A Euryarchaeal Lysyl-tRNA Synthetase: Resemblance to Class I Synthetases

See allHide authors and affiliations

Science  07 Nov 1997:
Vol. 278, Issue 5340, pp. 1119-1122
DOI: 10.1126/science.278.5340.1119


The sequencing of euryarchaeal genomes has suggested that the essential protein lysyl–transfer RNA (tRNA) synthetase (LysRS) is absent from such organisms. However, a single 62-kilodalton protein with canonical LysRS activity was purified from Methanococcus maripaludis, and the gene that encodes this protein was cloned. The predicted amino acid sequence of M. maripaludis LysRS is similar to open reading frames of unassigned function in bothMethanobacterium thermoautotrophicum and Methanococcus jannaschii but is unrelated to canonical LysRS proteins reported in eubacteria, eukaryotes, and the crenarchaeote Sulfolobus solfataricus. The presence of amino acid motifs characteristic of the Rossmann dinucleotide-binding domain identifies M. maripaludis LysRS as a class I aminoacyl–tRNA synthetase, in contrast to the known examples of this enzyme, which are class II synthetases. These data question the concept that the classification of aminoacyl–tRNA synthetases does not vary throughout living systems.

Lysyl–tRNA synthetase is essential for the translation of lysine codons during protein synthesis. In spite of the necessity for this enzyme in all organisms and the high degree of conservation among aminoacyl–tRNA synthetases (1), genes encoding a LysRS homolog have not been found by sequence similarity searches in the genomes of two Archaea,Methanococcus jannaschii (2) andMethanobacterium thermoautotrophicum (3). This raises the possibility that LysRS, like the asparaginyl– and glutaminyl–tRNA synthetases (4), is not present and that lysyl-tRNA (Lys-tRNA) is synthesized by tRNA-dependent transformation of a misacylated tRNA (5). Alternatively, these organisms may contain a LysRS activity encoded by a gene that is sufficiently different from those previously identified to prevent its detection by sequence similarity searches.

To investigate the formation of Lys-tRNA in cell-free extracts prepared from Methanococcus maripaludis, we used [14C]lysine and homologous total tRNA as substrates (6). Amino acid analysis of the 14C-labeled aminoacyl-tRNA produced in this reaction (7) indicated that lysine was directly acylated onto tRNA and therefore that Lys-tRNA was not the product of a tRNA-dependent amino acid transformation. This finding was confirmed by the observation that [14C]Lys-tRNA synthesis was inhibited by only one of the 20 canonical amino acids, lysine (Fig.1). It is consistent with the presence of LysRS, because Lys-tRNA synthesis via a mischarged tRNA would result in a nonlabeled amino acid other than lysine being able to inhibit lysylation of tRNA.

Figure 1

Direct attachment of lysine to tRNA by M. maripaludis protein extracts. Aminoacylation reactions were performed as described in the presence of 200 μg of RNA-free total protein prepared from an S160 extract in a reaction volume of 130 μl. Samples (20 μl) were periodically removed and analyzed. Reactions were performed in the presence of 20 μM [14C]lysine (○), 20 μM [14C]lysine and the 19 canonical amino acids (800 μM labeled with 12C) except for lysine (•), and 20 μM [14C]lysine and 800 μM [12C]lysine (▵).

On the basis of these observations, we sought to purify LysRS from M. maripaludis. By standard chromatographic procedures (8), we isolated a single 62-kD protein, which we purified to homogeneity as judged by both silver staining after denaturing SDS– polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 2) and the appearance of a single symmetrical peak during size exclusion chromatography. The LysRS activity eluted as a single discrete fraction during all purification steps, indicating that M. maripaludis contains only a single LysRS activity. The enzyme could acylate unfractionated tRNA with lysine to the same extent as theM. maripaludis crude extract (Fig. 1), confirming that the purified protein is LysRS. In our kinetic characterization of our LysRS for the aminoacylation of M. maripaludistotal tRNA with lysine (9), we calculated a Michaelis constant (K m) for lysine of 2.2 μM and a catalytic rate constant (k cat) of 1.6 s 1, values comparable with those of other LysRS proteins (10). Investigation of the species specificity of Lys-tRNA formation in cell-free extracts showed reactivity between tRNA and RNA-free protein extracts prepared fromEscherichia coli, M. thermoautotrophicum, andM. maripaludis (11). The only exception wasM. maripaludis tRNA, which is not a substrate for theE. coli enzyme, suggesting that it may lack 2-thiouridine and its derivatives, which are required for the LysRS·tRNALys interaction (12). These modified nucleotides have previously been detected at position 34 in all known tRNAGlu and tRNALys isoacceptors (13). In that these modifications are critical for recognition of the tRNALys anticodon by LysRS (12), their possible absence suggests that tRNA recognition by the M. maripaludis enzyme is substantially different from known examples and may explain the inability of an E. coliextract to aminoacylate M. maripaludis tRNA with lysine.

Figure 2

Purification and NH2-terminal sequencing of M. maripaludis LysRS. Protein purification was monitored by SDS-PAGE: Lane 1: S160 protein extract; lanes 2, 3, 4, and 5: active eluates from Q-sepharose, Mono-Q, Mono-S, and Superose 12, respectively; and lane M: molecular size standards (in kilodaltons). Samples from lane 5 were blotted to polyvinylidene fluoride membranes and subjected to NH2-terminal sequencing, which yielded the indicated 22–amino acid sequence (27). Lanes 1 to 4 were stained with Coomassie blue and lanes 5 and M with silver nitrate.

A portion (22 amino acid residues) (Fig. 2) of the NH2-terminal sequence of LysRS was determined by protein analysis, and the sequence was used to clone lysS, the gene that encodes LysRS. This 22–amino acid sequence is similar to the NH2-terminus of a predicted protein-coding region (MJ0539) in the M. jannaschii genome and one in the M. thermoautotrophicum genome, but we find no sequence similarity elsewhere in any available database. The coding sequence MJ0539 has been tentatively identified as a putative aminoacyl–tRNA synthetase (14). The NH2-terminal sequence was used in conjunction with a conserved internal region in the M. jannaschii and M. thermoautotrophicum open reading frames (ORFs) to clone a 450–base pair (bp) fragment of M. maripaludis lysS from genomic DNA by polymerase chain reaction (PCR). This fragment was 32P-labeled and used to isolate a DNA fragment containing the 3′-terminal 1293 bp of the lysSgene from a genomic library. The remainder of lysS (306 bp) was cloned by PCR from genomic DNA with a 5′ primer based on the NH2-terminal amino acid sequence of LysRS. After being sequenced, the two portions of lysS were used to derive a full-length clone (15). The lysS gene is 1599 bp long and encodes a 61.3-kD protein, in good agreement with the molecular mass deduced for the native protein from SDS-PAGE (Fig.3). To confirm the identity of the cloned gene as lysS, we subcloned it into pET15b, which allowed the overproduction and subsequent chromatographic purification of M. maripaludis LysRS as a hexahistidine (His6) fusion protein (16). This purified His6-LysRS could aminoacylate with lysine M. maripaludis total tRNA to the same extent and at the same rate as the native enzyme (Fig. 4).

Figure 3

Alignment of euryarchaeal LysRS amino acid sequences. The signature sequences characteristic of class I aminoacyl–tRNA synthetases are indicated by underlining. The sequences were aligned with the CLUSTAL W program (26). The sequences shown are from A. fulgidus(AF), M. jannaschii (MJ), M. maripaludis (MM), and M. thermoautotrophicum(MT).

Figure 4

Aminoacylation of M. maripaludistRNA by purified M. maripaludis His6-LysRS (16). Aminoacylation reactions were performed as described (20-μl samples) in the presence of 20 nM enzyme and the following amino acids: 20 μM [14C]lysine (○); 20 μM [14C]lysine and the 19 canonical amino acids (800 μM labeled with 12C) except for lysine (•); and 20 μM [14C]lysine and 800 μM [12C]lysine (▵). The squares represent an aminoacylation reaction performed with 20 μM [14C]lysine and 50 nM control protein preparation (His6–glutamyl-tRNA reductase). The amount of product formation at 30 min indicates that the M. Maripaludis total tRNA preparation contains about 32 pmol of tRNALys perA 260 unit, in good agreement with the value for commercial E. coli total tRNA reagents (40 pmol perA 260 unit; Boehringer Mannheim).A 260, absorbance at 260 nm.

The predicted amino acid sequence of M. maripaludis LysRS is similar to putative proteins in Archaeoglobus fulgidus(14), M. thermoautotrophicum, and M. jannaschii but otherwise appears to have little or no similarity to any known sequences outside of the Euryarchaeota, including that for LysRS from the crenarchaeote Sulfolobus solfataricus (a normal class II enzyme). The only exception is a putative ORF in the crenarchaeote Cenarchaeum sp., which shares about 28% amino acid identity with the euryarchaeal LysRS protein that we describe, including extensive sequence similarity in the regions containing the HIGH and KMSKS motifs (17). Thus, although this protein performs the enzymatic function of a conventional LysRS—the specific esterification of tRNA with lysine—catalysis is accomplished in the context of an amino acid landscape that lacks any sequences corresponding to motifs 1, 2, or 3 (18), which are found in known LysRS proteins now classified as class II aminoacyl–tRNA synthetases (Fig. 5). On the basis of extensive similarities in their NH2-terminal domains, LysRS, aspartyl–tRNA (AspRS), and asparaginyl–tRNA (AsnRS) synthetases were grouped as paralogous enzymes in subclass IIb (19). However, the euryarchaeal LysRS bears no similarity to AspRS (the latter is normal), and AsnRS is absent, which leaves AspRS as the only member of this subclass. Thus, there appears to be considerably more evolutionary variation of the aminoacyl–tRNA synthetases than previously thought, a proposal supported by the apparent absence of a recognizable cysteinyl–tRNA synthetase in at least some of the Archaea (2, 3) and the existence in theM. jannaschii genome of an ORF (MJ1660) encoding an unidentified class II aminoacyl–tRNA synthetase similar to the α subunit of phenylalanyl–tRNA synthetase (14).

Figure 5

Alignment of motifs 1, 2, and 3 (19,20) from class II LysRS enzymes against the most similar regions in euryarchaeal LysRS enzymes. Class II and euryarchaeal sequences were separately aligned with CLUSTAL W, and the aligned class II motifs (1, 2, or 3) were used individually to search for similar regions in the euryarchaeal enzymes. Conserved residues are shown in blue; those found only in class II LysRS enzymes are shown in red. The class II LysRS sequences shown are from Cricetulus longicaudatus (GenBank accession number Z31711), Homo sapiens (D31890),Caenorhabditis elegans (U41105), Saccharomyces cerevisiae cytoplasm (X56259), Lycopersicon esculentum (X94451), E. coli lysS (U28375),E. coli lysU (X16542), Haemophilus influenzae (P43825), Acinetobacter calcoaceticus(Z46863), Bacillus subtilis (P37477),Staphylococcus aureus (L36472), Synechocystissp. (D90906), Campylobacter jejuni (M63448), Thermus thermophilus (P41255), Mycoplasma genitalium(P47382), Mycoplasma pneumoniae (AE000055),Mycoplasma fermentans (U50826), Mycoplasma hominis (P46191), S. solfataricus (Y08257), and S. cerevisiae mitochondria (X57360).

The euryarchaeal LysRS proteins show variations of the HIGH and KMSKS nucleotide-binding motifs (Fig. 3) characteristic of class I aminoacyl–tRNA synthetases. The occurrence of such sequence motifs has been correlated by structural studies with the topology of the catalytic domain, class I aminoacyl–tRNA synthetases containing a Rossmann fold, and class II aminoacyl–tRNA synthetases containing an antiparallel β sheet. The fact that euryarchaeal LysRS proteins show the defining motifs of class I aminoacyl–tRNA synthetases forces the unexpected conclusion that the catalytic domains of these enzymes are structurally unrelated to those of their bacterial, eukaryotic, and even certain crenarchaeal counterparts that belong to class II (20). Phylogenetic analysis of an overall class I alignment (21) did not indicate any unequivocal specific relation between euryarchaeal LysRS and any other class I aminoacyl–tRNA synthetase. Thus, it is unclear whether the euryarchaeal-type LysRS was present in the last common ancestor or arose later through recruitment of another class I enzyme within the Archaea. This uncertainty is surprising because LysRS [like other aminoacyl–tRNA synthetases (21)] is conserved through evolution both in the other living kingdoms and in certain Crenarchaeota. To date, archaeal LysRS appears to represent the only known example of class switching among aminoacyl–tRNA synthetases and confirms the unexpected evolutionary origin of euryarchaeal LysRS. LysRS has so far been detected in both major branches of the Archaea: four euryarchaeotes (Fig. 4) and two crenarchaeotes,Cenarchaeum sp. and S. solfataricus. All show the unexpected class I LysRS except for S. solfataricus, which, like all other examples, contains a class II LysRS (whether it also contains a class I LysRS is not known). This dichotomy is sufficiently complex and the number of archaeal examples is sufficiently small that speculation concerning LysRS evolution is not warranted at this stage. However, our data do cast doubt on previous evolutionary proposals based on the presumption of constancy of distribution of any given aminoacyl–tRNA synthetase. The observation that archaeal LysRS is a class I aminoacyl–tRNA synthetase also represents a functional demonstration of nonorthologous displacement (22) of a gene required for a core process in gene expression. It differs, for example, from archaeal histone-like proteins (23) and the recently discovered DNA topoisomerase VI of Sulfolobus shibatae (24), both of which have sequence similarity to eukaryotic enzymes of known function.

Our data raise several questions concerning evolution, the first being the aminoacyl–tRNA synthetases themselves. If these enzymes are not evolving as orthologs despite the orthology of the translation apparatus [and the genetic code is constant, with exceptions that are relatively recent compared with our findings (25)], then we must compare the relation between evolution of the aminoacyl–tRNA synthetases with that of translation in general and the structure of the genetic code in particular. It appears that the aminoacyl–tRNA synthetases had not achieved as settled an evolutionary condition at the time the universal ancestor of all life gave rise to the primary lineages as has been generally assumed. Second, if the existing universal phylogeny is indeed representative of organismal lineages, then our data begin to speak to the general issue of which functions in the cell were firmly established when the universal ancestor gave rise to the primary lineages and which were not.

  • * To whom correspondence should be addressed. E-mail: soll{at}


View Abstract

Navigate This Article