doi:10.1046/j.1432-1033.2002.03241.x

Eur. J. Biochem. 269, 5271–5279 (2002) (cid:2) FEBS 2002

tRNA-dependent amino acid discrimination by yeast seryl-tRNA synthetase

Ita Gruic-Sovulj1,2, Irena Landeka1,2, Dieter So¨ ll3 and Ivana Weygand-Durasevic1,2 1Department of Chemistry, Faculty of Science, University of Zagreb, Croatia; 2Rudjer Boskovic Institute, Zagreb, Croatia; 3Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut, USA

tRNA-induced conformational change in the enzyme’s act- ive site. The 3¢-terminal adenosine of tRNASer is not important in effecting the rearrangement of the serine binding site. Our results do not provide an indication for a readjustment of ATP binding in a tRNA-assisted manner. The stoichiometric analyses of the complexes between the enzyme and tRNASer revealed that two cognate tRNA molecules can be bound to dimeric SerRS, however, with very different affinities.

Sequence-specific

tRNA–protein

Keywords: tRNASerÆSerRS complexes; tRNA-dependent amino acid recognition; amino acid selection; tRNA bind- ing; covalent cross-linking.

The ability of aminoacyl-tRNA synthetases to distinguish between similar amino acids is crucial for accurate trans- the genetic code. Saccharomyces cerevisiae lation of seryl-tRNA synthetase (SerRS) employs tRNA-dependent recognition of its cognate amino acid serine [Lenhard, B., Filipic, S., Landeka, I., Skrtic, I., So¨ ll, D. & Weygand- Durasevic, I. (1997) J. Biol. Chem. 272, 1136–1141]. Here we show that dimeric SerRS enzyme complexed with one molecule of tRNASer is more specific and more efficient in catalyzing seryl-adenylate formation than the apoenzyme alone. interactions enhance discrimination of the amino acid substrate by yeast SerRS and diminish the misactivation of the structurally similar noncognate threonine. This may proceed via a

Ile–tRNAPhe

Accurate translation of genetic information is dependent on the high fidelity of several molecular processes. Different quality control mechanisms are adapted to prevent or correct naturally occurring mistakes [1]. Measurements of the selectivity of aminoacyl-tRNA synthetases in vitro indicate that the upper limit for error in the selection of correct amino acid for protein synthesis is in the range of 10)4)10)5. The frequency of errors involving noncognate tRNA aminoacylation is, in most cases, 10)6 or lower [2]. While the selection of the correct tRNA is assumed to occur as a result of preferential reaction kinetics for the formation of cognate proteinÆRNA complexes, the differences of the side chains of amino acids are often sufficient to allow their specific binding [3]. The basic challenge in achieving high specificity for the amino acid substrates lies in the rejection of smaller substrates with similar side chain chemistry or isosteric amino acids [2]. A number of different pathways have been proposed to correct misactivation of amino acids in vivo. The noncognate aminoacyl-adenylate can be

hydrolyzed by several class I synthetases by tRNA- dependent or tRNA-independent (cid:2)pre-transfer(cid:3) editing, whereas in the (cid:2)post-transfer(cid:3) pathway a mischarged tRNA is rapidly deacylated in a synthetase-dependent manner [4–12]. The editing reactions of class II synthetases have been studied much less than those of class I. Class II phenylalanyl-tRNA synthetase (PheRS) specifically deacy- lates [13], and alanyl-tRNA synthetase (AlaRS) has been shown to hydrolyze misactivated serine and glycine [14]. Escherichia coli lysyl-tRNA synthetase (LysRS) hydrolyzes misactivated homocysteine, homoser- ine, cysteine, threonine and alanine [15]. Recent reports on threonyl-(ThrRS) [16,17] and prolyl-tRNA synthetase (ProRS) [18–20] show clearly that these enzymes misactivate smaller noncognate amino acids, serine and alanine, respectively, and therefore require proofreading activity. However, during the editing process, some of the correct products are also destroyed [21]. This makes corrective pathways energetically costly, and they are disfavored in the cell. Alternatively, the quality of aminoacyl-tRNA synthesis in the cell can be improved by the tRNA-mediated mechanisms that enhance the accuracy of amino acid discrimination. These are based on conformational changes in the enzymes, induced by the formation of macromolec- ular complex [22–24] or possibly by the interaction with a nonsynthetase protein [25,26].

The active site of class II aaRSs contains the motif 2 loop which is involved in binding of ATP, an amino acid, and the acceptor end of tRNA. Our earlier investigation revealed that accurate seryl-tRNA synthesis in Saccharomyces cere- visiae is accomplished via tRNA-assisted optimization of amino acid binding to the enzyme active site [27]. To understand the mechanism by which this occurs, we have generated a number of SerRS mutants with altered motif 2

Correspondence to I. Weygand-Durasevic, Department of Chemistry, Faculty of Science, University of Zagreb, Strossmayerov trg 14, 10000 Zagreb, Croatia. Fax: + 385 1 4561177, Tel.: +385 1 4561197, E-mail: weygand@rudjer.irb.hr Abbreviations: aaRS, aminoacyl-tRNA synthetase with amino acid representing Ala, Cys, Gln, Lys, Phe, Pro, Ser and Thr, thus for alanyl- (EC 6.1.1.7), cysteinyl- (EC 6.1.1.16), glutaminyl- (EC 6.1.1.18), lysyl- (EC 6.1.1.6), phenylalanyl- (EC 6.1.1.20), prolyl- (EC 6.1.1.15), seryl- (EC 6.1.1.11) and threonyl-tRNA synthetase (EC 6.1.1.3); aaRSÆtRNA and aaRS–tRNA, noncovalent and covalent complexes between synthetase and tRNA, respectively. (Received 27 May 2002, revised 9 August 2002, accepted 9 September 2002)

5272 I. Gruic-Sovulj et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

0.5 mM [32P]PPi (4–7 c.p.m.Æpmol)1), 10 mM KF. For the Km determination of serine the concentration varied from 50 lM to 900 lM, while ATP–MgCl2 was kept constant at 2 mM. The concentrations of wild-type and mutant enzymes were between 50 and 100 nM, and nonchargeable tRNAs were in the range of 100–500 nM. The Km for ATP was determined in the standard reaction mixture with ATP– MgCl2 varied from 0.4 lM to 50 lM, and serine was kept constant at 900 lM.

Misactivation

loop residues and showed that these mutations affect tRNA-dependent amino acid recognition, possibly by interfering with the flexibility of the motif 2 loop [27]. The same mode of conformational adjustment has been recently proposed for maize organellar SerRS [28]. To get better insight into contribution of tRNA in the amino acid selection, we have generated nonchargeable, 3¢-truncated tRNA analogs, which retain the ability of complex forma- tion with SerRS. These tRNAs significantly influence enzyme affinity for serine, as revealed by steady-state kinetic studies in the reaction of pyrophosphate exchange and decrease the misactivation of threonine by yeast seryl-tRNA synthetase. The stoichiometric analyses of the complexes between the enzyme and tRNASer revealed that two cognate tRNA molecules can be bound to dimeric SerRS, however, with very different affinities. The binding of the first tRNASer is sufficient for complete readjustment of serine binding site(s).

To study the misactivation of threonine by wild-type and mutated SerRS (141 nM), 0.5 mM)50 mM threonine was substituted for serine in the standard reaction mixture for the pyrophosphate exchange. In the experiment designed to determine the ability of tRNASer CC to suppress threonine misactivation, SerRS and tRNASer CC were 100 nM.

Inactivation of SerRS by oxidized tRNASer

E X P E R I M E N T A L P R O C E D U R E S

The overexpression and purification of S. cerevisiae wild- type and mutant (E281D, G291A) SerRS has been described [27]. [14C]Serine (166.1 mCiÆmmol)1) and tetra- (2.44 CiÆmmol)1) were sodium [32P]pyrophosphate purchased from DuPont NEN. The amino acids were from Sigma. The purity of threonine was checked by ESI-Ion Trap mass spectrometry. The signal at m/z value that corresponds to serine was not recorded.

The inactivation of the SerRS pyrophosphate exchange activity by tRNASer ox was studied with 100 nM enzyme and tRNASer ox varied from 0 to 1500 nM. In order to preform the macromolecular complex before binding of small substrates, SerRS and tRNASer were preincubated under conditions favorable for noncovalent complex formation (5 min, 30 (cid:3)C) and the activation reaction was started by addition of serine and ATP in the standard reaction mixture for pyrophosphate exchange.

Gel mobility shift assay

tRNASer

SerRSÆtRNA complexes were prepared by incubation of the enzyme with variable amounts of freshly renatured tRNA, for 5 min at 30 (cid:3)C, in buffer containing 30 mM Hepes/KOH pH 7.0 (or Mes/KOH pH 6.0), 10 mM MgCl2 and 1.6% glycerol, followed by cooling on ice. Glycerol was added to a final concentration of 7.5% and the preformed complexes were subjected to electrophoresis on 6% acrylamide/ bisacrylamide (40 : 1) gel containing 5% glycerol in elec- trophoresis buffer (25 mM Tris, 25 mM acetic acid, 10 mM magnesium acetate; pH 7.2, or 25 mM Mes/KOH, 10 mM magnesium acetate, pH 6.0). Electrophoresis was carried out at 4 (cid:3)C for 3–4 h at 100 V, and the gels ware stained by standard silver staining procedure.

Covalent SerRS–tRNASer complexes

tRNAs Yeast tRNASer and tRNATyr, purified from total brewer’s yeast tRNA as described previously [29], accepted 1.2 nmol of serine and 1.4 nmol of tyrosine per A260 unit of tRNA, respectively. tRNA integrity was checked by MALDI-TOF mass spectrometry. A mass resolution m/Dm ¼ 220 (for tRNATyr) and 170 (for tRNASer) was achieved in the linear mode of operation [30]. Nonchargeable substrate analogs were prepared by stepwise modifications of yeast tRNASer at the 3¢-end. Intact tRNASer CCA was first denatured to enable better exposition of the 3¢-adenosine diol groups and then oxidized by periodate treatment essentially as described [31], except that the concentration of periodate was increased five-fold. As the oxidized product is known as synthetase inhibitor, ox was extensively dialyzed prior to removal of 3¢-terminal nucleoside by b-elimination with lysine (pH 8.0), followed by alkaline phosphatase treatment [32]. The acceptor activity of truncated tRNASer CC was not detectable in the standard aminoacylation assay [27]. All tRNA substrates were carefully renatured prior the use in the kinetic assays and complex formation experiments by heating to 80 (cid:3)C and the temperature was then allowed to decrease to 50 (cid:3)C. Then, MgCl2 was added, to a final concentration of 10 mM, and the tRNA sample was cooled to 30 (cid:3)C. E. coli tRNASer 1 was purchased from Subriden RNA (Rolling Bay, WA, USA).

Synthesis of seryl-adenylate

Cross-linking reactions were carried out essentially accord- ing to [33]. The reaction mixture contained 20 mM Hepes pH 8.5, 10 mM MgCl2, 1.7% (v/v) glycerol. SerRS was 0.65 lM, and tRNASer ox /SerRS ratio was varied between 0.5 and 3.0. SerRS and freshly renatured tRNASer ox were incubated for 5 min at 30 (cid:3)C to allow noncovalent complex formation and the cross-linking reaction was started with the addition of 1 lL NaCNBH3 (0.5 mM). The addition of NaCNBH3 was repeated after 20 and 40 min. The final amount of NaCNBH3 added was 1.5 nmol. The reaction was stopped after 60 min by the addition of 0.2 volumes of 4.8 mM NaBH4 in 10 mM NaOH. After 10 min glycerol was added to a final concentration of 8%. The whole reaction mixture was loaded on the 6% nondenaturing

The extent of adenylate formation was measured by ATP- PPi exchange at 30 (cid:3)C [27]. Standard reaction mixtures contained 100 mM Hepes/KOH pH 7.2, 10 mM MgCl2,

tRNA-dependent amino acid discrimination (Eur. J. Biochem. 269) 5273

(cid:2) FEBS 2002

acrylamide/bisacrylamide gel (40 : 1) and the electrophor- esis took place at 4 (cid:3)C (3 h, 100 V) with 50 mM Tris, 25 mM boric acid and 10 mM magnesium acetate, pH 8.0, as running buffer.

Stoichiometric analysis

ox , tRNASer

CC, and E. coli tRNASer

Ferguson plot analysis on disc-PAGE [34] was used for mass determination of covalent SerRS–tRNASer ox complexes. Preformed covalent complexes between SerRS and tRNASer ox were separated on a series of gels containing 5, 6, 7, 8, 9 and 10% acrylamide/bisacrylamide (29 : 1) in 375 mM Tris, 5 mM MgCl2; pH 8.0 alongside 1–10 lg of native protein molecular weight standards (Sigma). Stack- ing gel was 4% acryamide/bisacrylamide (40 : 1) in 62 mM Tris, 5 mM MgCl2 pH 6.3. Electrophoresis was performed at 4 (cid:3)C, 100 V with 50 mM Tris, 5 mM MgCl2 pH 8.0 as anode buffer and 10 mM Tris, 90 mM glutamine pH 8.0 as cathode buffer. For each species, 100 [log(100l)] was determined and plotted against gel concentration, where l is equal to the mobility of the species relative to that of the bromophenol blue tracking dye. Logarithm of negative slope or retardation coefficient (Kr) was then plotted as a function of log molecular mass (kDa) for each protein standard, and from the evaluated retardation coefficients of covalent SerRS–tRNASer ox complexes their molecular mass was determined.

R E S U L T S

tRNASer species form complexes of equal mobility (lanes 1–3, respectively). Although the input ratio of enzyme to tRNA was the same (1 : 2) in all three cases, the shifted band that corresponds to the enzymeÆtRNA complex is somewhat weaker in lane 3. This may suggest that the removal of 3¢-terminal nucleotide slightly influences the complex stability. The oxidation of 3¢-terminal adenosine in tRNASer does not interfere with complex formation and its stability, in agreement with previously shown results for the tyrosyl–tRNA synthetase system [31].

ox or tRNASer

Complex formation with modified cognate and nonmodified heterologous tRNAs The integrity of the truncated tRNASer CC was first checked by high-resolution polyacrylamide gel electrophoresis (Fig. 1). Modified tRNASer CC appeared intact and its migration, with respect to mature tRNASer CCA, was in satisfying agreement with the removal of one nucleotide. Additional bands of lower intensity, visible in both lanes, are probably due to the presence of other tRNASer isoacceptors of different length. In order to determine the ability of modified tRNAs to participate in complex formation with the cognate synthe- tase, preformed complexes of SerRS with either tRNASer CCA, tRNASer CC were analyzed by the gel mobility shift assay (Fig. 2). Under the conditions described in the Experimental procedures, native, oxidized and truncated

When SerRS was mixed with an excess amount of E. coli tRNASer and electrophoresis performed as above (Fig. 2, lane 4), the gel mobility shift analysis revealed a lack of low mobility bands. This indicates that heterologous complex was either not formed or its low stability did not allow detection under the conditions used. This is an interesting finding, as E. coli tRNASer contains a conserved long extra arm which is considered to be the main recognition element for all cytosolic SerRS enzymes [35]. Furthermore, this tRNA is recognized by yeast SerRS in vivo and in vitro [36], showing that some recognition elements in tRNA are shared between bacteria and yeast.

Nonchargeable yeast tRNASer affects serine activation by SerRS

Fig. 2. Gel mobility shift assay of SerRS complexes with different tRNAs. SerRS (9 pmol) was incubated with 18 pmol of tRNASer CCA, tRNASer CCA and subjected to poly- acrylamide gel under native conditions (lanes 1, 2, 3, and 4, respect- ively). Noncomplexed tRNASer (5 pmol) was loaded onto the gel as electrophoretic mobility marker (lane 5). Full line arrow, noncom- plexed tRNA; dashed line arrow, SerRSÆtRNA complexes. The bands migrating between noncomplexed tRNAs and SerRSÆtRNA com- plexes could be assigned to tRNA oligomers and they also appear in tRNA lane when a higher amount is loaded (see Fig. 6A, lane 1). This additional tRNA band appeared when complex formation and elec- trophoresis were performed at pH 7.0 and above.

Kinetic parameters for serine were determined in pyrophos- phate exchange reaction, catalyzed by yeast SerRS (Table 1). Under standard conditions in the absence of tRNA, kinetic parameters are in a very good agreement with previously reported data [27]. The addition of modified

Fig. 1. Preparation of 3¢-truncated tRNASer. The size difference between intact tRNASer CCA and truncated tRNASer CC was confirmed by gel electrophoresis (left). The reaction scheme is shown on the right.

5274 I. Gruic-Sovulj et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

Table 1. The influence of nonchargeable tRNAsSer and heterologous tRNAs on seryl-adenylate synthesis by yeast SerRS. Kinetic parameters were determined in the PPi exchange reaction. Numbers in parenthesis denote the concentration of tRNA. The enzymes were at a concentration of 100 nM.

Km for ATP (lM) Km for Ser (lM) kcat (s)1) kcat/Km, Ser (sÆlM))1

ox (500 nM) CC (100 nM)

SerRSwt

14 13 16 20 ND 9 500 250 70 60 260 300 3.9 0.94 4.9 1.8 10 10 8 · 10)3 4 · 10)3 70 · 10)3 30 · 10)3 39 · 10)3 33 · 10)3 Without tRNA With tRNASer With tRNASer With tRNASer CC With tRNASer CCA E. coli (100 nM) With tRNASer CCA E. coli (500 nM) SerRS (E281D; G291A)

CC (100 nM) CC (500 nM)

fact

consistent with reduced stability of the heterologous tRNAÆSerRS complex and consequently facilitates substrate turnover, as previously observed in other systems [23,37]. It is also important to note that bacterial tRNASer does not fully optimize the serine binding site of yeast SerRS, as the Km for serine never reaches the value of 6.5 · 10)5 M, characteristic for aminoacylation of yeast tRNASer. Never- theless, heterologous tRNASerÆSerRS interactions are suffi- cient to promote conformational change of SerRS, leading to increased kcat/Km value. Thus, tRNA optimized amino acid activation can occur irrespectively of aminoacylation, as shown previously for GlnRS [37]. This finding can be explained by the that discrimination between cognate and noncognate tRNA occurs predominantly during the transfer reaction [24].

Binding of ATP to yeast SerRS is not mediated by tRNA

yeast tRNASer species, which cannot be charged, influences the kinetics with respect to serine. tRNASer ox decreases the Km for serine twofold. The reduction of kcat may reflect the interference between the tRNA’s modified 3¢-terminal adenosine with other substrates in the active site. In the presence of tRNASer CC, which lacks the 3¢-end nucleotide, the binding of serine is significantly enhanced, as revealed by a Km value decreased by almost an order of magnitude. Moreover, the Km value of 6.0 · 10)5 M is very close to that obtained in the aminoacylation reaction with wild-type SerRS (6.3 · 10)5 M) [27]. An order of magnitude increased affinity toward serine in the presence of truncated tRNASer CC suggests that the 3¢-terminal adenosine of tRNASer is not important in effecting the rearrangement of the serine binding site. It is worth noting that an equimolar concen- tration of tRNA per dimeric enzyme is capable of inducing a complete rearrangement of amino acid binding sites, as judged by full range change of the Km value for serine. This suggests that a2Æ(tRNASer)1 is a functional complex in seryl- tRNA formation. The effect of 3¢-truncated tRNASer on the activation of serine by the mutated yeast SerRS, bearing two amino acid changes in the motif 2 loop of the active site (E281D; G291A), is much less pronounced.

The analysis of kinetic parameters for ATP in seryl- adenylate formation did not reveal considerable differences between the Km values for ATP in the presence and absence of tRNAsSer (Table 1). Furthermore, the nature of non- chargeable or heterologous tRNASer did not affect the ATP binding properties of yeast SerRS. Thus, contrary to the observation based on crystallographic studies of Thermus thermophilus SerRS complexes [38], our biochemical experi- ments do not provide indications for tRNA-assisted rear- rangement of ATP binding site. We have previously observed a 6.3-fold decrease in the apparent affinity of ATP for the yeast enzyme in the aminoacylation reaction compared to the PPi exchange [27]. This could be explained by possible interference of correctly positioned Ser–tRNASer with ATP binding. If so, the transfer of acyl moiety to tRNA decreases the affinity of SerRS toward ATP.

Misactivation of threonine by wild-type and mutant SerRS

Slight misactivation of threonine by wild-type yeast SerRS (about 0.4%, based on the comparison of turnover numbers for cognate and noncognate amino acid) has been observed in the pyrophosphate exchange reaction, in which 0.5– 50 mM threonine was substituted for serine (Fig. 3).

The effect of heterologous tRNASer on serine binding Despite the observation that E. coli tRNASer neither is 1 charged well by yeast enzyme [36], nor forms a stable noncovalent complex with yeast SerRS (Fig. 2, lane 4), it affects serine binding properties of yeast synthetase. In the presence of E. coli tRNASer 1 , the Km for serine is decreased two-fold and the velocity of seryl-adenylate synthesis is elevated 2.5-fold (Table 1). Thus, interaction of yeast SerRS with heterologous tRNASer, that contains the long extra arm identity element, improves the overall catalytic prop- erties of the RNAÆsynthetase complex in the pyrophosphate exchange reaction. Accordingly, even though the binding of tRNASer analogs slightly decreased the kcat value, which may be due to the interference of 3¢-CCAox and 3¢-CC ends with the accessibility of the active site, an observed increase in kcat in the presence of E. coli tRNA suggests that the tRNAÆSerRS noncovalent complex is catalytically more efficient in seryl-tRNA formation than noncomplexed SerRS. On the other hand, the increase in kcat may be

Without tRNA With tRNASer With tRNASer 40 ND ND 850 490 350 0.056 0.030 0.028 6.59 · 10)5 6.12 · 10)5 8.00 · 10)5

tRNA-dependent amino acid discrimination (Eur. J. Biochem. 269) 5275

(cid:2) FEBS 2002

Although threonine concentration used in vitro corresponds to approximately 1–100 · Km for serine, these data may be relevant in vivo, as the average amino acid concentrations in yeast cells are in the millimolar range [39]. Glycine and alanine showed no detectable activation in PPi exchange assay. Mutated SerRS, bearing the alterations at two amino acid positions in the motif 2 loop of the active site (E281D; G291A) exhibits more pronounced misactivation with threonine (Fig. 3, inset). This was expected, as the kinetic analysis of the aminoacylation and amino acid activation reactions revealed that substitution of the class II invariant glycine has dramatic effect on seryl-adenylate formation, resulting in severe reduction of catalytic efficiency [27]. The other altered amino acid, E281, is conserved in the active site of seryl-tRNA synthetases and also has an important functional role in yeast SerRS [27]. Based on our experi- mental data, amino acid substitutions in the active site of mutant SerRS (E281D; G291A) decrease the ability of the synthetase to discriminate against noncognate substrates.

Fig. 3. tRNASer CC decreases the misactivation of threonine by yeast SerRS. Threonyl-adenylate formation was compared with and without tRNASer CC. Noncognate aminoacyl-adenylate formation (pmol amino acid-AMP per pmol SerRS per minute) was measured at different amino acid concentrations (0.5–50 mM, which corresponds approximately to 1–100 · Km for serine) in a standard pyro- phosphate exchange reaction. Glycine and alanine showed no detectable activation. The inset shows misactivation of threonine by wild-type and mutant SerRS (E281D; G291A).

CC decreases the misactivation

The addition of tRNASer by yeast SerRS

Figure 3 compares the extent of threonyl-adenylate synthe- sis by yeast SerRS with and without tRNA addition. The decreased level of misactivation observed in the presence of cognate, but nonchargeable tRNA, could be due to hydrolytic editing as described for several other aaRSs [16,18,40,41] or the consequence of tRNA-optimized amino acid discrimination. For the reasons discussed below, we favor the latter hypothesis. Nevertheless, our results point out the importance of tRNAÆsynthetase complex formation in the accuracy of amino acid selection, which may contribute to fidelity of translation in vivo.

Analysis of the complexes between SerRS and tRNASer

(a) Inactivation of SerRS by oxidized tRNASer. The dependence of the velocity of the pyrophosphate exchange reaction on the concentration of the tRNASer ox is shown in Fig. 4. The concentration of oxidized tRNASer was from 0 to 1500 nM, while the other substrates were held at constant concentration (2 mM ATP and 1 mM serine), in the standard pyrophosphate exchange reaction. The velocity decreased to 17% when the concentration of the tRNASer ox reached the value of about 1200 nM. As the active site titration performed without tRNA [42] showed that dimeric enzyme possesses two active sites at which the activation of serine occurs, almost total inactivation of the a2-dimeric SerRS with a nonchargeable tRNASer suggests that two tRNAs could be simultaneously bound per dimeric protein. (b) Gel mobility shift assay. SerRS forms only one type of complex which was sufficiently stable to be detected by the gel mobility shift assay, as shown in Fig. 5. The complex of the same electrophoretic mobility was detected with SerRS/tRNASer ratios 2 : 1, 1 : 1, 1 : 2, 1 : 3, 1 : 4 and 1 : 5 (Fig. 5, lanes 2–7, respectively). Uncomplexed tRNASer was subjected on the same gel as electrophoretic mobility marker (Fig. 5, lane 1). SerRS in complex with two bound tRNASer was not detected by the mobility shift assay despite the broad variation in pH and ionic strength. To test the specificity of the method,

Our kinetic experiments show that equimolar concentration of tRNA per dimeric enzyme induces a complete rearrange- ment of amino acid binding sites, as judged by full range change of the Km value for serine (Table 1). This is an indication that a2Æ(tRNASer)1 is a functional complex in seryl- tRNA formation. In order to determine whether SerRS can bind more than one cognate tRNA molecule under various conditions, several experimental approaches were employed.

Fig. 4. Dependence of seryl-adenylate formation velocity on the con- centration of tRNASer ox . To facilitate formation of noncovalent complex, SerRS and tRNASer ox were incubated for 5 min at 30 (cid:3)C prior to addition to the reaction mixtures for pyrophosphate exchange.

5276 I. Gruic-Sovulj et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

mined Kr values for covalent complexes were 9.2 (lower mobility band) and 10.9 (higher mobility band), corres- ponding to molecular masses of 136 kDa and 183 kDa., i.e. to one and two tRNASer bound to SerRS dimer, respectively (Fig. 6C). Molecular masses of SerRSÆtRNASer and SerR- SÆ(tRNASer)2 complexes obtained by MALDI-MS analysis were 136 kDa and 163 kDa, respectively [29], in very good agreement with those determined by native electrophoresis. Percentage of error, calculated as (estimated molecular mass – reported molecular mass/reported molecular mass) · 100, was in the range of 0 and 12%, i.e. the same as in previously published experiments performed by this method [43].

D I S C U S S I O N

tRNA-dependent amino acid discrimination as a possible quality control mechanism in seryl-tRNA formation

ox and SerRS–(tRNASer

SerRS was incubated with the noncognate yeast tRNATyr present either in the equimolar amount or in high molar excess (Fig. 5, lanes 8 and 9, respectively), followed by the electrophoresis on the native polyacrylamide gel. No band of retarded electrophoretic mobility relative to noncomplexed tRNATyr was detected. (c) Covalent cross-linking. Oxidized tRNA possesses 3¢-terminal dialdehyde groups able to form Schiff bases with lysine side chain amino groups. Considering the nucleophilicity of the lysine side chain amino groups, cross-linking experiments were performed in the pH range 8.0–9.0. In all cases, two complexes, presumably SerRS– tRNASer ox )2 were detected (data not shown). The largest amount of both complexes was obtained at pH 8.5. SerRS and tRNASer ox were incubated in ratios 2 : 1, 1 : 1, 1 : 2 and 1 : 3 (Fig. 6A, lanes 2, 3, 4 and 5, respectively) under conditions described in the Experimental procedures. A higher mobility band, appearing in lanes with SerRS/ tRNASer ox ratios 1 : 1–1 : 3 (Fig. 6A, lanes 3–5) could be assigned to SerRS–(tRNASer ox )2. Uncomplexed SerRS and tRNASer ox were used as mobility markers (Fig. 6A, lanes 6 and 1, respectively). Our results indicate that Schiff bases formed between yeast SerRS and tRNASer ox were stable even without reduction, as approximately the same amount of both cross- linked products were formed both with and without reducing reagents (NaCNBH3 and NaBH4) (data not shown). On the contrary, Madore et al. [33] could not detect the GlnRS– (tRNAGln)2 complex without treatment with NaCNBH3.

Stoichiometric analysis

An order of magnitude of increased affinity toward serine in the presence of nonchargeable 3¢-truncated tRNASer con- firms our previous finding that SerRS modulates its affinity for serine in a tRNA-dependent manner [27]. We have shown in this paper that cognate tRNAÆSerRS interactions increase the stringency of amino acid discrimination. This may proceed via a tRNA-induced conformational change in the enzyme’s active site. In agreement with the observation of Cusack et al. [38] that the alteration of glycines may influence loop flexibility in T. thermophilus SerRS, our results revealed that the activation of serine by mutated yeast SerRS (E281D; G291A) was less significantly affected by 3¢-truncated tRNASer. This indicates that the flexibility of the motif 2 loop is important for structural readjustment of the amino acid binding site. Based on the structural and functional resemblance among seryl-tRNA synthetases [27], we suggest that in yeast, like in T. thermophilus [44], docking interactions between the long variable arm of tRNASer and the N-terminal coiled coil of SerRS govern the positioning of the acceptor stem in the active site of the enzyme. Either these docking interactions or the interaction with the acceptor stem, induce conformational changes in the serine binding sites, which optimize the enzyme for seryl-adenylate formation. As a consequence, the misactivation of structur- ally similar noncognate amino acid is prevented. As recently shown for the aspartyl-system, hierarchical recognition of the synthetase and its cognate tRNA is crucial for formation of the active macromolecular complex [45]. In T. thermo- philus SerRS, serine specificity is guaranteed by two hydrogen bond interactions with the side chain hydroxyl group and by the size of the binding pocket [46]. Model building studies on this enzyme [46] revealed that binding of glycine and alanine would be unfavorable because of the absence of hydrogen bonding capacity, and many other amino acids would be too large. The hydroxyl group of structurally similar threonine also makes an important contribution to specificity of recognition by ThrRS [16,46]. Besides a protein side chain, a zinc ion serves as a specific recognition cofactor for the hydroxyl group of threonine in ThrRS. Although serine is weakly misactivated by ThrRS, the enzyme needs proofreading or editing activity to correct this misactivation. On the other hand, mechanistically analogous zinc ion-mediated amino acid discrimination by cysteinyl-tRNA synthetase (CysRS) fully assures specific

Ferguson plot analysis [34] was performed in order to confirm predicted stoichiometry of detected covalent com- plexes between SerRS and tRNASer ox . Relative mobilities of the protein standards and two covalent complexes were determined on a series of nondenaturing discontinuous gels of increasing polyacrylamide concentration having constant acrylamide/bisacrylamide ratio and their logarithm was plotted as a function of the gel concentration (Fig. 6B). The negative slope of each line [representing the retardation coefficient (Kr) for species] was defined and log Kr was plotted against log molecular mass (kDa) (Fig. 6C). Deter-

Fig. 5. Gel mobility shift assay with different amounts of tRNA. SerRS (9 pmol) was incubated with 4.5, 9, 18, 27, 36 and 45 pmol of tRNASer and with 9 and 45 pmol of yeast tRNATyr prior to electrophoresis on polyacrylamide gel under native conditions (lanes 2, 3, 4, 5, 6, 7 and 8, 9, respectively). Noncomplexed tRNASer was loaded on the gel as electrophoretic mobility marker (lane 1). Complex formation and electrophoresis were performed under pH 6.0. Full line arrow, non- complexed tRNA; dashed line arrow, SerRSÆtRNA complexes.

tRNA-dependent amino acid discrimination (Eur. J. Biochem. 269) 5277

(cid:2) FEBS 2002

ox and SerRS–(tRNASer

ox and SerRS–(tRNASer

ox and SerRS–(tRNASer

which seems to be negligible with any amino acid tested, including threonine [2]. Additionally, SerRS displays AMP- and pyrophospate-independent deacylation of cog- nate aminoacyl-tRNA in the presence of thiols, which mimics editing of homocysteine [15]. Our experiments show that the 3¢-terminal adenosine of cognate tRNASer does not seem to be important in bringing about the rearrangement of the serine binding site. However, less specific noncovalent interactions between yeast enzyme and E. coli tRNASer induce less pronounced conformational change in the active site, revealing the importance of cognate complex formation for accurate amino acid recognition.

Anticooperative binding of tRNASer

synthesis of cysteinyl-adenylate, without serine or threonine activation [47]. The discrimination against threonine by SerRS may rely partly on the fact that its methyl group would be in an unfavorable hydrophilic environment in the active site of enzyme [46]. However, this amino acid is still misactivated by yeast SerRS apoenzyme in vitro, especially when the enzyme is altered in the active site. The misacti- vation by wild-type enzyme is diminished in the presence of cognate tRNA. In the model, we suggested that such tRNA-optimized amino acid recognition may act as a quality control mechanism that lowers the extent of misactivation and thus editing would not be necessary. Another possibility, although much less likely, is that tRNA directs erroneous threonyl-adenylate from the active site to the center for editing. This needs to be further investigated. Thus far, the only experimental finding regarding the ability of SerRS to correct errors is related to pretransfer editing,

The finding that a high molar surplus of yeast-oxidized tRNASer in a PPi exchange reaction deactivated more

Fig. 6. Covalent complexes between SerRS and tRNASer ox . (A) Detection on the gel. SerRS (9 pmol) was incubated with 4.5, 9, 18 and 27 mol of tRNASer ox in the presence of NaCNBH3 and subjected to polyacrylamide gel under native conditions (lanes 2, 3, 4, and 5, respectively). Non- complexed tRNASer ox (5 pmol) and SerRS (9 pmol) was loaded onto the gel as an electrophoretic mobility marker (lanes 1 and 6, respectively). Covalent cross-linking reactions were stopped by the addition of NaBH4. (B) Representative Ferguson analysis. A logarithmic function of relative mobility for each of the protein standards and two covalent complexes, SerRS–tRNASer ox )2, was plotted against the poly- acrylamide concentration and fitted to the regression. Protein standards are indicated: (a) a-lactalbumin (molecular mass 14.2 kDa), (b and c) carbonic anhydrase (two charge isomers with molecular mass 29.0 kDa), (d) chicken egg albumin (molecular mass 45.0 kDa), (e and f) bovine serum albumin monomer (molecular mass 66 kDa) and dimer (molecular mass 132.0 kDa), respectively, (g and h) urease trimer (molecular mass ox )2. R2 values were 272.0 kDa) and hexamer (molecular mass 545.0 kDa), respectively. 1 : 1 and 1 : 2 denote SerRS–tRNASer in the range 0.9914–0.999 for all protein standards and both covalent complexes. (C) Representative plot of log Kr vs. log of molecular mass (kDa). Kr values for protein standards were derived by a Ferguson analysis and log Kr was plotted as a function of log molecular mass (kDa). Protein standards (a–h) are as indicated in (B). Interpolations of log Kr values derived for the SerRS–tRNASer ox )2 complexes (dashed lines) indicate molecular weights of 136 and 183 kDa, respectively. R2 value was 0.993.

5278 I. Gruic-Sovulj et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

5. Farrow, M.A., Nordin, B.E. & Schimmel, P. (1999) Nucleotide determinants for tRNA-dependent amino acid discrimination by class I tRNA synthetase. Biochemistry 38, 16898–16903.

6. Hendrickson, T.L., Nomanbhoy, T.K. & Schimmel, P. (2000) Errors from selective disruption of the editing center in a tRNA synthetase. Biochemistry 39, 8180–8186.

7. Nureki, O., Vassylyev, D.G., Tateno, M., Shimada, A., Nak- ama, T., Fukai, S., Konno, M., Hendrickson, T.L., Schimmel, P. & Yokoyama, S. (1998) Enzyme structure with two catalytic sites for double-sieve selection of substrate. Science 280, 578–582.

8. Schmidt, E. & Schimmel, P. (1995) Residues in a class I tRNA synthetase which determine selectivity of amino acid recognition in the context of tRNA. Biochemistry 34, 11204–11210. 9. Lin, L., Hale, S.P. & Schimmel, P. (1996) Aminoacylation error correction. Nature 384, 33–34.

10. Fukai, S., Nureki, O., Sekine, S., Shimada, A., Tao, J., Vassylyev, D.G. & Yokoyama, S. (2000) Structural basis for double-sieve discrimination of 1-valine from 1-isoleucine and 1-threonine by the complex of tRNA (Val) and valyl-tRNA synthetase. Cell 103, 793–803.

11. Chen, J.F., Guo, N.N., Li, T., Wang, E.D. & Wang, Y.L. (2000) CP1 domain in Escherichia coli leucyl-tRNA synthetase is crucial for its editing function. Biochemistry 39, 6726–6731.

12. Mursinna, R.S., Lincecum, T.L. & Martinis, S.A. (2001) A con- served threonine within Escherichia coli leucyl-tRNA synthetase prevents hydrolytic editing of leucyl-tRNALeu. Biochemistry 40, 5376–5381.

than 80% of yeast SerRS activity led to the conclusion that two tRNASer ox molecules could be simultaneously bound to dimeric SerRS. This was confirmed by covalent cross-linking. However, SerRSÆ(tRNASer)2 complex was not stable enough to be detected by the gel mobility shift assay under the conditions of relatively broad pH and ionic strength ranges. This suggests that two tRNASer binding sites on yeast SerRS are nonequivalent, which is in agreement with previously reported studies of others which revealed that two binding sites for tRNASer differ by about two orders of magnitude in their binding constants [48,49]. Both, SerRSÆtRNASer and SerRSÆ(tRNASer)2, complexes were also detected by MALDI-MS [29]. The specificity of the MALDI-MS measurements was confirmed in the competition assay. The SerRSÆ(tRNASer)2 complex was not detected on the control gel mobility shift assay conducted under the same conditions as the MALDI-MS measurements. This could be the consequence of increased electrostatic forces as the dielectric constant of the environ- ment is lowered during solvent evaporation and in vacuum. For prokaryotic systems, crystallographic studies showed that T. thermophilus SerRS form two types of noncovalent complexes with cognate tRNASer; SerRSÆtRNASer and SerRSÆ(tRNASer)2 [44,50], while Escherichia coli SerRS forms only the SerRSÆ(tRNASer)2 complex [51]. Borel et al. [52] were able to detect both 1 : 1 and 1 : 2 E. coli SerRSÆ tRNASer complexes by zone interference gel electrophoreses and concluded that binding of the tRNA was positively cooperative.

13. Yarus, M. (1972) Phenylalanyl-tRNA synthetase and isoleucyl- tRNA Phe: a possible verification mechanism for aminoacyl- tRNA. Proc. Natl Acad. Sci. USA 69, 1915–1919.

tRNASer

14. Tsui, W.-C. & Fersht, A.R. (1981) Probing the principles of amino acid selection using the alanyl-tRNA synthetase from Escherichia coli. Nucleic Acids Res. 9, 4627–4637.

15. Jakubowski, H. (1997) Aminoacyl thioester chemistry of class II aminoacyl-tRNA synthetases. Biochemistry 36, 11077–11085. 16. Sankaranarayanan, R., Dock-Bregeon, A.-C., Rees, B., Bovee, M., Caillet, J., Romby, P., Francklyn, C.S. & Moras, D. (2000) Zinc ion mediated amino acid discrimination by threonyl-tRNA synthetase. Nat. Struct. Biol. 7, 461–465.

It is tempting to speculate as to why yeast SerRS binds two tRNASer molecules with nonequivalent affinity. Are these binding sites a priori nonequivalent, or does binding of the first tRNA leads to a conformational change which lessens the affinity towards tRNA at the second site? Results presented in this paper show undoubtedly that binding of the equimolar amount of induces complete conformational optimization of the serine binding sites which results in efficient and accurate seryl-adenylate formation. The binding of one tRNA may promote structural rearrangement of SerRS which decreases the affinity for the second tRNA.

17. Musier-Forsyth, K. & Beuning, P.J. (2000) Role of zinc ion in translational accuracy becomes crystal clear. Nat. Struct. Biol. 7, 435–436.

18. Beuning, P.J. & Musier-Forsyth, K. (2000) Hydrolytic editing by a class II aminoacyl-tRNA synthetase. Proc. Natl Acad. Sci. USA 97, 8916–8920.

A C K N O W L E D G M E N T S

19. Ahel, I., Stathopoulos, C., Ambrogelly, A., Sauerwald, A., Toogood, H., Hartsch, T. & So¨ ll, D. (2002) Cystein activation is an inherent in vitro property of prolyl-tRNA synthetases. J. Biol. Chem. 277, 34743–34748.

20. Ambrogelly, A., Ahel, I., Polycarpo, C., Bunjun-Srihari, S., Krett, B., Jacquin-Becker, C., Ruan, B., Koehrer, C., Stathopoulos, C., (2002) Methanocaldococcus RajBhandary, U.L. & So¨ ll, D. jannaschii prolyl-tRNA synthetase charges tRNAPro with cysteine. J. Biol. Chem. 277, 34749–34784. We thank Sinisa Stipanicic for recording ESI-MS spectrum of the threonine sample and Tomislav Kamenski and Marko Mocibob for assistance in complex formation studies and Ferguson plot analysis. We are indebted to Professor Kucan for critical discussions. This work was supported by grants from International Centre for Genetic Engineering and Biotechnology, Trieste, the Ministry of Science and Technology of the Republic of Croatia, and National Institutes of Health (NIH/ FIRCA).

21. Jakubowski, H. (1993) Energy cost of proofreading in vivo: the charging of methionine tRNAs in Escherichia coli. FASEB J. 15, 168–172.

R E F E R E N C E S

1. Ibba, M. & So¨ ll, D. (1999) Quality control mechanisms during translation. Science 286, 1893–1897.

22. Hong, K.-W., Ibba, M., Weygand-Durasevic, I., Thomann, H.-U. & So¨ ll, D. (1996) Transfer RNA-dependent cognate amino acid recognition by an aminoacyl-tRNA synthetase. EMBO J. 15, 1983–1991. 2. Jakubowski, H. & Goldman, E. (1992) Editing of errors in selection of amino acids for protein synthesis. Microbiol. Rev. 56, 412–429. 3. First, E.A. (1998) Comprehensive Biological Catalysis (Sinnott, M. L., ed), pp. 573–607. Academic Press, London, UK.

23. Ibba, M., Hong, K.-W., Sherman, J.M., Sever, S. & So¨ ll, D. (1996) Interactions between tRNA identity nucleotides and their recognition sites in glutaminyl-tRNA synthetase determine the cognate amino acid affinity of the enzyme. Proc. Natl Acad. Sci. USA 93, 6953–6958. 4. Silvian, L.F., Wang, J. & Steitz, T.A. (1999) Insights into editing from an Ile-tRNA synthetase structure with tRNAIle and mupir- ocin. Science 285, 1074–1077.

tRNA-dependent amino acid discrimination (Eur. J. Biochem. 269) 5279

(cid:2) FEBS 2002

24. Ibba, M., Sever, S., Praetorius-Ibba, M. & So¨ ll, D. (1999) Transfer RNA identity contributes to transition state stabilization during aminoacyl-tRNA synthesis. Nucleic Acids Res. 27, 3631–3637.

38. 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. 39. Watson, T.G. (1976) Amino-acid pool composition of Saccharo- myces cerevisiae as a function of growth rate and amino-acid nitrogen source. J. Gen. Microbiol. 96, 263–268. 25. Deinert, K., Fasiolo, F., Hurt, E.C. & Simos, G. (2001) Arc1p organizes the yeast aminoacyl-tRNA complex and stabilizes its interaction with the cognate tRNAs. J. Biol. Chem. 276, 6000– 6008. 40. Schmidt, E. & Schimmel, P. (1994) Mutational isolation of a sieve for editing in a transfer RNA synthetase. Science 264, 265–267.

41. Nomanbhoy, T.K. & Schimmel, P. (2000) Misactivated amino acids translocate at similar rates across surface of a tRNA syn- thetase. Proc. Natl Acad. Sci. USA 97, 5119–5122.

42. Weygand-Durasevic, I., Lenhard, B., Filipic, S. & So¨ ll, D. (1996) The C-terminal extension of yeast seryl-tRNA synthetase affects stability of the enzyme and its substrate affinity. J. Biol. Chem. 271, 2455–2461. 43. Bryan, J.K. (1977) Molecular weights of protein multimers from polyacrylamide gel electrophoresis. Anal. Biochem. 78, 513–519. 26. Rocak, S., Landeka, I. & Weygand-Durasevic, I. (2002) Identi- fying Pex21p as a protein that specifically interacts with yeast seryl-tRNA synthetase. FEMS Microbiol. Lett. 214, 101–106. 27. Lenhard, B., Filipic, S., Landeka, I., Skrtic, I., So¨ ll, D. & Weygand- Durasevic, I. (1997) Defining the active site of yeast seryl-tRNA synthetase: mutations in motif 2 loop residues affect tRNA- dependent amino acid recognition. J. Biol. Chem. 272, 1136–1141. 28. Rokov-Plavec, J., Lesjak, S., Landeka, I., Mijakovic, I. & Weygand-Durasevic, I. (2002) Maize seryl-tRNA synthetase: specificity of substrate recognition by the organellar enzyme. Arch. Biochem. Biophys. 397, 40–50.

44. Biou, V., Yaremchuk, A., Tukalo, M. & Cusack, S. (1994) The 2.9 A˚ crystal structure of T. Thermophilus seryl-tRNA synthetase complexed with tRNASer. Science 263, 1404–1410.

29. Gruic-Sovulj, I., Luedemann, H.C., Hillenkamp, F., Weygand- Durasevic, I., Kucan, Z. & Peter-Katalinic, J. (1997) Detection of noncovalent tRNA: aminoacyl-tRNA synthetase complexes by matrix assisted laser desorption/ionization mass spectrometry. J. Biol. Chem. 272, 32084–32091. 45. Mouliner, L., Eiler, S., Eriani, G., Gangloff, J., Thierry, J.-C., Gabriel, K., McClain, W.H. & Moras, D. (2001) The structure of an AspRS-tRNAAsp complex reveals a tRNA-dependent control mechanism. EMBO J. 20, 5290–5301.

30. Gruic-Sovulj, I., Luedemann, H.-C., Hillenkamp, F., Weygand- Durasevic, I., Kucan, Z. & Peter-Katalinic, J. (1997) Matrix- assisted laser desorption/ionisation mass spectrometry of transfer ribonucleic acids isolated from yeast. Nucleic Acids Res. 25, 1859– 1861.

46. Belrhali, H., Yaremchuk, A., Tukalo, M., Larsen, K., Berthet- Colominas, C., Leberman, R., Beijer, B., Sproat, B., Als-Nielsen, J., Grubel, G., Legrand, J.-F., Lechmann, M. & Cusack, S. (1994) Crystal structures at 2.5 A˚ ngstrom resolution of seryl-tRNA synthetase complexed with two analogs of seryl adenylate. Science 263, 1432–1436.

31. Gruic-Sovulj, I., Weygand-Durasevic, I. & Kucan, Z. (2001) Influence of modified tRNATyr on the activation of tyrosine catalyzed by tyrosyl-tRNA synthetase from Saccharomyces cerevisiae. Croatica Chem. Acta. 74, 161–171. 47. Newberry, K.J., Hou, Y.-M. & Perona, J.J. (2002) Structural origins of amino acid selection without editing by cysteinyl-tRNA synthetase. EMBO J. 21, 2778–2787.

48. Krauss, G., Pingoud, A., Boehme, D., Riesner, D., Peters, F. & Maass, G. (1975) Equivalent and non-equivalent binding sites for tRNA on aminoacyl-tRNA synthetases. Eur. J. Biochem. 55, 517– 529. 32. Hountondji, C., Schmitter, J.-M., Beauvallet, C. & Blanquet, S. (1990) Mapping of the active site of Escherichia coli methionyl- tRNA synthetase: identification of amino acid residues labeled by periodate-oxidized tRNAfMet molecules having modified lengths at the 3¢-acceptor end. Biochemistry 29, 8190–8198. 49. Rigler, R. & Pachmann, U.

(1976) On the interaction of seryl-tRNA synthetase with tRNASer. A contribution to the problem of synthetase-tRNA recognition. Eur. J. Biochem. 65, 307–315. 33. Madore, E., Lipman, R.S.A., Hou, Y.-M. & Lapointe, J. (2000) Evidence for unfolding of the single-stranded GCCA-3¢-end of a tRNA on its aminoacyl-tRNA synthetase from a stacked helical to a foldback conformation. Biochemistry 39, 6791–6798.

34. Orchard, K. & May, G.E. (1993) An EMSA-based method for determining the molecular weight of a protein – DNA complex. Nucleic Acids Res. 21, 3335–3336.

50. Yaremchuk, A.D., Tukalo, M.A., Krikliviy, I.A., Malchenko, N., Biou, V., Berthet-Colominas, C. & Cusack, S. (1992) A new crystal form of the complex between seryl-tRNA synthetase and tRNASer from Thermus Thermophilus that diffracts to 2.8 A˚ resolution. FEBS Lett. 310, 157–161. 35. Lenhard, B., Orellana, O., Ibba, M. & Weygand-Durasevic, I. (1999) tRNA recognition and evolution of determinants in seryl- tRNA synthesis. Nucleic Acids Res. 27, 721–729.

implication of

36. Weygand-Durasevic, I., Ban, N., Jahn, D. & So¨ ll, D. (1993) Yeast seryl-tRNA synthetase expressed in E. coli recognizes bacterial serine-specific tRNAs in vivo. Eur. J. Biochem. 214, 869–877. 37. Agou, F., Quevillon, S., Kerjan, P. & Mirande, M. (1998) Switching the amino acid specificity of an aminoacyl-tRNA synthetase. Biochemistry 37, 11309–11314. 51. Price, S., Cusack, S., Borel, F., Berthet-Colominas, C. & Leber- man, R. (1993) Crystallization of the seryl-tRNA synthetase: tRNASer complex of Escherichia coli. FEBS Lett. 324, 167–170. 52. Borel, F., Vincent, C., Leberman, R. & Ha¨ rtlein, M. (1994) Seryl- tRNA synthetase from Escherichia coli: its N-terminal domain in aminoacylation activity and specificity. Nucleic Acids Res. 22, 2963–2969.