Structural flexibility of the methanogenic-type seryl-tRNA synthetase active site and its implication for specific substrate recognition Silvija Bilokapic 1 , Jasmina Rokov Plavec 1 , Nenad Ban 2 and Ivana Weygand-Durasevic 1 1 Department of Chemistry, Faculty of Science, University of Zagreb, Croatia 2 Department of Biology, Institute of Molecular Biology and Biophysics, Swiss Federal Institute of Technology, Zu ¨ rich, Switzerland The fidelity of protein synthesis depends on the correct attachment of amino acids to the 3¢-ends of their cog- nate tRNA species by aminoacyl-tRNA synthetases (aaRS), a family of enzymes that bridges the informa- tion gap between nucleic acids and proteins. Although strategies for the specific recognition of amino acid and tRNA are unique to each enzyme, two reaction steps in the esterification of amino acids to their cog- nate tRNAs are conserved among all aaRS: (a) the activation of amino acids with ATP by formation of aminoacyl-adenylate, and (b) the transfer of the aminoacyl-moiety from the aminoacyl-adenylate to the cognate tRNA substrate [1]. Sequence alignment and later structural data showed that aaRSs constitute a family of enzymes in which the same catalytic reaction is performed at two topologi- cally different structural domains. Class I aaRSs are built around a canonical dinucleotide-binding fold (Rossmann fold) with the consensus motifs HIGH and KMSKS, which define two regions of sequence conser- vation for all class I aaRSs. Class II synthetases share an antiparallel b sheet partly enclosed by helices, thus Keywords conformational flexibility; motif 2 loop; seryl- tRNA synthetase; specificity of substrate recognition; synthetase:tRNA model Correspondence I. Weygand-Durasevic, Department of Chemistry, Faculty of Science, University of Zagreb, Horvatovac 102a, 10000 Zagreb, Croatia Fax: +385 1 460 6401 Tel: +385 1 460 6230 E-mail: weygand@chem.pmf.hr (Received 20 February 2008, revised 21 March 2008, accepted 26 March 2008) doi:10.1111/j.1742-4658.2008.06423.x Seryl-tRNA synthetase (SerRS) is a class II aminoacyl-tRNA synthetase that catalyzes serine activation and its transfer to cognate tRNA Ser . Previ- ous biochemical and structural studies have revealed that bacterial- and methanogenic-type SerRSs employ different strategies of substrate recogni- tion. In addition to other idiosyncratic features, such as the active site zinc ion and the unique fold of the N-terminal tRNA-binding domain, metha- nogenic-type SerRS is, in comparison with bacterial homologues, charac- terized by a notable shortening of the motif 2 loop. Mutational analysis of Methanosarcina barkeri SerRS (mMbSerRS) was undertaken to identify the active site residues that ensure the specificity of amino acid and tRNA 3¢-end recognition. Residues predicted to contribute to the amino acid spec- ificity were selected for mutation according to the crystal structure of mMbSerRS complexed with its cognate aminoacyl-adenylate, whereas those involved in binding of the tRNA 3¢-end were identified and mutagenized on the basis of modeling the mMbSerRS:tRNA complex. Although mMbSerRSs variants with an altered serine-binding pocket (W396A, N435A, S437A) were more sensitive to inhibition by threonine and cyste- ine, none of the mutants was able to activate noncognate amino acids to greater extent than the wild-type enzyme. In vitro kinetics results also suggest that conformational changes in the motif 2 loop are required for efficient serylation. Abbreviations aaRS, aminoacyl-tRNA synthetase; HTH, helix–turn–helix; mMbSerRS, archaeal (Methanosarcina barkeri) SerRS; Ni-NTA, nickel-nitrilotriacetic acid; SerRS, seryl-tRNA synthetase; SOL, serine ordering loop. FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2831 creating a fold found only in the class II synthetases and in their paralogues. Three short conserved sequence motifs (motifs 1, 2 and 3) are characteristic of all class II synthetases [1]. The separation of aaRSs into two classes correlates with the modes of tRNA binding: class I synthetases approach the acceptor helix from the minor groove side, whereas class II synthetases approach it from the opposite, major groove side [2,3]. The functional differ- ences between the two classes can also be seen in the mode of ATP binding: the ATP molecule adopts an extended structure in the class I active site, whereas bound ATP in the class II catalytic core has a bent conformation. Recent completion of representative crystal structures for all 20 aaRSs has created a new context to determine if synthetases of the same class share kinetic features that parallel those based on struc- ture [4,5]. Again, a distinct mechanistic signature has been shown to divide the two classes of synthetase [6]. Seryl-tRNA synthetase (SerRS), which catalyzes serylation of the corresponding set of isoaccepting tRNAs Ser , belongs to the class II aaRSs [7]. The crys- tal structure of Escherichia coli SerRS was the first example of an aaRS that does not possess the Ross- mann dinucleotide-binding fold [8]. The enzyme is homodimeric, with motif 1 involved in dimer interface contacts. As revealed by the crystal structures of two prokaryotic seryl-tRNA synthetases from E. coli and Thermus thermophilus, each subunit possesses a C-ter- minal active site domain typical for class II aaRSs, whereas the first 100 N-terminal residues form an anti- parallel a-helical coiled-coil crucial for the selection and binding of tRNA [7]. In the crystal structure of T. thermophilus SerRS complexed with tRNA Ser , the N-terminal helical arm is buried between the TwC arm and the long extra arm of tRNA Ser [9,10]. The crystal structure of mammalian mitochondrial SerRS from Bos taurus has an N-terminal domain that also consists of a long a-helical arm, and extensions in the N- and C-termini of the enzyme ensure recognition of two unusual mitochondrial tRNAs Ser [11]. In general, the active sites of seryl-tRNA synthetases contain a very long loop in motif 2 (the longest of all class II synthe- tases), involved in binding ATP and the acceptor end of tRNA [8,10,12]. With the first sequence of archaeal genome (Methan- ococcus jannaschii) [13], it became apparent that it encodes an atypical SerRS, later found in all methano- genic archaea, and thus named methanogenic-type SerRS [14,15], to be distinguished from SerRSs found in all other organisms (bacterial-type SerRSs). The N-terminal region of this enzyme is significantly longer than the corresponding domain of its bacterial-type counterparts. Accordingly, the crystal structure of Methanosarcina barkeri SerRS (mMbSerRS) [12] revealed two idiosyncratic features: a novel N-terminal tRNA-binding domain and a zinc ion in the active site. Biochemical analysis confirmed and absolute require- ment of zinc for enzymatic activity. In addition, meth- anogenic-type SerRS is, in comparison with bacterial homologues, characterized by a notable shortening of the motif 2 loop, questioning the mode of tRNA bind- ing. Evidently, bacterial- and methanogenic-type SerRSs have diverse modes of substrate recognition [10,15–19]. Although the widely distributed bacterial- type SerRS has been extensively characterized, the unusual methanogenic-type still represents an intrigu- ing puzzle in structure–functional and evolutionary terms. Using mutational and kinetic analysis, we eva- luated the importance of mMbSerRS active site flexibility in the discriminating between substrates and the proper positioning of the tRNA CCA-end. Results Structure-based design of mMbSerRS variants The positions of particular motifs along the mMbSerRS polypeptide are shown in Fig. 1. Each mMbSerRS subunit consists of two domains linked by a short flexible oligopeptide (L). The N-terminus (residues 1–165) is a mixed a ⁄ b domain composed of a six-stranded antiparallel b sheet capped by a bundle of four helices (H1, H2, H3, H4). This fold is distinct from the coiled-coil of the tRNA-binding domain in bacterial-type SerRS, but presumably also interacts with the tRNA extra arm (see later). The catalytic module (residues 174–502), built upon eight anti- parallel strands surrounded by three a helices, repeats the general structure of the catalytic core in the bacte- rial enzyme. The catalytic domain contains three class II signature motifs, designated M1, M2 and M3 in Fig. 1A. M1 motifs participate at the dimer inter- face, whereas M2 and M3 contain residues involved in serine, ATP and tRNA 3¢-end interactions. The tetra- coordinated Zn 2+ ion is bound to three conserved protein ligands (Cys306, Glu355 and Cys461; marked with asterisks in Fig. 1A) and a water molecule, which dissociates from the zinc ion to allow coordination of the amino group of the serine substrate (Fig. 1B). Accordingly, biochemical analysis revealed a loss of enzyme function as a consequence of alterations in zinc-binding residues [12]. In this study, we aimed to characterize the other active site residues, which according to the crystal structure may influence the specificity of substrate recognition. These are marked mMbSerRS active site mutants S. Bilokapic et al. 2832 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS with a black lozenge in Fig. 1A. The mutations were generated using site-specific in vitro mutagenesis of the mMbSerRS gene, and N-terminally His-tagged vari- ants with alanines (or valine in one case) at the alter- ation sites were expressed and purified from E. coli. All mutated enzymes possessed a characteristic class II dimeric nature, as confirmed by gel-filtration chroma- tography (not shown). CD spectra of the wild-type and all mutants were similar, confirming that the over- all fold is maintained (Fig. 2). In addition, selected mutants were crystallized and the structures were essentially identical to that of the wild-type enzyme (not shown). Kinetic analysis of amino acid activation by mMbSerRS variants with an altered serine-binding pocket Serine binding in wild-type mMbSerRS is accompanied by a significant localized conformational change in the A B Fig. 1. Structural and functional features of mMbSerRS. (A) Idiosyncratic and general structural–functional motifs along the polypeptide chain of one subunit. H1–H4 represent four idiosyncratic helices in the N-terminal domain, L is a linker region, M1, M2 and M3 are class II signa- ture motifs, HTH and SOL denote idiosyncratic oligopeptide insertions into the mMbSerRS catalytic core. Zinc-ion ligands are marked by asterisks (*). A black lozenge denotes other active site residues that are, according to a crystal or model structure, presumed to be impli- cated in the specificity of substrate recognition. These are changed to alanine (or valine) in this study. (B) View of the active site-bound seryl-adenylate analogue. The characteristic extended conformation of seryl-adenylate in the active site of mMbSerRS can be observed. Residues that interact with the analogue are indicated. The active site zinc ion is in cyan. 200 210 220 230 240 250 260 –12 –10 –8 –6 –4 –2 0 2 4 6 8 [Θ]X 10 3 (deg·cm 2 · dmol –1 ) Wavelen g th (nm) Fig. 2. CD spectra of wild-type mMbSerRS (h) and its mutants: N435A (s), S437A (n), W396A ⁄ N435A S437A (,), R267A (X), E338A (e), R347A (+). S. Bilokapic et al. mMbSerRS active site mutants FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2833 ‘serine ordering loop’ (SOL, residues 394–410; Fig. 1), which brings the loop into the proximity of the zinc ion and enables direct contact between Gln400 and the carbonyl oxygen of the serine substrate [12]. These movements are required to position the carboxylate oxygen for nucleophilic attack of the a-phosphate of ATP. Thus, according to the crystal structure, the spec- ificity of serine recognition depends on: (a) the zinc ion, (b) the size of the active site and (c) the hydrophilic nature of the serine-binding pocket. Despite precise recognition of the cognate amino acid, functional assays have shown slight, but notable, misactivation of threonine by mMbSerRS, which does not seem to be edited in vitro [12]. To obtain further insight into the discrimination strategy, we designed several mMbS- erRS mutants and tested their ability to activate serine and misactivate noncognate amino acids. Because the size of the active site depends predominantly on Trp396, which is located at the bottom of the serine- binding subsite and packs above the amino group of the serine substrate (Fig. 1B), by substituting alanine, we expected to produce a variant with an enlarged active site that was capable of accommodating a larger noncognate amino acid. However, formation of a hydrophilic environment around Asn435 and Ser437 (which coordinate a structurally conserved water mole- cule) may reduce the binding of amino acids with hydrophobic side chains (e.g. threonine). Accordingly, the mMbSerRS triple-mutant (W396A ⁄ N435A ⁄ S437A) was expected to combine important misactivating features, namely a larger binding site and increased hydrophobicity. The kinetic parameters for serine activation were determined using the pyrophosphate-exchange reaction for the wild-type and five mMbSerRS variants with alterations in the amino acid-binding pocket (W396A, N435A, S437A, the double-mutant N435Al ⁄ S437A and the triple-mutant W396A ⁄ N435A ⁄ S437A). All mutants showed significant reductions in catalytic effi- ciency (k cat ⁄ K m ) for serine compared with wild-type mMbSerRS (Table 1). In accordance with structural data and the sequence conservation of Trp396 (replacement by Phe in some methanogenic-type SerRSs is presumably functionally equivalent), muta- tion of this residue had deleterious effects on k cat and K m for serine in seryl-adenylate synthesis (Table 1). Replacing Asn435 does not affect the affinity for serine in the first step of the aminoacylation reaction as much as mutation of Trp396. However, catalytic efficiency is seriously hampered. The alteration S437A leads to the K m value for serine being elevated by two orders of magnitude. This residue is highly conserved in all SerRSs, although its direct interaction with serine has not been observed in bacterial- and methanogenic-type SerRS structures [12,16]. CD spectrometry and X-ray analysis showed that the observed kinetic effect (Table 1) is not due protein misfolding (Fig. 2). Having established the effect of alterations in both Asn435 and Ser437 on serine recognition, we investigated the catalytic properties of the enzyme using double substitutions. The variant N435A ⁄ S437A displayed kinetic parameters similar to those deter- mined for S437A (Table 1), suggesting that the large reduction in the affinity for serine was caused primar- ily by the mutation S437A. Although mutations N435A ⁄ S437A altered the chemical properties of the amino acid binding pocket significantly, the substrate specificity of the double mMbSerRS mutant was not relaxed. Activation of noncognate amino acids in the ATP–PP i exchange reaction was not increased in com- parison with wild-type enzyme (not shown). This mutant showed decreased activation efficiency towards both cognate and noncognate substrates (serine and threonine, respectively). Although the triple-mutant W396A ⁄ N435A ⁄ S437A showed a 10 3 -fold decrease in serine affinity (Table 1) its k cat value was only three times lower than for the wild-type enzyme, suggesting a compensatory effect for the mutations. It seems that in an enlarged amino acid-binding subsite orientation of the carboxylate group is preserved and the attack on ATP is presumably facilitated. An electron-density map for the mutated enzyme in complex with serine Table 1. Kinetic parameters for serine in ATP–PP i exchange reaction. mMbSerRS K m (lM) k cat (s )1 ) k cat ⁄ K m (s )1 ÆlM )1 ) k cat ⁄ K m (s )1 ÆM )1 )(k cat ⁄ K m ) rel a Wild-type 6.61 ± 1.325 2.06 ± 0.208 0.31143 311430.7 1 W396A 264.5 ± 12.55 0.287 ± 0.0631 0.00108 1083.932 287.32 N435A 70.28 ± 0.280 0.172 ± 0.0167 0.00245 2448.776 127.18 S437A 573.8 ± 69.50 0.729 ± 0.169 0.00127 1271.349 244.96 N435A S437A 689.5 ± 15.25 0.576 ± 0.0283 0.00084 835.388 372.79 N435A ⁄ S437A ⁄ W396A 2879 ± 303.5 1.32 ± 0.2421 0.00046 459.535 677.71 a (k cat ⁄ K m ) wt ⁄ (k cat ⁄ K m ) mut . mMbSerRS active site mutants S. Bilokapic et al. 2834 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS (data not shown) revealed the ordering of 394–410 loop residues and interactions with Gln400, despite replacing Trp396 with alanine. The influence of mutations in the serine-binding pocket on magnesium binding The characteristic conformation of ATP in the class II active site is usually stabilized by hydrogen bonds with motif 2 and motif 3 conserved arginine residues and three Mg 2+ ions. In addition to phosphate groups, acidic residues from the active site and water molecules coordinate Mg 2+ ions. Interestingly, an exception has been observed in the structure of HisRS: two magne- sium cations are bound in the active site and a HisRS- specific arginine residue occupies the same position as the third Mg 2+ and takes over its role in binding the a-phosphate of the ATP molecule [20]. Magnesium ions play a crucial role in the aminoa- cylation reaction because they decrease and delocalize the negative charge of the triphosphate moiety of ATP, which is in close proximity to the negatively charged carboxylate group of the amino acid sub- strate [21]. The binding of ATP and the amino acid substrate in the active site are linked, and incorrect binding of ATP can affect serine binding and vice versa [22]. Asn435, together with Asp416 and Glu432, contributes to binding of the Mg 2+ ion between the a- and b-phosphates of ATP. Indeed, a 10-fold increase in the K m value for serine has been observed for the N435A mutant compared with the wild-type enzyme. Therefore, we assayed the activity depen- dence of mMbSerRS variants comprising N435A replacement on magnesium concentration in the ATP–PP i exchange reaction at saturating levels of serine and ATP substrates. The double-mutant N435A ⁄ S437A and the triple-mutant N435A ⁄ S437A ⁄ W396A showed maximal reaction activity at magne- sium concentrations that were higher than for the wild-type enzyme or W396A mutant, supporting the role of Asn435 in magnesium binding (Fig. 3). Kinetic analysis of aminoacylation by mMbSerRS variants with an altered serine-binding pocket The catalytic parametars of W396A, N435A, S437A, the double-mutant N435Al ⁄ S437A and the triple- mutant W396A ⁄ N435A ⁄ S437A were significantly dif- ferent from those of the wild-type mMbSerRS in serine activation. The kinetic properties of the active site mutants were further investigated in the aminoacyla- tion reaction. The production and characterization of different tRNAs Ser substrates is described below. The data reveal that alteration in the SOL (W396A) has a more pronounced effect on adenylate synthesis than on tRNA aminoacylation (Table 2). Also, mutations in Asn435 and Ser437 have a less deleterious influence on tRNA aminoacylation. This suggests that tRNA bind- ing stabilizes the active site despite perturbations caused by the mutations. Although only four class I aaRSs, glutamyl-, glutaminyl-, arginyl- and lysyl-tRNA synthetases (GluRS, GlnRS, ArgRS and LysRS, respectively), are known to require their cognate tRNA for amino acid activation, a number of other synthetases (including several class II representatives) [1], use tRNA to optimize the amino acid binding pocket. Indeed, tRNA-mediated amino acid recognition has been documented for yeast SerRS [23,24]. Our results show that tRNA also con- tributes to optimization of the active site in methano- genic-type SerRS, which can be detected when kinetic parameters for the mutated proteins are measured. Selectivity of the active site towards amino acid substrates Based on previous biochemical and crystallographic studies [12], we assumed that of the 20 amino acids alanine, glycine, threonine and cysteine may potentially bind into the active site of methanogenic-type SerRS and cause mischarging problems. The ability of these noncognate amino acids to impair the specific aminoa- cylation of tRNA Ser with serine was measured for wild-type and mutant mMbSerRS enzymes (Fig. 4). Alanine and glycine were chosen because they are smaller than serine; yet they do not have similar chem- ical properties. Valine was used as a positive control 0 0.2 0.4 0.6 0.8 1 1.2 0 5 10 15 20 25 c(Mg)/mM Relative reaction activity Fig. 3. Comparison of the activity dependence of wild-type and mMbSerRS mutants on magnesium concentration. Activity was assayed in an ATP–PP i exchange reaction: wild-type (r), W396A ( ), N435A ⁄ S437A ( ), W396A ⁄ N435A ⁄ S437A (s). S. Bilokapic et al. mMbSerRS active site mutants FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2835 because it is bigger and cannot form the H-bonds that are characteristic of serine in the active site. Threonine and cysteine were used because of their chemical simi- larity with serine. Binding of threonine may be particu- larly likely because of the similarity between the mMbSerRS and threonyl-tRNA synthetase (ThrRS) catalytic domains, both of which contain a zinc ion. Serylation activity of the wild-type protein was 10% lower in the presence of cysteine, and the inhibition of N435A ⁄ S437A and W396A ⁄ N435A ⁄ S437A mutants was 55% and 70%, respectively (Fig. 4). Interestingly, the noncognate amino acid was not activated in the ATP–PP i assay, indicating that the inhibition of seryla- tion was not caused by cysteinyl-tRNA Ser formation. Cysteine may be a competitive inhibitor of aminoacy- lation in the presence of tRNA or the –SH group of cysteine may attack serylated-tRNA, resulting in the formation of a serine–cysteine dipeptide and deacylated tRNA, in accordance with the hypothesis that aaRSs are capable of performing thioester-depen- dent peptide synthesis [25,26]. It remains to be deter- mined which of two proposed scenarios is actually taking place in methanogenic-type SerRS. In addition, inhibition of serylation with 6 mm threonine was 10–20% higher with mutated mMbSerRS variants than with wild-type enzyme. Again, the mutants did not misactivate threonine (or other noncognate amino acids) more efficiently than the wild- type enzyme (data not shown). It seems that noncognate substrates are not bound in the proper orientation, which would allow their carboxyl group to attack the a-phosphate of bound ATP and lead to the completion of the first step of the aminoacylation reaction. tRNA acceptor-stem binding by mMbSerRS motif 2 loop residues and the design of the mutants Multiple sequence alignments of members of two distinct groups of SerRS showed that the motif 2 loop sequence, which according to crystal structures participates in acceptor-stem contact, is highly con- served within each type, but differs between them [12]. The crystal structure of T. thermophilus SerRS in com- plex with cognate tRNA revealed that Phe262 from this long loop interacts with the fifth base pair of the acceptor stem [10]. In contrast to bacterial-type SerRS, the discriminator base (G73) and the first base pair in the acceptor stem were shown to be important determinants of specific tRNA Ser recognition in metha- nogenic-type enzyme [15]. We assumed that the differ- ences in acceptor-stem recognition between the two SerRS types are due to differences in the motif 2 loop sequence and that each motif 2 loop is capable of performing different but specific interactions with cognate tRNA. Because the acceptor end of tRNA Ser in the T. ther- mophilus SerRS:tRNA co-crystal structure is disor- dered, we made a mMbSerRS:tRNA docking model using the crystal structures of ThrRS in complex with the cognate tRNA. Because of the similar size of the motif 2 loop in methanogenic-type SerRS and ThrRS and the involvement of zinc ions in the recognition of amino acid substrates, the M. barkeri loop was homol- Table 2. Kinetic parameters for serine in aminoacylation reaction. The tRNA used in these experiments was M. barkeri tRNA Ser overpro- duced in E. coli. mMbSerRS K m (lM) k cat (s )1 ) k cat ⁄ K m (s )1 ÆlM )1 ) k cat ⁄ K m (s )1 ÆM )1 )(k cat ⁄ K m ) rel a Wild-type 11.13 ± 1.310 0.473 ± 0.0323 0.04247 42470.80 1 W396A 760.2 ± 92.74 0.239 ± 0.0320 0.00032 315.049 134.80 N435A 283.1 ± 14.91 0.435 ± 0.0322 0.00154 1535.853 27.65 S437A 412.9 ± 75.04 0.396 ± 0.0297 0.00096 958.828 44.29 N435A ⁄ S437A 509.9 ± 48.49 0.361 ± 0.0113 0.00071 707.394 60.038 N435A ⁄ S437A ⁄ W396A 1191 ± 299.2 0.175 ± 0.0348 0.00015 146.600 289.70 a (k cat ⁄ K m ) wt ⁄ (k cat ⁄ K m ) mut . 0 20 40 60 80 100 120 Activity % val g ly ala cys thr Fig. 4. Inhibition of mMbSerRS with noncognate amino acids. Activity of wild-type mMbSerRS (black bar) and different mutants (W396A, dark gray; N435A ⁄ S437A, gray; W396A ⁄ N435A ⁄ S437A, white) in the presence of other amino acids. The level of inhibition was determined as a ratio of initial velocities of inhibited (6 m M noncognate amino acid) and uninhibited reactions. mMbSerRS active site mutants S. Bilokapic et al. 2836 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS ogy modeled based on the structure of the E. coli ThrRS loop. We observed that the positioning of the cognate tRNA in the active site of mMbSerRS would be facilitated upon the conformational change of the motif 2 loop, which was the only proximal protein region in the proposed mMbSerRS:tRNA complex able to mediate the interactions with the tRNA accep- tor stem. The model used to predict the amino acids that might participate in the binding of the 3¢-end of tRNA, which were therefore subjected to mutational and kinetic analyses, is shown in Fig. 5. Kinetic parameters for M. barkeri tRNA Ser amino- acylation by wild-type and mutant mMbSerRS enzymes are given in Table 3. The results are consistent with the involvement of several motif 2 residues (Glu338, Arg347, Gly340 and Gly341) in the serylation reaction. Loss of the catalytic efficiency of mutant E338A is essentially the consequence of a decreased k cat value (Table 3), and is in agreement with previous studies on aspartyl-tRNA synthetase (AspRS) [27,28] and LysRS [29]. Residue Glu338 in mMbSerRS is presumably functionally equivalent to Glu258 in T. thermophilus SerRS. The crystal structure of bacterial binary SerRS complexes reveals an interaction between Glu258 and N6 of ATP or Ser-AMS. Upon tRNA binding, this residue forms a hydrogen bond with N2 of G73. How- ever, the side chain of Glu338 may adopt a different conformation, which would facilitate interactions with the O2¢ of C74 ribose, analogous to those observed in the structure of the E. coli AspRS:tRNA complex. Because G73 is a discriminator base in mMbSerRS, specific interactions, in addition to Glu338, must con- tribute to the importance of G73. The class-invariant Arg347 from the motif 2 loop in mMbSerRS is posi- tioned to interact with N1 of G73. It cannot be excluded, however, that upon tRNA binding the orien- tation of Arg347 side chain changes, allowing recogni- tion of C74 as in other class II synthetases. The crystal structure of mMbSerRS shows that Arg347 and Glu338 also form contacts with ATP (Fig. 1B). Our kinetic results revealed that replacement of these two residues with alanine affects tRNA aminoacylation (Table 3), which is consistent with the involvement of these two residues in both steps of the aminoacylation reaction in all class II synthetases [30]. The R347A variant shows a greater reduction in k cat than the E338A variant (8-fold for E338A vs. 175-fold for R347A). Moreover, Arg347 also contributes slightly to the binding of tRNA, as deduced from the twofold decrease in K m (Table 3). Strikingly, two glycines in succession, Gly340 and Gly341, in the motif 2 loop are conserved among methanogenic-type SerRSs. We analyzed the functional significance of this apparent flexibility for tRNA CCA-end binding and recognition of the G1:C72 iden- tity determinant. In agreement with our model, the G340V ⁄ G341A mutant showed a 410-fold diminished Fig. 5. Model of the acceptor end of tRNA Ser bound in the active site of mMbSerRS. The view is focused on the acceptor part of tRNA in the active site of mMbSerRS. The tRNA is shown as an orange tube with the CCA-end and G73 shown as sticks. The cata- lytic domain of mMbSerRS in complex with Ser-AMS is shown in blue. The structure of mMbSerRS with the proposed motif 2 loop conformation upon tRNA binding is depicted in gray. Residues that participate in the interaction with tRNA and that were tested are shown as sticks. It can be seen that the side chain of Ile342 points away from tRNA molecule. All molecular depictions were produced using PYMOL (http://pymol.sourceforge.net). Table 3. Kinetic parameters for M. barkeri tRNA Ser aminoacylation with wild-type and mutant mMbSerRS enzymes. The tRNA used in these experiments was M. barkeri tRNA Ser overproduced in E. coli. mMbSerRS K m (lM) k cat (s )1 ) k cat ⁄ K m (s )1 ÆlM )1 ) k cat ⁄ K m (s )1 ÆM )1 ) k cat ⁄ K m rel Wild-type 2.99 ± 0.31 0.454 ± 0.018 0.152076 152076.4 1 R267A 3.40 ± 0.39 0.032 ± 0.0020 0.009436 9435.79 16.12 E338A 2.17 ± 0.14 0.059 ± 0.0043 0.027221 27220.68 5.59 R347A 6.38 ± 1.12 0.0026 ± 0.00046 0.000403 403.13 377.23 G340V ⁄ G341A 6.76 ± 0.55 0.0011 ± 9.42 · 10 )5 0.000171 171.28 887.9 S. Bilokapic et al. mMbSerRS active site mutants FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2837 catalytic efficiency and a twofold higher K m (Table 3). Thus, the flexibility of this region might be important in avoiding steric clashes with tRNA and allowing hydrogen bonding of the first base pair. The crystal structure of mMbSerRS revealed that the unique  30-residue insertion between motifs 1 and 2 in methanogenic-type SerRSs adopts a helix– turn–helix (HTH) fold. The HTH fold of one mono- mer is positioned above the catalytic core of the other. Interestingly, Arg267 from the beginning of helix 9 in the HTH fold of one monomer is positioned towards the active site of the other monomer and can contrib- ute to interactions with the tRNA acceptor end. In the proposed model, Arg267 coordinates the ribose O2¢ of C74. Substitution of Arg267 with alanine significantly decreased the k cat value for tRNA aminoacylation (Table 3). Amino acids subjected to alterations in G340V ⁄ G341A and R267A mutants presumably interact with tRNA in the transition state and contribute to structural rearrangements that result in new contacts between the acceptor end of tRNA and the active site of synthetase. Furthermore, Gly340 and Gly341 from the motif 2 loop also contribute to the affinity for tRNA. Importantly, our kinetic data indicate that Arg267 from the HTH-motif, belonging to one monomer of the dimeric mMbSerRS, is essen- tial for tRNA aminoacylation in the active site of the other monomer. The crystal structures of mMbSerRS in complex with ATP or the seryl-adenylate analogue do not reveal any interactions between this arginine and small substrates. Therefore, detected changes in the kinetic parameters for this mutant can only be due to tRNA CCA-end binding (Table 3). Thus, Arg267 from the HTH-motif seems to be essential for the second, but not the first, step in the aminoacylation reaction. Wild-type enzyme and mutants were tested for their ability to bind tRNA using a gel-shift assay. All pro- teins were able to shift the cognate tRNA, although some mutants form weaker complexes, especially R347A, as shown in Fig. 6. Variability of tRNA substrates for mMbSerRS We encountered many difficulties in preparing suffi- cient quantities of tRNA substrates to determine the kinetic parameters with mMbSerRS variants. Because of the very inefficient in vitro transcription of M. bark- eri synthetic tRNA Ser genes we decided to attempt to overproduce M. barkeri tRNAs Ser in E. coli. Both tRNA CGA and tRNA GGA isoacceptors were found to be well expressed in vivo. K m values for serylation were estimated for the expressed tRNAs and for in vitro transcribed tRNA CGA as a reference. The K m and k cat values for both isoacceptors were similar and compara- ble with tRNA Ser obtained by in vitro transcription (Table 4). It was thus evident that the M. barkeri tRNAs Ser were processed correctly in the bacterial host. tRNA GGA acceptor made up 50–60% of the total tRNA and was used for subsequent experiments because it was better expressed than tRNA CGA , which contrib- uted only 25–30%. Endogenous E. coli tRNA Ser , 3.5% in the same preparation, was negligible in comparison with the overexpressed M. barkeri tRNA GGA . Because of the 25-fold higher k cat value (Table 4), E. coli tRNA Ser was not expected to influence the kinetic results obtained with in vivo produced M. barkeri tRNA Ser .To confirm this, in vivo expressed M. barkeri tRNA GGA was purified on urea PAGE. The kinetic parameters obtained with pure M. barkeri tRNA GGA and mMbS- erRS were the same as for the ‘crude’ sample (data not shown). Methanococcus maripaludis tRNA GCU gene tran- script was not efficiently recognized by wild-type mMbSerRS (Table 4), which is consistent with struc- tural differences between homologous and heter- ologous serine-specific tRNAs in the D-loops. Although M. barkeri, in contrast to Me. jannaschii and Me. maripaludis, does not possess tRNA Sec [31], our experiments revealed charging of Me. jannaschii tRNA Sec by mMbSerRS, comparable with the recogni- tion of tRNA Sec in the human and bacterial systems (Table 4). Fig. 6. Native PAGE analysis of complex formation between over- expressed tRNA Ser and mMbSerRS mutants. (Upper) Wild-type mMbSerRS forms a more stable complex with overexpressed tRNA Ser in comparison with R347A mutant: lane 1, wild-type mMbSerRS (1.6 l M); lanes 2–6, wild-type and tRNA (lane 2, 0.12 l M; lane 3, 0.3 lM; lane 4, 0.6 lM; lane 5, 1.2 lM, lane 6, 2.4 l M); lanes 7–10, R347A and tRNA (lane 7, 0.12 lM; lane 8, 0.3 l M; lane 9, 0.6 lM; lane 10, 1.2 lM). (Lower) Complex formation between tRNA Ser (1.2 lM) and different mMbSerRS mutants (1.6 l M): lane 1, wild-type mMbSerRS; lane 2, N435A S437A; lane 3, W394A; lane 4, catalytic domain (does not make complex with tRNA; unpublished results); lane 5, catalytic domain mutant; lane 6, R347A; lane 7, G340V ⁄ G341A. The gel was stained with Coomas- sie Brilliant Blue. Arrow denotes the position of the complex. mMbSerRS active site mutants S. Bilokapic et al. 2838 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS Discussion Switching the specificity for amino acid substrate Although alteration of the three residues (W396A, N435A and S437A) in the amino acid binding pocket of mMbSerRS was expected to increase the level of threonine misrecognition, no improvement in threonine activation was observed with mutated enzymes. Thus, it appears to be far easier to abolish cognate amino acid acivation than to replace it with noncognate activation, despite redesigning the active site in terms of size, polarity and hydrogen-bonding capacity. Specific amino acid recognition is therefore substan- tially more resistant to mutation than structure-based predictions would suggest, as observed previously for other systems [32]. Our results suggest that engineering methanogenic-type SerRS in an attempt to change its substrate specificity is not straightforward and may possibly be achieved only by multiple alterations of the active site residues. Furthemore, contemporary metha- nogenic-type SerRS and ThrRS may have diverged too far to readily allow the switching of substrate specifici- ties between the two, as seen for tryptophanyl-tRNA synthetase (TrpRS) and tyrosyl-tRNA synthetase (TyrRS) [32]. By contrast, substrate-specificity switching was possible for GlnRS ⁄ GluRS, a much more closely related pair of aaRSs, which have only relatively recently diverged from a common ancestor [33,34]. Flexibility of the motif 2 loop is required for tRNA binding The studies described here were undertaken to test the role of the motif 2 loop in functional binding of cog- nate tRNA. Sequence variability in this loop which exists in different class II aaRSs clearly points to its role in the selectivity of tRNA recognition. In order to study binding of the tRNA acceptor end into the active site of mMbSerRS, we superimposed the ThrRS in complex with cognate tRNA onto the catalytic core of methanogenic-type SerRS. The most striking feature of the obtained model is the severe steric clash between the acceptor end of tRNA and the motif 2 loop of mMbSerRS, especially with residue Ile342 (Fig. 7). In the co-crystal structures of AspRS:tRNA and ThrRS:tRNA complexes, the conformation of motif 2 loop is closed, in comparison with the mMbSerRS motif 2 loop. The clash seen in the mMbSerRS:tRNA model, combined with the results of our mutational and kinetic analyses, suggest that a conformational change in this loop is required upon tRNA binding. Recognition by methanogenic SerRS relies, in addi- tion to the long extra arm, on G1:C72 in the cognate tRNA, which is achieved by the motif 2 residues (Figs 5 and 7). Two successive glycines in the motif 2 loop are conserved among all the methanogenic-type SerRSs. Our biochemical experiments point to the importance of these residues for the flexibility of the tRNA 3¢-end binding region. The ability of the loop to change its conformation upon tRNA binding is crucial for correct positioning of the tRNA acceptor end in the active site. However, whether these residues also contribute to specific interactions with the tRNA G1:C72 pair remains to be investigated. Role of motif 2 and the SOL in the two steps of the aminoacylation reaction Strucutural [12] and kinetic data (Table 3) showed that the SOL has a crucial role in cognate amino acid Table 4. Kinetic parameters for wild-type mMbSerRS enzyme with various tRNA substrates in aminoacylation reaction. Mb, Mm and Mj denote M. barkeri, M. maripaludis and Me. jannaschii, respectively. tRNA K m (lM) k cat (s )1 ) k cat ⁄ K m (s )1 ÆlM )1 ) k cat ⁄ K m (s )1 ÆM )1 ) k cat ⁄ K m rel a In vivo Mb tRNA Ser GGA 2.99 ± 0.31 0.454 ± 0.018 0.1518395 151839.5 1 In vivo Mb tRNA Ser CGA 3.07 ± 0.52 0.467 ± 0.038 0.1521173 152117.3 0.998 In vitro transcript Mb tRNA Ser 10.10 ± 1.67 0.439 ± 0.111 0.0434653 43465.3 3.49 In vitro transcript Mm tRNA Ser 3.316 ± 0.129 0.079 ± 0.027 0.0238239 2382.39 63.73 tRNA Ser from total E. coli tRNA 2.34 ± 0.37 0.0175 ± 0.0015 0.0074786 7478.63 19.74 In vitro transcript Mj tRNA Sec 2.41 ± 0.04 0.005 ± 0.0006 0.0020746 2074.69 73.19 a k cat ⁄ K m (in vivo Mb tRNA Ser GGA ) ⁄ k cat ⁄ K m (tRNA). S. Bilokapic et al. mMbSerRS active site mutants FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2839 binding to mMbSerRS. In addition, the motif 2 loop participates in ATP binding and recognition of the tRNA 3¢-end in all class II aaRSs. The motif 2 loop of T. thermophilus SerRS is disordered in the absence of substrate. It adopts two different conformations upon substrate binding: an ‘A-conformation’ in the presence of ATP or adenylate or a ‘T-conformation’ when tRNA Ser is bound. In mMbSerRS, the motif 2 loop is fully ordered in an apo-enzyme structure. However, in the presence of ATP or seryl-adenylate the motif 2 loop shifts (Fig. 8). In addition, the side chains of the conserved arginine residues in the motif 2 loop show a concerted movement following substrate binding in the active site. In the apo-enzyme, the motif 2 Arg336 occupies the adenine-binding site and upon binding of ATP or Ser-AMS shifts to form interactions with the a-phosphate group. Arg336 also interacts with the carbonyl oxygen of Ser-AMS, but not with serine. Therefore, this residue can sense the presence of both substrates needed for the first step of the amino- acylation reaction. Fig. 7. Proposed model of the mMbSerRS:tRNA complex. Overall view with ThrRS:tRNA Thr complex (1qf6) shown as a ribbon repre- sentation and colored in orange. The apo-structure mMbSerRS (2cim) is shown as a blue transparent surface. For clarity, only the tRNA CCA-end and the first two base pairs are shown as sticks. Two synthetases are superposed by means of their catalytic domains. The view reveals steric clashes between the mMbSerRS motif 2 loop and the acceptor end of tRNA (circled). The gray arrow indicates the likely movement of the mMbSerRS N-terminal domain upon tRNA binding. (Inset) Comparison of the motif 2 loops in AspRS:tRNA (1c0a; red). ThrRS:tRNA (1qf6; orange) and mMbS- erRS (2cim; blue) structures, resulting from the superpositions of their entire catalytic domains. The backbone of the tRNA Thr accep- tor arm is shown (tRNA Asp has been omitted for clarity). The open conformation of motif 2 loop in apo-structure of mMbSerRS causes clashes with tRNA, especially Ile342. r.m.s.d Serine AT P Ser-AMS Residues around serine loop Beginning of motif 3 Motif 2 loop Helix 9 r.m.s.dr.m.s.d Fig. 8. Conformational changes in the active site upon binding of the substrates. The r.m.s.d of C a atoms after superposition of the catalytic domains in complex with different substrates and the cata- lytic domain of the apo-enzyme. The binding of the small substrates in the active site leads to conformational changes in the motif 2 loop and SOL. The serine loop residues were excluded from calcu- lation of the r.m.s.d. because they are not visible in the apo-struc- ture or the structure of the enzyme in complex with ATP. However, it can be seen that the residues that are enclosing this loop are flexible and adopt a range of conformations. mMbSerRS active site mutants S. Bilokapic et al. 2840 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS [...]... binding, the SOL has to adopt a new conformation, different from the one with serine or Ser-AMS in the active site The side chain of Glu338 presumably forms interactions with the incoming tRNA substrate, as in bacterial-type SerRS, AspRS and phenyl-tRNA synthetase (PheRS) The conformational change may aid release of the by-product AMP after the second step of the aminoacylation reaction The new SOL conformation... breaking of the interactions between serine, Gln400 and Trp396, facilitating the release of charged tRNA Therefore, a conformational switch in the SOL in complexes with serine or tRNA bound to the active site, as well as the changes in the motif 2 loop residues with the respect to ATP and tRNA binding, control the two steps of the aminoacylation reaction Experimental procedures Site- directed mutagenesis and. .. were used, the kinetic parameters were determined from two independent experiments Active site titration To determine the amount of active enzyme in the preparation, active site titration was performed Formation of the enzyme tightly bound intermediate seryl-AMP was followed in the presence of pyrophosphatase (PPiase) and saturating amounts of 14C-serine and ATP at 37 °C The reaction was performed in... Cusack S (1994) ˚ The 2.9 A crystal structure of T thermophilus seryltRNA synthetase complexed with tRNASer Science 263, 1404–1410 10 Cusack S, Yaremchuk A & Tukalo M (1996) The crystal structure of the ternary complex of T thermophilus seryl-tRNA synthetase with tRNASer and a seryl-adenylate analogue reveals a conformational switch in the active site EMBO J, 15, 2834–2842 mMbSerRS active site mutants 11... (2004) Mutation and evolution of the magnesium-binding site of a class II aminoacyl-tRNA synthetase Biochemistry 43, 7028–7037 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2843 mMbSerRS active site mutants S Bilokapic et al 23 Lenhard B, Filipic S, Landeka I, Skrtic I, Soll D & ¨ Weygand-Durasevic I (1997) Defining the active site of yeast seryl-tRNA synthetase: mutations... 7 Weygand-Durasevic I & Cusack S (2005) Structure, function and evolution of seryl-tRNA synthetases In The Aminoacyl-tRNA Synthetases (Ibba M, Francklyn C & Cusack S, eds), pp 177–192 Landes Bioscience, Georgetown, TX 8 Cusack S, Berthet-Colominas C, Hartlein M, Nassar N ¨ & Leberman R (1990) A second class of synthetase structure revealed by X-ray analysis of Escherichia coli ˚ seryl-tRNA synthetase. .. interacts with the c-phosphate of the ATP molecule Thus, the pyrophosphate molecule can be released from the active site, terminating the amino acid activation Unlike Arg336, Arg347 does not stabilize Ser-AMP in the active site Interestingly, successful completion of the serine activation step releases Arg347 for interaction with the CCA-end of bound tRNASer Our tRNA-binding model suggests that conformational... 5¢-ATACTTATCTCCATTGATGGCCACAGCCTGGAACTCAAGCCATTC Combination of primers for N435A ⁄ S437A and W396A FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2841 mMbSerRS active site mutants S Bilokapic et al Synthetic genes for M barkeri tRNASerCGA and tRNASerGGA with the isopropyl thio-b-d-galactoside-inducible T7 promoter at their 5¢-end were constructed and cloned between the SalI and BamHI sites of pET3a transcription... Blue to analyze the components of the tRNASer:SerRS complex Crystalyzation of mMbSerRS mutants was performed as described previously [12] Acknowledgements This work was supported by grants from Scientific Co-operation between Eastern Europe and Switzerland (SCOPES) program of the Swiss National Science Foundation (SNSF), the Ministry of Science, Education and Sports of the Republic of Croatia (project... Bilokapic et al mMbSerRS active site mutants The conserved motif 2 loop arginine, Arg347, also contributes to ATP binding In the apo-structure, it interacts with Glu338 from the motif 2 loop Upon ATP binding, Arg347 swings toward the bound substrate and stabilizes its bend conformation by interacting with the c-phosphate Following the formation of seryl-adenylate, Arg347 moves from its previous position . Structural flexibility of the methanogenic-type seryl-tRNA synthetase active site and its implication for specific substrate recognition Silvija. Model of the acceptor end of tRNA Ser bound in the active site of mMbSerRS. The view is focused on the acceptor part of tRNA in the active site of mMbSerRS.