Aminoacyl-tRNA synthetases: A new image for a classical family

Biochimie 81 (1999) 683−700 © 1999 Société française de biochimie et biologie moléculaire / Éditions scientifiques et médicales Elsevier SAS. All rights reserved

Review

Aminoacyl-tRNA synthetases: A new image for a classical family Susan A. Martinisb, Pierre Plateauc, Jean Cavarellid, Catherine Florentza* a

UPR 9002 du CNRS, Institut de Biologie Moléculaire et Cellulaire, 15, rue René-Descartes, 67084 Strasbourg cedex, France b Department of Biology and Biochemistry, University of Houston, Houston, TX 77204-5513, USA c Laboratoire de Biochimie, UMR 7654, CNRS-École Polytechnique, 91128 Palaiseau cedex, France d Laboratoire de Biologie Structurale, Institut Génét. Biologie Moléculaire et Cellulaire, 1, rue Laurent-Fries, BP 163, 67404 Illkirch cedex, France (Received 29 April 1999; accepted 21 May 1999)

Abstract — The aminoacyl-tRNA synthetases (aaRSs) are a family of enzymes well known for their role in protein synthesis. More recent investigations have discovered that this classic family of enzymes is actually capable of a broad repertoire of functions which not only impact protein synthesis, but extend to a number of other critical cellular activities. Specific aaRSs play roles in cellular fidelity, tRNA processing, RNA splicing, RNA trafficking, apoptosis, transcriptional and translational regulation. A recent EMBO workshop entitled ‘Structure and Function of Aminoacyl-tRNA Synthetases’ (Mittelwihr, France, October 10–15, 1998), highlighted the diversity of the aaRSs’role within the cell. These novel activities as well as significant advances in delineating mechanisms of substrate specificity and the aminoacylation reaction affirm the family of aaRSs as pharmaceutical targets. © 1999 Société française de biochimie et biologie moléculaire / Éditions scientifiques et médicales Elsevier SAS aminoacyl-tRNA synthetases / molecular evolution / antibiotic / protein synthesis / RNA recognition / tRNA identity

1. Introduction The aminoacyl-tRNA synthetases (aaRS) were first discovered to ligate an amino acid to tRNA in 1958 [1, 2], signifying their important role in protein synthesis. Over thirty years of vigorous international research rendered this diverse family’s common aminoacylation activity a clear textbook example for defining general enzymeactivity relationships [3-5]. Here we present extensive progress in understanding the more classical features of aaRSs (appropriate reviews are referenced to provide more extensive information and further background to our brief descriptions of recent advances in the aaRS field). Moreover, as shown in figure 1, we review a new appreciation for the expanding role of the aaRSs based on the work of multiple laboratories and presented at the EMBO workshop entitled ‘Structure and Function of AminoacyltRNA Synthetases’ (the organizers of the meeting were Catherine Florentz (IBMC-CNRS, Strasbourg), Jean Cavarelli (IGBMC, Illkirch), Karin Musier-Forsyth (University of Minnesota, Minneapolis), and Pierre Plateau (CNRS-Ecole Polytechnique, Palaiseau)). We also discuss novel tRNA charging pathways discovered from mounting genomic information and intense biochemical investigations. These new pathways rely on distinct enzymes unrelated to their functional aaRS counterparts [6]. Fi* Correspondence and reprints

nally, extensive structural data for representatives of 19 of the 20 aaRSs (some of which span multiple species) continue to surprise and refine our understanding of structure-function relationships [7, 8] and substrate specificity [9]. The advances in the short period since the last meeting (Keystone Symposium on Aminoacyl-tRNA Synthetases in Biology and Disease: February 1–6, 1997, Taos, New Mexico, USA) on aaRSs also re-emphasize the family’s potential as a viable and essential target for drug development [10, 11]. 2. aaRSs proofread nascent tRNAs in the nucleus prior to export to the cytoplasm Eukaryotic pre-tRNAs are synthesized in the nucleus and require extensive processing to yield functional tRNAs [12-14]. The mature tRNA binds to a tRNAspecific receptor, exportin-t, and is exported to the cytoplasm via a Ran-GTPase-dependent mechanism [12, 15, 16]. In an elegant series of experiments in which tRNA transcripts were injected into the nuclei of Xenopus laevis oocytes, Lund and Dahlberg showed clearly that mature tRNAs (including methionine, tyrosine, alanine, asparagine, leucine, lysine, and phenylalanine specific tRNAs) are charged with their cognate amino acids in the nucleus prior to export to the cytoplasm [17]. Blockage of the tRNA export machinery resulted in accumulation of nuclear charged tRNA that could be isolated and detected

684

Martinis et al. with Arc1p (described below) which contains an NLS, it is possible that MetRS, for example, could be imported as part of a complex into the nucleus [18]. While nuclear import of tRNA synthetases remains to be experimentally tested and observed, the comparisons suggest discrete structure-function differences that emerged during the evolution of bacterial and eukaryotic tRNA synthetases. 3. The yeast protein, Arc1p, promotes tRNA binding to MetRS and GluRS

Figure 1. Cellular roles and functions of aaRS.

on acidic electrophoresis gels. In addition, introduction of inhibitors specific to a particular aaRS decreased export of the relevant uncharged tRNA. Thus, the aaRS-dependent aminoacylation reaction in the nucleus may serve as a stringent quality control mechanism prior to export of mature tRNA to the cytoplasm where it is functionally required for protein synthesis [17]. The discovery of a nuclear proofreading mechanism for tRNA maturation indicates that additional pathways or enzymes may also be enlisted to ensure accurate processing of other RNAs such as ribosomal and mRNAs. Ubiquitous cellular requirements for proofreading of biological molecules could account for the maturation and function of RNAs in separate subcellular compartments. Aminoacylation of tRNAs in the nucleus [15, 17] would presumably require import of tRNA synthetases from the cytoplasm where they are synthesized [18]. Proteins that are imported into the nucleus often contain a classical nuclear localization signal (NLS) consisting primarily of a pattern of basic residues. Careful analysis of Saccharomyces cerevisiae sequences determined that 15 of the cytoplasmic tRNA synthetases contained one or more potential classical NLSs [18]. Moreover, each of these putative NLSs were completely lacking in the corresponding E. coli tRNA synthetase and eight of the fifteen classical NLSs found in the yeast enzymes were located in a discrete domain or insertion that was also not found in the E. coli counterpart. MetRS (specific aaRSs are abbreviated using their three letter code followed by RS) was one of the five yeast tRNA synthetases which did not appear to have an NLS. However, through association

Genetic screening for components interacting functionally with the yeast nuclear pore-associated protein Los1p led to the discovery of the protein Arc1p which associates with MetRS and GluRS [19]. Further in vivo and in vitro studies have clearly shown that the N-terminal domain of Arc1p interacts with both aaRSs, while the C-terminal region binds tRNAs [20]. Upon binding to Arc1p, the affinity of the synthetases for their cognate tRNAs as well as their catalytic efficiency are enhanced. Therefore, Arc1p may serve as a mobile trans-acting tRNA binding domain. Los1p was recently determined to genetically interact with Arc1p as a nuclear export factor (exportin) for tRNA [21]. Thus, Arc1p may be involved in tRNA channeling by facilitating transport of tRNA out of the nucleus via the nuclear pores for delivery to the aaRSs in the cytoplasm [20]. The trans-acting tRNA binding domain may also be important in other organisms as it is conserved with several polypeptides, including the C-terminal domains of E. coli and C. elegans MetRSs, human and bovine TyrRS [19, 22] and p43 [23], a nonsynthetase protein which is part of the higher eukaryotic multimeric synthetase complex. Interestingly, proEMAPII (discussed below) which is homologous to these proteins was also found to interact with human ArgRS and stimulate aminoacylation activity by lowering the KM for tRNA (Park et al., unpublished results). 4. Mammalian TyrRS domains act as cytokines Components of the translation apparatus have not been previously identified to function as cytokines. However, Wakasugi and Schimmel reported that human TyrRS can be split into two distinct cytokines arising from the N- and C-terminal halves of the protein [24]. While the native enzyme is inactive in assays for a variety of cytokine activities, the sequestered TyrRS-derived cytokines are released upon cleaving the native enzyme at a site between the two domains. The endothelial-monocyte activating polypeptide II (EMAP II)-like C-terminal domain [19, 22] appears to lack an obvious role in aminoacylation, but when released from the fused N-terminal domain exhibits significant leukocyte and monocyte chemotaxis activity [24]. The free TyrRS C-terminal do-

Aminoacyl-tRNA synthetases main also induces production of myeloperoxidase, TNFα and tissue factor. Notably, Kornelyuk et al. have expressed and purified the recombinant EMAP II-like domain of bovine TyrRS. Likewise, the purified protein domain exhibits cytokine-like activities similar to those of EMAP II, namely: i) monocyte chemotaxis induction; ii) induction of tissue factor by endothelial cells; and iii) synergistic action in the up-regulation of tissue factor by TNFα [25]. Homologues to EMAP II are also found appended to MetRSs from some species, as well as aaRS-associated proteins (multi-aaRS complex) including human p43 [23] and yeast Arc1p [19] and also interact with human ArgRS (Park et al., unpublished results) as discussed above. A human mini-TyrRS was constructed which retained the catalytic N-terminal aminoacylation domain, but lacked the C-terminal EMAP II-like domain [24]. Surprisingly, the mini-TyrRS binds to the IL-8 type A receptor and acts as an IL-8-like chemokine. Certain chemokines contain a conserved Glu-Leu-Arg motif which was also found in the human TyrRS class I-specific Rossmann fold. Mutation of arginine to glutamine in this motif of the mini-TyrRS abolished leukocyte chemotaxis. Notably, E. coli and human full-length TyrRSs as well as the R→Q mini-Tyr RS mutant failed to stimulate chemotaxis, nor could competitively bind with the mini-TyrRS (KD = 21 nM) to the cytokine receptor. Under apoptotic conditions in culture, full-length TyrRS was shown to be specifically secreted and also cleaved by a protease such as leukocyte elastase to generate the two cytokines. Thus, in mammalian cells, TyrRS may play dual roles in apoptosis, halting translation while simultaneously stimulating cytokine production to recruit macrophages that remove cell corpses. 5. Some aaRSs bind specifically to mRNA and DNA For a number of years, a few aaRSs have been known to be involved in transcriptional and translational regulation [26]. For B. subtilis it is reported that expression of most aaRSs is regulated through a common mechanism involving transcriptional anti-termination where the principal effector is the uncharged tRNA [27, 28]. In E. coli, aaRS gene expression is regulated by a broad range of diverse mechanisms, and in certain cases, the aaRS plays a direct role in control. For example, E. coli ThrRS regulates its own expression at the translational level by binding to two hairpin domains of the mRNA operator which separately mimic the anticodon stem-loop of tRNAThr [29]. The enzyme represses its own translation by occluding ribosome binding in a steric hindranceguided process due to interspersed ribosome and ThrRS recognition sites [30]. E. coli AlaRS binds to a palindromic sequence in the promoter of its own gene, alaS, and represses transcription in vitro [31]. More recently, T. thermophilus PheRS has also been found to bind specifi-

685 cally to DNA, although the functional implications of this interaction remain undefined [32]. 6. LeuRS and TyrRS are essential to mitochondrial splicing While group I introns are autocatalytic and self-splice in vitro, additional protein factors are required for splicing activity in vivo. These proteins promote the folding of the intron RNA into the catalytically-active structure. Three mitochondrial aaRSs, TyrRSs from Neurospora crassa [33] and Podospora anserina [34] as well as LeuRS from S. cerevisiae [35, 36], facilitate group I intron RNA splicing [37, 38]. The role of LeuRS in yeast mitochondrial splicing appears to be quite distinct from TyrRS in N. crassa. Unlike the N. crassa TyrRS where an idiosyncratic domain contributes to splicing [39, 40], yeast LeuRS lacks obvious unusual insertion sequences that might be specialized to contribute to the splicing activity. Moreover, LeuRS-dependent splicing requires a second protein co-factor, a maturase [35, 41, 42] and its specificity is limited to two closely related introns, bI4 and aI4α. During the EMBO workshop, it was reported that human mitochondrial and Mycobacterium tuberculosis LeuRSs complement a yeast null strain lacking the endogenous mitochondrial LeuRS, (Houman et al., unpublished results). A C-terminal five-amino acid deletion and single site substitutions within a conserved CP-1 based region were previously shown to effect the splicing activity induced by yeast mitochondrial LeuRS [43]. Corresponding mutations in the M. tuberculosis LeuRS also decreased or abolished complementation of the yeast null strain. Thus, it appears that prokaryotic and eukaryotic sources of the leucine enzyme may contain universal determinants to facilitate essential mitochondrial splicing activities. Because of its lack of ubiquity, aaRS-dependent splicing has been proposed to be a relatively recent evolutionary event that occurred, for example, subsequent to the split of S. cerevisiae which has recruited LeuRS for group I intron splicing and N. crassa and P. anserina which rely on TyrRS [37, 38, 44]. Although it may be fortuitous, it is intriguing that both LeuRSs and TyrRSs recognize class II tRNAs which have extra long variable loops. Tertiary structure predictions indicate that group I introns contain at least remnants of a class II tRNA-like domain [44]. (Class II tRNAs include those tRNAs with long variable domains.) Alternative hypotheses suggest that either group I introns evolved from tRNAs or that group I introns and tRNAs may have evolved convergently to recruit tRNA synthetases for added splicing stability [37, 44]. One of the most important lessons from these examples of aaRS dual functionality, however, may be the interplay of protein and RNA structure-function roles within a cell while co-adapting to accommodate more than one essential activity.

686 7. A family of HisRS-like homologs influence bacterial histidine biosynthesis Some bacterial genomes (i.e., Bacillus subtilis, Synechocystis sp., and Lactococcus lactis) contain an open reading frame (HisZ, or Orf3 in L. lactis [45]) which resembles the catalytic domain of HisRS. It is intriguing that each of these HisZ homologs systematically lacks one or more essential catalytic residues rendering the purified protein inactive in catalyzing either ATP-PPi exchange or aminoacylation of tRNAHis (Francklyn, unpublished results). However, HisZ does appear to be involved in the biosynthesis of histidine. Upon disruption of hisZ, bacteria will not grow on minimal medium, unless supplemented with histidine or histidinol, the last intermediate in the biosynthesis of histidine (Delorme et al., unpublished results). Moreover, the hisZ gene correlates with the length of the hisG gene encoding ATP phosphoribosyltransferase (the first enzyme in the histidine biosynthesis pathway) such that when hisZ is present, the hisG gene product is about 100 amino acids shorter than that encoded by the hisG of bacterial species lacking hisZ. Although it is inactive in aminoacylation, purified HisZ binds tRNAs tightly with a dissociation constant of 100–200 nM, a property which is shared by the long hisG gene product from Salmonella typhimurium [46]. These combined results suggest that, when the hisG gene is short, the HisRS-like HisZ product can complement an incomplete ATP phosphoribosyltransferase (Sissler et al., unpublished results). Moreover, phylogenetic analysis indicates that the HisZ protein family is distinct from the functional HisRS enzymes, and likely diverged from them early in the evolution of the bacteria.

8. Design of unnatural aaRSs and new aminoacylation procedures In the past few years, in vitro selected RNA has been shown to catalyze a wide variety of activities. These include aminoacylation properties as pioneered by the laboratories of Szostak and Yarus [47-51]. Recent in vitro selection experiments detected an RNA that folds into a tRNA-like structure and can transfer amino acids from a short ribonucleotide to tRNA (Suga and Lee, unpublished results). These results, combined with a growing body of work, support models for the evolution of the genetic code based on primordial tRNA molecules [52]. Misaminoacylation of tRNAs with unnatural amino acids followed by site-specific incorporation of these amino acids into proteins is a challenging field which presents a wealth of opportunities and applications. It is now possible to chemically link protected amino acid analogues to tRNA. The novel charged tRNAs are em-

Martinis et al. ployed in in vitro protein synthesis systems to incorporate the amino acid into specific predetermined positions. Subsequent deblocking of the atypical amino acid yields the catalytically competent enzyme (Hecht, unpublished results). Another approach aims to allow site-specific incorporation of unnatural amino acids into proteins in vivo by creation of engineered aaRS/tRNA pairs specifically adapted to these amino acids. Based on the crystallographic structure of the glutamine specific complex, an ‘orthogonal’ suppressor tRNA has already been prepared [53, 54]. A mutant GlnRS has been evolved to functionally interact with the engineered tRNA and charge it with glutamine in vitro and in vivo. Experiments to select mutants of the novel engineered GlnRS/tRNA pair which accept unnatural amino acids are underway (Liu and Schultz, unpublished results). 9. Towards a complete set of aaRS crystal structures Extensive crystallographic studies of aaRSs have revealed the multi-domain modularity of these enzymes, including insertions and terminal modules appended to each class-specific active site [7, 8]. Subgroup partitions of the two classes of aaRSs, initially based on sequence analyses alone [55], have now been improved via X-ray crystallography to define common structural modules shared by several aaRSs [7, 8, 56]. New structures of aaRSs including yeast ArgRS complexed with arginine [57], T. thermophilus ArgRS (Shimada et al., unpublished results) and AsnRS [58], as well as four aaRSs complexed with cognate wild type or in vitro transcript tRNA including GluRS/tRNAGlu (Sekine et al., unpublished data), IleRS/tRNAIle (discussed in other sections below: Steitz, unpublished data), ProRS/tRNAPro ([59]; Cusack et al., unpublished data), and ThrRS/tRNAThr [60] were presented. A set of structures is shown in figure 2. This growing data base of structural information has yielded insights into enzymatic mechanisms and molecular interactions that facilitate cognate tRNA binding. Amino acid or aminoacyl-adenylate binding has now been examined in structural detail for several class I aaRSs (i.e., IleRS [61]; Steitz, unpublished data), ArgRS [57], GlnRS [62], TyrRS [63], TrpRS [64]) and nearly all class II aaRSs except AlaRS and PheRS [7, 8]. These unique snapshots of the first step of the aminoacylation reaction provide clues to: i) specific amino acid recognition; ii) conformational changes associated with amino acid and ATP binding; and iii) the mechanism of amino acid activation. In addition, catalytic and structural roles for three bound magnesium ions are now well established for several class II tRNA synthetases including SerRS, AsnRS, and AspRS [58, 65, 66]. Enzymes from both classes seem to use a lock and key mechanism for the side chain of the amino acid and an induced fit mechanism for other substrates including ATP,

Aminoacyl-tRNA synthetases

687

Figure 2. Structures of 13 aminoacyl-tRNA synthetases illustrating the modularity of these enzymes. The seven class I synthetases presented correspond to yeast ArgRS [57], Thermus termophilus MetRS (M. Konno, personal communication), T. thermophilus IleRS [61], T. thermophilus GluRS [146], E. coli GlnRS [150], Bacillus stearothermophilus TyrRS [63], and B. stearothermophilus TrpRS [64]. Their active site contains a canonical dinucleotide-binding fold shown in red. The six class II synthetases correspond to E. coli SerRS [56], E. coli ThrRS [60], E. coli HisRS [70], T. thermophilus GlyRS [69], yeast AspRS [151] and T. thermophilus AspRS [152]. Class II synthetases are built around a mixed β-sheet which contains three characteristic homologous motifs (shown in red). In both classes, various insertions and additional domains (shown in blue, green, orange and yellow) are appended to the active site. Most of these domains play a functional role either in tRNA recognition (anticodon binding modules shown in yellow, except for TyrRS, TrpRS, SerRS), or in proofreading mechanisms (so-called connecting peptide 1 (CP1) in IleRS; CP1 is colored in blue in the seven class I aaRSs). Only one monomer is shown for the six class II aaRSs, for TyrS and TrpRS, which are biologically active as dimers. Structures of yeast AspRS, E. coli GlnRS and E. coli ThrRS are those within complexes with cognate tRNAs.

688

Martinis et al.

Figure 2 (continued).

tRNA, and the peptidic moiety of the amino acid [7, 8, 67]. An exception is the class IIa enzyme HisRS (and to a lesser extent ProRS) where the histidine side-chain binding site is created by induced fit of a crucial active site loop which notably contains two conserved tyrosines interacting with the imidazole and a conserved catalytic arginine (Cusack et al., unpublished results). Such induced fit conformational changes can result in long-range coordination of multiple modules during catalysis in a more sophisticated way than was previously thought. This has been shown in a series of structures of HisRS and ProRS (Cusack et al., unpublished results) where the disposition

of tRNA binding elements depends on the contents of the active site (ATP, amino acid or aminoacyl-adenylate). In the class I dimeric TrpRS example, non-polar packing links the helical anticodon-binding domain to the highly conserved isoleucine from the Rossmann fold active site TIGN signature, coupling tRNA binding to active site chemistry (Carter Jr. et al., unpublished results). TrpRS assumes three distinct conformations during tryptophan activation, confirming in molecular detail the results of pre-steady state kinetic analyses of TyrRS active-site mutants [68]. Residues interacting with the nucleotide ribose in the closed substrate-bound conforma-

Aminoacyl-tRNA synthetases

689

Figure 2 (continued).

tion are widely separated in the ground-state enzyme, consistent with the observation that mutations of these residues exert their effect only in and following the

transition state. The first step of the ‘induced-fit’ TrpRS catalytic cycle is driven by ATP binding which suffices to close the active site forming a pre-transition state. A

690

Martinis et al.

Figure 2 (continued).

ternary product-analog:PPi:TrpRS crystal complex mimics the Michaelis complex for PPi-exchange and suggests

that transfer of the high-energy AMP-PPi bond to the tryptophan carboxylate stabilizes the product-complex

Aminoacyl-tRNA synthetases conformation, which binds the activated amino acid before its transfer to cognate tRNA. The induced-fit mechanism exploits conformational forces favoring transfer of the α-phosphoryl high-energy bond first from PPi (in ATP) to the amino acid and then from the activated amino acid to the cognate tRNA [67]. By implication, the related TyrRS and perhaps all class I aaRSs may have corresponding, yet to be identified, conformational states. Unexpected secondary and tertiary features illustrate additional strategies for specific substrate selection. For example, the architecture of ArgRS [57] is characterized by two nucleic acid binding modules including an anticodon binding domain with topology similar to that of MetRS and IleRS. The N-terminal region of the protein contains an RNA-binding module exhibiting overlapping topology with several other proteins including the S4 ribosomal protein. The class II ThrRS contains four domains [60]. In addition to its catalytic module and an anticodon binding domain common to the class IIa aaRSs except SerRS (i.e., ProRS [59], GlyRS [69], HisRS [70, 71], and ThrRS [72]), ThrRS has a novel N-terminal bi-modal domain. It also contains an unexpected catalytic zinc atom and, similar to other class II synthetases, interacts specifically with the tRNA acceptor stem from the major groove side. However, the unusual N-terminal module is also involved in acceptor stem recognition and interacts with the minor groove of the tRNA. A tRNA minor groove recognition module may also be present in three other class IIa synthetases including HisRS and GlyRS (based on structure superposition predictions) and possibly AlaRS as suggested by structure-based sequence analyses [60]. Recent structures of ProRS and ThrRS tRNA complexes have re-emphasized anticodon binding as a key mechanism for tRNA recognition and discrimination. Anticodon recognition modules range from exclusively α-helical or β-barrel to mixtures of secondary structures. ProRS and ThrRS possess the α/β fold class IIa anticodon binding domain. Nucleotides in positions 35 and 36 of the tRNA anticodon are major identity determinants of ProRS and ThrRS. However, ProRS only recognizes these two bases, while ThrRS interacts with all three bases of the anticodon of tRNAThr. RNA binding and recognition deforms the anticodon loop in each case, but a common pattern for base specific recognition has not yet been identified. Each new structure of aaRSs complexed with its cognate tRNA introduces novel and varied mechanisms through which bases are recognized by the protein. To understand the second step of the class II aminoacylation reaction at atomic resolution, aspartyl-adenylate was bound in situ to the co-crystal complex of E. coli AspRS and its cognate tRNA revealing its structural state just prior to transfer of the amino acid moiety to the tRNA (Moras et al., unpublished data). Comparison to the co-crystal structure of the heterologous complex between yeast tRNAAsp and E. coli AspRS reveals for the first time

691 a tRNA induced mechanism that dictates specificity of the reaction and underlines the active role of tRNA in the recognition process with key control played by the acceptor stem (Moras et al., unpublished data). Extensive structural information for the multiple AspRS examples has emphasized significant kingdom-dependent differences (eucarya, bacteria, and archaea) for each step of the aminoacylation reaction, thereby revealing the complexity of essential fine tuning at the molecular level for the aminoacylation reaction ( [66]; Moras et al., unpublished results). These species-specific distinctions may be effectively exploited for antibiotic development as described below [10, 11]. 10. Protein synthesis fidelity ensured by two aaRS catalytic sites The primary role of the aaRS is to covalently link the correct amino acid to its cognate tRNA. The aminoacylation reaction proceeds through a two-step mechanism via an aminoacyl adenylate intermediate. In 1958, Pauling first proposed that closely related, isosteric side chains of amino acids such as valine and isoleucine would be difficult for an enzyme to completely distinguish and predicted high error rates in protein synthesis [73]. Subsequently, Loftfield demonstrated the remarkable fidelity of isoleucine versus valine incorporation into proteins by measuring a low error rate of only 1 in 3000 [74, 75]. Baldwin and Berg determined that IleRS selectively hydrolyzes valyl adenylates when tRNAIle was added, rather than charging valine to the wrong tRNA [76]. This RNA-dependent enzymatic proofreading or editing mechanism was hypothesized by Fersht to be accomplished through a ‘double sieve’ [77-79], requiring two active sites for accurate molecular recognition. For example, Fersht predicted that IleRS would contain an activation site that would exclude amino acids larger than isoleucine. A second active site would select the activated isoleucyl adenylate based on size and hydrolyze smaller aminoacyl adenylates including the misactivated valyl adenylate, effectively editing the enzyme’s original mistake. Two decades since the double sieve model was proposed, X-ray crystallography and mutagenesis efforts have identified a specific editing active site, delineating a detailed molecular view of the first and second sieve [61, 80-82]. The ten class I aaRSs contain a Rossmann nucleotide binding fold [55, 83] which incorporates the first sieve and is responsible for amino acid recognition and activation as well as subsequent charging of tRNA. The Rossmann fold of the three class I aaRSs which recognize amino acids with aliphatic side chains, isoleucine, valine, and leucine, contain a large insertion of 270–300 amino acids called CP1 (connective polypeptide 1; [84, 85]). The editing activity of the isoleucine and

692 valine enzymes were localized by isolating recombinantly the CP1 domains of E. coli IleRS and B. stearothermophilus ValRS as fusion proteins [82]. In vitro kinetic assays demonstrated that each purified, stable CP1 domain specifically hydrolyzes the appropriate misactived adenylate. Nureki et al. solved the crystal structure of T. thermophilus IleRS and identified a bound valine in the Rossmann fold-based activation site and also in a cleft formed by two completely conserved peptides in the CP1 editing domain [61]. Mutational analysis established that specific conserved residues are important for editing [61, 86]. Notably, these homologous editing motifs are also present in the aligned CP1 domains of ValRS and LeuRS enzymes. Thus, the ancestors of these related enzymes which recognize aliphatic substrates appear to have acquired a common protein module capable of hydrolytic activity and then evolved to specifically recognize relevant misactivated amino acids. Editing has been shown to occur either pre- or posttransfer of the misactivated amino acid to the cognate tRNA. Both editing mechanisms are RNA-dependent. Systematic testing of chimeric tRNAs (tRNAIle and tRNAVal) identified the D-loop of tRNAIle as the specific RNA trigger stimulating valyladenylate hydrolysis by IleRS [87]. There are only three nucleotide differences between the D-loops of tRNAIle and tRNAVal including C →G and D (dihydrouridine)→G substitutions as well as a D insertion in the tRNAIle D-loop. A small RNA minihelix mimicking the tRNAIle D-loop failed to stimulate editing activity demonstrating that the D-loop based editing determinants required presentation in the context of the full-length tRNA [88]. Although the role of the D-loop in the editing activation of IleRS has been clearly established, it remains unclear how this outside corner of the tRNA interacts either directly or indirectly with the protein to stimulate editing of misactivated or mischarged amino acids. Modification of the terminal adenosine ribose of tRNAIle also appears to influence the proofreading reaction. While IleRS can transfer valine to a modified tRNA-CC(3’-deoxyA), it does not hydrolyze the resulting misaminoacylated tRNA [89]. It is possible that isomerization of the aminoacyl moiety from the 2’ to 3’ ribose position which is accompanied by a conformational change on the 3’ terminal ribose of the tRNA (Slosser et al., unpublished result) is required for editing. It has also been shown for other tRNAs that the terminal adenosine ribose undergoes a transition to the 2’-endo conformer upon aminoacylation which unstacks the adenine base (see below; [90]). Therefore, it is also possible that the ribose conformational change and/or migration of the linked amino acid may act as a switch to trigger proofreading by the CP1-based hydrolytic editing domain (Sprinzl, unpublished results).

Martinis et al. A co-crystal structure of Staphylococcus aureus IleRS complexed to E. coli tRNAIle was presented (Steitz, unpublished results). Interestingly, the 3’ end of the tRNA was in close proximity to the CP1 editing cleft, rather than the Rossman fold-based active site, illuminating a potential RNA-protein ‘editing complex’. Although electron density from the X-ray diffraction data only extended through the C74 base of tRNAIle, continuation of the α-helical terminus placed A76 near the critical amino acids of the hydrolytic editing cleft. One clear question that persists is the translocation of the misactivated amino acid or the 3’ end of the mischarged tRNA by over 25 Å from the synthetic active site to the CP1-based editing cleft. Steitz proposed that the tRNA shuttles its misacylated 3’ end between the editing cleft and Rossman fold analagous to the DNA polymerase mechanism of editing [91]. Thus, the second sieve near the hydrolytic site could sample the RNA-linked amino acid to determine if it is correct or mischarged. It remains unclear from both crystal structures ( [61, 86]; Steitz, unpublished results), however, how the misactivated valyl adenylate could be translocated to the editing site, and also the molecular role of the tRNA D-loop since it is distant from the RNAprotein interface.

11. Misacylated tRNAs are an essential prequisite for novel translation pathways The explosion of genomic information in the past few years has revealed that one or more of the 20 aaRSs found in the well characterized E. coli model are conspicuously absent in some organisms [6]. Yet their cognate amino acids are well represented in protein synthesis. Careful biochemical work has shown that asparagine and glutamine-charged tRNAs respectively require mischarged Asp-tRNAAsn [92, 93] and Glu-tRNAGln [92, 94] intermediates for some eubacteria, archae, and organelles. These mischarged tRNAs undergo a tRNA-dependent transamidation reaction to convert the linked amino acid prior to incorporation into protein. The three genes encoding the heterotrimeric tRNA-dependent amidotransferase (AdT) were cloned from both Bacillus subtilis [95] and Deinococcus radiodurans [96]. Interestingly, both proteins possess dual specificities and catalyze both AsntRNAAsp and Gln-tRNAGlu transamidase activities [96]. Sequence comparisons show that each of the three Glu/Asp-AdT subunits are distinct from other well characterized glutamine amidotransferases as well as GlnRSs and AsnRSs. N-terminal sequencing of a heterodimeric Asp-AdT purified from T. thermophilus [93] showed some homology to the β-subunit of B. subtilis Glu-AdT [95]. The other transamidase subunits were highly divergent and not similar to AdTs that lacked dependence on tRNA.

Aminoacyl-tRNA synthetases Some genomes have genes for both the tRNAdependent AdT and corresponding aaRS begging the question of functional duplication and the cell’s preference for aminoacylation pathways. Two excellent examples include D. radiodurans [96] and T. thermophilus [93, 97]. Both contain two distinct AspRSs, one which will only charge cognate tRNAAsp and the other which aspartylates both tRNAAsp and non-cognate tRNAAsn. T. thermophilus and D. radiodurans also have AsnRS and tRNA-dependent Asp-AdT as mentioned above, but notably, appear to lack a functional traditional pathway to synthesize asparagine. Thus, survival would require either import of asparagine or protein degradation to obtain the essential amino acid. Alternatively, it has also been proposed that the D. radiodurans tRNA-dependent transamidase pathway may actually serve as the asparagine prototroph’s sole route to asparagine synthesis, perhaps as a remnant from the RNA world [96]. The complete and novel amino acid biosynthetic pathway would require a yet to be identified tRNA hydrolase enzyme to release the asparagine product. Another proposal suggests that T. thermophilus elegantly balances its apparently redundant pathways by utilizing the more efficient direct AsnRS aminoacylation route when asparagine is present, yet in its absence the cell resorts to the indirect, transamidase pathway to produce Asn-tRNAAsn [93]. AdTs’ important role in protein synthesis builds upon a growing list of other essential enzymes, which carry out tRNA-dependent modifications of amino acids [6]. For example, fMet-tRNAfMet is required for the initiation of protein synthesis and is synthesized by Met-tRNA transformylase which formylates Met-tRNAfMet [98, 99]. Also, the rare amino acid selenocysteine is generated by selenocysteine synthase which modifies serine mischarged to tRNASec in a pyridoxyl-dependent reaction [100]. Although most methanogenic archae genomic sequences do contain an open reading frame encoding CysRS, Methanococcus jannaschii [101] and Methanobacterium thermoautotrophicum [102] lack a canonical sequence for CysRS. Thus, an alternate route to charging cysteine tRNA has been suggested such as thiolation of serine mischarged to tRNACys similar to the selenocysteine transformation [6, 103]. While purified native M. thermoautotrophicum SerRS could aminoacylate its cognate tRNASer, it did not misacylate tRNACys [103]. However, recombinant M. jannaschii SerRS can mischarge serine to a T7 transcript of M. jannaschii tRNACys (Saks, unpublished results), indicating that this aaRS has a relaxed tRNA specificity. Although neither the transformation pathway, nor enzyme(s) required for converting SertRNACys to Cys-tRNACys have been identified, it is possible that SerRS may play pivotal roles for the incorporation of serine, selenocysteine, and cysteine in some organisms during protein synthesis.

693 12. The acceptor termini of aminoacylated tRNAs act as identity elements for interaction with proteins The dependence of protein synthesis on mischarged tRNAs demands another level of biological scrutiny to prevent misincorporation of amino acids during translation. Binding and RNA-protection studies have clearly shown that T. thermophilus EF-Tu exhibits no affinity for the mischarged Asp-tRNAAsn [93]. In contrast, tRNAdependent amidation yielding correctly charged AsntRNAAsn binds well to EF-Tu. In a separate example, EF-Tu isolated from chloroplasts, which are dependent on Glu-AdT to convert Glu-tRNAGln, binds poorly to the mischarged tRNA [104]. Interestingly, EF-Tu from E. coli, which lacks a tRNA-dependent AdT pathway, essentially binds correctly and incorrectly charged chloroplast tRNAGln with similar affinities. Therefore, EF-Tu not only shuttles charged tRNA from the aaRS or AdT to the ribosome, but, when required by the cell, can also enhance fidelity by introducing an additional proofreading step based on selective binding to correctly charged tRNAs. Aminoacylation of a tRNA results in a conformational change of its 3’-terminal adenosine as demonstrated by multiple biophysical studies. These include: i) the determination of the NMR and X-ray crystal structures of tRNA terminus model compounds such as 3’-Oanthraniloyladenosine (3’-ant-Ado), 2’(3’)-Ado and 2’(3’)-ant-AMP [90]; (ii) fluorescence studies of a tRNA in which the 3’ adenosine residue was substituted by formycin (Sprinzl, unpublished results); and iii) determination of the crystallographic structure of the complex between Met-tRNAfMet formyltransferase and its fMettRNAfMet product [105]. According to the experimental data, the ribose of the 3’-terminal adenosine of an aminoacylated tRNA adopts the 2’-endo conformation and the terminal adenosine is destacked. Such a conformation at the 3’-end of tRNA could be recognized as a structural determinant by the various enzymes, protein factors and ribosomal proteins which are required to distinguish aminoacylated tRNAs from the pool of uncharged tRNAs [90]. The 5’-terminal phosphate of tRNA also plays a role in the recognition of aminoacylated tRNAs. Thus, peptidyltRNA hydrolase (PTH), the enzyme which recycles the products of premature termination of translation, requires the presence of the 5’-phosphate group to correctly select N-blocked aminoacylated elongator tRNAs and reject formylated aminoacylated initiator tRNA [106]. Because the environments of the 5’-phosphate are similar in the PTH-tRNA [106] and EF-Tu-tRNA [107] complexes, it was proposed that selection of elongator tRNAs by EF-Tu also involves the 5’-phosphate. 13. aaRSs are highly adapted to charge specific tRNA The primary function of aaRSs is to aminoacylate their cognate tRNAs in a very specific way and to charge them

694

Martinis et al.

Figure 3. Distribution of identity elements within the tRNA 3D structure for aminoacylation by the 10 E. coli synthetases from class I (a) and the 10 from class II (b). The size of the spheres is proportional to the frequency of identity nucleotides at a given position. In b the fully colored variable domain indicates its participation in Ser identity. Adapted from [109] with permission of Oxford University Press.

with the corresponding amino acid. Collective efforts to decipher the tRNA molecular signals that trigger specific aminoacylation have been successful (reviewed in [9]). Currently, tRNA identity sets recognized by all of the E. coli aaRSs, 14 yeast aaRSs and a couple of other enzymes from higher eukaryotes and organelles are known, providing an overall understanding of aaRS recognition mechanisms. In general, identity elements are limited in number and located within the two distal extremities of the tRNA, namely the anticodon loop and the amino acid accepting stem (figure 3). In a few exceptions, the major determinants reside in the central core of the molecules. Most identity elements are in direct contact with the aaRS and, in several systems, specific chemical groups of bases or riboses involved in the functional interactions with the synthetase have been identified. The importance of structural features as identity signals has also been highlighted. Identity elements can be of different strength (major and minor elements), act either independently or in a cooperative manner, and contribute mostly to the kinetic steps of the aminoacylation rather than to binding. Interestingly, antideterminants further enhance specificity by hindering false recognition of tRNAs by non-cognate aaRSs. Recent developments have highlighted the existence and importance of cryptic [108] and permissive elements [109, 110] as well as that of alternative identity sets for the same specificity [111]. One unique system concerns the identity of lysine-specific tRNAs which are recognized either by a class II LysRS (in most bacteria and all eukaryotes) or by a class I LysRS (in some archae and bacteria) [6, 112-114]. The unusual class I LysRSs from

Borrelia burgdorferi and Methanococcus maripaludis rescue an E. coli strain deficient in class II LysRS, indicating their ability to substitute for the more common class II enzyme. It turns out that the two types of synthetases, those of class I as well as class II, recognize the same tRNA identity elements, specifically the discriminator base and anticodon nucleotides. The spirochete tRNALys contains a G2:U71 wobble pair which is an antideterminant for class II LysRS but does not affect class I enzyme recognition, suggesting that class I and class II LysRSs approach and interact with the acceptor stem helix from the major and minor groove sides, respectively. Several aaRS extend their aminoacylation properties to tRNA-like substrates with non-canonical structures. These include the 3’-ends of a series of viral genomes (e.g. [115-118]) and the more recently discovered tmRNA acting both as a transfer RNA and a messenger RNA [119-121]. Specific aminoacylation of alanine tRNAs is dependent on synthetase recognition of particular acceptor stem nucleotides, especially a G3-U70 base-pair and A73 discriminator nucleotide [122-124]. These elements are conserved in all known tRNAAla sequences from prokaryotes, archae, eukaryotes (cytoplasmic), chloroplasts, and mitochondria of plants and single-cell organisms [125]. Remarkably in contrast, virtually all known tRNAAla sequences from animal mitochondria [125] lack a G3:U70 base pair [126]. In some cases, including insects, birds, and some mammals (including man), another G-U base pair is found in a nearby location in the acceptor helix, suggesting that the interaction between AlaRS and the acceptor stem may be translocated (Chihade and Schim-

Aminoacyl-tRNA synthetases mel, unpublished observations). However, in other examples (e.g., X. laevis and Asterina pectinifera, a starfish), the mitochondrial tRNAAla acceptor stem completely lacks a G:U base pair. Identity element evolution has also occurred for tRNACys. E. coli CysRS is unable to recognize efficiently yeast or human tRNACys [127]. By successively modifying yeast and human tRNACys until cross-species aminoacylation by E. coli CysRS was achieved, elements important for aminoacylation, but divergent during evolution, have been identified. New aspects on tRNA identity establish the evolutionary relationships between the identity elements for a given specificity. The positional location of determinants are often conserved amongst species, but the mechanisms by which they are expressed may vary [128]. Alternatively, identity elements may partially vary along the evolutionary scale such as in the case of species-specific differences for proline identity [129]. While tRNA anticodon recognition has been maintained through evolution, significant changes in acceptor stem recognition have occurred. Thus, all tRNAPro molecules from bacteria strictly conserve A73 and C1-G72, whereas all available cytoplasmic eukaryotic tRNAPro sequences have a C73 and a G1-C72 base-pair. In contrast to the E. coli enzyme which uses these acceptor stem elements as major recognition determinants, the human synthetase does not, but rather refers to the anticodon domain. Evolution of identity sets has been directly linked to the co-evolution of cognate aaRSs. Analysis of ProRS sequences covering all three taxonomic domains (bacteria, eucarya and archaea) reveal that the sequences are divided into two distant groups [129]. E. coli ProRS resides in one group and human ProRS belongs to the other. Interestingly, the motif 2 loop differs between the two groups, but is well conserved within each group. Based on cysteinescanning mutagenesis of the motif 2 loop of E. coli ProRS, as well as cross-linking experiments, it has been proposed that an arginine residue of this loop interacts with the G72 base (Yang et al., unpublished results) which is a major determinant for the E. coli enzyme, but not the human aaRS. These results suggest that a co-adaptation of the synthetase-tRNA contacts may have occurred during evolution, leading to species-specific differences in the tRNA recognition mechanisms. Moreover, this and other work identifying species-specific recognition by aaRSs [128, 130-133] validate the aaRSs as an important pharmaceutical target for antibiotic development [10, 11]. 14. Inhibitors targeting aaRSs may lead to novel antibiotics The commercially available antibiotic mupirocin, also known as pseudomonic acid, specifically targets IleRS in Gram-positive bacteria including S. aureus. Two separately solved crystal structures of IleRS complexed with

695 mupirocin, shows the drug molecule bound to the class I aaRS active site (Steitz, unpublished results; Nakama et al., unpublished results). Pharmaceutical efforts to identify other inhibitors of IleRS led to the design of very efficient new molecules [11]. Two classes of IleRS inhibitors have been identified: i) analogs of the aminoacyl-adenylate, such as the alkyl-adenylate, Ile-ol-AMP; and ii) analogs of the naturally occurring pseudomonic acid. Although each class of compounds competitively inhibits isoleucine and ATP binding, their mechanisms of interaction are mutually exclusive. Spectrofluorimetric techniques detected differences in each class’s modes of binding to IleRS [134, 135]. Novel chimeric compounds based on the two classes of inhibitors were designed to improve efficiency of the antibiotics. One particular synthetic chimeric compound exhibited an exceptional affinity for IleRS (KD < 10 fM), comparable to that of the most potent reversible enzyme inhibitors (Pope, unpublished results). Structural work with other aaRSs bound to inhibitors has also yielded valuable information that may be useful to antibiotic discovery and development. For example, the closed, ternary, pre-transition state TrpRS complex (see above) was first observed with the bacteriospecific TrpRS inhibitor, indolmycin and ATP. The affinity of TrpRS for indolmycin is strongly dependent on ATP, suggesting that the inhibited form is actually this ternary complex, which therefore probably resembles the transition state complex. Thus, development of these high affinity leads for antiinfective drug development can exploit knowledge of transition state conformations (Carter Jr., unpublished results). Other pharmaceutical efforts have selected peptides from combinatorial phage display libraries that bind specifically to a number of aaRSs. For example, one inhibitory peptide called Pro-3 binds to E. coli ProRS (Paige et al., unpublished results). The KI’s were measured for inhibition of ProRS aminoacylation activity and determined to be 0.5 µM and 0.4 µM for proline and ATP binding, respectively (Tao et al., unpublished results). Pro-3 was expressed intracellularly via an inducible expression system and shown to inhibit E. coli growth in rich media. Induction of the Pro-3 peptide was also capable of providing protection in an E. coli lethal infection model providing a novel technology to intracellularly test antibiotic targets of known and unknown function for drug discovery (Tao et al., unpublished results). Peptides were also isolated from phage display libraries that bind with nanomolar affinity to Haemophilus influenzae TyrRS (Paige et al., unpublished results). A fluorescence polarization assay was used to detect the interaction of these peptides with TyrRS. Competition of the aaRS-binding peptide with a small molecule can be detected with high sensitivity using this assay. In general, these small molecule inhibitors can potentially be used as lead compounds to develop antibiotics targeting aaRSs [11].

696 15. Evolution of the aaRSs and implications to the origins of the genetic code The aaRSs have often been experimentally recruited as an evolutionary model for transitions within and between the three kingdoms [136, 137]. Expanded cellular aaRS roles (figure 1) as well as the complete absence of specific aaRSs [6] within certain species have also invited questions and comments to understand the evolution of essential biochemical pathways and mechanisms. Speciesspecific sequence and biochemical differences have suppported recent horizontal gene transfer events [137]. For example, the molecular basis of certain types of bacterial resistance to mupirocin (discussed above), a bacterial IleRS-specific inhibitor, is due to the presence of an ‘eukaroytic-like’ IleRS [138, 139]. In S. aureus, the eukaryotic-like IleRS gene is borne on an extrachromosomal plasmid and is suggested to have been acquired by horizontal transfer from eukaryotes after the divergence of eucarya and archaea [140]. Interestingly, in other species, the eucaryotic-like gene appears to be integrated into the main chromosome. The emergence and evolution of the Glx RS family is particularly intriguing because of the apparent lack of GlnRS in archaea (as well as in some bacteria and organelles) and also the existence of an alternative transamidase pathway in some organisms which rely on a non-discriminatory GluRS to misacylate tRNAGln (discussed above) [6, 137, 141-143]. It has been proposed that GlnRS arose via duplication of the GluRS gene in eukaryotes and was subsequently transferred to bacteria [141]. GlnRS genes have been identified in lower eukaryotes showing that it existed deep in the branches of the eucarya kingdom [143]. In addition, phylogenetic analyses of the catalytic Rossmann fold core demonstrate that archaea GluRS are more closely related to both GlnRS and GluRS from eucarya than to bacteria GluRS [142, 143]. Interestingly, the substitution of just two amino acids in human GlnRS confers glutamic acid aminoacylation to yeast tRNA [144]. The combined results from these studies suggest that the core GluRS existed in the last common ancestor which relied on the transamidation pathway to produce Gln-tRNAGln. Moreover, since the anticodon-binding modules of GlnRS [145] and GluRS [146] are completely distinct, the C-terminal domain would have been appended subsequent to the bacteria and archaea/eucarya split [142]. The remarkable discovery of two completely distinct class I and II LysRSs has also opened new evolutionary windows [113]. Representatives of both class I and II LysRSs have been found in bacteria, albeit not simultaneously in the same organism. All known eukaryotic LysRSs are class II, while most of the examples of archaea LysRSs are class I. Although lateral gene transfers between kingdoms could account for the presence of two completely unrelated LysRSs in bacteria, there are no

Martinis et al. apparent remnants of displaced genes. Phylogenetic analysis suggests that it is possible that establishment of the two classes of LysRSs occurred prior to separation of the three kingdoms [147, 148]. Interestingly, the evolutionary distinctions of the class I and class II LysRSs are not mirrored by their cognate extant tRNAs which are phylogenetically quite similar. This unique model provided the first opportunity to substantiate existence of a tRNA and its identity elements prior to the emergence of its cognate aaRS [147, 148]. It is also consistent with the origins of the genetic code in the context of tRNA, perhaps in an RNA world. Evolution of the aaRSs has resulted in exceptional specificity which is critical to accurate translation of the genetic code and the fidelity of protein synthesis. Moreover, existing aaRS binding sites, particularly for the tRNA, have also adapted to aid critical alternate cellular functions, such as in group I intron RNA splicing by N. crassa Tyr RS and S. cerevisiae LeuRS [37, 38] and translational regulation by E. coli ThrRS [29, 30]. Maizels and Weiner also point out that dual cellular roles exhibited by some tRNA synthetases are excellent examples of ‘exaptation’, where a molecule with one function is recruited to also carry out an entirely different activity [149]. Human TyrRS not only has an appended EMAPII-like domain with cytokine-like activity, but remarkably, the catalytic core for aminoacylation also appears to function as a specific cytokine [24]. Interestingly the cell has safeguarded itself from potential havoc, similar to other important cascade mechanisms, by requiring proteolytic cleavage [24] to release the cytokine-like activities of the two exapted domains [149]. Other EMAPII-like domains are fused to or in association with additional aaRSs in lower and higher eukaryotes as well as prokaryotes suggesting further novel functions that extend beyond protein synthesis [19, 22, 23, 149]. It is also intriguing to note that many of the aaRSs contain insertions and appendages with as of yet unassigned responsibilities.

16. Conclusion A wealth of information contributed by many laboratories from diverse disciplines has yielded exquisite insight into the aaRSs’ interactions with substrates and catalytic mechanisms. We continue to elaborate our understanding of this remarkable family of enzymes’ adaptation and exaptation to varied cellular environments and evolutionary pressures and are beginning to appreciate the aaRSs expanded roles within the cell. These fascinating new roles reflect a vast network of functional activities that extend far beyond the protein synthesis aminoacylation reaction. Full realization of the aaRSs’ diverse and essential functional roles will not only present evolution-

Aminoacyl-tRNA synthetases ary models for the development of the cell, but provide enormous opportunities for pharmaceutical research and advances. Acknowledgments We thank Drs. Sylvain Blanquet, Richard Giegé, and Paul Schimmel for their comments on the manuscript. We are also indebted to the many investigators who reviewed specific sections citing their work and also those who graciously permitted citation of their unpublished data. We apologize to the numerous participants at the EMBO workshop whose excellent work we were unable to reference due to space constraints. Finally, we gratefully acknowledge main financial support to the meeting from EMBO and NSF.

References [1] Hoagland M., Stephenson M., Scott J., Hecht L., Zamecnik P., A soluble ribonucleic acid intermediate in protein synthesis, J. Biol. Chem. 231 (1958) 241–256. [2] Hoagland M., Biochemistry or molecular biology? The discovery of ‘soluble RNA’, Trends Biochem. Sci. 21 (1996) 77–80. [3] Martinis S.A., Plateau P., Cavarelli J., Florentz C., AminoacyltRNA synthetase: A family of expanding functions, EMBO J. (1999), in press. [4] Carter Jr. C.W., Cognition, mechanism, and evolutionary relationships in aminoacyl-tRNA synthetases, Annu. Rev. Biochem. 62 (1993) 715–748. [5] Schimmel P.R., Söll D., Aminoacyl-tRNA synthetases: General features and recognition of transfer RNAs, Annu. Rev. Biochem. 48 (1979) 601–648. [6] Ibba M., Curnow A.W., Söll D., Aminoacyl-tRNA synthesis: divergent routes to a common goal, Trends Biochem. Sci. 22 (1997) 39–42. [7] Arnez J.G., Moras D., Structural and functional considerations of the aminoacylation reaction, Trends Biochem. Sci. 22 (1997) 211–216. [8] Cusack S., Aminoacyl-tRNA synthetases, Curr. Opin. Struct. Biol. 7 (1997) 881–889. [9] Giegé R., Sissler M., Florentz C., Universal rules and idiosyncratic features in tRNA identity, Nucleic Acids Res. 26 (1998) 5017–5035. [10] Francklyn C., Musier-Forsyth K., Martinis S.A., AminoacyltRNA synthetases in biology and disease: New evidence for structural and functional diversity in an ancient family of enzymes, RNA 3 (1997) 954–960. [11] Schimmel P., Tao J., Hill J., Aminoacyl tRNA synthetases as targets for new anti-infectives, FASEB J. 12 (1998) 1599–1609. [12] Wolin S.L., Matera A.G., The trials and travels of tRNA, Genes Dev. 13 (1999) 1–10. [13] Simos G., Hurt E., tRNA aminoacylation: a visa to leave the nucleus?, Curr. Biol. 1 (1999) 235–242. [14] Hopper A., Nuclear functions charge ahead, Science 282 (1998) 2082–2085. [15] Arts G.J., Kuersten S., Romby P., Ehresmann B., Mattaj I.W., The role of exportin-t in selective nuclear export of mature tRNAs, EMBO J. 17 (1998) 7430–7441. [16] Kutay U., Lipowsky G., Izaurralde E., Bischoff F.R., Schwarzmaier P., Hartmann E., Gorlich D., Identification of a tRNAspecific nuclear export receptor, Mol. Cell 1 (1998) 359–369. [17] Lund E., Dahlberg J.E., Proofreading and aminoacylation of tRNAs before export from the nucleus, Science 282 (1998) 2082–2085.

697 [18] Schimmel P., Wang C.C., Getting tRNA synthetases into the nucleus, Trends Biochem. Sci. 24 (1999) 127–128. [19] Simos G., Segref A., Fasiolo F., Hellmuth K., Shevchenko A., Mann M., Hurt E.C., The yeast protein Arc1p binds to tRNA and functions as a cofactor for the methionyl and glutamyl-tRNA synthetases, EMBO J. 15 (1996) 5437–5448. [20] Simos G., Sauer A., Fasiolo F., Hurt E.C., A conserved domain within Arc1p delivers tRNA to aminoacyl-tRNA synthetases, Mol. Cell 1 (1998) 235–242. [21] Hellmuth K., Lau D.M., Bischoff F.R., Kunzler M., Hurt E., Simos G., Yeast Los1p has properties of an exportin-like nucleocytoplasmic transport factor for tRNA, Mol. Cell Biol. 18 (1998) 6374–6386. [22] Kleeman T.A., Wei D., Simpson K.L., First E.A., Human tyrosyl-tRNA synthetase shares amino acid sequence homology with a putative cytokine, J. Biol. Chem. 272 (1997) 14420–14425. [23] Quevillon S., Agou F., Robinson J.C., Mirande M., The p43 component of the mammalian multi-synthetase complex is likely to be the precursor of the endothelial monocyte-activating polypeptide II cytokine, J. Biol. Chem. 272 (1997) 32573–32579. [24] Wakasugi K., Schimmel P., Two distinct cytokines released from a human aminoacyl-tRNA synthetase, Science 284 (1999) 147–151. [25] Kornelyuk A., Tas M.P.R., Dubrovsky A., Murray J.C., Cytokine activity of the non-catalytic EMAP-2-like domain of mammalian tyrosyl-tRNA synthetase, Biopolymery i kleta (Kiev) 15 (1999) 168–172. [26] Putzer H., Grunberg-Manago M., Springer M., Bacterial aminoacyl-tRNA synthetases: genes and regulation of expression, in: Söll D., RajBhandary U. (Eds.), tRNA: Structure, Biosynthesis, and Function, ASM Press, Washington, DC, 1995, pp. 293–333. [27] Henkin T.M., tRNA-directed transcription antitermination, Mol. Microbiol. 13 (1994) 381–387. [28] Condon C., Grunberg-Manago M., Putzer H., Aminoacyl-tRNA synthetase gene regulation in Bacillus subtilis, Biochimie 78 (1996) 381–389. [29] Romby P., Caillet J., Ebel C., Sacerdot C., Graffe M., Eyermann F., Brunel C., Moine H., Ehresmann C., Ehresmann B., Springer M., The expression of E. coli threonyl-tRNA synthetase is regulated at the translational level by symmetrical operatorrepressor interactions, EMBO J. 15 (1996) 5976–5987. [30] Sacerdot C., Caillet J., Graffe M., Eyermann F., Ehresmann B., Ehresmann C., Springer M., Romby P., The Escherichia coli threonyl-tRNA synthetase gene contains a split ribosomal binding site interrupted by a hairpin structure that is essential for autoregulation, Mol. Microbiol. 29 (1998) 1077–1090. [31] Putney S.D., Schimmel P., An aminoacyl tRNA synthetase binds to a specific DNA sequence and regulates its gene transcription, Nature 306 (1981) 441–447. [32] Lechler A., Kreutzer R., The phenylalanyl-tRNA synthetase specifically binds DNA, J. Mol. Biol. 278 (1998) 897–901. [33] Akins R.A., Lambowitz A.M., A protein required for splicing group I introns in Neurospora mitochondria is mitochondrial tyrosyl-tRNA synthetase or a derivative thereof, Cell 50 (1987) 331–345. [34] Kamper U., Kuck U., Cherniack A.D., Lambowitz A.M., The mitochondrial tyrosyl-tRNA synthetase of Podospora anserina is a bifunctional enzyme active in protein synthesis and RNA splicing, Mol. Cell Biol. 12 (1992) 499–511. [35] Labouesse M., The yeast mitochondrial leucyl-tRNA synthetase is a splicing factor for the excision of several group I introns, Mol. Gen. Genet. 224 (1990) 209–221. [36] Herbert C.J., Labouesse M., Dujardin G., Slonimski P.P., The NAM2 proteins from S. cerevisiae and S. douglasii are mitochondrial leucyl-tRNA synthetases and are involved in mRNA splicing, EMBO J. 7 (1988) 473–483.

698 [37] Lambowitz A.M., Caprara M.G., Zimmerly S., Perlman P.S., Group I and group II ribozymes as RNPs: Clues to the past and guides to the future, in: Gesteland R.F., Cech T.R., Atkins J.F. (Eds.), The RNA World, 2nd Edition, Cold Spring Harbor Laboratory Press, New York, 1999, pp. 451–485. [38] Dujardin G., Herbert C.J., Aminoacyl tRNA synthetases involved in group I intron splicing, in: Green R., Schroeder R. (Eds.), Ribosomal RNA and group I introns, Landes Bioscience, Austin, TX, 1997, pp. 179–198. [39] Cherniak A.D., Garriga G., Kittlejr J.D., Akins R.A., Lambowitz A.M., Function of Neurospora mitochondrial tyrosyl-tRNA synthetase in RNA splicing requires an idiosyncratic domain not found in other synthetases, Cell 62 (1990) 745–755. [40] Kittlejr J.D., Mohr G., Gianelos J.A., Wang H., Lambowitz A.M., The Neurospora mitochondrial tyrosyl-tRNA synthetase is sufficient for group I intron splicing in vitro and uses the carboxyterminal tRNA-binding domain along with other regions, Genes Dev. 5 (1991) 1009–1021. [41] Labouesse M., Netter P., Schroeder R., Molecular basis of the ‘box effect’: A maturase deficiency leading to the absence of splicing of two introns located in two split genes of yeast mitochondrial DNA, Eur. J. Biochem. 144 (1984) 85–93. [42] De LaSalle H., Jacq C., Slonimski P.P., Critical sequences within mitochondrial introns: Pleiotrophic mRNA maturase and cisacting signals of the box intron controlling reductase and oxidase, Cell 28 (1982) 721–732. [43] Li G.Y., Becam A.M., Slonimski P.P., Herbert C.J., In vitro mutagenesis of the mitochondrial leucyl-tRNA synthetase of Saccharomyces cerevisiae shows that the suppressor activity of the mutant proteins is related to the splicing function of the wild type protein, Mol. Gen. Genet. 252 (1996) 667–675. [44] Caprara M.G., Lehnert V., Lambowitz A.M., Westhof E., A tyrosyl-tRNA synthetase recognizes a conserved tRNA-like structural motif in the group I intron catalytic core, Cell 87 (1996) 1135–1145. [45] Delorme C., Ehrlich S.D., Renault P., Histidine biosynthesis genes in Lactococcus lactis subsp. lactis, J. Bacteriol. 174 (1992) 6571–6579. [46] Blasi F., Barton R.W., Kovach J.S., Goldberger R.F., Interaction between the first enzyme for histidine biosynthesis and histidyl transfer ribonucleic acid, J. Bacteriol. 106 (1971) 508–513. [47] Suga H., Lohse P.A., Szostak J.W., Structural and kinetic characterization of an acyl transferase ribozyme, J. Am. Chem. Soc. 120 (1998) 1151–1156. [48] Suga H., Cowan J.A., Szostak J.W., Unusual metal ion catalysis in an acyl-transferase ribozyme, Biochemistry 37 (1998) 10118–10125. [49] Lohse P.A., Szostak J.W., Ribozyme-catalysed amino-acid transfer reactions, Nature 381 (1996) 442–444. [50] Yarus M., Amino acids as RNA ligands: a direct-RNA-template theory for the code’s origin, J. Mol. Evol. 47 (1998) 109–117. [51] Illangasekare M., Kovalchuke O., Yarus M., Essential structures of a self-aminoacylating RNA, J. Mol. Biol. 274 (1997) 519–529. [52] Schimmel P., Alexander R., Diverse RNA substrates for aminoacylation: clues to origins?, Proc. Natl. Acad. Sci. USA 95 (1998) 10351–10353. [53] Liu D.R., Magliery T.J., Schultz P.G., Characterization of an ‘orthogonal’suppressor tRNA derived from E. coli tRNA2(Gln), Chem. Biol. 4 (1997) 685–691. [54] Liu D.R., Magliery T.J., Pastrnak M., Schultz P.G., Engineering a tRNA and aminoacyl-tRNA synthetase for the site-specific incorporation of unnatural amino acids into proteins in vivo, Proc. Natl. Acad. Sci. USA 94 (1997) 10092–10097. [55] Eriani G., Delarue M., Poch O., Gangloff J., Moras D., Partition of tRNA synthetases into two classes based on mutually exclusive sets of sequence motifs, Nature 347 (1990) 203–206. [56] Cusack S., Berthet-Colominas C., Härtlein M., Nassar N., Leberman R., A second class of synthetase structure revealed by x-ray analysis of Escherichia coli seryl-tRNA synthetase at 2.4 Å, Nature 347 (1990) 249–255.

Martinis et al. [57] Cavarelli J., Delagoutte B., Eriani G., Gangloff J., Moras D., L-arginine recognition by yeast arginyl-tRNA synthetase, EMBO J. 17 (1998) 5438–5448. [58] Berthet-Colominas C., Seignovert L., Hartlein M., Grotli M., Cusack S., Leberman R., The crystal structure of asparaginyltRNA synthetase from Thermus thermophilus and its complexes with ATP and asparaginyl-adenylate: the mechanism of discrimination between asparagine and aspartic acid, EMBO J. 17 (1998) 2947–2960. [59] Cusack S., Yaremchuk A., Krikliviy I., Tukalo M., tRNA (Pro) anticodon recognition by Thermus thermophilus prolyl-tRNA synthetase, Structure 6 (1998) 101–108. [60] Sankaranarayanan R., Dock-Bregeon A.C., Romby P., Caillet J., Springer M., Rees B., Ehresmann C., Ehresmann B., Moras D., The structure of E. coli threonyl-tRNA synthetase complexed with tRNAThr enlightens its translational repressor activity and reveals an essential catalytic zinc, Cell 97 (1999) 1–20. [61] Nureki O., Vassylyev D.G., Tateno M., Shimada A., Nakama T., Fukai S., Konno M., Hendrickson T.L., Schimmel P., Yokoyama S., Enzyme structure with two catalytic sites for double-sieve selection of substrate, Science 280 (1998) 578–582. [62] Rath V.L., Silvian L.F., Beijer B., Sproat B.S., Steitz T.A., How glutaminyl-tRNA synthetase selects glutamine,Structure 6 (1998) 439–449. [63] Brick P., Bhat T.N., Blow D.M., Structure of tyrosyl-tRNA synthetase refined at 2.3 Å resolution. Interaction of the enzyme with the tyrosyl adenylate intermediate, J. Mol. Biol. 208 (1988) 83–98. [64] Doublié S., Bricogne G., Gilmore C., Carter Jr. C.W., Tryptophanyl-tRNA synthetase crystal structure reveals an unexpected homology to tyrosyl-tRNA synthetase, Structure 3 (1995) 17–31. [65] Belrhali H., Yaremchuk A., Tukalo M., Larsen K., BerthetColominas C., Leberman R., Beijer B., Sproat B., Als-Nielsen J., Grübel G., Legrand J.F., Lehmann M., Cusack S., Crystal structures at 2.5 Å resolution of seryl-tRNA synthetase complexed with two analogs of seryl adenylate, Science 263 (1994) 1432–1436. [66] Schmitt E., Moulinier L., Fujiwara S., Imanaka T., Thierry J.C., Moras D., Crystal structure of aspartyl-tRNA synthetase from Pyrococcus kodakaraensis KOD: archaeon specificity and catalytic mechanism of adenylate formation, EMBO J. 17 (1998) 5227–5237. [67] Carter Jr. C.W., Ilyin V.A., Huang X., Yin Y., Retailleau P., Vachette P., Longo A., Hu M., TRPRS: Three-state behavior, induced fit, conformational coupling, and free-energy transduction, Acta Crystallographica A55 (1999), in press. [68] Fersht A.R., Dissection of the structure and activity of the tyrosyl-tRNA synthetase by site-directed mutagenesis, Biochemistry 26 (1987) 8031–8037. [69] Logan D.T., Mazauric M.H., Kern D., Moras D., Crystal structure of glycyl-tRNA synthetase from Thermus thermophilus, EMBO J. 14 (1995) 4156–4167. [70] Arnez J.G., Harris D.C., Mitschler A., Rees B., Francklyn C.S., Moras D., Crystal structure of histidyl-tRNA synthetase from Escherichia coli complexed with histidyl-adenylate, EMBO J. 14 (1995) 4143–4155. [71] Åberg A., Yaremchuk A., Tukalo M., Rasmussen B., Cusack S., Crystal structure analysis of the activation of histidine by T. thermophilus histidyl-tRNA synthetase, Biochemistry 36 (1997) 3084–3094. [72] Cura V., Kern D., Mitschler A., Moras D., Crystallization of threonyl-tRNA synthetase from Thermus thermophilus and preliminary crystallographic data, FEBS Lett. 374 (1995) 110–112. [73] Pauling L., The probability of errors in the process of synthesis of protein molecules, Festschrift Arthur Stoll, Birkhauser Verlag, Basel, 1958, p. 597. [74] Loftfield R.B., Vanderjagt D., The frequency of errors in protein biosynthesis, Biochem. J. 128 (1972) 1353–1356.

Aminoacyl-tRNA synthetases [75] Loftfield R.B., The frequency of errors in protein biosynthesis, Biochem. J. 89 (1963) 82–92. [76] Baldwin A.N., Berg P., Transfer ribonucleic acid-induced hydrolysis of valyladenylate bound to isoleucyl ribonucleic acid synthetase, J. Biol. Chem. 241 (1966) 839–845. [77] Fersht A.R., Enzyme Structure and Mechanism, Freeman, San Francisco, 1977. [78] Fersht A.R., Sieves in sequence, Science 280 (1998) 541. [79] Fersht A., Enzyme Structure and Mechanism, W.H. Freeman and Company, New York, 1985. [80] Schmidt E., Schimmel P., Residues in a class I tRNA synthetase which determine selectivity of amino acid recognition in the context of tRNA, Biochemistry 34 (1995) 11204–11210. [81] Schmidt E., Schimmel P., Mutational isolation of a sieve for editing in a transfer RNA synthetase, Science 264 (1994) 265–267. [82] Lin L., Hale S.P., Schimmel P., Aminoacylation error correction, Nature 384 (1996) 33–34. [83] Rould M.A., Perona J.J., Söll D., Steitz T.A., Structure of E. coli glutaminyl-tRNA synthetase complexed with tRNAGln and ATP at 2.8 Å resolution, Science 246 (1989) 1135–1142. [84] Starzyk R.M., Webster T.A., Schimmel P., Evidence for dispensable sequences inserted into a nucleotide fold, Science 237 (1987) 1614–1618. [85] Hou Y.M., Shiba K., Mottes C., Schimmel P., Sequence determination and modeling of structural motifs for the smallest monomeric aminoacyl-tRNA synthetase, Proc. Natl. Acad. Sci. USA 88 (1991) 976–980. [86] Nureki O., Vassylyev D.G., Tateno M., Shimada A., Nakama T., Fukai S., Konno M., Hendrickson T.L., Schimmel P., Yokoyama S., Proofreading by isoleucyl-tRNA synthetase: response, Science 283 (1999) 453. [87] Hale S.P., Auld D.S., Schmidt E., Schimmel P., Discrete determinants in transfer RNA for editing and aminoacylation, Science 276 (1997) 1250–1252. [88] Nordin B.E., Schimmel P., RNA determinants for translational editing. Mischarging a minihelix substrate by a tRNA synthetase, J. Biol. Chem. 274 (1999) 6835–6838. [89] von derHaar F., Cramer F., Hydrolytic action of aminoacyl-tRNA synthetases from baker’s yeast: ‘Chemical proofreading’ preventing acylation of tRNAIle with misactivated valine, Biochemistry 15 (1976) 4131–4138. [90] Nawrot B., Milius W., Ejchart A., Limmer S., Sprinzl M., The structure of 3’-O-anthraniloyladenosine, an analogue of the 3’-end of aminoacyl-tRNA, Nucleic Acids Res. 25 (1997) 948–954. [91] Joyce C.M., Steitz T.A., Function and structure relationships in DNA polymerases, Annu. Rev. Biochem. 63 (1994) 777–822. [92] Curnow A.W., Ibba M., Söll D., tRNA-dependent asparagine formation, Nature 382 (1996) 589–590. [93] Becker H.D., Kern D., Thermus thermophilus: a link in evolution of the tRNA-dependent amino acid amidation pathways, Proc. Natl. Acad. Sci. USA 95 (1998) 12832–12837. [94] Schon A., Hottinger H., Söll D., Misaminoacylation and transamidation are required for protein biosynthesis in Lactobacillus bulgaricus, Biochimie 70 (1988) 391–394. [95] Curnow A.W., Hong K., Yuan R., Kim S., Martins O., Winkler W., Henkin T.M., Söll D., Glu-tRNAGln amidotransferase: a novel heterotrimeric enzyme required for correct decoding of glutamine codons during translation, Proc. Natl. Acad. Sci. USA 94 (1997) 11819–11826. [96] Curnow A.W., Tumbula D.L., Pelaschier J.T., Min B., Söll D., Glutamyl-tRNA(Gln) amidotransferase in Deinococcus radiodurans may be confined to asparagine biosynthesis, Proc. Natl. Acad. Sci. USA 95 (1998) 12838–12843. [97] Becker H.D., Reinbolt J., Kreutzer R., Giegé R., Kern D., Existence of two distinct aspartyl-tRNA synthetases in Thermus thermophilus. Structural and biochemical properties of the two enzymes, Biochemistry 36 (1997) 8785–8797.

699 [98] RajBhandary U.L., Initiator transfer RNAs, J. Bacteriol. 176 (1994) 547–552. [99] RajBhandary U.L., Chow C.M., Initiator tRNAs and initiation of protein synthesis, in: Söll D., RajBhandary U.L. (Eds.), tRNA: Structure, Biosynthesis, and Function, ASM Press, Washington, D.C., 1995. [100] Baron C., Böck A., The selenocysteine-inserting tRNA species: Structure and function, in: Söll D., RajBhandary U.L. (Eds.), tRNA: Structure, Biosynthesis, and Function, ASM Press, Washington, D.C., 1995, pp. 529–544. [101] Bult C.J., White O., Olsen G.J., Zhou L., Fleischmann R.D., Sutton G.G., Blake J.A., Fitzgerald L.M., Clayton R.A., Gocayne J.D., Kerlavage A.R., Dougherty B.A., Tomb J.F., Adams M.D., Reich C.I., Overbeek R., Kirkness E.F., Weinstock K.G., Merrick J.M., Glodek A., Scott J.L., Geoghagen N.S.M., Venter J.C., Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii, Science 273 (1996) 1058–1073. [102] Smith D.R., Doucette-Stamm L.A., Deloughery C., Lee H., Dubois J., Aldredge T., Bashirzadeh R., Blakely D., Cook R., Gilbert K., Harrison D., Hoang L., Keagle P., Lumm W., Pothier B., Qiu D., Spadafora R., Vicaire R., Wang Y., Wierzbowski J., Gibson R., Jiwani N., Caruso A., Bush D., Reeve J.N. et al Complete genome sequence of Methanobacterium thermoautotrophicum deltaH: functional analysis and comparative genomics, J. Bacteriol. 179 (1997) 7135–7155. [103] Kim H., Vothknecht U.C., Hedderich R., Celic I., Söll D., Sequence divergence of seryl-tRNA synthetases in archaea, J. Bacteriol. 180 (1998) 6446–6449. [104] Stanzel M., Schon A., Sprinzl M., Discrimination against misacylated tRNA by chloroplast elongation factor Tu, Eur. J. Biochem. 219 (1994) 435–439. [105] Schmitt E., Panvert M., Blanquet S., Mechulam Y., Crystal structure of methionyl-tRNAfMet transformylase complexed with the initiator formyl-methionyl-tRNAfMet, EMBO J. 17 (1998) 6819–6826. [106] Fromant M., Plateau P., Schmitt S., Mechulam Y., Blanquet S., Receptor site for the 5’-phosphate of elongator tRNAs governs substrate selection by peptidyl-tRNA synthetase, Biochemistry 38 (1999) 4982–4987. [107] Nissen P., Kjeldgaard M., Thirup S., Polekhina G., Reshetnikova L., Clark B.F., Nyborg J., Crystal structure of the ternary complex of Phe-tRNAPhe, EF-Tu, and a GTP analog, Science 270 (1995) 1464–1472. [108] Saks M.E., Sampson J.R., Variant minihelix RNAs reveal sequence-specific recognition of the helical tRNA(Ser) acceptor stem by E. coli seryl-tRNA synthetase, EMBO J. 15 (1996) 2843–2849. [109] Sissler M., Giegé R., Florentz C., The RNA sequence context defines the mechanistic routes by which yeast arginyl-tRNA synthetase charges tRNA, RNA 4 (1998) 647–657. [110] Frugier M., Helm M., Felden B., Giegé R., Florentz C., Sequences outside recognition sets are not neutral for tRNA aminoacylation. Evidence for nonpermissive combinations of nucleotides in the acceptor stem of yeast tRNAPhe, J. Biol. Chem. 273 (1998) 11605–11610. [111] Sissler M., Giegé R., Florentz C., Arginine aminoacylation identity is context-dependent and ensured by alternate recognition sets in the anticodon loop of accepting tRNA transcripts, EMBO J. 15 (1996) 5069–5076. [112] Ibba M., Losey H.C., Kawarabayasi Y., Kikuchi H., Bunjun S., Substrate recognition by class I lysyl-tRNA synthetases: A molecular basis for gene displacement, Proc. Natl. Acad. Sci. USA 96 (1999) 418–423. [113] Ibba M., Morgan S., Curnow A.W., Pridmore D.R., Vothknecht U.C., Garnder W., Lin W., Woese C.R., Söll D., A euryarchaeal lysyl-tRNA synthetase: Resemblance to class I synthetases, Science 278 (1997) 1119–1122. [114] Ibba M., Bono J.L., Rosa P.A., Söll D., Archaeal-type lysyl-tRNA synthetase in the Lyme disease spirochete Borrelia burgdorferi, Proc. Natl. Acad. Sci. USA 94 (1997) 14383–14388.

700 [115] Florentz C., Giegé R., tRNA-like structures in plant viral RNAs, in: Söll D., RajBhandary U.L. (Eds.), tRNA: Structure, Biosynthesis, and Function, ASM Press, Washington, D.C., 1995, pp. 141–163. [116] Dreher T.W., Goodwin J.B., Transfer RNA mimicry among tymoviral genomic RNAs ranges from highly efficient to vestigial, Nucleic Acids Res. 26 (1998) 4356–4364. [117] Goodwin J.B., Dreher T.W., Transfer RNA mimicry in a new group of positive-strand RNA plant viruses, the furoviruses: differential aminoacylation between the RNA components of one genome, Virology 246 (1998) 170–178. [118] Goodwin J.B., Skuzeski J.M., Dreher T.W., Characterization of chimeric turnip yellow mosaic virus genomes that are infectious in the absence of aminoacylation, Virology 230 (1997) 113–124. [119] Himeno H., Nameki N., Tadaki T., Sato M., Hanawa K., Fukushima M., Ishii M., Ushida C., Muto A., Escherichia coli tmRNA (10Sa RNA) in trans-translation, Nucleic Acids Symp. Ser. 37 (1997) 185–186. [120] Felden B., Hanawa K., Atkins J.F., Himeno H., Muto A., Gesteland R.F., McCloskey J.A., Crain P.F., Presence and location of modified nucleotides in Escherichia coli tmRNA: structural mimicry with tRNA acceptor branches, EMBO J. 17 (1998) 3188–3196. [121] Himeno H., Sato M., Tadaki T., Fukushima M., Ushida C., Muto A., In vitro trans translation mediated by alanine-charged 10Sa RNA, J. Mol. Biol. 268 (1997) 803–808. [122] Hou Y.M., Schimmel P., A simple structural feature is a major determinant of the identity of a transfer RNA, Nature 333 (1988) 140–145. [123] McClain W.H., Foss K., Changing the identity of a tRNA by introducing a G-U wobble pair near the 3’acceptor end, Science 240 (1988) 793–796. [124] Beuning P.J., Yang F., Schimmel P., Musier-Forsyth K., Specific atomic groups and RNA helix geometry in acceptor stem recognition by a tRNA synthetase, Proc. Natl. Acad. Sci. USA 94 (1997) 10150–10154. [125] Sprinzl M., Horn C., Brown M., Ioudovitch A., Steinberg S., Compilation of tRNA sequences and sequences of tRNA genes, Nucleic Acids Res. 26 (1998) 148–153. [126] Chihade J.W., Hayashibara K., Shiba K., Schimmel P., Strong selective pressure to use G:U to mark an RNA acceptor stem for alanine, Biochemistry 37 (1998) 9193–9202. [127] Lipman R.S., Hou Y.M., Aminoacylation of tRNA in the evolution of an aminoacyl-tRNA synthetase, Proc. Natl. Acad. Sci. USA 95 (1998) 13495–13500. [128] Becker H.D., Giegé R., Kern D., Identity of prokaryotic and eukaryotic tRNA(Asp) for aminoacylation by aspartyl-tRNA synthetase from Thermus thermophilus, Biochemistry 35 (1996) 7447–7458. [129] Stehlin C., Burke B., Yang F., Liu H., Shiba K., Musier-Forsyth K., Species-specific differences in the operational RNA code for aminoacylation of tRNAPro, Biochemistry 37 (1998) 8605–8613. [130] Quinn C., Tao N., Schimmel P., Species-specific microhelix aminoacylation by a eukaryotic pathogen tRNA synthetase dependent on a single base pair, Biochemistry 34 (1995) 12489–12495. [131] Hipps D., Shiba K., Henderson B., Schimmel P., Operational RNA code for amino acids: Species-specific aminoacylation of minihelices switched by a single nucleotide, Proc. Natl. Acad. Sci. USA 92 (1995) 5550–5552. [132] Shiba K., Stello T., Motegi H., Noda T., Musier-Forsyth K., Schimmel P., Human lysyl-tRNA synthetase accepts nucleotide 73 variants and rescues Escherichia coli double-defective mutant, J. Biol. Chem. 272 (1997) 22809–22816. [133] Shiba K., Motegi H., Schimmel P., Maintaining genetic code through adaptations of tRNA synthetases to taxonomic domains, Trends Biochem. Sci. 22 (1997) 453–457.

Martinis et al. [134] Pope A.J., Moore K.J., Mc Vey M., Mensah L., Benson N., Osbourne N., Broom N., Brown M.J., O’Hanlon P., Characterization of isoleucyl-tRNA synthetase from Staphylococcus aureus. II. Mechanism of inhibition by reaction intermediate and pseudomonic acid analogues studied using transient and steadystate kinetics, J. Biol. Chem. 273 (1998) 31691–31701. [135] Pope A.J., McVey M., Fantom K., Moore K.J., Effects of substrate and inhibitor binding on proteolysis of isoleucyl- tRNA synthetase from Staphylococcus aureus, J. Biol. Chem. 273 (1998) 31702–31706. [136] Brown J.R., Doolittle W.F., Root of the universal tree of life based on ancient aminoacyl-tRNA synthetase gene duplications, Proc. Natl. Acad. Sci. USA 92 (1995) 2441–2445. [137] Brown J.R., Doolittle W.F., Archaea and the prokaryote-toeukaryote transition,Microbiol. Mol. Biol. Rev. 61 (1997) 456–502. [138] Gilbart J., Perry C.R., Slocombe B., High-level mupirocin resistance in Staphylococcus aureus: evidence for two distinct isoleucyl-tRNA synthetases, Antimicrob. Agents Chemother. 37 (1993) 32–38. [139] Sassanfar M., Kranz J.E., Gallant P., Schimmel P., Shiba K., A eubacterial Mycobacterium tuberculosis tRNA synthetase is eukaryote-like and resistant to a eubacterial-specific antisynthetase drug, Biochemistry 35 (1996) 9995–10003. [140] Brown J.R., Zhang J., Hodgson J.E., A bacterial antibiotic resistance gene with eukaryotic origins, Curr. Biol. 8 (1998) 365–367. [141] Lamour V., Quevillon S., Diriong S., N’guyen V.C., Lipinski M., Mirande M., Evolution of the Glx-tRNA synthetase family: The glutaminyl enzyme as a case of horizontal gene transfer, Proc. Natl. Acad. Sci. USA 91 (1994) 8670–8674. [142] Siatecka M., Rozek M., Barciszewski J., Mirande M., Modular evolution of the Glx-tRNA synthetase family--rooting of the evolutionary tree between the bacteria and archaea/eukarya branches, Eur. J. Biochem. 256 (1998) 80–87. [143] Brown J.R., Doolittle W.F., Gene descent, duplication, and horizontal transfer in the evolution of glutamyl- and glutaminyltRNA synthetases, J. Mol. Evol. (1999), in press. [144] Agou F., Quevillon S., Kerjan P., Mirande M., Switching the amino acid specificity of an aminoacyl-tRNA synthetase, Biochemistry 37 (1998) 11309–11314. [145] Rould M.A., Perona J.J., Söll D., Steitz T.A., Structural basis of anticodon loop recognition by glutaminyl-tRNA synthetase, Nature 352 (1991) 213–218. [146] Nureki O., Vassylyev D.G., Katayanagi K., Shimizu T., Sekine S., Kigawa T., Miyazawa T., Yokoyama S., Morikawa K., Architectures of the class-defining, subclass-defining, and specific domains of glutamyl-tRNA synthetase from Thermus thermophilus, Science 267 (1995) 1958–1965. [147] Schimmel P., Ribas dePouplana L., Genetic code origins: experiments confirm phylogenetic predictions and may explain a puzzle, Proc. Natl. Acad. Sci. USA 96 (1999) 327–328. [148] Ribas dePouplana L., Turner R.J., Steer B.A., Schimmel P., Genetic code origins: tRNAs older than their synthetases?, Proc. Natl. Acad. Sci. USA 95 (1998) 11295–11300. [149] Weiner A.M., Maizels N., A deadly double life, Science 284 (1999) 63–64. [150] Rould M.A., Perona J.J., Söll D., Steitz T.A., Structure of E. coli glutaminyl-tRNA synthetase complexed with tRNAGln and ATP at 2.8 Å resolution, Science 246 (1989) 1135–1142. [151] Ruff M., Krishnaswamy S., Boeglin M., Poterszman A., Mitschler A., Podjarny A., Rees B., Thierry J.C., Moras D., Class II aminoacyl transfer RNA synthetases: Crystal structure of yeast aspartyl-tRNA synthetase complexed with tRNAAsp, Science 252 (1989) 1682–1689. [152] Poterszman A., Delarue M., Thierry J.C., Moras D., Synthesis and recognition of aspartyl-adenylate by Thermus thermophilus aspartyl-tRNA synthetase, J. Mol. Biol. 244 (1994) 158–167.