (even if not as precise as that found in modern cells) that an established code would have been required to produce such proteins. Many authors have concurred with such a perspective (e.g., ref. 12), imagining that the tRNA synthetases evolved and replaced a now-extinct ribozyme-based aminoacylation system. The likelihood that such a system could operate was demonstrated with RNA-based aaRSs (1315). Other scientists contend that the genetic code emerged from coevolution of aaRSs with tRNAs (16), invoking either a restricted set of coded amino acids or an ambiguouscoding scheme at the outset of aaRS evolution. This viewpoint does not favor or disfavor the notion that the genetic code evolved to its current state (17) as opposed to being a “frozen accident” (18).

In this article we shall consider whether the evolution of the genetic code predates the emergence of the aaRSs. One experimental approach relies on the study of tRNA identity, a key “molecular fossil.” Identity elements reside in tRNA and cause specific aminoacylation. These identity determinants are specific bases or portions of the secondary/tertiary structure that are recognized by the cognate aaRS; antideterminants help to reject association with noncognate synthetases. In most tRNAs, the major identity elements are bases located in the anticodon, the discriminator (19) position, and the acceptor helix (20). Identity elements are also found in other regions of the tRNA or in conserved structural features, such as the core region (21) or the extra arm, but these appear to depend on the particular synthetase or can have species-specific differences (reviewed in refs. 20 and 22). Here we examine a case where distinct aaRSs acylate the same tRNA. If the aaRSs established the genetic code, then we would expect to observe two distinct tRNAs with unique identity elements, one recruited to construct the genetic code by each aminoacylation system. If, on the other hand, we observe a common set of major identity determinants in the same tRNA isoacceptor, this would favor a situation where two evolutionarily distinct aaRSs evolved to recognize a common preestablished tRNA identity.

The system of choice for this work is the archaeal tRNACys and its acylation involving enzymes from two different pathways. The absence of the canonical class I cysteinyl-tRNA synthetase (CysRS) in the Methanocaldococcus jannaschii genome (23) was unresolved for several years (24). An initial suggestion (25), later supported by bioinformatic analysis (26), predicted a class II tRNA synthetase (encoded by MJ1660) would cysteinylate tRNACys. However, biochemical work established that in M. jannaschii Cys-tRNA is formed by a two-step pretranslational amino acid modification (27). In the first step, O-phosphoseryl-tRNA synthetase (SepRS), a novel class II aaRS encoded by MJ1660, forms Sep-tRNACys. Then a second enzyme, Sep-tRNA:Cys-tRNA synthase (SepCysS), converts this tRNA to correctly charged Cys-tRNACys (27). This indirect mechanism of Cys-tRNA formation is confined to certain members of the Euryarchaea (Fig. 1A, organisms in bold text), and in some cases (e.g., Methanococcus maripaludis) coexists with CysRS, responsible for direct charging of tRNACys. Comparative phylogenetics showed that both cysteine coding systems were extant at the time of the last universal common ancestor (28). Furthermore, SepRS was vertically inherited in the Archaea and subsequently lost in some archaeal lineages, and CysRS was vertically inherited in the bacteria and only later horizontally transferred in several independent events to some of the archaeal and all eukaryotic lineages (28, 29).

Fig. 1.
tRNA phylogeny and secondary structure diagrams. (A) A distance-based phylogenetic tree (see Materials and Methods) of the tRNACys sequences from all known Archaea, along with representative bacterial (red) and eukaryotic (green) examples. The Archaea ...

Here we present the biochemical characterization of the M. jannaschii tRNACys identity elements for charging by Me. maripaludis SepRS and CysRS, and also the results of a comparative phylogenetic tRNACys sequence analysis. The answers suggest that the universal genetic code evolved before the evolution of the modern aminoacylation systems.


tRNA Identity of SepRS and CysRS.

Because Me. maripaludis possesses both SepRS and CysRS (30) we selected these enzymes and determined their identity elements in tRNACys. The source of the tRNA was the single isoacceptor species from M. jannaschii, which as in vitro transcript (Mj tRNACys) was well chargeable by both synthetases. The corresponding Me. maripaludis tRNACys transcript appears to fold poorly and could not be charged (data not shown). The in vitro transcribed Mj tRNACys had comparable levels of catalytic efficiency (kcat/KM) for both SepRS and CysRS relative to the Mj tRNACys expressed in Escherichia coli, indicating that modified nucleotides do not play a critical role in the tRNACys recognition. Earlier studies determining E. coli tRNACys identity were also performed with unmodified tRNA transcript (31). Therefore, we determined steady-state kinetic parameters (KM and kcat) in the aminoacylation reaction with Sep (for SepRS) and Cys (for CysRS) of wild-type and mutant Mj tRNACys species.

Discriminator Base and Anticodon.

The discriminator base U73 and anticodon GCA are universally conserved in tRNACys species from all three domains of life, and these residues have been suggested to be (32) or experimentally shown to function as major identity determinants for the specific aminoacylation of tRNACys by CysRS (33). To assess the importance of these residues for SepRS, they were replaced and kinetic parameters for each mutant tRNACys were determined (Fig. 1B and Table 1). Mutation of U73 to any other base showed the largest detrimental effect on charging by CysRS and SepRS. Changes in the anticodon nucleotides G34, C35, and A36 greatly impeded recognition (a 20- to 50-fold decrease in kcat/KM). For CysRS, the only exception is the C35A change, which leads to a 4-fold increase in KM and a slight elevation in kcat. Whereas A36C and A36U have a mild effect on SepRS recognition, charging is barely detectable for A36G. Thus, the anticodon and discriminator base constitute major tRNACys identity determinants for both enzymes. Recognition of the anticodon region by SepRS is slightly less sensitive to alteration than that by CysRS.

Table 1.
Aminoacylation of mutant tRNACys species by Me. maripaludis SepRS and CysRS

tRNA Core Region.

Additional critical identity elements in E. coli and Haemophilus influenzae tRNACys include the G15:G48 base pair and A46:[A13:A22] base triple constituting the tRNA core region (34). The Mj tRNACys core includes a G15:C48 pair and the U46:[U13:A22] base triple. Whereas mutation at position 15 in E. coli tRNACys caused a 100-fold decrease in kcat/Km for CysRS (34), replacement of G15 by A in Mj tRNACys resulted in smaller decreases for CysRS (7-fold) and SepRS (3-fold). The core structure of Mj tRNACys is more similar to those of yeast and human, where multiple mutations in this region gave only slight decreases in CysRS catalytic efficiency (34). Unlike E. coli tRNACys, the core region of Mj tRNACys does not appear to contain any major identity elements.

Acceptor Stem.

Base pair 1–72 is typically not a tRNACys identity element. Although this is true for CysRS, we observed a 10-fold decrease in catalytic efficiency for SepRS when G1-C72 was replaced by A1-U72. Base pairs 2–71 and 3–70 do not contribute significantly to tRNA identity for SepRS and CysRS, yet these residues function as minor identity elements in E. coli (35).

Other Identity Elements.

The catalytic domain of SepRS is most closely related to its counterpart in the α-subunit of PheRS (28). E. coli tRNAPhe has identity determinants U20, U59, G10:C25, A73, and the anticodon GAA (36). Specific mutations were made to assess whether SepRS and PheRS have similar tRNA recognition features. Surprisingly, the C20U alteration led to a 2.5-fold increase in aminoacylation by SepRS, but it did not affect CysRS. An A59U change caused a 3-fold decrease in aminoacylation by SepRS but did not alter CysRS recognition. Replacement of the G10:C25 base pair by A:U reduced the phosphoserine (Sep) acylation efficiency 2-fold and increased CysRS efficiency slightly.

Evolutionarily Conserved SepRS Identity Elements.

To uncover possible SepRS-specific identity elements that may constitute a “tRNASep” signature, all known archaeal tRNACys sequences were aligned with representative eukaryotic and bacterial ones. In a distance-based phylogenetic tree (Fig. 1A) basic features of the canonical phylogenetic pattern are evident (29), e.g., congruence with rRNA phylogeny. Because of its short length, caution must be exercised in interpreting the result of any tRNA phylogeny (see the supporting information, which is published on the PNAS web site, for alternate treeing methods). The tRNACys sequences cluster into distinct archaeal, eukaryotic, and bacterial groups, with the eukaryotic sequences more similar to the archaeal versions. The deep divide between Crenarchaea and Euryarchaea is also present. The euryarchaeal groupings are surprisingly similar to established taxonomy (37), and Nanoarchaeum equitans and “Korarchaeota” sp. branch deeply in the archaeal portion of the tree. If the tree even roughly represents the evolutionary history of tRNACys, the presence of canonical pattern indicates a single common origin of tRNACys in the last universal common ancestor and subsequent coevolution, dominated by vertical gene flow, with the organismal lineages. Regardless of whether this interpretation is correct, it is striking that we do not observe distinct “tRNASep” and tRNACys groups.

“tRNASep” signature residues were defined by comparing patterns of conservation in five groups: Bacteria, Eucarya, Archaea, euryarchaeal organisms that contain SepRS, and a final group of putative “tRNASep” sequences. Grouping the sequences by domain of life was not decided a priori; rather, these are the major clusters that emerge in the tree. Some of the SepRS-containing organisms have multiple copies of tRNACys. Sequences Methanosarcina mazei 1 and Methanosarcina barkeri 3 (Fig. 1A), which are excluded from the “tRNASep” group, have mutations at several positions that are otherwise completely conserved in the SepRS containing organisms.§

We found that six residues, G6:C67, C20, G37, A47, and A59, are strictly conserved among the putative “tRNASep” species that are only partially conserved, are highly variable, or are conserved but as a different base in each of the other groups mentioned (see the supporting information). Position 47 is least conserved in the other groups. In Archaea, A or U are typical, U is dominant in eukaryotes, and the position is variable in Bacteria. Mutation A47U causes a 7-fold reduction in aminoacylation efficiency for both SepRS and CysRS. A59 is well conserved in other Archaea, and Bacteria and eukaryotes usually have A or U. Whereas C20 is usually conserved in Archaea, U is often present in eukaryotes and Bacteria. A59 and C20 coincide with identity elements in tRNAPhe (see above results). G37 is conserved in most Archaea. It is exclusively A in Bacteria and either A or G in eukaryotes. The G37A mutant decreased catalytic efficiency of SepRS to 14%, yet CysRS was mildly affected. Finally, G6:C67 is well conserved in Archaea but variable in Bacteria and eukaryotes. Because mutation to a CG pair has no effect on SepRS or CysRS charging, conservation here is purely historical. In summary, of the six residues suggested by evolutionary analysis three, G37, A47, and A59, are minor identity elements for SepRS.

Although this work was performed with Mj tRNACys, it should apply to Me. maripaludis tRNACys, which is 80% sequence-identical and contains all of the identity elements described here.

Transplantation of Sep Identity Elements into E. coli tRNAGly.

To probe whether the tRNACys identity set had been completely determined, we transplanted the nucleotides we assumed were necessary for recognition by SepRS into the E. coli tRNAGly framework. This tRNA was chosen because it is the only E. coli tRNA whose loops and helical stems are the same size as those of Mj tRNACys (Fig. 1C). Wild-type E. coli tRNAGly displayed only 5% charging efficiency with SepRS compared with Mj tRNACys (Table 2). This basal charging level is caused by the presence of U73 in tRNAGly. Mutation U73G (MG1 transcript) abolishes activity. Slight increases in efficiency were observed for anticodon mutations C34G or C36A (in transcripts MG2 and MG3). When the Mj tRNACys anticodon and discriminator base were present (transcript MG4) aminoacylation efficiency by SepRS was increased (32%) and a further increase (82%) was obtained with the A37G mutation (transcript MG5). Transcript MG6 incorporates multiple recognition elements: U73, anticodon GCA, A15, A37, A59, and C6:G67 in the E. coli tRNAGly; it could be acylated by SepRS with enhanced activity compared with Mj tRNACys.

Table 2.
Aminoacylation of E. coli tRNAGly mutants by Me. maripaludis SepRS

From the data above it is evident that SepRS and CysRS share the major and some minor identity elements, but that tRNACys in addition harbors some “tRNASep” signature residues. SepCysS in its productive interaction with Sep-tRNACys may depend on very different identity elements in the tRNA, as its catalytic action only concerns the modification of the aminoacyl residue. In line with this argument, Glu-tRNAGln amidotransferase (38), the enzyme that converts Glu-tRNAGln to the cognate Gln-tRNA, has very different identity elements from glutamyl-tRNA synthetase (39, 40).


tRNA Identity in Convergent Aminoacylation Systems.

Most aaRSs evolved from a common ancestral gene with the same amino acid and tRNA specificity as its descendents. Even when common identity elements are observed for such aaRSs one cannot tell whether the code assignments predated the aaRSs or not. The only informative cases are the four in which evolutionarily distinct aaRSs converged to recognize a common tRNA. In each of these cases, the most critical major identity elements are common whereas some other identity elements are idiosyncratic. Four cases exist where unrelated or distantly related aaRSs aminoacylate the same tRNA species. Both tRNACys (described above) and tRNALys are recognized by unrelated aaRSs, one from each of the two classes. The case of tRNAGly is different. The two glycyl-tRNA synthetases (GlyRSs), a strictly bacterial α2 homodimer and a principally archaeal/eukaryotic (αβ)2 heterotetramer, both share the class II aaRS catalytic core domain but are as distantly related to each other as ProRS is related to PheRS (41). Finally, there are two distantly related archaeal seryl-tRNA synthetases (SerRSs). Although both SerRS types are related to other subclass IIa aaRSs, the rare and common forms have no specific relationship (42). For both GlyRS and SerRS, phylogenetic evidence indicates that the two distantly related forms did not share a common ancestor with the same enzymatic specificity. Thus, like the class I and class II examples mentioned above, the two forms converged to the same function. With the knowledge of tRNACys, the identity elements for all these systems have now been determined.

The first evidence that convergent aminoacylation systems might share identity elements was the finding that Borrelia burgdorferi class I LysRS (LysRS1) efficiently charged tRNALys from E. coli, an organism that contains only LysRS2 (43). Phylogenetic analysis supported the notion, showing that the two LysRSs recognized a single monophyletic group of tRNAs (44). Both LysRSs share common major identity elements U35 and U36 in the anticodon (45). It may appear as a trivial similarity that common identity elements are found in the anticodon, yet the anticodon bases are not recognized by all aaRSs, e.g., SerRS (46) and in some LeuRSs (47). In transplantation experiments, the UUU anticodon along with the A73 discriminator base are enough to ensure lysylation by LysRS2 (48, 49), yet only with the addition of the acceptor stem base pairs G2-U71 and G3-C70 is the identity set for LysRS1 complete (45).

Both GlyRSs share the common major identity elements C35 and C36 in the anticodon and the G1-C72 base pair in the acceptor stem. The discriminator base is typically A in archaeal and eukaryotic tRNAGly and usually U in bacterial examples, where in the E. coli dimeric GlyRS it is a critical identity element (50). However, the dimeric Thermus thermophilus GlyRS is able to efficiently glycylate archaeal, eukaryotic, or bacterial tRNAGly. The discriminator base is not an identity element in this case (51). Both SerRSs share major identity elements in the extra arm, discriminator G73 and the G30:C40 base pair, whereas the rare archaeal form has G1:C72 and additional unpaired bases in the extra arm as additional identity elements (52). Although tRNASer from Me. maripaludis is a poor substrate for yeast SerRS, the rare SerRS in Me. maripaludis effectively serylates the yeast tRNASer (53).

Evolutionary Significance of an RNA Record.

The RNA record is a catalog of evolutionary events that can be described by using the comparative analysis of RNA sequences. Changes in rRNA sequences provide a map of the evolutionary history of life on Earth (42). RNA relics (54); ubiquitous RNAs that participate in basic cellular functions, e.g., the ribosome, RNase P and tRNAs, provide a glimpse of some biomolecular species that must have been extant in the RNA world. tRNA identity, yet another facet of the RNA record, encodes the final events of the evolution of the genetic code. That major tRNA identity elements are not dependent on the details of the aminoacylation system, in each of the four informative cases, attests to the antiquity of tRNA identity. Several lines of evidence support the hypothesis that early cellular evolution occurred in an RNA world (55), which transitioned into the modern cellular world via the evolution of translation and then, to a large extent, the takeover of protein enzymes. The fundamental importance of the RNA record is that it can unveil a more distant evolutionary past than is accessible by molecular phylogenetics of proteins, i.e., the protein record.

Evolutionary Stages in tRNACys Recognition.

The development of tRNACys recognition by the modern aminoacylation systems is also captured in the RNA record, which tells of two stages and three kinds of evolutionary change. The initial stage involves adaptation of protein to preexisting RNA. The common identity elements, recognized by both CysRS and SepRS, are likely the most ancient features of tRNACys. These include the anticodon and discriminator base as the common major identity elements, and G15 and A47 are common minor identity elements. That tRNA identity is not a sensitive property of the aminoacylation systems is the primary evidence that SepRS and CysRS independently adapted to a preestablished tRNA identity. Two pieces of phylogenetic evidence, one RNA and one protein-based, further support this notion: (i) canonical phylogenetic pattern in tRNACys suggests vertical descent of these sequences from a single common ancestral gene that was present in last universal common ancestor; and (ii) even though CysRS has been horizontally transferred to the archaea (29), none of these transfers appear to have involved an accompanying bacterial type tRNACys. The above two points underscore the fact that we do not observe markedly different tRNASep and tRNACys sequences, which would have been the hallmark of tRNA identity being purely a product of coevolution of tRNAs with cognate aaRSs.

In the second stage, RNA adapts to protein and the two coevolve. As mentioned above, SepRS and PheRS are related by sharing a more recent common ancestor with each other than with other class II aaRSs. SepRS and PheRS have minor identity elements at homologous positions in their cognate tRNAs. The origin of these elements may have resulted from the adaptation of the archaeal tRNACys to the protein architecture that SepRS inherited from its shared ancestor with PheRS. These data imply that SepRS and PheRS approach and bind their cognate tRNAs in a similar (homologous) orientation. Even though the position of some minor identity elements is conserved in SepRS and PheRS, the bases involved are different (except for G1-C72). The fine-level detail of the SepRS:tRNACys interaction, therefore, is the result of coevolution between tRNA and aaRS. Coevolution also accounts for both the position and nature of the minor identity determinant G37, which is SepRS-specific. Finally, coevolution of CysRS with tRNACys best explains the location of CysRS-specific minor identity elements in the tRNA core region (21). Although the later evolutionary stages of tRNACys recognition are characterized by tRNA adaptation to CysRS and coevolution between the two, the most important identity elements and the earliest stage of the process involved the adaptation of the proteins to the RNA.

Evolution of the Genetic Code.

Evolutionary aspects of this work deal largely with the point at which the universal genetic code emerged and events thereafter. We conclude that the defining features of tRNA identity, and thus of the universal genetic code, were established before the modern class I or class II aminoacylation systems. This further implies that the evolution of the genetic code occurred in the RNA world. That ribozymes could have been responsible for specific aminoacylation of tRNAs has been demonstrated. By using in vitro evolution experiments, ribozymes that specifically recognize glutamine and that include the discriminator base in tRNA recognition have been produced (14). In addition, aminoacylating ribozymes have been constructed that interact with the acceptor stem and recognize the anticodon through base pairing (56). Therefore, the evolution of the code must be understood in terms of the kinds of molecular species that would have been extant in the RNA world, and also in terms of the development of tRNA identity, which goes beyond the codon–anticodon interaction to include other major and minor identity elements.

Materials and Methods


[35S]Cysteine [1,075 Ci/mmol (1 Ci = 37 GBq)] was purchased from PerkinElmer Life Sciences (Boston, MA). [14C]O-phospho-l-serine (55 mCi/mmol) was obtained from American Radiolabeled Chemicals (St. Louis, MO). Oligonucleotide synthesis and DNA sequencing were performed at the Keck Foundation Research Biotechnology Resource Laboratory at Yale University.

Production of Recombinant Me. maripaludis SepRS and CysRS.

The sequence of Me. maripaludis sepS (MMP0688) was used to design specific primers, containing NdeI and BamHI restriction sites, for amplification of the gene from Me. maripaludis S2 genomic DNA. The PCR product was digested with NdeI and BamHI and ligated into the pET 15b vector (Novagen, Madison, WI) for expression of N-terminal six-histidine His6-tagged protein in the E. coli BL21-Codon Plus(DE3)-RIL strain. Cultures were grown at 37°C in LB medium supplemented with 100 μg/ml ampicillin and 34 μg/ml chloramphenicol. When the cultures reached A600 of 0.6, 1 mM isopropyl-β-d-thiogalactoside was added and expression was induced for 6 h. After harvesting, cells were disrupted by sonication in a buffer containing 50 mM Tris·HCl (pH 8.0), 5 mM MgCl2, 14.3 mM 2-mercaptoethanol, and 10% glycerol.

The soluble fraction of the cell extract containing the recombinant N-terminal His6-tagged protein was loaded onto a DEAE Sepharose (Amersham, Pittsburgh, PA) column, washed with buffer A [50 mM Tris·HCl (pH 8.0)/5 mM MgCl2/14.3 mM 2-mercaptoethanol], and protein-eluted with buffer B [50 mM Tris·HCl (pH 8.0)/75 mM NaCl/5 mM MgCl2/14.3 mM 2-mercaptoethanol]. The eluate was then applied to a Ni-NTA agarose (Qiagen, Valencia, CA) column, and His6-SepRS was further purified according to the vendor's instructions. The protein containing eluate fraction was dialyzed against aminoacylation buffer (see below) containing 50% glycerol and stored at −20°C. The Me. maripaludis SepRS was >99% pure, as judged by Coomassie brilliant blue staining after SDS/PAGE.

Me. maripaludis CysRS was produced as an N-terminal His6-tagged protein and purified to apparent homogeneity as described previously (30).

Preparation and Purification of tRNA Gene Transcripts.

Me. maripaludis tRNACys and E. coli tRNAGly variants were synthesized by in vitro T7 RNAP run-off transcription as described (57). Each tRNA gene together with the T7 promoter was constructed from overlapping oligonucleotides and was cloned in the vector pUC18. To generate tRNA transcripts with a 3′-CCA end, a BstNI restriction site was placed at the 3′ end of each tRNA gene sequence. The in vitro transcription reaction was performed at 37°C for 3–5 h in a buffer containing 40 mM Tris·HCl (pH 8.1), 22 mM MgCl2, 5 mM DTT, 1 mM spermidine, 0.05% Triton X-100, 16 mM guanosine monophosphate, 4 mM of each nucleoside triphosphate, BstNI-digested vector containing the template DNA (0.1 μg/μl), and 1 mM T7 RNAP. The tRNA transcripts were purified by electrophoresis on denaturing polyacrylamide gels. Full-length tRNAs were eluted and desalted on Sephadex G25 Microspin columns (Amersham). Before use, the tRNA transcripts were heated for 10 min at 85°C, cooled down slowly to 45°C, followed by addition of 5 mM MgCl2 before placing on ice.

Aminoacylation of tRNA.

Sep-tRNA and Cys-tRNA formation were assayed at 37°C in aminoacylation buffer containing 50 mM Hepes·KOH (pH 7.0), 50 mM NaCl (for SepRS) or 50 mM KCl (for CysRS), 20 mM MgCl2, 10 mM ATP, 5 mM DTT, 200 μM [14C]phosphoserine (55 mCi/mmol), or 50 μM [35S]cysteine (1,075 Ci/mmol), with various concentrations of tRNA transcripts (1–200 μM), and purified enzyme (500 nM to 1.5 μM SepRS or 50–200 nM CysRS). Aliquots from the reaction mixture were removed periodically and spotted on Whatman 3MM paper filter disks (Whatman). After three washing steps in 10% trichloroacetic acid, the radioactivity was measured by liquid scintillation counting. Kinetic parameters for aminoacylation of the tRNAs were obtained from the corresponding Eadie–Hofstee plots. Where very little aminoacylation occurred and a line could not be accurately determined, the kcat/Km value relative to M. jannaschii wild-type tRNACys transcript was estimated to be <0.01.

Phylogenetic Analyses.

The “Korarchaeota” sp. sequence was kindly provided ahead of publication by James Elkins and Karl Stetter (personal communication). The source of the remaining sequences, alignments details (58), and phylogenetic methods are described in the supporting information. Because of the weak phylogenetic signal in tRNA sequences, owing to their short length, phylogenetic methods with the fewest assumptions, such as distance-based, are preferred (Fig. 1A). For comparison, maximum parsimony/maximum likelihood-based trees are given in the supporting information.

Supplementary Material

Supporting Information:


We thank Nigel Goldenfeld and Carl Woese for inspired discussions and to James Elkins and Karl Stetter for sharing unpublished data. M.J.H. held a Feodor Lynen Postdoctoral Fellowship of the Alexander von Humboldt Stiftung, H.-S.P. was the recipient of a postdoctoral fellowship of the Korean Science Foundation (KRF-2005-000-10184), and P.O. holds a National Science Foundation Postdoctoral Fellowship in Biological Informatics. This work was supported by grants from the National Institute of General Medical Sciences (GM22854), the National Science Foundation (DBI-0535566), and the Department of Energy (DE-FG02-98ER20311).


O-phosphoseryl-tRNA synthetase
cysteinyl-tRNA synthetase
lysyl-tRNA synthetase
aminoacyl-tRNA synthetase
glycyl-tRNA synthetase


The authors declare no conflict of interest.

§A second putative Methanospirillum hungatei tRNACys, which is aberrant throughout the molecule and is the only tRNACys that lacks the discriminator base U, has been excluded from our analysis.


1. Ibba M, Söll D. Annu Rev Biochem. 2000;69:617–650. [PubMed]
2. Eriani G, Delarue M, Poch O, Gangloff J, Moras D. Nature. 1990;347:203–206. [PubMed]
3. Cusack S, Härtlein M, Leberman R. Nucleic Acids Res. 1991;19:3489–3498. [PMC free article] [PubMed]
4. Ibba M, Morgan S, Curnow AW, Pridmore DR, Vothknecht UC, Gardner W, Lin W, Woese CR, Söll D. Science. 1997;278:1119–1122. [PubMed]
5. Dale T, Uhlenbeck OC. Trends Biochem Sci. 2005;30:659–665. [PubMed]
6. Park SG, Ewalt KL, Kim S. Trends Biochem Sci. 2005;30:569–574. [PubMed]
7. Brown MJ, Mensah LM, Doyle ML, Broom NJ, Osbourne N, Forrest AK, Richardson CM, O'Hanlon PJ, Pope AJ. Biochemistry. 2000;39:6003–6011. [PubMed]
8. Lee JW, Beebe K, Nangle LA, Jang J, Longo-Guess CM, Cook SA, Davisson MT, Sundberg JP, Schimmel P, Ackerman SL. Nature. 2006;443:50–55. [PubMed]
9. Roy H, Ibba M. Nature. 2006;443:41–42. [PubMed]
10. Wang L, Xie J, Schultz PG. Annu Rev Biophys Biomol Struct. 2006;35:225–249. [PubMed]
11. Woese CR. The Genetic Code. New York: Harper & Row; 1967.
12. Nagel GM, Doolittle RF. J Mol Evol. 1995;40:487–498. [PubMed]
13. Illangasekare M, Sanchez G, Nickles T, Yarus M. Science. 1995;267:643–647. [PubMed]
14. Lee N, Bessho Y, Wei K, Szostak JW, Suga H. Nat Struct Biol. 2000;7:28–33. [PubMed]
15. Kourouklis D, Murakami H, Suga H. Methods. 2005;36:239–244. [PubMed]
16. Schimmel P. Trends Biochem Sci. 1991;16:1–3. [PubMed]
17. Vetsigian K, Woese C, Goldenfeld N. Proc Natl Acad Sci USA. 2006;103:10696–10701. [PMC free article] [PubMed]
18. Crick FH. J Mol Biol. 1968;38:367–379. [PubMed]
19. Crothers DM, Seno T, Söll D. Proc Natl Acad Sci USA. 1972;69:3063–3067. [PMC free article] [PubMed]
20. Giege R, Sissler M, Florentz C. Nucleic Acids Res. 1998;26:5017–5035. [PMC free article] [PubMed]
21. Hou YM, Motegi H, Lipman RS, Hamann CS, Shiba K. Nucleic Acids Res. 1999;27:4743–4750. [PMC free article] [PubMed]
22. Beuning PJ, Musier-Forsyth K. Biopolymers. 1999;52:1–28. [PubMed]
23. Bult CJ, White O, Olsen GJ, Zhou L, Fleischmann RD, Sutton GG, Blake JA, FitzGerald LM, Clayton RA, Gocayne JD, et al. Science. 1996;273:1058–1073. [PubMed]
24. Ruan B, Nakano H, Tanaka M, Mills JA, DeVito JA, Min B, Low KB, Battista JR, Söll D. J Bacteriol. 2004;186:8–14. [PMC free article] [PubMed]
25. Das R, Vothknecht UC. Biochimie. 1999;81:1037–1039. [PubMed]
26. Sethi A, O'Donoghue P, Luthey-Schulten Z. Proc Natl Acad Sci USA. 2005;102:4045–4050. [PMC free article] [PubMed]
27. Sauerwald A, Zhu W, Major TA, Roy H, Palioura S, Jahn D, Whitman WB, Yates JR, III, Ibba M, Söll D. Science. 2005;307:1969–1972. [PubMed]
28. O'Donoghue P, Sethi A, Woese CR, Luthey-Schulten ZA. Proc Natl Acad Sci USA. 2005;102:19003–19008. [PMC free article] [PubMed]
29. Woese CR, Olsen GJ, Ibba M, Söll D. Microbiol Mol Biol Rev. 2000;64:202–236. [PMC free article] [PubMed]
30. Stathopoulos C, Kim W, Li T, Anderson I, Deutsch B, Palioura S, Whitman W, Söll D. Proc Natl Acad Sci USA. 2001;98:14292–14297. [PMC free article] [PubMed]
31. Hou YM, Westhof E, Giege R. Proc Natl Acad Sci USA. 1993;90:6776–6780. [PMC free article] [PubMed]
32. Mallick B, Chakrabarti J, Sahoo S, Ghosh Z, Das S. DNA Res. 2005;12:235–246. [PubMed]
33. Komatsoulis GA, Abelson J. Biochemistry. 1993;32:7435–7444. [PubMed]
34. Lipman RS, Hou YM. Proc Natl Acad Sci USA. 1998;95:13495–13500. [PMC free article] [PubMed]
35. Sterner T, Jansen M, Hou YM. RNA. 1995;1:841–851. [PMC free article] [PubMed]
36. Peterson ET, Uhlenbeck OC. Biochemistry. 1992;31:10380–10389. [PubMed]
37. Brochier C, Forterre P, Gribaldo S. BMC Evol Biol. 2005;5:36. [PMC free article] [PubMed]
38. Nakamura A, Yao M, Chimnaronk S, Sakai N, Tanaka I. Science. 2006;312:1954–1958. [PubMed]
39. Sylvers LA, Rogers KC, Shimizu M, Ohtsuka E, Söll D. Biochemistry. 1993;32:3836–3841. [PubMed]
40. Sekine S, Nureki O, Sakamoto K, Niimi T, Tateno M, Go M, Kohno T, Brisson A, Lapointe J, Yokoyama S. J Mol Biol. 1996;256:685–700. [PubMed]
41. O'Donoghue P, Luthey-Schulten Z. Microbiol Mol Biol Rev. 2003;67:550–573. [PMC free article] [PubMed]
42. Woese CR, Fox GE. Proc Natl Acad Sci USA. 1977;74:5088–5090. [PMC free article] [PubMed]
43. Ibba M, Bono JL, Rosa PA, Söll D. Proc Natl Acad Sci USA. 1997;94:14383–14388. [PMC free article] [PubMed]
44. Ribas De Pouplana L, Turner RJ, Steer BA, Schimmel P. Proc Natl Acad Sci USA. 1998;95:11295–11300. [PMC free article] [PubMed]
45. Ambrogelly A, Frugier M, Ibba M, Söll D, Giege R. FEBS Lett. 2005;579:2629–2634. [PubMed]
46. Achsel T, Gross HJ. EMBO J. 1993;12:3333–3338. [PMC free article] [PubMed]
47. Soma A, Kumagai R, Nishikawa K, Himeno H. J Mol Biol. 1996;263:707–714. [PubMed]
48. McClain WH, Foss K, Jenkins RA, Schneider J. Proc Natl Acad Sci USA. 1990;87:9260–9264. [PMC free article] [PubMed]
49. Tamura K, Himeno H, Asahara H, Hasegawa T, Shimizu M. Nucleic Acids Res. 1992;20:2335–2339. [PMC free article] [PubMed]
50. Mazauric MH, Keith G, Logan D, Kreutzer R, Giege R, Kern D. Eur J Biochem. 1998;251:744–757. [PubMed]
51. Mazauric MH, Roy H, Kern D. Biochemistry. 1999;38:13094–13105. [PubMed]
52. Korencic D, Polycarpo C, Weygand-Durasevic I, Söll D. J Biol Chem. 2004;279:48780–48786. [PubMed]
53. Gruic-Sovulj I, Jaric J, Dulic M, Cindric M, Weygand-Durasevic I. J Mol Biol. 2006;361:128–139. [PubMed]
54. Jeffares DC, Poole AM, Penny D. J Mol Evol. 1998;46:18–36. [PubMed]
55. Gesteland RF, Cech TR, Atkins JF, editors. The RNA World: The Nature of Modern RNA Suggests a Prebiotic RNA. Cold Spring Harbor, NY: Cold Spring Harbor Lab Press; 2006.
56. Bessho Y, Hodgson DR, Suga H. Nat Biotechnol. 2002;20:723–728. [PubMed]
57. Sampson JR, Uhlenbeck OC. Proc Natl Acad Sci USA. 1988;85:1033–1037. [PMC free article] [PubMed]
58. Roberts E, Eargle J, Wright D, Luthey-Schulten Z. BMC Bioinformatics. 2006;7:382. [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...