Methanocaldococcus jannaschii Prolyl-tRNA Synthetase Charges tRNAPro with Cysteine*

Methanocaldococcus jannaschiiprolyl-tRNA synthetase (ProRS) was previously reported to also catalyze the synthesis of cysteinyl-tRNACys(Cys-tRNACys) to make up for the absence of the canonical cysteinyl-tRNA synthetase in this organism (Stathopoulos, C., Li, T., Longman, R., Vothknecht, U. C., Becker, H., Ibba, M., and Söll, D. (2000) Science 287, 479–482; Lipman, R. S., Sowers, K. R., and Hou, Y. M. (2000)Biochemistry 39, 7792–7798). Here we show by acid urea gel electrophoresis that pure heterologously expressed recombinantM. jannaschii ProRS misaminoacylates M. jannaschii tRNAPro with cysteine. The enzyme is unable to aminoacylate purified mature M. jannaschiitRNACys with cysteine in contrast to facile aminoacylation of the same tRNA with cysteine by Methanococcus maripaludiscysteinyl-tRNA synthetase. Although M. jannaschii ProRS catalyzes the synthesis of Cys-tRNAPro readily, the enzyme is unable to edit this misaminoacylated tRNA. We discuss the implications of these results on the in vivo activity of the M. jannaschii ProRS and on the nature of the enzyme involved in the synthesis of Cys-tRNACys in M. jannaschii.

genic archaea. The reports that purified recombinant M. jannaschii prolyl-tRNA synthetase (ProRS) was able to aminoacylate tRNA Cys with cysteine provided an apparent answer to this question (6 -11). This led to the notion of ProRS as an unusual dual specificity enzyme capable of attaching proline to tRNA Pro and cysteine to tRNA Cys (6,7); thus, the enzyme was named ProCysRS (12). The finding (6,9) that unmodified tRNA Cys , produced by in vitro transcription of the corresponding tRNA Cys gene, could not be aminoacylated with cysteine by M. jannaschii ProCysRS suggested that nucleotide modifications present in the mature M. jannaschii tRNA Cys were essential for correct aminoacylation with cysteine (11) as shown before for some tRNAs (13)(14)(15). However, much of the previous work was done with unfractionated M. jannaschii tRNA. Our recent finding that cysteine activation is an inherent property of ProRS enzymes (43) prompted us to undertake an in-depth investigation of the tRNA specificity of archaeal prolyl-tRNA synthetases as exemplified by the M. jannaschii enzyme. Here we show that a purified recombinant form of the M. jannaschii enzyme, previously called ProCysRS, misaminoacylates tRNA Pro with cysteine but is unable to aminoacylate tRNA Cys with cysteine. These properties of M. jannaschii ProRS raise the question of the true in vivo activities of this enzyme. In view of these results, we have renamed the enzyme ProRS in this paper.

EXPERIMENTAL PROCEDURES
General-Oligonucleotide synthesis and DNA sequencing were performed by the Keck Foundation Research Biotechnology Resource Laboratory at Yale University. The TOPO-TA cloning kit was from Invitrogen. pET-15b expression vector was purchased from Novagen. Epicurian coli © BL21-CodonPlus TM competent cells were from Stratagene. [ 35  Isolation of tRNA and tRNA Transcripts-Isolation of T7 RNA polymerase transcripts of tRNA Pro and M. jannaschii total tRNA was carried out as described previously (6,8).
Purification of tRNA Pro and tRNA Cys by Affinity Chromatography-Individual Cys-tRNA Pro and Cys-tRNA Cys samples were purified by affinity chromatography on immobilized Thermus thermophilus EF1A-GTP as described previously (16). Unfractionated M. jannaschii tRNA was aminoacylated with [ 35 S]cysteine using either M. jannaschii ProRS or M. maripaludis CysRS. The resulting Cys-tRNA was separated from the uncharged tRNA by the formation of a ternary complex with T. thermophilus EF1A-GTP immobilized on a Ni-NTA-agarose column. After elution of Cys-tRNA and removal of cysteine by deacylation (30 min at 37°C in 0.2 M Tris acetate, pH 9.0), the tRNAs were found to be highly purified as judged by amino acid acceptor activity (1440 pmol/ A 260 for tRNA Cys and 1040 pmol/A 260 for tRNA Pro ).
Wild type and Mutant ProRS and CysRS Gene Constructs and Purification of Enzymes-The M. jannaschii proS gene and the M. maripaludis cysS gene previously cloned into pET-15b expression vector (6,17) were used. A carboxyl-terminal deletion of M. jannaschii ProRS (named ProRS⌬50) was constructed by the removal of 50 carboxylterminal amino acids. The PCR product was cloned into TOPO-TA vector, sequenced, and finally subcloned into pET-15b vector for overexpression. The His 6 -tagged M. jannaschii ProRS and ProRS⌬50 as well as His 6 -tagged M. maripaludis CysRS (17) were obtained by overexpression in Escherichia coli and purified by Ni-NTA chromatography. These enzymes are referred to as recombinant enzymes, because they are derived by heterologous expression of cloned genes. Protein concentration was measured by the Bradford assay (Bio-Rad) using bovine serum albumin as standard. The enzyme preparations were Ͼ95% pure as judged by SDS-PAGE followed by staining with Coomassie Brilliant Blue.
Aminoacylation of tRNA-The standard reaction mixture contained 50 mM HEPES-KOH, pH 7.0, 50 mM KCl, 15  Acid Urea Gel Electrophoresis of tRNA and Aminoacyl-tRNA-This method (19) allows the separation of charged from uncharged tRNA attributed to a difference in electrophoretic mobility between the two species. Hybridization of a sequence-specific probe permits the determination of the identity of the aminoacylated tRNA on the gel. Unfractionated M. jannaschii tRNA was aminoacylated either with proline (by M. jannaschii ProRS) or cysteine (by M. jannaschii ProRS or M. maripaludis CysRS) as described previously (6,19). After phenol extraction and ethanol precipitation, the aminoacyl-tRNAs were dissolved in 10 mM sodium acetate, pH 4.5, 1 mM Na-EDTA at a final concentration of 8 g/l and stored at Ϫ80°C. Half of the preparation was deacylated for 30 min at 37°C with 0.2 M Tris acetate, pH 9.0. For the electrophoresis, 20 g of unfractionated tRNA (aminoacylated or deacylated) were loaded on a 65-cm long, 0.4-mm thick 6.5% polyacrylamide gel containing 7 M urea and run at 4°C, 600 V in 0.1 M sodium acetate, pH 5.0, for 48 h. The portion of the gel containing the tRNAs was electroblotted onto a Hybond N ϩ membrane (Amersham Biosciences) using a Hoefer Electroblot apparatus at 10 V for 15 min and then at 30 V for 2 h with 10 mM Tris acetate, pH 8.0, 5 mM sodium acetate, and 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 to a 5Ј-32 P-labeled oligodeoxyribonucleotide probe. The probes were complementary to nucleotides 1-21 and 26 -51 of tRNA Cys and tRNA UGG Pro , respectively. Replacement of the 3Ј-terminal Adenosine of tRNAs by 2Ј-Deoxyadenosine or 3Ј-Deoxyadenosine-Replacement of the terminal adenosine of M. jannaschii tRNA Pro transcript (A 260 ϭ 12.5 units) by 2Ј-deoxyadenosine, 3Ј-deoxyadenosine, or adenosine was performed as described previously (20 -22) and consisted of the following steps: periodate oxidation (50 mM sodium acetate, pH 6.5, 1 mM sodium periodate) for 2 h in the dark at 37°C followed by the removal of excess periodate (addition of 0.2% glucose at 25°C for 30 min in the dark), ␤-elimination (250 mM lysine, pH 9.0, for 4 h at 25°C), dephosphorylation (50 mM Tris-HCl, pH 9.0, 1 mM MgCl 2 , 25 units of bacterial alkaline phosphatase for 90 min at 37°C), and reconstruction of the 3Ј-terminal residue (20 mM 2Ј-dATP, 3Ј-dATP, or ATP in presence of yeast tRNA CCA-nucleotidyl transferase in 100 mM Tris-HCl, pH 9.0, 50 mM KCl, 25 mM MgCl 2 , 2 mM dithiothreitol, and 0.1 mg/ml bovine serum albumin). tRNAs were extracted with phenol/chloroform (1:1), precipitated with ethanol in the presence of glycogen (American Scientific), and resuspended in sterile water. The same procedure was followed to isolate the 2Ј-deoxyadenosine or 3Ј-deoxyadenosine containing analogues of tRNA starting from unfractionated M. jannaschii tRNA. The tRNAs reconstructed with ATP had similar charging levels compared with the original tRNA Pro transcript (proline charging, 300 pmol/A 260 ) or the unfractionated tRNA (cysteine charging, 20 pmol/A 260 ). Assays

M. jannaschii ProRS Aminoacylates tRNA Pro with Cysteine-
Acid urea gel electrophoresis has been a reliable tool to determine the aminoacylation specificity of aminoacyl-tRNA synthetases (19). The method is based on the electrophoretic separation of tRNA and aminoacyl-tRNA and subsequent hybridization with a sequence-specific oligonucleotide to determine the identity of the aminoacylated tRNA. To identify which of the M. jannaschii tRNAs were being aminoacylated by the recombinant M. jannaschii ProRS, we aminoacylated unfractionated M. jannaschii tRNAs with the ProRS in the presence of proline or cysteine. Following acid urea gel electrophoresis, the tRNAs that were aminoacylated were identified by Northern blot analysis using probes specific for tRNA Pro (Fig. 1A) or tRNA Cys (Fig. 1B). As a positive control for aminoacylation of tRNA Cys , recombinant M. maripaludis CysRS was also used in a separate aminoacylation reaction. Fig. 1 shows the results. M. jannaschii ProRS charges proline onto tRNA Pro (Fig. 1A, lane 6) but not onto tRNA Cys (Fig. 1B, lane 10). M. jannaschii ProRS also aminoacylates tRNA Pro essentially quantitatively with cysteine (Fig. 1A, lane 5) but does not aminoacylate tRNA-Cys with cysteine (Fig. 1B, lane 12). The tRNA Cys in the unfractionated tRNA preparation used is fully active, because it could be aminoacylated essentially quantitatively by M. maripaludis CysRS (Fig. 1B, lane 13). Thus, M. jannaschii ProRS misaminoacylates tRNA Pro with cysteine but does not aminoacylate tRNA Cys . As expected, all of the aminoacyl-tRNAs could be deacylated to tRNA by treatment with base (Fig. 1A, lanes 4  and 7, and Fig. 1B, lane 14).
The above conclusion that M. jannaschii ProRS misaminoacylates tRNA Pro with cysteine but not tRNA Cys was also confirmed by aminoacylation of purified preparations of M. jannaschii tRNA Pro and tRNA Cys isolated by EF1A affinity chromatography (16). For this experiment, unfractionated M. jannaschii tRNA was aminoacylated with cysteine using M. jannaschii ProRS or M. maripaludis CysRS. The aminoacyl-tRNA was separated from all of the other tRNAs by chromatography on GTP-activated T. thermophilus EF1A immobilized on a Ni-NTA-agarose column. Following elution of the aminoacyl-tRNA, the amino acid attached to the tRNA was removed by treatment with base. The resulting tRNA Pro sample could be aminoacylated by M. jannaschii ProRS with proline or with cysteine to approximately equal levels, whereas it could not be aminoacylated with cysteine by M. maripaludis CysRS (Fig.  2A). The tRNA Cys sample could be aminoacylated by M. maripaludis CysRS, but it could not be aminoacylated with either cysteine or proline by M. jannaschii ProRS (Fig. 2B). These results demonstrate that M. jannaschii ProRS catalyzes the synthesis of Cys-tRNA Pro but not of Cys-tRNA Cys . Modified nucleosides in M. jannaschii tRNA Pro , suggested to be a prerequisite for accurate aminoacylation (11), are not essential as the transcript of the tRNA Pro gene can be aminoacylated with proline or cysteine (data not shown) (11).
Site of Aminoacylation of the tRNA Pro with Proline and Cysteine-tRNA Pro molecules in which the terminal adenosine was replaced by 2Ј-or 3Ј-deoxyadenosine were used as substrates for the M. jannaschii ProRS to identify the site of attachment of the amino acid in Pro-tRNA Pro and Cys-tRNA Pro . The results (Fig. 3) show that both for aminoacylation with proline ( Fig.  3A) and with cysteine (Fig. 3B), the tRNA Pro transcript with the 2Ј-deoxyA at the 3Ј-end is a better substrate than the transcript with the 3Ј-deoxyA. Thus, the site of attachment of the amino acid is primarily the 3Ј-hydroxyl group. There is also amino attachment on the 2Ј-hydroxyl group, in line with the earlier data on E. coli ProRS (20 -22). This conclusion agrees with the fact that class II aminoacyl-tRNA synthetases transfer the amino acid onto the 3Ј-hydroxyl group of the terminal adenosine of tRNA, whereas the class I enzymes transfer it to the 2Ј-hydroxyl group (25,26).
Recognition of Cysteine by ProRS-Because M. jannaschii ProRS synthesizes Cys-tRNA Pro and because misaminoacyla- tion of tRNA Pro with cysteine in vitro appears to be a general feature of ProRSs (43), it was of interest to investigate the recognition of cysteine analogues by the enzyme. Two complementary methods were used. First, the ability of non-radioactive cysteine analogues to inhibit Cys-tRNA Pro formation was evaluated. As seen in Table I, compounds with an altered carboxyl group (cysteamine, L-cysteine methyl ester, and Lcysteine ethyl ester) inhibited aminoacylation. In contrast, oxidation or reduction of the sulfhydryl group yielded compounds (L-sulfinic acid, L-cysteic acid, and S-methyl-L-cysteine), which did not inhibit aminoacylation. N-Acetyl-L-cysteine, D-cysteine, and DL-homocysteine were also found to be inhibitors. Second, the question of whether the inhibitors could be activated by M. jannaschii ProRS was investigated (Fig. 4). It was found that D-cysteine and DL-homocysteine were substrates in the ATP-PP i exchange reaction. These results suggest that the sulfhydryl group and to a lesser extent the ␣-amino group of cysteine are recognized by M. jannaschii ProRS. E. coli CysRS, co-crystallized with cysteine, shows a clear interaction of the amino group with the enzyme and the ϪSH group with a protein-bound zinc atom (5).
Does ProRS Deacylate Cys-tRNA Pro by Editing?-Editing of the misacylated Ala-tRNA Pro by E. coli ProRS was shown to be highly efficient and ascribed to the existence of an insertion domain in the bacterial-type ProRSs compared with the archaeal ProRS (27). The archaeal type ProRS does not contain this insertion sequence but has an extension at the carboxyl terminus. Thus, the latter region could be a possible editing domain. Hydrolysis of a mutant Ala-tRNA Pro by M. jannaschii ProRS, although not as efficient as by the bacterial ProRS, was recently reported (23). Because this editing function might be extended to the hydrolysis of other mischarged tRNAs, we investigated whether the deletion of the carboxyl terminus of M. jannaschii ProRS would affect either the formation or the hydrolysis of Cys-tRNA Pro and Ala-tRNA Pro . We reasoned that the deletion of a possible editing domain would lead to an increased mischarging of cysteine and alanine onto unfractionated M. jannaschii tRNA.
First, we checked the ability of the M. jannaschii ProRS⌬50 enzyme (lacking the 50 carboxyl-terminal amino acids) to charge proline, cysteine, or alanine onto M. jannaschii tRNA Pro . There was no difference between the wild-type and the truncated ProRS in an initial velocity of aminoacylation for proline, although the mutant enzyme was less active for cysteine (ϳ3-fold) and alanine (ϳ10-fold) (Fig. 5). Second, we checked the editing activity of the wild-type ProRS or the ProRS⌬50 enzyme toward Cys-tRNA Pro (Fig. 6A) or Ala-tRNA Pro (Fig. 6B) in post-transfer editing experiments. There was no measurable deacylation of either of the substrates. Thus, under our experimental conditions, we could not show in M. jannaschii ProRS an editing mechanism for mature Cys-tRNA Pro or Ala-tRNA Pro . Whether the difference between our results and those of Beuning and Musier-Forsyth (23) is because of their use of a mutant Ala-tRNA Pro for the deacylation experiments is not known.

DISCUSSION
Based on the nature of the final aminoacyl-tRNA products, aminoacyl-tRNA synthetases may be divided into three groups. The first group, which comprises the majority of these enzymes, attaches the cognate amino acid to the cognate set of isoacceptor RNAs. These enzymes correct any misactivation of the amino acid or any misacylation (mischarging) of the tRNA by pre-transfer and/or post-transfer editing mechanisms (28). Another group consisting of the non-discriminating aspartyl-, glutamyl-, and seryl-tRNA synthetases recognizes the cognate amino acid yet charges it onto two different sets of tRNAs. These non-discriminating AspRS and GluRS enzymes provide the crucial first step, the production of mischarged tRNA, in the transamidation pathway that leads to formation of Asn-tRNA and Gln-tRNA (29,30). SerRS similarly is involved in the formation of selenocysteinyl-tRNA (31). Lastly, the archaeal ProRS described in this study may belong to a third type as it charges two canonical amino acids onto the same set of tRNAs, generating a correctly charged and a mischarged aminoacyl-tRNA. Mutant aminoacyl-tRNA synthetases with a compromised editing function (a valyl-tRNA synthetase forming Cys-tRNA Val and a leucyl-tRNA synthetase forming Ile-tRNA Leu and Met-tRNA Leu ) are known to do the same (32-34) and may also be placed in the third group.
How does the archaeal prolyl-tRNA synthetase recognize two canonical amino acids? Cysteine is well recognized by the M. jannaschii ProRS, which is significantly more mischarging than the ProRS enzymes of many other organisms (43). Earlier biochemical experiments showed that in M. jannaschii ProRS, the binding sites for proline and cysteine are overlapping in that the mutation of amino acid 100 greatly affected CysRS activity, whereas the mutation of amino acid 103 greatly affected ProRS activity (8). The crystal structure of M. jannaschii ProRS is consistent with this conclusion and shows that the proline binding pocket of this enzyme readily accommodates cysteine. 2 Further support is derived from suggestions based on a modeling experiment that the archaeal-type T. thermophilus ProRS could also bind cysteine (35). Proline and cysteine have similar molecular volumes, although they have different shapes. This particular feature may explain the lack of speci-2 S. Kamtekar, D. Kennedy, and T. Steitz, personal communication.  ficity of M. jannaschii ProRS for the two amino acids. Also, these two amino acids may be too similar to be distinguished by a double sieve editing mechanism (36). CysRS also must face the challenge of discriminating proline. The crystal structure of E. coli CysRS (5) shows that this enzyme has developed a different strategy to ensure accuracy of aminoacylation. The enzyme selects cysteine specifically by coordination of the sulfhydryl group with a zinc atom. As this mechanism ensures better specificity, the enzyme can dispense with an editing domain (5). A metal ion (Zn 2ϩ )-based mechanism is also used by threonyl-tRNA synthetase to discriminate against the isosteric valine (37).
An alternative mechanism to increase discrimination for the cognate amino acid in aminoacylation of tRNA Pro was suggested for T. thermophilus ProRS (38). Based on the crystal structures of T. thermophilus ProRS complexed with tRNA Pro , proline, ATP, and prolyl-adenylate, it was proposed that the binding and activation of proline triggers structural changes in the active site that promote the correct binding of the 3Ј-end of tRNA Pro at the active site. Such an induced fit mechanism in which binding or activation of the cognate amino acid results in the enzyme adopting a fully ordered active site conformation necessary for aminoacylation also seems to be used by some other class IIa enzymes, e.g. E. coli histidyl-tRNA synthetase (39). In T. thermophilus ProRS, Ala-206 is one of the amino acids in the "proline binding loop" close to the active site, which undergoes rearrangement upon the activation of proline. This rearrangement results in the formation of a hydrogen bond between the main chain of Ala-206 and His-83 in the "ordering loop." The replacement of Ala-206 by proline in the archaeal ProRSs may explain the much poorer discrimination observed in the archaeal enzymes.
Our unexpected finding that M. jannaschii ProRS misaminoacylates native tRNA Pro with cysteine in vitro raises the question of whether this also occurs in vivo and, if so, whether any Cys-tRNA Pro that is formed is edited. Because proline and cysteine inhibit the aminoacylation of tRNA with each other by ProRS (6), the relative affinity of these amino acids for ProRS, their intracellular concentrations, their rates of aminoacyl adenylate formation, and rates of aminoacylation of tRNAs will all be important determinants for the relative amounts of Pro-tRNA Pro and Cys-tRNA Pro formed in vivo. In vitro studies with the M. jannaschii ProRS indicate that k cat /K m in proline activation is 107-fold higher than in cysteine activation (43). Unfortunately, no data are available on the intracellular concentration of these amino acids in M. jannaschii (or in any other prokaryote). However, it is known for enzymes in glycolysis that the physiological substrate concentrations are generally lower than the K m values for these substrates presumably to maximize reaction rates (40). Assuming that the same applies to aminoacyl-tRNA synthetases, a consideration of the K m values of ProRS and CysRS of the archaeon M. maripaludis could provide an indication of the intracellular levels of proline and cysteine. The ProRS of this organism, similar to that of M. jannaschii, is known to catalyze the synthesis of Cys-tRNA Pro in vitro (data not shown). The enzyme has a K m for proline of 4.6 M, whereas that for cysteine is 74.5 M (17). Based on these K m values, it is logical to assume that the cysteine concentration should be at least 15 times as high as that of proline for Cys-tRNA Pro synthesis to compete efficiently with Pro-tRNA Pro synthesis. However, the K m value of M. maripaludis CysRS for cysteine (9.7 M) suggests that the concentration of free cysteine would only be twice as high as that of proline. Thus, it is plausible that appropriate intracellular levels of proline and cysteine will minimize the mischarging of tRNA Pro with cysteine without recourse to editing.
Finally, our finding that the recombinant M. jannaschii ProRS is unable to aminoacylate tRNA Cys in vitro with cysteine reopens the question of how Cys-tRNA Cys is synthesized in M. jannaschii. Much of the recent work with this enzyme including that described in this paper has been carried out with the recombinant ProRS made in E. coli. The possibility that ProRS isolated from M. jannaschii contains a certain modification(s) that allows it to aminoacylate tRNA Cys with cysteine cannot, therefore, be ruled out. Alternatively, other RNA and/or protein factors along with M. jannaschii ProRS may be involved in the formation of Cys-tRNA Cys (7). Attempts in identifying such components by rescue of an E. coli cysS knock-out strain have not been successful. 3 Additional possibility is a recently described M. jannaschii protein that has the ability to aminoacylate tRNA with cysteine (41) or an unusual CysRS, which still remains to be discovered. In view of the finding in this paper that M. jannaschii ProRS can misaminoacylate tRNA Pro with cysteine, assays for CysRS activity in vitro must now go beyond relying on the incorporation of radioactive cysteine into unfractionated tRNA. Such assays will have to ensure that cysteine is being incorporated into tRNA Cys and not into other tRNAs. This will require the use of either purified M. jannaschii tRNA Cys or, as described in here and elsewhere (42), acid urea gel electrophoresis followed by Northern blot hybridization to identify the tRNA that is aminoacylated.