Report

RNA-Dependent Cysteine Biosynthesis in Archaea

See allHide authors and affiliations

Science  25 Mar 2005:
Vol. 307, Issue 5717, pp. 1969-1972
DOI: 10.1126/science.1108329

Abstract

Several methanogenic archaea lack cysteinyl–transfer RNA (tRNA) synthetase (CysRS), the essential enzyme that provides Cys-tRNACys for translation in most organisms. Partial purification of the corresponding activity from Methanocaldococcus jannaschii indicated that tRNACys becomes acylated with O-phosphoserine (Sep) but not with cysteine. Further analyses identified a class II–type O-phosphoseryl-tRNA synthetase (SepRS) and Sep-tRNA:Cys-tRNA synthase (SepCysS). SepRS specifically forms Sep-tRNACys, which is then converted to Cys-tRNACys by SepCysS. Comparative genomic analyses suggest that this pathway, encoded in all organisms lacking CysRS, can also act as the sole route for cysteine biosynthesis. This was proven for Methanococcus maripaludis, where deletion of the SepRS-encoding gene resulted in cysteine auxotrophy. As the conversions of Sep-tRNA to Cys-tRNA or to selenocysteinyl-tRNA are chemically analogous, the catalytic activity of SepCysS provides a means by which both cysteine and selenocysteine may have originally been added to the genetic code.

The translation of cysteine codons in mRNA during protein synthesis requires cysteinyl-tRNA (Cys-tRNACys). Cys-tRNACys is normally synthesized from the amino acid cysteine and the corresponding tRNA isoacceptors (tRNACys) in an adenosine triphosphate (ATP)–dependent reaction catalyzed by cysteinyl-tRNA synthetase (CysRS). Genes encoding CysRS, cysS, have been detected in hundreds of organisms encompassing all three living domains (1). The only exceptions are certain methanogenic archaea, the completed genome sequences of which encode no open reading frames (ORFs) with obvious homology to known cysS sequences (1). Because of the discovery that the genomes of a number of methanogenic archaea either lack cysS (Methanocaldococcus jannaschii, Methanothermobacter thermautotrophicus, and Methanopyrus kandleri) or can dispense with it (Methanococcus maripaludis), the formation of Cys-tRNACys in these organisms has been a much studied and increasingly contentious topic (2, 3). A noncognate aminoacyl-tRNA synthetase [aaRS (46)] and a previously unassigned ORF (7) were variously implicated in Cys-tRNACys formation. Recent studies failed to provide conclusive support for either of these routes, leaving the mechanism of Cys-tRNACys formation still in doubt (2).

Previous investigations of archaeal Cys-tRNACys biosynthesis have been hampered by the significant levels of noncognate tRNA routinely cysteinylated and detected by conventional filter binding assays. This problem was circumvented with a more stringent assay of Cys-tRNACys formation: gel-electrophoretic separation of uncharged tRNA from aminoacyl-tRNA (aa-tRNA) and subsequent detection of the tRNA moieties by sequence-specific probing (8). Given that M. jannaschii is a strict anaerobe, and considering that earlier aerobic purification erroneously identified prolyl-tRNA synthetase (4, 5), we used anaerobic conditions for all procedures unless otherwise indicated. When these procedures were used to monitor acylation of total M. maripaludis tRNA by an undialyzed M. jannaschii cell-free extract (S-100), tRNACys was charged with an amino acid that gave rise to the same mobility shift (9) exhibited by standard M. maripaludis Cys-tRNACys generated by M. maripaludis CysRS (1) (Fig. 1A, lanes 7 and 8). Further optimization of the reaction at this stage showed that Zn2+ and ATP were also required for the successful formation of charged tRNACys. When the S-100 fraction was dialyzed, all enzyme activity was lost and could not be recovered by addition of a mixture of the 20 canonical amino acids (Fig. 1A, lanes 3 and 4). These data established that tRNACys charging took place in the S-100 extract but not as a result of direct acylation of cysteine to tRNACys and not by a Ser-tRNACys–dependent conversion mechanism (10). In contrast, the dialyzed S-100 extract supplemented with 20 amino acids formed Ser-tRNASec (Fig. 1B, lanes 3 and 4), as did M. maripaludis seryl-tRNA synthetase (Fig. 1B, lanes 1 and 2). This result is consistent with a tRNA-dependent transformation of serine to selenocysteine (Sec) as seen in bacteria (11). On the basis of these results, we reasoned that the Cys-tRNACys–forming activity consisted of one or more enzymes and some low-molecular-weight substrates that together participated in a tRNA-dependent amino acid biosynthesis pathway.

Fig. 1.

Acid urea gel electrophoresis and Northern blot analysis of total M. maripaludis tRNA charged with M. maripaludis SerRS, dialyzed M. jannaschii S-100, M. maripaludis CysRS, and M. jannaschii SepRS in the presence of 20 amino acids (20 AA), phosphoserine, or a M. jannaschii S-100 cell-free extract filtrate (Y3). Half of each tRNA sample was deacylated by mild alkaline hydrolysis (OH). The blots were probed with 32P-labeled oligonucleotides complementary to M. maripaludis tRNACys (A) and M. maripaludis tRNASec (B). Total M. maripaludis tRNA charged with dialyzed or undialyzed M. maripaludis DcysS S-100 cell-free extract (20) in the presence of 20 amino acids and Na2S, or Sep and NasS (C). The blot was analyzed with 32P-labeled oligonucleotides complementary to M. maripaludis tRNACys.

To identify the components of the Cys-tRNACys biosynthetic pathway, the M. jannaschii S-100 extract was separated into two fractions: a low-molecular-mass “filtrate”(Y3) derived by a membrane filtration step (cutoff at 3 kD) and a protein fraction. Addition of Y3 to the dialyzed M. jannaschii S-100 restored activity (Fig. 1A, lanes 5 and 6). Both the protein and the filtrate fractions were purified individually by various chromatographic procedures; the activity was assayed by reconstitution of purified fractions from both sources [see supporting online material (SOM)]. Chromatographic analysis of the filtrate initially implicated O-phosphoserine (Sep) as one of the components in Y3 necessary for formation of Cys-tRNACys. This was subsequently verified using the l-enantiomer of this amino acid (see SOM for details). Significant advancement in the protein purification strategy was derived from a proteomic analysis of various partially purified column chromatographic fractions (12). Repeated liquid chromatography (LC)–mass spectrometry (MS) analysis in the pattern LC-LC-MS-MS identified 20 proteins in the most active fractions, of which 13 were excluded because of their predicted functions or inconsistent phylogenetic distribution. Of the remaining seven proteins, two of the most abundant (Mj1660 and Mj1678) were consistently observed in genomes lacking cysS. Although Mj1660 is a paralog of the α subunit of phenylalanyl-tRNA synthetase (PheRS), it is inactive in Phe-tRNA formation (13). Mj1678 has been annotated as a putative pyridoxal phosphate–dependent enzyme. On the basis of its high homology to known class II aaRSs, we speculated that Cys-tRNACys biosynthesis could be initiated by Mj1660 with Sep as one of the substrates. His6-Mj1660, produced and purified heterologously from Escherichia coli, was found to stably attach Sep to tRNACys in an efficient aerobic ATP-dependent reaction, which suggested that it could function as an aaRS (Fig. 1A, compare lanes 9 and 10 with lanes 11 and 12, and Fig. 2). However, tRNASec was not a substrate for Mj1660 (Fig. 1B, lanes 9 to 12). Specificity for Sep was further supported by the observation that His6-Mj1660 and its M. thermautotrophicus counterpart His6-Mth1501 both catalyzed Sep-dependent and tRNA-independent ATP-[32P]pyrophosphate exchange, a reaction characteristic of aaRSs (Fig. 3A) (14). No pyrophosphate exchange activity was detected with either His6-Mj1660 or His6-Mth1501 when Sep was replaced by phenylalanine. Sep was unable to stimulate ATP-[32P]pyrophosphate exchange by E. coli PheRS, which indicated that it is a specific substrate for Mj1660-type proteins. Analysis of the position of aminoacylation by using M. thermautotrophicus total tRNA labeled with [32P] in the terminal pA residue showed that Sep was attached to the 3′ terminus, the normal site for aminoacylation by aaRSs (Fig. 3B). A similar conclusion came from the protection against periodate oxidation of charged tRNACys (9). In light of these various enzymatic activities and their specificities, we propose that Mj1660-type proteins are classified as aaRSs and are consequently renamed O-phosphoseryl-tRNA synthetase (SepRS, encoded by sepS). Like pyrrolysyl-tRNA synthetase (PylRS), which acylates a suppressor tRNA with pyrrolysine, SepRS belongs to an emerging set of synthetases that use modified amino acids but not their canonical counterparts (15, 16). Amino acid sequence similarities indicate that both PylRS and SepRS are subclass IIc aaRSs most closely related to the canonical PheRS. The relative scarcity and narrow phylogenetic distributions of both PylRS and SepRS make it unclear whether these enzymes recently diverged from PheRS or, instead, coevolved with PheRS from a common ancestor.

Fig. 2.

Amino acid specificity of M. jannaschii SepRS. Aminoacylation by the recombinant M. jannaschii SepRS was tested with the filter binding assay (as described in SOM). M. jannaschii unfractionated tRNA charged with Sep (squares), total M. maripaludis tRNA and M. jannaschii SepRS incubated with Sep (circles), or with a 20–amino acid mixture (diamonds).

Fig. 3.

Amino acid activation and aminoacylation by SepRS. (A) ATP–inorganic pyrophosphate (PPi) exchange catalyzed by M. jannaschii SepRS and Sep or Phe; M. thermautotrophicus SepRS and Sep or Phe; and Escherichia coli PheRS and Sep or Phe. (B) 3′-Aminoacylation of M. thermautotrophicus total tRNA with Sep by SepRS (right panel). [α-32P]A76 total tRNA was aminoacylated with Sep by using SepRS (0.1 μM) and was subjected to RNase P1 digestion; the products were separated by thin-layer chromatography (TLC) and then visualized by phosphor imaging. Quantification of Sep∼[α-32P] indicated that about 3% of the total tRNA can be aminoacylated with Sep. The position of migration of Sep∼pA was independently confirmed using [14C]Sep (left).

Attachment of Sep to tRNACys by SepRS is a chemically plausible first step in Cys-tRNACys synthesis, as Sep-tRNA could feasibly be converted to Cys-tRNA in the presence of a synthase and the appropriate sulfur donor. Analogous pretranslational amino acid modifications have been described for the synthesis of asparaginyl-, formylmethionyl-, glutaminyl-, and selenocysteinyl-tRNAs (17). To investigate whether such a transformation accounts for Cys-tRNACys formation, preformed Sep-tRNACys was incubated with a dialyzed M. jannaschii S-100 extract in the presence of Na2S. Electrophoretic analysis of the resulting aa-tRNA indicated formation of a product whose mobility was consistent with Cys-tRNACys (Fig. 4A). On the basis of the above proteomic analysis, we postulated that Mj1678 encoded the enzymatic component responsible for converting Sep-tRNACys to Cys-tRNACys. His6-Mj1678, produced heterologously in E. coli, was found to efficiently convert preformed Sep-tRNACys into Cys-tRNACys in an anaerobic reaction in the presence of pyridoxal phosphate (PLP) and Na2S (Fig. 4B, lane 5). The natural sulfur donor of the reaction remains uncharacterized. On the basis of the conversion activity, we suggest that Mj1678 is a Sep-tRNA:Cys-tRNA synthase (SepCysS; encoded by pscS). SepRS and SepCysS, both of which are encoded in all archaea lacking cysS, together provide a facile two-step pathway for the synthesis of Cys-tRNACys by means of Sep-tRNACys (Fig. 4C). This route is consistent with the earlier observation that Sep is a precursor of cysteine in M. jannaschii (18). As in other organisms (19), the proposed route of Sep formation involves r-3-phosphoglycerate dehydrogenase (MJ1018) and an as yet unidentified phosphoserine aminotransferase.

Fig. 4.

Conversion of in vitro synthesized Sep-tRNACys to Cys-tRNACys. (A) Aminoacylation of tRNACys monitored by acid urea gel electrophoresis and Northern blotting. Lane 1, total M. maripaludis tRNA; lane 2, tRNACys charged with Sep by recombinant M. jannaschii SepRS; lane 3, Sep-tRNACys incubated with dialyzed M. jannaschii cell-free S-100 extract in the presence of dithiothreitol (DTT) and Na2S; lane 4, tRNACys charged with cysteine by M. maripaludis CysRS. (B) Phosphorimages of TLC separation of [14C]Sep and [14C]Cys recovered from the aa-tRNAs of the SepCysS activity assay (see SOM). Cysteine was analyzed in its oxidized form as cysteic acid (Cya). Lane 1, Ser marker; lane 2, cysteine from Cys-tRNACys generated with M. maripaludis CysRS; lane 3, Sep from Sep-tRNACys made with M. jannaschii SepRS; lane 4, Sep-tRNACys incubated with E. coli S-100 cell-free extract in the presence of DTT and Na2S (see SepCysS assay in SOM); lane 5, Sep-tRNACys converted to Cys-tRNACys with recombinant MJ1678 protein in the presence of DTT and Na2S (see SepCysS assay in SOM). (C) Scheme of Cys-tRNACys formation in methanogenic archaea.

From available genome sequences, the organismal distributions of SepRS and SepCysS are apparently coupled. To date, sepS and pscS have only been detected in the genomes of the methanogenic archaea M. jannaschii, M. maripaludis, M. thermautotrophicus, M. kandleri, Methanococcoides burtonii, the Methanosarcinaceae, and in Archaeoglobus fulgidus. Although some of these organisms lack cysS, others, such as M. maripaludis, also encode a canonical CysRS and thus contain two potentially functional pathways for Cys-tRNACys synthesis (20). Comparable redundancy is seen for Asn-tRNAAsn synthesis in many bacteria, where the tRNA-dependent route is the sole pathway for asparagine biosynthesis (21). Present knowledge of the genes required for archaeal amino acid biosynthesis suggests that the SepRS/SepCysS pathway may provide the only means for de novo production of cysteine in a number of organisms (e.g., M. jannaschii, M. maripaludis), whereas other organisms (e.g., Methanosarcinaceae) have both tRNA-dependent and tRNA-independent routes to cysteine. In contrast, most nonmethanogenic archaea with known genomes (e.g., Aeropyrum, Sulfolobus, Pyrococcus, Pyrobaculum, Thermoplasma, Picrophilus, Halobacteria) encode O-acetylserine sulfhydrylase (22) or cysteine synthase, which suggests that cysteine biosynthesis is tRNA-independent in these organisms.

To investigate whether the SepRS/SepCysS pathway can act as the sole route for cysteine biosynthesis we used M. maripaludis, which has a facile genetic system. This organism has both a dispensable CysRS (20) and the sepS and pscS genes but no known pathway for de novo biosynthesis of free cysteine. Biochemical evidence of a functional SepRS/SepCysS pathway in M. maripaludis extracts is presented in Fig. 1C. In dialyzed extracts of a cysS deletion mutant, Cys-tRNACys biosynthesis is dependent on the addition of Sep and Na2S (Fig. 1C, lane 2). To test if the SepRS/SepCysS pathway is necessary for cysteine biosynthesis, the sepS gene was deleted from the chromosome of the wild type of M. maripaludis. The resulting DsepS strain was a cysteine auxotroph (Fig. 5). Although it grew at a rate comparable to that of wild type on complete medium, it was unable to grow in the absence of exogenous cysteine. These findings indicate that under certain conditions the SepRS/SepCysS pathway can provide the sole source of cysteine for the cell via Cys-tRNACys. Reliance on such a route clearly satisfies the requirements for cysteine during protein synthesis, but how cysteine is made available for other metabolic processes is less clear. One possibility is that hydrolysis of Cys-tRNACys directly provides free cysteine, as previously proposed for free Asn synthesis via Asn-tRNAAsn in certain bacteria (21). In addition, protein turnover in the cell would be expected to contribute more significantly to the cellular cysteine pool when CysRS is absent, as the free amino acid is not itself a substrate for protein synthesis in such cases. Finally, most of the organisms harboring the SepRS/SepCysS pathway are methanogens, which, even in the absence of glutathione, may not require a large pool of free cysteine for redox buffering in the cytoplasm. Methanogens contain high levels of the essential coenzyme 2-mercaptoethanesulfonate (23), which may fulfill the redox buffering function of free cysteine. For thermophilic organisms, replacement of the heat-labile cysteine with the thermostable 2-mercaptoethanesulfonate may be an additional benefit.

Fig. 5.

Growth response of the DsepS mutant and wild-type M. maripaludis strain S2 to the presence and absence of cysteine in mineral media containing acetate. About 2 × 103 cells were inoculated into prewarmed McAV medium containing 3 mM coenzyme M for a final cysteine concentration of <0.16 μM ([Cys]low) or into the same medium with 3 mM cysteine ([Cys]high). Wild-type S2 (circles), and sepS mutant S210 (squares).

The discovery of the SepRS/SepCysS pathway raises the question as to whether this mechanism predates direct charging by CysRS and tRNA-independent cysteine biosynthesis. Similar scenarios have been suggested for Asn-tRNA and Gln-tRNA biosynthesis, where the tRNA-dependent pathways have been proposed as the original routes for synthesis of both the aa-tRNAs and the corresponding amino acids (2426). If SepRS/SepCysS was indeed the ancestral pathway for cysteine synthesis via Cys-tRNA, a lack of alternative cysteine biosynthetic capacity may explain why certain organisms have retained this route. This would be consistent with earlier proposals that CysRS (27, 28) and cysteine itself (22, 29, 30) were the last—or very late—canonical additions to the genetic code. The recent demonstration in mammalian cells (31) of the Ser-tRNASec to Sep-tRNASec conversion by a special kinase [present also in archaea (31)] implicates the Sep moiety as an intermediate in Sec synthesis. As the conversions of Sep-tRNA to Cys-tRNA or Sec-tRNA are chemically analogous (using suitable sulfur or selenium donors, respectively), the addition of selenocysteine to the genetic code may have been patterned on an accepted route for cysteine formation, the SepRS/SepCysS pathway.

Supporting Online Material

www.sciencemag.org/cgi/content/full/307/5717/1969/DC1

Materials and Methods

Figs. S1

References and Notes

References and Notes

View Abstract

Navigate This Article