Báo cáo khoa học: Kinetic and mechanistic characterization of Mycobacterium tuberculosis glutamyl–tRNA synthetase and determination of its oligomeric structure in solution

Kinetic and mechanistic characterization of Mycobacterium tuberculosis glutamyl–tRNA synthetase and determination of its oligomeric structure in solution Stefano Paravisi1, Gianluca Fumagalli1, Milena Riva1, Paola Morandi1, Rachele Morosi1, Peter V. Konarev2,3, Maxim V. Petoukhov2,3, Ste´ phane Bernier4, Robert Cheˆ nevert4, Dmitri I. Svergun2,3, Bruno Curti1 and Maria A. Vanoni1

1 Dipartimento di Scienze Biomolecolari e Biotecnologie, Universita` degli Studi di Milano, Italy 2 European Molecular Biology Laboratory, Hamburg, Germany 3 Institute of Crystallography, Russian Academy of Sciences, Moscow, Russia 4 De´ partment de Chimie, CREFSIP, Universite´ Laval, Canada

Keywords glutamyl–tRNA reductase; glutamyl–tRNA synthetase; Mycobacterium tuberculosis; protein synthesis; tetrapyrrole synthesis

Correspondence M. A. Vanoni, Dipartimento di Scienze Biomolecolari e Biotecnologie, Universita’ degli Studi di Milano, Via Celoria 26, 20131 Milan, Italy Fax: +39 025 031 4895 Tel: +39 025 031 4901 E-mail: maria.vanoni@unimi.it

(Received 15 October 2008, revised 23 December 2008, accepted 24 December 2008)

doi:10.1111/j.1742-4658.2009.06880.x

glutamyl–tRNA synthetase

Abbreviations aaRS, aminoacyl–tRNA synthetases; ALA, d-amino levulinic acid; ALAS, d-amino levulinic acid synthase; ArgRS, arginyl–tRNA synthetase; Bs-GluRS, Bacillus subtilis glutamyl–tRNA synthetase; D-GluRS, discriminating glutamyl–tRNA synthetase; DLS, dynamic light scattering; E, total enzyme concentration; Ec-GluRS, Escherichia coli glutamyl–tRNA synthetase; GlnRS, glutaminyl–tRNA synthetase; GluRS, glutamyl– tRNA synthetase; GluTR, glutamyl–tRNA reductase; GluTR-His, GluTR carrying a C-terminal His6-tag; GoA, glutamol-AMP; GSA, glutamate 1-semialdehyde; GSA-AM, GSA aminomutase; His6-GluRS, GluRS carrying a N-terminal His6-tag; IPTG, isopropyl thio-b-D-galactoside; LysRS, lysyl–tRNA synthetase; Mt-GluRS, M. tuberculosis GluRS; ND-GluRS, nondiscriminating GluRS; PPi, pyrophosphate; SAXS, small angle X-ray scattering; Te-GluRS, Thermosynechococcus elongatus GluRS; Tt-GluRS, Thermus thermophilus GluRS; b-ME, 2-mercaptoethanol.

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1398

Mycobacterium tuberculosis (Mt-GluRS), encoded by Rv2992c, was overproduced in Escherichia coli cells, and puri- fied to homogeneity. It was found to be similar to the other well-character- ized GluRS, especially the E. coli enzyme, with respect to the requirement for bound tRNAGlu to produce the glutamyl-AMP intermediate, and the steady-state kinetic parameters kcat (130 min)1) and KM for tRNA (0.7 lm) and ATP (78 lm), but to differ by a one order of magnitude higher KM value for l-Glu (2.7 mm). At variance with the E. coli enzyme, among the several compounds tested as inhibitors, only pyrophosphate and the glut- amyl-AMP analog glutamol-AMP were effective, with Ki values in the lm range. The observed inhibition patterns are consistent with a random bind- ing of ATP and l-Glu to the enzyme–tRNA complex. Mt-GluRS, which is predicted by genome analysis to be of the non-discriminating type, was not toxic when overproduced in E. coli cells indicating that it does not catalyse the mischarging of E. coli tRNAGln with l-Glu and that GluRS ⁄ tRNAGln recognition is species specific. Mt-GluRS was significantly more sensitive than the E. coli form to tryptic and chymotryptic limited proteolysis. For both enzymes chymotrypsin-sensitive sites were found in the predicted tRNA stem contact domain next to the ATP binding site. Mt-GluRS, but not Ec-GluRS, was fully protected from proteolysis by ATP and glutamol- AMP. Small-angle X-ray scattering showed that, at variance with the E. coli enzyme that is strictly monomeric, the Mt-GluRS monomer is pres- ent in solution in equilibrium with the homodimer. The monomer prevails at low protein concentrations and is stabilized by ATP but not by gluta- mol-AMP. Inspection of small-angle X-ray scattering-based models of Mt-GluRS reveals that both the monomer and the dimer are catalytically

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

active. By using affinity chromatography and His6-tagged forms of either GluRS or glutamyl–tRNA reductase as the bait it was shown that the M. tuberculosis proteins can form a complex, which may control the flux of Glu–tRNAGlu toward protein or tetrapyrrole biosynthesis.

Mycobacterium tuberculosis infects over two-thirds of the world population and causes 1.6 million deaths every year, according to World Health Organization estimates [1]. The intrinsic resistance of M. tuberculo- sis to most antibiotics and the spread of multidrug- resistant strains prompted the study of M. tuberculosis metabolism and the identification of novel anti- tubercular drug targets through the in vitro character- ization of essential enzymes. With this goal in mind we focused on the production and characterization of M. tuberculosis glutamyl–tRNA synthetase (Mt- GluRS).

L-Glu + ATP + tRNAGlu $ Glu (cid:2) tRNAGlu

amidotransferase that converts the glutamyl moiety into a glutaminyl residue correcting the misacylation and providing the Gln–tRNAGln needed for protein synthesis [2,6,7]. In these cells, the GlnRS is missing. Furthermore, in most bacteria and plants GluRS also plays a role in tetrapyrrole biosynthesis, which requires GluRS (Eqn 3), Glu–tRNA reductase (GluTR; Eqn 4) and glutamate 1-semialdehyde aminomutase (GSA- AM; Eqn 5) for synthesis of d-aminolevulinic acid (ALA), the first common precursor of all tetrapyrroles [8,9]. This C5 pathway of tetrapyrrole biosynthesis dif- fers from that of most eukaryotes and other bacteria, which uses succinyl-CoA, glycine and ALA synthase (ALAS; Eqn 6), the so-called C4 pathway of tetrapyr- role biosynthesis.

ð3Þ

Glu (cid:2) tRNAGlu + NADPH + Hþ $ GSA + NADPþ ð4Þ

GSA $ ALA ð5Þ the activation of succinyl-CoA + glycine $ ALA + CoA ð6Þ

Glutamyl–tRNA synthetases (GluRS) belong to the broad class of aminoacyl–tRNA synthetases (aaRS), which catalyse the essential charging reaction of tRNA with the cognate amino acid ensuring correct transla- tion of the mRNA into the corresponding polypeptide [2]. The ubiquity and essentiality of aaRS makes them of interest as targets of new anti-infectives [3]. Their reaction formally consists of the amino acid by adenylation (Eqn 1) followed by transfer of the amino acyl residue to the 2¢-OH or 3¢-OH posi- tion of the 3¢-OH end of the cognate tRNA (Eqn 2).

amino acid þ ATP $ aminoacyl-AMP + pyrophosphate ð1Þ

interfering with synthesis protein amino acyl-AMP + tRNAaa $ AMP + aminoacyl (cid:2) tRNAaa ð2Þ

[10]. As isoforms

How the flux of Glu–tRNAGlu is directed toward protein or tetrapyrrole biosynthesis has not been fully clarified. Most likely, different mechanisms operate in different organisms. the low levels of In general, GluTR, catalysing the rate-limiting step of ALA biosynthesis, may be sufficient to ensure ALA supply without [9]. However, GluTR may distinguish between different Glu–tRNAGlu an alternative, complex formation between GluRS and GluTR, as a function of the cell requests, may divert Glu–tRNAGlu toward tetrapyrrole biosynthesis [11]. Finally, GluRS isoforms differing in tRNAGlu specificity [10] or, in principle, in their ability to interact with GluTR may be expressed.

is (Tt-GluRS) structural model

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1399

Most aaRS catalyse the formation of the aminoacyl- AMP intermediate in the absence of tRNA. However, GluRS, glutaminyl–tRNA synthetase (GlnRS), argi- nyl–tRNA synthetase (ArgRS) and class I lysyl–tRNA synthetase (LysRS) are exceptions in that activation of the amino acid requires the presence of the cognate tRNA [2,4]. In these aaRS the binding of tRNA induces an ATP productive binding mode [5]. GluRS are also distinguished on the basis of their ability to discriminate between tRNAGlu and tRNAGln . The dis- criminating GluRS (D-GluRS) only catalyses the tRNAGlu with l-Glu yielding charging reaction of Glu–tRNAGlu. However, the nondiscriminating GluRS (ND-GluRS) also charges the tRNAGln forming a mis- acylated Glu–tRNAGln. The organisms containing the ND-GluRS also contain a specific Glu–tRNAGln Only the structures of GluRS from thermophilic bacteria have been solved. The Thermus thermophilus for the enzyme D-GluRS [5,12–14], and the Thermosynechococcus elongatus form (Te-GluRS) is the structural model for the ND-GluRS class [15]. Thus, details of the struc- ture, flexibility, oligomeric state and conformational states of a mesophilic enzyme are not known, limiting the to some thorough understanding of extent

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

rational design of the for structure–function relationship in this enzyme with consequences specific inhibitors.

E. coli (Fig. S1A). Similar results were obtained with cells transformed with pETGTS2, which encodes a tag and the fusion between an N-terminal His6 Rv2992c coding region (510 residues and a predicted mass of 55 876 Da; Fig. S1B).

three ORFs

Analysis of the M. tuberculosis genome sequence led to the identification of one ORF encoding a putative GluRS (Rv2992c, 1473 bp). Upstream of Rv2992c the sequences of one tRNAGlu (gluU) and one tRNAGln (glnU) gene are found. Additional tRNAGlu (gluT) and tRNAGln (glnT) genes have been annotated, but they are in chromosome regions far from the putative GluRS gene and the tetrapyrrole biosynthetic genes (see below). No ORF encoding a putative GlnRS was (Rv3011c, Rv3009c and found, but Rv3012c) predict the presence of the three subunits of the Glu–tRNAGln amidotransferase. These observa- tions suggest that the Mt-GluRS is of the nondiscrimi- nating type.

Finally, genes encoding the putative GluTR (hemA, Rv0509), GSA-AM (hemL, Rv0524) and other enzymes of the tetrapyrrole biosynthetic pathway have been annotated in the M. tuberculosis genome. No ORF encoding proteins similar to ALAS have been found. Thus, Mt-GluRS is predicted to provide Glu– tRNAGlu for both protein and tetrapyrrole biosynthe- sis, the latter occurring via the C5 pathway. Furthermore, indicate that [16]

Up to 100 mg of homogeneous protein, as judged by SDS ⁄ PAGE (Fig. S1A), were obtained from 20 g of E. coli BL21 (DE3) cells harboring pETGTS1. The the putative GluRS (His6– His6-tagged variant of GluRS) could be purified to homogeneity (Fig. S1B) using a single nitrilotriacetic acid–Sepharose column ((cid:3) 10 mgÆg)1 of cells). Both protein species could be concentrated to up to 40 mgÆmL)1 without observable precipitation. They were stable for up to 2 years when stored at )20 (cid:2)C in 50% glycerol, as judged by SDS ⁄ PAGE, determination of the protein concentra- tion after centrifugation, activity (see below) and dynamic light scattering (DLS) measurements. Glyc- erol removal by either dialysis or gel filtration led to soluble protein that maintained activity for up to 1 week when stored at 4 (cid:2)C. Freezing samples from which glycerol had been removed caused the aggrega- tion of a small fraction (< 5%) of the protein, as determined by DLS without, however, causing detect- able activity loss. N-Terminal sequencing and mass determination by MALDI-TOF confirmed the identity of the proteins and that the N-terminal Met residue had been correctly removed by post-translational processing. the genome-wide gene inactivation experiments of Sassetti et al. the putative GluRS, as well as glutamyl–tRNAGln amido- transferase and the enzymes of the C5 pathway of ALA biosynthesis are essential for M. tuberculosis.

Identification of Rv2992c as the Mt-GluRS and steady-state kinetic characterization

in order

For these reasons, with the dual goal of contributing to understanding of the metabolism of this pathogen and providing the enzyme for the identification and development of selective inhibitors, we cloned and expressed Rv2992c in E. coli. With the purified pro- tein we carried out a kinetic, mechanistic and struc- the resulting Mt-GluRS. tural characterization of Rv0509, encoding the putative M. tuberculosis GluTR (Mt-GluTR) was also cloned in vectors for protein production in E. coli to ask questions about GluRS–GluTR complex formation for the mycobacterial proteins.

Results

Formation of Glu–tRNAGlu was monitored by mea- suring the increase in acid-precipitable radioactivity upon incubation of the enzyme with E. coli tRNAGlu, l-[U14C]Glu, ATP, MgCl2 at pH 7.3. The increase in l-[U14C]Glu–tRNAGlu concentration was linear for up to 10 min when 0.2–1.85 pmol Mt-GluRS was used in the reactions (Fig. S2). Under these conditions the enzyme had an apparent turnover number of 17.7 ± 0.3 min)1 at 0.5 mm l-Glu and 32.0 ± 1.2 min)1 at 2 mm l-Glu. Similar activity was measured with the homogeneous His6-GluRS form (Fig. S2). These values are lower than that of (cid:3) 100 min)1 calculated from the specific activity reported for the E. coli enzyme by Lin et al. [17]. Expression of Rv2992c in E. coli BL21(DE3) and purification of the putative Mt-GluRS

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1400

The predicted ORF Rv2992c was cloned in pET-based vectors for production of the corresponding protein product in E. coli BL21(DE3). The pETGTS1 plasmid coded for a 490-residue protein (53 831 Da), which was produced at high levels and in a soluble form in The activity was found to increase hyperbolically with MgCl2 concentrations up to 5 mm. At concentra- tions > 10 mm the activity decreased. Thus, in all assays MgCl2 concentration was held constant at 10 mm, well above ATP concentrations (or its analogs, see below), but below the onset of inhibition.

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

Table 1. Steady-state kinetic parameters of the Mt-GluRS reaction. The kcat and KM values for ATP, L-Glu and E. coli tRNAGlu were deter- mined for the aminoacylation reaction catalysed by Mt-GluRS (6.3 nM) at 37 (cid:2)C in the presence of 35 mM Hepes ⁄ NaOH buffer, pH 7.3, 25 mM KCl, 10% glycerol, 2 mM dithiothreitol, 10 mM MgCl2 and 0.1% BSA and the indicated concentrations or concentration ranges of the enzyme substrates. For comparison, published KM values for ATP, L-Glu and tRNAGlu for the E. coli enzyme are shown [18].

ATP (mM)

L-Glu (mM)

tRNAGlu (lM)

kcat (min)1)

KM

Mt-GluRS

0.5 0.03–2.0 0.5

3.6 3.6 0.45–4.0

16.3 ± 0.4 129.0 ± 28.0 24.5 ± 1.5

0.08 ± 0.01 mM 2.7 ± 0.8 mM 0.7 ± 0.2 lM

Ec-GluRS

0.01–2.0 1.0 1.0 Varied

Varied

Varied

0.25 mM 0.10 mM 0.16 lM

1000

100

10 10

t a c k

1

0.1

Determination of

100

l

u G - L K

/

10

t a c k

the apparent kcat and KM for ATP, l-Glu (Kl-Glu) and tRNAGlu was carried out at 37 (cid:2)C under conditions detailed in Materials and meth- ods and in the legend to Table 1. The quality of data is shown in Fig. S3. KM values for ATP and tRNA are of the same order of magnitude as those reported for E. coli glutamyl–tRNA synthetase (Ec-GluRS) [18], which we here use as the reference GluRS. The Kl-Glu value is very high so that it remains poorly defined and it may be ‡ 2–3 mm, i.e. at least (cid:3) 20-fold higher than the corresponding value reported for Ec-GluRS.

1

6 6

7 7

8 8

9 9

pH

reported that The kcat extrapolated at infinite l-Glu concentration at saturating concentrations of the other substrates (129 ± 28 min)1), within the limits imposed by the high value of the KL-Glu that prevents an accurate estimate of this parameter, is now of the same order for Ec-GluRS of of magnitude ((cid:3) 100 min)1) [17].

Fig. 1. pH dependence of the steady-state kinetic parameters kcat and kcat ⁄ KL-Glu of Mt-GluRS. The apparent kcat (in min)1) and kcat ⁄ KL- Glu (in min)1ÆmM )1) values of the reaction catalysed by Mt-GluRS (6.3 nM) were determined at 37 (cid:2)C in 35 mM Hepes ⁄ NaOH buffer at the indicated pH values in the presence of 1 mM ATP, 3.6 lM tRNAGlu, 10 mM MgCl2, 25 mM KCl, 2 mM dithiothreitol, 10% glyc- erol 0.1% BSA and varying L-[U14C]Glu. kcat values were fitted to Eqn (13), assuming that kcat increases to a limiting value of 110 ± 1.0 min)1 at high pH as a single group with pKa of 6.2 ± 0.03 deprotonates. kcat ⁄ KL-Glu values fitted well fitted with Eqn (14) assuming that the parameter decreases from a limiting value of 70 ± 5.0 min)1ÆmM )1 as a group with a pKa value of 8.7 ± 0.23 deprotonates.

The high Kl-Glu value of Mt-GluRS, compared with Ec-GluRS, did not depend on the pH at which the activity assays were carried out. Indeed, kcat and Kl-Glu values were determined between pH 6.5 and 8.5 in the presence of fixed concentrations of the other substrates (Fig. 1). kcat values were found to increase as a group with an apparent pKa < 7 dissociated to reach a con- stant value above pH 7.3. The kcat ⁄ Kl-Glu profile instead showed a plateau at pH values between 6.5 and 7.5 and decreased at high pH as a group with a pKa value > 8 deprotonated.

Because of the high cost of l-[U14C]Glu and the need to maintain the ionic strength of the assay rela- tively low, enzyme activity was routinely measured in the presence of 0.5 or 2 mm l-Glu.

Alternate substrates and inhibitors of Mt-GluRS

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1401

Several analogs of the enzyme substrates were tested as alternate substrates or inhibitors of Mt-GluRS (Tables S1 and S2). Mt-GluRS was found to be very specific for the amino acid substrate. l-Gln (2–5 mm) and 2-oxoglutarate (1–5 mm) did not inhibit the reac- tion (Table S1). Furthermore, l-Gln could not effi- ciently substitute for l-Glu as the substrate. Indeed, in the presence of 2 mm l-Gln the apparent turnover number was 0.03 min)1, i.e. 0.1% of that measured in the presence of 2 mm l-Glu (Table S1). Mt-GluRS is

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

4

M µ

]

A N R

l

2

t - u G

[ ;

M µ

, ]

also highly specific for the nucleotide substrate. ATP could not be substituted as the substrate by b,c-methyl- ene-ATP, despite the presence of the hydrolysable a,b-phoshoanhydride bond. a,b-Methylene-ATP and b,c-methylene-ATP were not inhibitors of the reaction (Table S2) nor were AMP and its analog decoyinine [19]. For comparison a,b-methylene-ATP was found to be an inhibitor of Ec-GluRS, competitive with ATP (Ki (cid:3) 0.45 mm) [20]. With Ec-GluRS AMP was a non- competitive inhibitor with respect to ATP and l-Glu with Ki values in the mm range, as deduced by the data presented in Kern and Lapointe [20]. On the contrary, the glutamyl-AMP analog glutamol-AMP (GoA) [21] and pyrophosphate (PPi) (but not a series of PPi ana- logs) were potent inhibitors of Mt-GluRS (Table S2). Several divalent cations were also tested as substitutes for Mg2+ or inhibitors. None could replace Mg2+ in the reaction, and they all acted as mild inhibitors (Table S3). in order to form the aminoacyl-AMP intermediate from ATP and the free amino acid [2,4]. To establish the requirement of tRNA for the aminoacyl–adenyla- tion reaction, Mt-GluRS was incubated with either l-[14C]Glu or [3H]ATP under various conditions, and the reaction components were identified and quantified after chromatographic separation on a MonoQ column (Fig. S5). Only in the presence of both tRNA and l-Glu was the appearance of [3H]AMP observed. That the radioactivity associated with the elution position of AMP did not correspond to Glu-AMP was tested by carrying out the same experiments in the presence of l-[U14C]Glu (not shown). The kinetics of [3H]AMP formation were also determined (Fig. 2). The amount of [3H]AMP formed at early incubation times matched well with that of l-[U14C]Glu–tRNAGlu formed in par- allel filter-binding assays. At later reaction times the amount of AMP formed exceeded that of Glu–tRNA, presumably due to recycling of tRNA derived from

P M A

[

0 0

The inhibitory effect of PPi and GoA was studied in greater detail. PPi was found to be a noncompetitive inhibitor with respect to both ATP and l-Glu with Ki values in the 10–100 lm range (Table 2 and Fig. S4). GoA was a competitive inhibitor with respect to both l-Glu and ATP with Ki values of (cid:3) 4 and 1.5 lm, respectively. These values are similar to those reported for the Ec-GluRS ((cid:3) 3 lm with respect to both sub- strates) [21]. GoA was instead uncompetitive with respect to tRNA (Ki (cid:3) 4 lm).

0

5

15

10 Time (min)

Requirement of tRNA for the adenylation of l-Glu in MtGluRS

Table 2. Inhibition of Mt-GluRS by glutamol-AMP and pyrophos- phate. Activity assays were carried out at 37 (cid:2)C in 35 mM Hepes ⁄ NaOH, pH 7.3, in the presence of 2 mM dithiothreitol, 10% glycerol, 0.1% BSA, 10 mM MgCl2, 25 mM KCl, Mt-GluRS (6.3 nM in the 150 lL assay mixture). When the substrate concentrations were held constant they were: 1 mM ATP, 0.5 mM L-Glu and 3.6 lM tRNA. The inhibition pattern was established throught the best fit of the data to Eqns (10–12) describing competitive (C), noncompeti- tive (NC) and uncompetitive (UC) inhibition, respectively.

Inhibitor

Varied substrate

Pattern

Kis (lM)

Kii (lM)

Glutamol-AMP

1.5 ± 0.4 3.9 ± 1.0

Pyrophosphate

3.9 ± 0.7 27 ± 15.4

ATP L-Glu tRNA ATP L-Glu

C C UC NC NC

31.4 ± 7.6 12.5 ± 3.7

101 ± 54

Fig. 2. Kinetics of [3H]AMP formation during Mt-GluRS reaction as determined by chromatographic separation of the reaction com- ponents. GluRS (0.67 lM) was incubated at 37 (cid:2)C in 35 mM Hepes ⁄ NaOH buffer, pH 7.3, 10% glycerol, 2 mM dithiothreitol, 1 mM [2,5¢,83H]ATP (33 300 dpmÆnmol)1), 10 mM MgCl2, 25 mM KCl, 2 mM L-Glu and 3.6 lM tRNAGlu (s) in a final volume of 150 lL. After 1–10 min, cold water was added (2 mL) and a 2 mL sample was rapidly injected onto a MonoQ column equilibrated in 20 mM tri- ethanolamine ⁄ HCl buffer, pH 7.7, and developed by increasing the KCl concentration in the same buffer. Fractions (1 mL) were col- lected and the radioactivity was measured by scintillation counting. The concentration of [3H]AMP formed in the assays at any given time was calculated from the amount of radioactivity present in the In separate experiments the time- AMP elution peak (Fig. S5). course of [3H]AMP formation in assays lacking L-Glu (d) or L-Glu and tRNAGlu (h) was also measured. The kinetics of Glu–tRNAGlu forma- ) determined using the filter-binding assay in the presence of tion ( L-[U14C]Glu and unlabelled ATP under identical conditions is shown for comparison. Note that at long incubation times, the amount of AMP formed exceeds that of Glu–tRNA due to recycling of tRNA because of the spontaneous hydrolysis of Glu–tRNA.

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1402

together with GlnRS, ArgRS and class I GluRS, LysRS, are the only aaRS that require bound tRNA

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

25

A

20

15

) M µ (

10

P M A

5

0

–5

25

B

20

15

) M µ (

10

P D A

5

0

–5

[3H]AMP and [3H]ADP formation from ATP. Fig. 3. Kinetics of Assays were set up in 35 mM Hepes ⁄ NaOH buffer, pH 7.3, 10% glycerol, 2 mM dithiothreitol, 1 mM [2,5¢,83H]ATP (33 300 dpmÆ- nmol)1), 10 mM MgCl2, 25 mM KCl, 2 mM L-Glu, 0.004% BSA and 3.6 lM tRNAGlu in a final volume of 150 lL and incubated at 37 (cid:2)C in the presence of different Mt-GluRS concentrations [6.7 nM (cir- cles), 76.3 nM (squares) and 822 nM (triangles)]. At different times, 10 lL aliquots were rapidly applied onto poly(ethyleneimine)–cellu- lose sheets, subjected to TLC and quantification of radiolabelled ADP and AMP. The kinetics of [3H]AMP and [3H]ADP formation in tRNAGlu (black symbols) or of both L-Glu and the absence of tRNAGlu (grey symbols) were also determined in parallel samples. (A) Time-course of AMP formation in the complete assay mixture at the three different enzyme concentrations (open symbols). In the absence of tRNAGlu no AMP formation above background was detected (closed symbols). Similar results were obtained in the absence of both tRNAGlu and L-Glu (not shown). Note that the reac- tion velocity is independent of the Mt-GluRS concentration because the formation of AMP is monitored at long incubation times and tRNA with high Mt-GluRS concentrations when recharging of derived from hydrolysis of Glu–tRNAGlu is being observed. (B, C) At increasing Mt-GluRS concentrations, formation of ADP from ATP could be detected at rates that were essentially independent from the presence of the enzyme substrates. The time-course of ADP formation in the complete assay mixture (not shown) was similar to that obtained in the absence of tRNAGlu (B).

25

C

20

15

) M µ (

10

P D A

5

0

–5

0

60

180

120 Time (min)

[32P]PPi ⁄ ATP exchange

increasing concentrations of

by studying the reaction [2,4,23,24]. Thus, Mt-GluRS was incubated under vari- ous conditions with [32P]PPi and the reaction compo- nents were separated by TLC. [32P] associated with the various compounds was quantified using a phosphoim- ager (Fig. S6). No radioactive species other than PPi and minor amounts of Pi were observed when the enzyme was incubated in solutions lacking one of the three substrates were enzyme substrates. When all present, only formation of [32P]ATP was observed. In kinetic experiments, the rate of formation of [32P]ATP [32P]PPi increased at (Fig. 4). As expected from the observed inhibitory effect of PPi on the tRNA charging reaction, the velo- city of [32P]ATP formation was inversely proportional to that of l-[U14C]Glu–tRNA production in parallel filter-binding assays (not shown). The kcat of [32P]ATP formation was 1100 ± 174 min)1. This value should be compared with that calculated for the tRNAGlu charging reaction under similar conditions (90 min)1).

spontaneous hydrolysis of Glu–tRNA [22]. These results were confirmed by separating the reaction com- ponents using TLC (Fig. 3A). In these experiments, we also observed ADP formation (Fig. 3B,C) at a rate that was dependent on the enzyme concentration, but independent of the presence of l-Glu and tRNA. This ATP hydrolysing activity is very low (0.135 min)1 when calculated from the kinetics of ADP formation tRNA with 0.8 lm Mt-GluRS) in the absence of (Fig. 3B) and it is unlikely to represent a physiologi- cally relevant side reaction. Is Mt-GluRS a discriminating or a non-discriminating GluRS?

PPi ⁄ ATP exchange reaction of Mt-GluRS

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1403

Evidence for the presence of the Glu-AMP intermedi- ate in GluRS and in other aaRS has been obtained According to analyses of the M. tuberculosis genome, Mt-GluRS is predicted to be of the nondiscriminating type (see above for details). Attempts to produce tRNAGlu and tRNAGln (in vivo or M. tuberculosis

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

800 800

2 2

B

A

) l o m n (

i

) 1 – n m

400

1

(

]

production in both Luria–Bertani (Fig. S1) and M9 medium (not shown).

E / v

P T A A P 2 3 [

0

0 0

6 6

0 0.0

0.5

1.0

4 2 4 2 Time (min)

[PPi] (mM)

Structural studies on Mt-GluRS in solution

ligands) we carried out

Fig. 4. Kinetic parameters of PPi ⁄ ATP exchange reaction. The time-course of the incorporation of [32P]PPi into ATP was deter- mined as a function of PPi concentration (empty circle, 0.01 mM; full circle, 0.1 mM; empty square, 0.5 mM; full square, 1.0 mM) in 100 mM Hepes ⁄ NaOH buffer, pH 7.2, containing 0.3% glycerol, 2 mM ATP, 16 mM MgCl2, 6 mM L-Glu, 25 mM KCl, 0.01–1 mM (231 113 dpmÆnmol)1), 3.6 lM tRNAGlu and Mt-GluRS Na-[32P]PPi (7.4 nM) in a final volume of 50 lL, at 37 (cid:2)C. Aliquots (1 lL) of the reaction mixtures were withdrawn before and at different times after tRNA addition. They were rapidly applied onto the poly(ethyl- eneimine)–cellulose sheets, which were developed immediately. After quantitation of [32P]ATP formed at the different times (left), calculation of the initial reaction velocity and correction for the amount of enzyme present, the rates of [32P]ATP formation at the different PPi concentrations were fitted to the Michaelis–Menten equation to calculate the apparent kcat (1101 ± 174 min)1) and KM for PPi (0.4 ± 0.17 mM) of the reaction (right). For comparison, the velocity of Glu–tRNA formation under similar substrate concentra- tions was calculated to be 89 min)1.

Several unsuccessful attempts were made to obtain crystals of the protein for X-ray diffraction studies. With the aim of gathering structural information (i.e. oligomeric structure, conformational flexibility, effect of limited proteolysis and small-angle X-ray scattering (SAXS) measurements on Mt-GluRS, using the Ec-GluRS species as a reference protein and the available high-resolution structures of Tt-GluRS [5,12,13] and Te-GluRS [15] as models.

of proteolytic indicated that

(to establish the toxicity of Both Mt- and Ec-GluRS were found to be more sen- sitive to trypsin than to chymotrypsin, with Mt-GluRS being significantly more sensitive than Ec-GluRS to the given protease. Incubation of Mt-GluRS with chymotrypsin 0.1% (w ⁄ w) led to the formation of a limited number of protein fragments with five main species (bands M1-M5 of Fig. 5), of which M2 (27.5 kDa), M4 (18.4 kDa) and M5 ((cid:3) 9 kDa) were stable to further proteolytic attack. From the N-termi- nal sequence and the mass of the fragments, and from the kinetics of the process (Figs S7–S9 and Table S5) we concluded that the main sites of proteolytic cleav- age are at the C-terminus of the predicted catalytic domain and in the stem-contact domain [5,12,13,15]. By projecting these cleavage sites on the Tt-GluRS structures available, they are found to be close to the shown). Accordingly, GoA ATP-binding site (not fully protected Mt- (Fig. 5) and ATP (not shown) GluRS from chymotryptic degradation. Interestingly, MgCl2 was not required for the binding of these nucle- otides to GluRS. Ec-GluRS was less sensitive than Mt-GluRS to chymotrypsin and a limited number of fragments could be observed using 1% chymotrypsin fragments the (Fig. 5). Analysis (Figs S7–S9 and Table S5) the main proteolytic site in Ec-GluRS is S238. From sequence comparisons (Fig. S7), this residue is in the ‘KMSK’ fingerprint of GluRS, which identifies the ATP-binding site [2]. At variance with the Mt-GluRS, GoA and ATP had no effect on the proteolytic pattern observed with Ec-GluRS. l-Glu did not have an effect on prote- olysis with any of the enzymes.

and revealing

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1404

in vitro) in quantities sufficient to carry out kinetic assays have not been successful, yet. Thus, in order to study the discriminating or nondiscriminating nature of Mt-GluRS we used the toxicity test developed by Baick et al. [25]. Overproduction of the nondiscrimi- nating Bacillus subtilis GluRS (Bs-GluRS) in E. coli cells, which lack the Glu–tRNAGln amidotransferase, was found to be toxic. Supplementing the medium with l-Gln protected the cells, presumably by allowing the endogenous GlnRS to saturate the tRNAGln with l-Gln, thus avoiding the misacylation reaction. Diluted cultures of E. coli BL21(DE3) cells containing pET- GTS1 were plated on Luria–Bertani or M9 medium (to measure the cells vitality) or medium containing ampicillin (to count cells containing the plasmid) in the absence or presence of isopropyl thio-b-d-galacto- side (IPTG) the over- effects of l-Gln production of Mt-GluRS). The (2.5–25 mm, to relieve toxicity) and of l-Glu (2.5– 25 mm, to enhance the hypothesized misacylation reac- tion of tRNAGln) were also tested. In none of the conditions (Table S4) was pETGTS1 toxic, nor was the induction of the gene expression. This agrees well with the fact that large amounts of soluble Mt-GluRS were produced in E. coli BL21(DE3) cells for protein In order to extract structural information, although at a low resolution, protein samples were analysed by SAXS. DLS provided an important set of information preliminary to the SAXS experiments, allowing us to that conditions establish working Ec- and Mt-GluRS likely differed for their aggregation for details). state (Fig. S10 and accompanying text

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

Fig. 5. SDS ⁄ PAGE analysis of E. coli (upper) and M. tuberculosis GluRS (lower) limited proteolysis products. The enzymes (1 mgÆmL)1) (w ⁄ w) Na-tosyl-L-lysyl chloromethyl ketone-treated chymotrypsin in 50 mM were incubated with 1% (Ec-GluRS) or 0.1% (Mt-GluRS) Hepes ⁄ NaOH buffer, pH 8.0, in the absence or presence of 0.5 mM GoA at 25 (cid:2)C in a final volume of 100–200 lL. At different incuba- tion times, aliquots (10 lL) were analysed by SDS ⁄ PAGE after chymotrypsin inactivation. The mass of the peptides was calculated by comparison with a calibration curve built with the 14–202 kDa molecular mass protein standard mix (*). E1 (45.3 kDa), E2 (30.3 kDa) and E3 (26.6 kDa) are the main proteolysis products obtained with Ec-GluRS (54.0 kDa). From N-terminal sequencing E2 corresponds to the N-terminal fragment of Ec-GluRS and E3 starts at position 238. M1 (28.9 kDa), M2 (27.5 kDa), M3 (23.4 kDa) and M4 (18.4 kDa) derive from Mt-GluRS (54.3 kDa). The N-terminal sequences of M1 and M2 corresponds to the N-terminus of intact Mt-GluRS. M4 starts at position 319.

the structure of

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1405

Ec-GluRS solutions (0.5–11 mgÆmL)1) yielded scatter- ing patterns consistent with the presence of one species in solution. The scattering curves computed from the atomic coordinates of the Tt- or Te-GluRS monomers by program crysol [26] yielded reasonable fits to the experimental patterns (Table S6 and Fig. 6A, curve 1). Mt-GluRS yielded more complex SAXS patterns (Fig. 6A, curves 2–16). The calculated radius of gyra- tion (Rg) and molecular mass (MM) increased at increasing protein concentration (Table S6), indicating the presence of multiple species in solution. At protein concentrations > 4 mgÆmL)1 the calculated Rg stabi- lized at 3.8 nm suggesting the presence of a dimeric (Table S6 and Fig. 6A, curves 2 and 3). species Te-GluRS is a dimer in the crystal form [15], as opposed to Tt-GluRS [5,12–14]. However, a poor fit to the data was obtained by assuming for Mt-GluRS a structure similar to that of the Te-GluRS dimer, even taking into account a monomer–dimer equilibrium. Therefore, a model Mt-GluRS dimer was built on the the Tt-GluRS subunit basis of extracted from that of the Tt-GluRS ⁄ tRNA complex (PDB code 1g59). The scattering curve of the symmetric homodimer shown in Fig. 6B clearly yielded the best fits to the SAXS data of Mt-GluRS at concentrations > 4 mgÆmL)1. Based on this dimeric model of Mt-GluRS, and using the program oligomer [27], it was possible to quantify the relative distribution of Mt-GluRS monomers and dimers at various protein concentrations, confirming the concentration depen- dence of the monomer–dimer equilibrium of Mt-GluRS solutions (Fig. 6A, curves 2–6 and Table S6). The pres- ence of l-Glu or GoA had no effect on the scattering curves of Mt-GluRS solutions (not shown), whereas ATP, either alone or in the presence of l-Glu, shifted the equilibrium towards the monomer, without causing full conversion (Fig. 6A, curves 7–10 and Table S6). By contrast, MgCl2 appears to stabilize the dimer (Fig. 6A,

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

B

A

Fig. 6. SAXS analysis of GluRS oligomeriza- tion in solution. (A) Scattering profiles of (1) Ec-GluRS (the pattern merged from different concentrations, no oligomerization effect observed); (2–6) Mt-GluRS with no MgCl2 and no ATP for concentrations (mg.mL)1) c = 6.25, 4.16, 3.7, 1.8 and 0.9 (from top to bottom); (7–10) Mt-GluRS with ATP (1 mM) and no MgCl2 at c = 7.85, 4.16, 1.63 and 0.86; (11–14) Mt-GluRS with MgCl2 (0.2 mM) and no ATP, c = 4.4, 3.96, 3.75 and 3.56; (15,16) Mt-GluRS with MgCl2 (0.2 mM) and ATP (1 mM) (c = 3.9 and 3.67). Experimental data are denoted by black dots and the fits from OLIGOMER [27] (or CRYSOL [26] for Ec-GluRS) are shown as red solid lines. The curves are appropriately displaced in logarithmic scale for better visualization. (B) The dimer composed by two adjacent monomers (shown in red and blue), which yields the best fit to Mt-GluRS data at concentrations > 4 mgÆmL)1 in the absence of ligands. The monomer structure was extracted from the crystal structure of Tt-GluRS in complex with tRNA (PDB ID 1g59).

curves 11–14 and Table S6) and reduces dissociation into monomers when ATP is added (Fig. 6A, curves 15–16 and Table S6).

Interaction between M. tuberculosis GluRS and GluTR

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1406

In those bacteria and plants that use the C5 pathway for ALA biosynthesis it has been proposed that GluRS and GluTR may form a complex in order to commit Glu–tRNAGlu to tetrapyrrole biosynthesis. Complex formation between GluRS and GluTR has been shown with purified Chlamydomonas reinhardtii enzymes [11]. We tested complex formation between the M. tubercu- losis enzymes by using affinity chromatography. Although homogeneous preparations of Mt-GluRS can be obtained as described above, all attempts to the putative Mt-GluTR produce large amounts of (Rv0509) in a soluble form in E. coli or M. smegmatis cells were unsuccessful (not shown). However, clon- ing of Rv0509 in pET11a or in pET23b (to generate a C-terminally His6-tagged version of Mt-GluTR, Mt-GluTR–His) led to the production of a small amount of soluble protein, which could be increased by co-producing the E. coli chaperon proteins DnaJ, DnaK and GrpE from plasmid p20 [28]. The identity of Rv0509 with the Mt-GluRS was established indi- rectly. E. coli cells overexpressing Rv0509 were red due to the accumulation of heme, which was in part released in the culture medium, as established from the

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

Fig. 7. GluRS–GluTR interaction. (Upper) The crude extract obtained from the homogenization of E. coli BL21 (DE3, pGTR, p20) cells (2 g) that contained the native Mt-GluTR, was incubated with Mt–His6GluRS (2 mg) for 30 min at 4 (cid:2)C and 10 r.p.m. on a rotary shaker. Two milli- tres of a 50% Ni-nitrilotriacetic acid-Sepharose suspension in 20 mM Hepes ⁄ NaOH buffer, pH 8.0, 10% glycerol, 1 mM b-ME was added. After 1 h the suspension was poured into a chromatographic column and the packed resin was extensively washed with the equilibration buffer. The column was developed with a 0–100 mM imidazole gradient in 10 mM steps followed by a final wash with 500 mM imidazole. Aliquots of the collected fractions were denatured for SDS ⁄ PAGE. The gels were stained with Coomassie Brilliant Blue and destained. (Mid- dle) The cell extract was substituted by a crude extract of E. coli BL21 (DE3, pET23b, p20) cells. (Lower) The His6–GluRS solution was substituted by the same volume of buffer. In all gels the fractions eluted with 50–100 mM imidazole showed no detectable proteins so that the corresponding lanes are not shown. The column flow-through has also been omitted. The dots mark an E. coli protein that migrates just below Mt-GluTR. In the upper gel, the white box highlights the fraction containing both GluRS and GluTR, whose spectrum and DLS signal are shown in Fig. S11. The migration positions of GluTR and GluRS, were determined by comparison with those of a homogeneous sample of GluRS (S) and of a sample enriched in GluTR (R) obtained by solubilizing inclusion bodies from overproducing cells. The band correspond- ing to GluTR has also been identified by western blots and immunodecoration with anti-GluTR IgG. The star indicates the standard proteins, with the corresponding mass shown on the side of the gels (in kDa).

tetrapyrrole biosynthesis

interaction Therefore, the

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1407

had produced Mt-GluTR followed by affinity chroma- tography on Ni-nitriloacetate–Sepharose allowed us to demonstrate formation of the GluRS-GluTR complex (Fig. 7). Similar results were obtained by adding homogenous GluRS to a crude extract of cells that had produced Mt-GluTR–His (not shown). Interestingly, all column fractions containing Mt-GluTR (Fig. 7) or its His6-tagged variant (not shown), exhibited an absor- bance spectrum consistent with the presence of a pro- tein-bound heme cofactor (Fig. S11A). Furthermore, the DLS signal of the fractions containing Mt-GluTR (Fig. 7, top, boxed lane) showed a single, although broad, peak (r = 5.2 nm) corresponding to a mass of only 180 kDa (Fig. S11B). This finding suggests that complex formation with GluRS and GluTR may isolate and stabilize a soluble GluTR form. absorbance spectra of crude extracts and culture medium. Such a red phenotype is expected for the overproduction of the enzyme catalysing the first and rate-limiting step of [8]. Interestingly, the red phenotype was lost when the C50S and C50A variants of Rv0509 were overpro- duced (not shown). C50 of Rv0509 corresponds to the catalytically essential C48 of Methanopyrus kandlerii GluTR, which together with the E. coli form is the best characterized GluTR [29–32]. Attempts to purify Mt-GluTR led to the isolation of aggregates of > 1 MDa. between M. tuberculosis GluRS and GluTR had to be studied using purified Mt-GluRS and Mt-GluTR forms con- tained in crude extracts of overproducing cells. Addi- tion of His6–GluRS to a crude exctract of cells that

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

Discussion

capable of is (cid:3) 10-fold faster than tRNA charging at pH 7.2, while supporting a similar reaction mechanism for the two enzymes, suggests a different pH dependence for the individual reactions steps, perhaps reflecting differ- ences in the fine structure of their active sites.

Rv2992c gene product was demonstrated to encode Mt-GluRS, which is charging E. coli tRNAGlu with l-Glu. The enzyme can be obtained in large quantities and in a soluble and stable form using a four-step purification procedure based on previously described methods [17,18,33].

similar is

Mt-GluRS was found to be very specific for the amino acid and the nucleotide substrate. It could use l-Gln instead of l-Glu to charge tRNAGlu only at a very low rate. Neither l-Gln nor 2-oxoglutarate are inhibitors. Despite the presence of the a-b hydrolysable bond, b,c-methylene-ATP could not replace ATP as the substrate. Furthermore, the a,b- and b,c-methylene analogs of ATP tested, and AMP and its analog decoi- nine did not inhibit Mt-GluRS. This is at variance with Ec-GluRS which was inhibited by both a,b-meth- ylene-ATP and AMP, although with Ki values in the mm range [20]. Mt-GluRS exhibits properties similar, but not identi- cal, to those of the well-characterized GluRS from E. coli, which we used as the prototype of bacterial GluRS. The turnover number to that reported for Ec-GluRS, as are the KM values for ATP and tRNAGlu. However, KL-Glu was found to be (cid:3) 20- fold higher for Mt-GluRS than for the E. coli enzyme. Such a difference cannot be ascribed to a different pH dependence of the reaction, but rather to a difference between the enzymes. the M. tuberculosis Like other GluRS,

the steps reaction

As for the E. coli enzyme, PPi, which binds to the enzyme ⁄ tRNA ⁄ Glu-AMP intermediate, was found to be a noncompetitive inhibitor with respect to both l-Glu and ATP. GoA was competitive with respect to both Glu and ATP, but uncompetitive with respect to tRNAGlu, with Ki values of the same order of magni- tude as those reported for Ec-GluRS [21]. GoA is one of the glutamyl-AMP analogs being developed as a GluRS inhibitor, and it will be of interest to test them on Mt-GluRS in future studies of potential novel anti- tubercular drugs, which are the long-term aim of this project [21,34,35].

Scheme 1. Minimal kinetic scheme of the Mt-GluRS reaction.

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1408

The precise definition of the kinetic mechanism of Mt-GluRS was outside the scope of this work, particu- larly in light of the complexity of the reaction, as shown by the elegant work done with the E. coli enzyme [20,22–24] and the expected overall similarity between Mt- and Ec-GluRS. However, our data are consistent with a minimal reaction scheme in which tRNA binding to the free enzyme is followed by an activatory conformational change [5] and random binding of l-Glu and ATP to the latter species (Scheme 1). In particular, the competitive inhibition pattern observed with GoA versus l-Glu and ATP is diagnostic for the random sequential portion of the enzyme requires bound tRNAGlu to carry out the formation of for the Glu-AMP intermediate indicating that Mt-GluRS also binding of tRNA might induce the conformational change observed with Tt-GluRS that switches the binding mode of ATP to a productive one [5]. Mt-GluRS is similar to other GluRS in that it indicating catalyses the PPi ⁄ ATP exchange reaction, that enzyme ⁄ tRNA ⁄ linking Glu ⁄ ATP complex to yield the enzyme ⁄ tRNA ⁄ Glu- AMP complex are reversible. The kcat value measured during the PPi ⁄ ATP exchange reaction is (cid:3) 10-fold higher than that measured for the tRNA aminoacyla- indicating that transfer of l-Glu from tion reaction, Glu-AMP to the tRNA is slower than pyrophosphor- olysis of the intermediate to yield ATP and l-Glu. A similar conclusion was reached for Ec-GluRS. How- ever, for Ec-GluRS, it has been reported that the pH dependence of the PPi ⁄ ATP exchange reaction follows an inverse profile with respect to that of the tRNA charging activity, so that the ratio between the veloci- ties of the exchange and charging reactions was (cid:3) 30 at pH 6.2, but only 1.5 at pH 7.4 and 0.3 at pH 8.6 [22,24]. Thus, our finding that the PPi ⁄ ATP exchange

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

kinetic mechanism as discussed recently [36] for bisub- strate inhibitors.

which lack the Glu–tRNAGln amidotransferase needed to correct the misacylation of tRNAGln caused by a ND- GluRS, is not toxic. These results lead to the conclusion that Mt-GluRS ⁄ tRNAGln recognition is species specific. Sequence analyses may provide a rationale for the discriminating behaviour of Mt-GluRS in E. coli.

In the absence of tRNA and l-Glu, Mt-GluRS was found to hydrolyse ATP to ADP + Pi, although the reaction velocity was only 0.1% that of the physiologi- cal tRNAGlu charging reaction, under the same condi- tions, leading to the conclusion that this reaction is not biologically relevant.

active. By

and of Helicobacter pylori the overall shape of

ligands and that of

and B. subtilis

is Val460, which corresponds

(as

Mt-GluRS differs from Ec-, Tt- and Te-GluRS for the oligomeric state in that its monomer exists in solu- tion in equilibrium with the dimeric species. At the low concentrations used in the activity assays, the enzyme monomer should prevail, indicating that this species is contrast, Ec-GluRS and catalytically Tt-GluRS [5,12–14] appear to be strictly monomeric, whereas the crystal structure of the Te-GluRS shows dimers [15]. By using SAXS and the Tt- and Te-GluRS structures for rigid body modeling, it was determined the Mt- and Ec-GluRS that subunits are similar to each other and to those of Tt- and Te-GluRS. SAXS sensitivity is not sufficient to distinguish among the conformations of Tt-GluRS bound to the different the Te-GluRS subunit. The crystallographically detected conformations are indeed catalytically significant, but structurally minor, implying rotations of domains of just a few degrees (e.g. 7(cid:2) interdomain rotations upon tRNA binding to the Tt-GluRS and local limited rearrangements in the active site) [5]. Interestingly, the Mt-GluRS dimer found in solution appears to differ from that found in the Te-GluRS crystals [15]. How- ever, by fitting the SAXS curves a model could be built, which indicates that this species may be catalytically active because the tRNA binding surface and ATP and l-Glu binding sites are solvent accessible. Despite the structural similarity of the enzyme subunits, Mt-GluRS is significantly more sensitive than Ec-GluRS to prote- olysis, suggesting greater conformational flexibility. In both enzymes, the sites sensitive to chymotryptic attack are next to the ATP-binding site. However, only in the case of Mt-GluRS are the chymotrypsin-sensitive sites protected by ATP and GoA, highlighting another dif- ference between the enzymes. Interestingly, ATP and GoA had a similar effect on the proteolytic pattern of Mt-GluRS, but only ATP appeared to stabilize the monomeric form, as established by SAXS. The flexibil- ity of Mt-GluRS coupled to the monomer ⁄ dimer equi- librium may be the reason for the failure to obtain crystals suitable for determination of the Mt-GluRS structure by X-ray diffraction. Studies on Tt-GluRS indicated that discriminating and nondiscriminating GluRS can be distinguished on the basis of the presence of a specific Arg residue (Arg358 in Tt-GluRS, Arg350 in Ec-GluRS) in the anticodon recognition region of GluRS [12] (Fig. S7). Indeed, its substitution with a Gln, the residue found at the equivalent position in Bs-GluRS, conferred non- discriminating properties on Tt-GluRS [12]. Further- more, Te-GluRS, a ND-GluRS, contains Gly366 in the position equivalent to Arg358 of Tt-GluRS [15]. However, site-directed mutagenesis of the two GluRS isoforms comparative sequence analyses [37] indicated that the Arg residue is not sufficient to distinguish between D- and ND- GluRS but, more likely, several residues are important. One was identified as a Thr (Thr444 in Tt-GluRS) found in most D-GluRS, which is often substituted by Gly, Ala, Ser (e.g. Gly454 in Te-GluRS) in ND- GluRS. A third candidate was found by Schultze et al. [15] who observed that in Tt-GluRS Arg358 forms a salt bridge with Glu443, which is not found in several ND-GluRS even when they have an Arg residue equiv- to Tt-GluRS Arg358. Comparison of alent the sequence of Mt-GluRS with those of the T. thermophi- lus, T. elongatus, E. coli enzymes showed that Arg372 of Mt-GluRS is at a position equivalent to that of Arg358 in Tt-GluRS, and Ser461 substitutes Thr444 (Fig. S7). In Mt-GluRS the pre- ceeding residue to Glu443 of Tt-GluRS and His453 of Te-GluRS. Thus, Mt-GluRS seems to obey to the rules established pre- viously [15,37] for a ND-GluRS. However, these rules do not seem sufficient to predict the discriminating properties of a GluRS. In particular, the discriminat- ing Ec-GluRS has a Gln437–Ser438 pair where a Glu–Thr is expected. That Mt-GluRS may be of the nondiscriminating type in M. tuberculosis sup- ported by genome analyses), but its sequence shares features with the discriminating Ec-GluRS (Arg372, Val460 and Ser461 in Mt-GluRS versus Arg350, Gln437 and Ser438 in Ec-GluRS), might explain the absence of toxicity of its overproduction in E. coli where it behaves as a D-GluRS like the endogenous enzyme. Finally, we

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1409

Despite M. tuberculosis genome analysis indicating that Mt-GluRS is of the discriminatory type (see above for details), overproduction of Mt-GluRS in E. coli cells, found that M. tuberculosis GluRS and GluTR can form a complex confirming the results obtained with C. reinhardtii enzymes [11] and

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

the same restriction enzyme and purified, yielding pET- GTS1. The NdeI fragment was also cloned into pET28b (Novagen) digested with the same enzyme. The resulting plasmid (pETGTS2) encoded a fusion between an N-termi- nal His6 tag and the Rv2992c coding region with a ten resi- dues spacer between the sixth His residue and the start codon of the predicted Rv2992c gene product. The result- ing protein is indicated as His6-GluRS. The insert of all plasmids and the adjacent regions were sequenced by PRIMM srl (Milan, Italy).

supporting the concept that in M. tuberculosis forma- tion of this complex may regulate the flux of Glu– tRNAGlu toward tetrapyrrole biosynthesis as opposed to that of proteins. During the course of these studies we also demonstrated that Mt-GluTR contains bound heme, a property previously ascribed only to the plant- type enzyme [38,39]. Finally, the fact that Mt-GluTR isolated using Mt-GluRS as bait does not seem to aggregate or precipitate also opens the way to the iso- lation and subsequent characterization of this enzyme responsible for heme biosynthesis, which has also been demonstrated to be essential in M. tuberculosis [16]. Production of M. tuberculosis GluRS in E. coli BL21(DE3) cells

Materials and methods

pETGTS1 and pETGTS2 were used to produce the Rv2992c gene product or the N-terminally His6-tagged vari- ant, respectively, in E. coli BL21(DE3) cells grown at 25 (cid:2)C in Luria-Bertani medium containing 0.1 mgÆmL)1 ampicil- lin. Overexpression of the heterologous gene was induced at an D600 value of 0.7 by adding IPTG to a final concentra- tion of 0.1 mm. After 19 h, cells were harvested by centrifu- gation at 6000 g and 4 (cid:2)C for 15 min. The cell pellet was washed with 0.9% NaCl and stored at )20 (cid:2)C until protein purification.

Restriction endonucleases were obtained from GE Health- care (Chalfont St Giles, UK) and Promega (Madison, WI, USA). Unless otherwise stated, chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA) or Merck (Whitehouse Station, NJ, USA). TLC poly(ethyleneimine)– cellulose sheets with fluorescent indicator (254 nm) were from Macherey-Nagel (Du¨ ren, Germany).

Chemicals and materials

M. tuberculosis Rv2992c, corresponding to the putative gltX gene encoding GluRS, was amplified by PCR using cosmid BAC Rv30 from the Institut Pasteur collection as the template in the presence of synthetic oligonucleotides pairs.

Primer 1: 5¢-AAGAAGAAGCATATGTCACCGTGCCCG

ACCAGCTG-3¢

Primer 2: 5¢-AAGAAGAAGCATATGACCGCCACGG

Plasmid pET-ERS was a kind gift of J. Lapointe (Univer- site´ Laval, Que´ bec, Canada). It was transformed into E. coli BL21(DE3, pLysS) cells, which were grown in Luri- a–Bertani medium containing 60 lgÆmL)1 kanamycin and 34 lgÆmL)1 chloramphenicol at 37 (cid:2)C until D600 reached 0.3. Overproduction of the Ec-GluRS was induced by add- ing IPTG at a final concentration of 0.1 mm. After 5 h, cells were harvested and stored as described above.

AAACAGTCCGG-3¢

Cloning of M. tuberculosis GluRS gene Production of E. coli GluRS in E. coli BL21(DE3) cells

(15 U)

reported for

to that

similar

Protein concentration of crude extracts was determined by the biuret method [40] and that of purified samples using the Bradford Reagent (Amresco, Solon, OH, USA) [41]. BSA (Sigma) was used as the standard. Using an electro- phoretically homogeneous protein preparation it was deter- mined that a 1 mgÆmL)1 Mt-GluRS solution absorbs 0.79 ± 0.064 at 280 nm (average of 10 determinations), a the Ec-GluRS value (e280 = 0.87) [18]. To calculate the enzyme concentration a mass of 53 685 was used for Mt-GluRS by taking into account the post-translational removal of Met-1 to yield the predicted 489 residues protein. A mass of 55 876 was used for His6–GluRS (509 residues after removal of Met-1). A mass of 53 669 was used for Ec-GluRS (470 residues for the mature protein).

The primers introduced NdeI sites (underlined) for clon- ing of the amplified fragment into pET11a (Novagen, San Diego, CA, USA) digested with NdeI. The GTG start codon was also changed into an ATG (bold in Primer 1) by the insertion of the NdeI restriction site. PCR was set up by mixing BAC Rv30 (30 ng), dNTPs (50 lm each), pri- mer 1 and 2 (24 pmol each) and PfuTurbo Taq polymerase in 20 mm (Stratagene, La Jolla, CA, USA) Tris ⁄ HCl buffer, pH 8.8, 10 mm KCl, 10 mm (NH4)2SO4, 2 mm MgSO4, 0.1% Triton X-100 and 0.1 mgÆmL)1 BSA. PCR conditions were as follows: cycle 1, 5 min at 95 (cid:2)C; cycles 2–36, 1 min at 95 (cid:2)C, 30 s at 60 (cid:2)C and 4 min at 72 (cid:2)C; cycle 37, 8 min at 72 (cid:2)C. The amplified 1500 bp fragment was purified using the QiaQuik Gel Extraction Kit (Qiagen, Venlo, NL, USA) according to the manufac- turer’s instructions, precipitated and digested with NdeI. After purification by agarose gel electrophoresis the frag- ment was ligated with pET11a that had been digested with

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1410

Determination of protein concentration

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

in buffer + 0.5 m NaCl 10–100 mm imidazole gradient (15 vol, 1 mLÆmin)1). The enzyme eluted from this column at (cid:3) 70 mm imidazole. The His6–GluRS-containing fractions were pooled on the basis of their electrophoretic pattern, concentrated by ultrafiltration and dialysed against buffer B (2 L, for 5 h) and buffer B + 50% glycerol (0.5 L, 19 h) as described for the native enzyme preparation. Also in this case, the enzyme (20–40 mgÆmL)1) was stable for years when stored at )20 (cid:2)C.

the crude extract;

Purification of Mt-GluRS

SDS ⁄ PAGE was performed according to Laemmli [42] using 12% minigels and a GE Healthcare SE280 apparatus. Protein samples were denatured by incubation at 100 (cid:2)C for 10 min in SDS sample buffer [62.5 mm Tris ⁄ HCl, pH 6.8, 2% (w ⁄ v) SDS, 0.001% (w ⁄ v) bromophenol blue, 10% (w ⁄ v) glycerol, 0.8 mm b-ME] added from two- or fourfold concentrated stock solutions (2· SDS sample buffer or 4· SDS sample buffer). After the run, the gels were stained by immersion in 0.1% Coomassie Brilliant Blue in 40% methanol, 10% acetic acid and destained by diffusion in 40% methanol and 10% acetic acid.

Electrophoretic techniques and western blots

(medium, GE Healthcare)

The procedures of Lin et al. [17], Lapointe et al. [18] and Kern et al. [18,33] for Ec-GluRS were combined to obtain homogeneous preparations of Mt-GluRS. Purification con- sisted of: (a) resuspension of 10–20 g cells in 20–40 mL 10 mm (K)PO4 buffer, pH 7.5, 10% glycerol, 1 mm phen- ylmethanesulfonyl fluoride, 1 mm dithiothreitol, and cell disruption by sonication and centrifugation at 25 500 g for 1 h at 4 (cid:2)C; (b) poly(ethylene glycol) 6000 (7%, w ⁄ v) ⁄ dex- tran (1.4%, w ⁄ v) partitioning of (c) recovery of the top poly(ethylene glycol)-rich phase after centrifugation at 17 500 g for 20 min; (d) chromatography on a first Q-Sepharose ion-exchange column (1.5 · 11.3 cm, 20 mL; GE Healthcare) equilibrated in buffer A (20 mm Tris ⁄ HCl, pH 7.4 25 (cid:2)C, 10% glycerol, 1 mm dithiothrei- tol), eluted with buffer A + 0.2 m NaCl (5 vol) and a 0.2– 1.0 m NaCl gradient in buffer A (20 vol, 1 mLÆmin)1); (e) concentration of the pooled GluRS-containing fractions by ultrafiltration in an Amicon apparatus (Millipore, Billerica, MA, USA) equipped with a YM10 membrane, and dialysis against buffer B (50 mm Hepes ⁄ NaOH, pH 8.0, 10% glyc- erol, 1 mm dithiothreitol, 2 L); (f) chromatography on a second Q-Sepharose column (1.5 · 8.49 cm, 15 mL) equili- brated in buffer C (20 mm (K)PO4 buffer, pH 7.5, 10% glycerol, 1 mm dithiothreitol), and eluted with buffer C (1 vol) followed by a gradient in which the (K)PO4 concen- tration was varied from 20 to 250 mm and the pH from 7.5 to 6.5 in 20 vol at a flow-rate of 1 mLÆmin)1; (g) concentra- tion of the GluRS-containing fractions by ultrafiltration to (cid:3) 20 mgÆmL)1 and 4 mL; and and (h) dialysis against 1 L of buffer B (5 h) followed by dialysis against 0.5 L of buf- fer B containing 50% glycerol (14 h). The enzyme (typically 40 mgÆmL)1) was stored at )20 (cid:2)C without significant activ- ity loss for up to 2 years.

GluRS-containing fractions were pooled after each step on the basis of their electrophoretic pattern. Mt-GluRS eluted from the first column between 0.2 and 0.3 m NaCl, and from the second at (cid:3) 150 mm (K)PO4 and pH 7.2. The same purification procedure was used to obtain Ec-GluRS preparations.

E. coli BL21 (DE3, pETGTS2) cells (12 g)

A Mt-GluRS aliquot was gel filtered through a Sephadex G25 column equilibrated in 10 mm Hepes ⁄ KOH, pH 7.5, and concentrated to 5–10 mgÆmL)1 using a Centricon-10 (Millipore) microcon- centrator. The protein mass was determined on diluted samples by MALDI-TOF with a Bruker Daltonics Reflex IV instrument (Brucker Daltonics, Bremen, Germany) equipped with a nitrogen laser. N-terminal sequencing of GluRS and proteolytic fragments was carried out with an Applied Biosystems (Foster City, CA, USA) Procise Model 491 sequencer using aliquots of the Mt-GluRS solution or protein samples resolved by SDS ⁄ PAGE and electrotrans- ferred onto a Immobilon-PSQ (Millipore) membrane, stained with Coomassie Brilliant Blue and thoroughly destained [43].

DLS measurements were carried out using a DynaPro instrument (Protein Solutions, Charlottesville, VA, USA) in a 50 lL quartz cuvette at 17 (cid:2)C (average of 20 30 s acquisi- tions, sensitivity 70–100% depending on protein concentra- tion). Data were analysed using the dynapro software (version V5 or V6). The software uses Eqn (7) [44] to calcu- late the mass of globular protein of 24–110 kDa.

ð

Þ3=SpecVol

ð7Þ

MW ¼ 4=3

ð

ÞpNA Rh=FricRatio

that had overproduced His6–GluRS, were resuspended in 10 mm Hepes ⁄ NaOH, pH 8.0, 10% glycerol, 5 mm b-mercaptoetha- nol (b-ME), 1 mm phenylmethanesulfonyl fluoride (24 mL), disrupted by sonication and centrifuged. Twenty milliliters of a 50% Ni-nitrilotriacetic acid–Sepharose (Novagen- Merck, Darmstadt, Germany) suspension equilibrated in the homogenization buffer were added to the crude extract and the suspension was incubated for 1 h at 12 r.p.m. and 4 (cid:2)C on a rotary shaker. The resin was packed into a chromato- graphic column (inner diameter, 1.5 cm), washed with one column volume of the equilibration buffer, 1 vol of the same buffer containing 0.5 m NaCl, 5 vol of buffer containing 0.5 m NaCl and 10 mm imidazole and then developed with a

where MW is the molecular mass, NA is Avogadro’s num- ber (6.022 · 1023 mol)1), Rh is the radius in cm, SpecVol is the specific volume (0.726 cm3Æg)1) and FricRatio is the

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1411

N-terminal sequence, mass and aggregation state of GluRS

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

frictional ratio (1.25707). Protein samples were centrifuged in a microfuge at 15 000 g for 10 min at 4 (cid:2)C before each measurement.

Steady-state kinetic analyses and inhibition studies

inspection of

after

The apparent kcat and KM values of Mt-GluRS for tRNAGlu, ATP and l-Glu were determined in the tRNA aminoacyla- tion reaction at 37 (cid:2)C as described above, except that the concentration of one of the substrates was varied and the lev- els of the others fixed. In these assays the amount of enzyme was chosen in order to observe linearity up to 10 min, and aliquots were typically withdrawn at 1, 2, 3, 5 and 10 min. The grafit 4.3 software package (Erithacus Software Ltd, East Grinstead, UK) was used to fit the v ⁄ E values as a func- tion of the varied substrate concentration (S) to the Michael- is–Menten equation (Eqn 8) the Lineweaver–Burk (double reciprocal) plot (Eqn 9), and to obtain the values and the associated errors of the steady-state kinetic parameters [45].

ð8Þ

v/E = (kcatS)/(KM + S)

Activity assays

Inhibition studies were performed by measuring the initial velocity of reactions that contained fixed levels of the inhibitor (I), varying concentration of one of the substrates and constant concentration of others. After inspection of the Dixon or the double-reciprocal plots, the data were fit- ted to the equation describing competitive (Eqn 10), non- competitive (Eqn 11) or uncompetitive (Eqn 12) inhibition. In Eqn (10), Kis and Kii are the inhibition constants affect- ing the slopes and the intercepts of the double reciprocal plots, respectively [45].

ð10Þ

v/E = (kcatS)/[S + KM(1 + I/KiÞ(cid:4)

ð9Þ v/E = (1/ kcatÞþðKM=kcat)(1/S)

ð11Þ v/E = (kcatS)/[S(1 + I/KiiÞþKM(1 + I/KisÞ(cid:4)

The pH dependence of the apparent kcat and kcat

⁄ Kl-Glu values was measured at fixed levels of ATP (1 mm) and tRNAGlu (3.6 lm) and varying l-Glu (0.2–2.0 mm) in 35 mm Hepes ⁄ NaOH buffer at pH 6.5–8.5. All other condi- tions were as stated for the standard activity assay. The cal- culated values of kcat and kcat ⁄ Kl-Glu were fitted to Eqns (13,14), respectively [45].

þ 1

ð13Þ

Y ¼

Limit (cid:5) 10ðpH(cid:2)pKaÞ 10ðpH(cid:2)pKaÞ

ð12Þ v/E = (kcatS)/[S(1 + I/KiiÞþKM(cid:4)

In Eqns (13,14), Limit is the pH independent value of the

Y ¼ þ 1 ð14Þ Limit (cid:5) 10ðpKa(cid:2)pHÞ 10ðpKa(cid:2)pHÞ

steady-state kinetic parameter under analysis (Y).

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1412

tRNA charging reaction The tRNA aminoacylation activity of Mt-GluRS was deter- mined at 37 (cid:2)C by measuring the rate of formation of acid- precipitable l-[U14C]Glu–tRNA as described previously [18,33]. Standard assays contained 35 mm Hepes ⁄ NaOH, pH 7.3, 25 mm KCl, 2 mm dithiothreitol (buffer C), 0.5– 2 mm l-[U14C]Glu (11 431 dpmÆnmol)1; GE Healthcare), 10 mm MgCl2, 1 mm ATP, 3.6 lm E. coli tRNAGlu, 0.1% BSA and enzyme (typically, 10–100 ng, 0.19–1.9 pmol, 1.24–12.4 nm) in a volume of 150 lL. E. coli tRNAGlu (55– 60% specific tRNAGlu; Sigma) was resuspended in buffer C to yield a 160 lm stock solution, which was stored in small aliquots at )20 (cid:2)C. The amount of tRNAGlu present in each batch and its stability were checked by quantifying the amount of l-[U14C]Glu–tRNA obtained in assays in which the charging of the tRNA present was brought to comple- tion. Mt-GluRS stock solutions (typically 20–40 mgÆmL)1) were first diluted to 1 mgÆmL)1 in 40 mm Hepes ⁄ NaOH, pH 8.0, 10% glycerol. The protein concentration was deter- mined with the Bradford reagent at this stage. The enzyme solution was then serially diluted to up to 10 lgÆmL)1 in the same buffer containing 0.1% BSA. For each assay, the reac- tion mixture (145.5 lL) lacking tRNA was equilibrated at 37 (cid:2)C for 5 min. A 20 lL aliquot was withdrawn and spot- ted on a 1 · 1 cm square of Whatman 3MM filter paper (GE Healthcare), which was immediately transferred to a beaker containing 10% trichloroacetic acid and kept under vigorous stirring until the end of the assay. The reaction was started by adding tRNAGlu (4.5 lL). At different times 20-lL aliquots were withdrawn, spotted on the Whatman 3MM filters, which were transferred into 10% trichloroace- tic acid with magnetic stirring. At the end of the assay, all filters were transferred to fresh 10% trichloroacetic acid (500 mL, 10 min). Washings in 5% trichloroacetic acid and 95% ethanol, with interval stirring (10 min each) followed. Dried filters were placed in an 8 mL plastic vial. Radioactiv- ity was determined by scintillation counting in a TriCarb 2100-TR (Perkin–Elmer, Wellesley, MA, USA) after addi- tion of 5 mL of Ultima Gold (Perkin–Elmer) scintillation fluid. dpm were calculated from cpm using a calibration curve made with a [14C] standard (Perkin–Elmer). The amount of l-[U14C]Glu–tRNAGlu (in nmol) formed in the 20 lL aliquot at the different times was calculated. The ini- tial velocity (v) of reactions was determined by interpolating linear portion of the curve of Glu–tRNAGlu the initial formed as a function of time. Activity was expressed as apparent turnover number (v ⁄ E in min)1) by taking into account the amount of enzyme (E) present in the 20 lL aliquot (in nmol).

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

squares and transferred to 8 mL vials for scintillation counting.

Chromatographic separation of reaction components using [2,5¢,83H]ATP and L-[U14C]Glu

10% glycerol,

2 mm dithiothreitol,

3.2 lm tRNAGlu

231 113 dpmÆnmol)1),

recorded on a storage phosphor

using

Incorporation of [32P]PPi into ATP was determined in reac- tion mixtures containing 100 mm Hepes ⁄ NaOH, pH 7.2, 2 mm ATP, 16 mm MgCl2, 6 mm l-glutamate, 0.3% glyc- erol, 25 mm KCl, 1 mm Na-[32P]-PPi (from a 85 mm stock and solution, Mt-GluRS (0-0.7 lm) [23]. After incubation at 37 (cid:2)C for different times, 1–10 lL aliquots were applied onto the poly(ethyleneimine)–cellulose sheets, which were developed as described above. The migration position of the nucleo- tides was determined by irradiation with UV light. It was marked on the sheet by spotting 1 lL of the radioactive PPi solution (2000 dpm) on a lane in which a mixture of AMP, ADP and ATP was resolved. The autoradiographic image was screen (Molecular Dynamics, Sunnyvale, CA, now GE Healthcare) for 5–10 min, and analysed with a Typhoon 9400 (GE the manufacturer’s Healthcare) phosphoimager software. Quantitation of [32P]ATP was carried out with imagequant 5.2 (GE Healthcare) software. Phosphoimager counts were converted into dpm using a calibration curve made for each TLC sheet by spotting aliquots of [32P]PPi solutions of known radioactivity onto a free lane of the sheet before data collection.

⁄ ATP exchange assay [32P]PPi

GluRS (0.5–61 lg; final concentration, 0.06–7.5 lm) from a stock solution prepared as described for the tRNA charging reaction was incubated in 35 mm Hepes ⁄ NaOH buffer 1 mm pH 7.3, [2,5¢,83H]ATP (33 300 dpmÆnmol)1; Perkin–Elmer), 10 mm MgCl2, 25 mm KCl, 2 mm l-Glu in the presence or absence of 3.0 lm tRNAGlu (final volume: 150 lL). After 1–20 min at 37 (cid:2)C, aliquots were injected directly or after dilution with cold water onto a MonoQ column connected to an AKTA apparatus (GE Healthcare) equipped with an absor- bance detector set at 254 nm. The column was equilibrated with 20 mm triethanolamine ⁄ HCl buffer, pH 7.7 [46]. Elu- tion was performed by washing the column with 14 vol of the equilibrating buffer and then increasing KCl concentra- tion in the buffer from 0 to 0.3 m in 30 vol and from 0.3 to 1 m in 8 vol. After 3 vol of buffer containing 1 m KCl, the column was re-equilibrated in the starting buffer. The flow rate was 1 mLÆmin)1 and fractions (1 mL) were directly col- lected in 8 mL plastic vials. The radioactivity was deter- mined by scintillation counting. Assays were also carried out under the same conditions using unlabeled ATP and l-[U14C]Glu. In preliminary experiments, 200 lL aliquots of ATP, ADP and AMP (1 mm each), l-[U14C]Glu (2 mm) and tRNAGlu (3.0 lm) were loaded onto the MonoQ col- umn to identify their elution positions. Control experiments also showed that addition of unlabeled nucleotides, as carrier, to the samples under analyses did not alter the the distribution and recovery of chromatography nor radioactivity.

Limited proteolysis

20 mm,

Ec-GluRS or Mt-GluRS (1 mgÆmL)1) were incubated with 0.1–10% (w ⁄ w) Na-tosyl-l-phenyl chloromethyl ketone- in 50 mm Hepes ⁄ NaOH buffer, treated trypsin (Sigma) pH 8.0, at 25 (cid:2)C, in a final volume of 100–200 lL. Before trypsin addition, and at different incubation times, 10 lL aliquots of the reaction mixture were transferred into eppendorf tubes containing 20 lL 2· SDS sample buffer, Na-tosyl-l-lysyl 10 lL). chloromethyl ketone, Proteins were immediately denatured by incubation at 100 (cid:2)C for 10 min. This procedure was found to be suffi- cient to block proteolysis. Similar experiments were carried out using Na-tosyl-l-lysyl chloromethyl ketone-treated chy- motrypsin instead of trypsin, except for the fact that the reaction was blocked by transferring the 10 lL aliquots in eppendorf tubes containing Na-tosyl-l-phenyl chloromethyl ketone instead of Na-tosyl-l-lysyl chloromethyl ketone.

(Millipore) membranes

Immobilon

Proteolysis products were resolved by SDS ⁄ PAGE, visu- alized by Coomassie Brilliant Blue staining. Their mass was calculated from a calibration curve built with the 14– 202 kDa molecular mass protein standard mix (Sigma). To identify the sites of proteolysis, the proteins were blotted PSQ after onto SDS ⁄ PAGE, and subjected to N-terminal sequencing. Sequences were compared with the known sequences of

AMP, ADP, ATP, Pi and PPi were resolved by TLC on poly(ethyleneimine)–cellulose sheets as described previously [47]. The enzyme (0.05–6.4 lg, 0.9–118 pmol, final concen- tration, 6–790 nm) was incubated in 35 mm Hepes ⁄ NaOH buffer, pH 7.3, 10% glycerol, 2 mm dithiothreitol, 25 mm KCl, 1 mm [2, 5¢,83H]ATP (33 300 dpmÆnmol)1; Perkin- Elmer), 10 mm MgCl2, 2 mm l-Glu, 0.004% BSA in the presence or absence of 3.6 lm tRNAGlu (final volume: 150 lL). Parallel samples, lacking one or more of the com- ponents were also set up. At different times of incubation at 37 (cid:2)C 10 lL aliquots were removed and applied onto the poly(ethyleneimine)-cellulose sheets, which were immedi- ately developed with 0.75 m NaH2PO4 ⁄ H3PO4 buffer, pH 3.4, as the mobile phase. For optimal resolution and minimal background, the poly(ethyleneimine)–cellulose sheets had to be pre-developed in 1.75 m K2PO4 ⁄ H3PO4 buffer, pH 3.4, washed with water, dried and stored at 4 (cid:2)C for at least 19 h [48]. Spots corresponding to AMP, ADP and ATP were identified with a UV lamp at 254 nm. Strips corresponding to the sample lanes were cut into 1.5 cm

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1413

TLC separation of reaction components

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

the plasmid-encoded proteins was obtained with 0.1 mm IPTG. Cells were harvested after 17 h.

Mt-GluRS (UniProtKB accession number P0A636) and Ec-GluRS (UniProtKB accession number P04805).

in

were

resuspended

acid-Sepharose

suspension in the

the

Synchrotron X-ray scattering data from solutions of Mt- and Ec-GluRs in the presence or absence of ligands the X33 beamline (DESY, Hamburg, were collected at Germany) [49] at protein concentrations (c) ranging from 11 to 0.5 mgÆmL)1. At a sample-detector distance of 2.7 m, the range of momentum transfer 0.1 < s < 5 nm)1 was covered [s = 4p sin(h) ⁄ k, where 2h is the scattering angle and k = 0.15 nm is the X-ray wavelength]. The data were processed using standard procedures by the program pack- age primus [27]. The forward scattering I(0) and Rg were evaluated using the indirect transform package gnom [50]. The effective molecular mass of the solute (MM) was esti- mated by comparison of the forward scattering I(0) with that from reference solutions of BSA (MM = 66 kDa). The scattering intensities for monomeric and dimeric mod- els of Mt-GluRS were computed by crysol [26] from the atomic coordinates of the discriminating Tt-GluRS (PDB files: 1n75 for the complex with ATP; 1j09 for the complex with ATP and Glu; 1n77 for the complex with tRNA and ATP; 1n78 for the complex with tRNA and GoA; 1g59 for the nondiscriminating the complex with tRNA) and of Te-GluRS (PDB file 2cfo for the complex with Glu) and were used to analyse the oligomeric composition of all the samples. The program oligomer [27] was used to find the volume fractions of components minimizing the discrepancy v2 (normalized sum of the reduced standard deviations) between the linear superposition of the weighted intensities of the components and the experimental data from the mixture.

Small angle X-ray scattering data collection and modeling

affinity

under

To study the GluTR-GluRS interaction, 2 g of E. coli BL21 (DE3, pGTR, p20) cells that had produced the 20 mm native Mt-GluTR, Hepes ⁄ NaOH buffer, pH 8.0, 10% glycerol, 5 mm b-ME and 1 mm phenylmethanesulfonyl fluoride (4 mL). Glass beads (12 g; 0.3 mm diameter) were added and cells were disrupted by applying five cycles of vigorous vortexing (1 min) followed by 1 min on ice. After twofold dilution, the homogenate and a 2 mL rinse of the glass beads were centrifuged for 1 h at 22 500 g at 4 (cid:2)C. The crude extract (13–15 mL) was diluted fivefold in 20 mm Hepes ⁄ NaOH buffer, pH 8.0, 10% glycerol. Mt-His6-GluRS (2 mL of a 1 mgÆmL)1 solution in the same buffer + 1 mm b-ME) was added. After incubation for 30 min at 4 (cid:2)C and 10 r.p.m. on a rotary shaker, 2 mL of a 50% Ni-nitrilotri- same buf- acetic fer + 1 mm b-ME was added. After 1 h the suspension was poured into a small chromatographic column (inner diameter, 1.6 cm) and the packed resin was extensively washed with the equilibration buffer. The column was developed with a stepwise 0–500 mm imidazole gradient in 20 mm Hepes ⁄ NaOH buffer, pH 8.0, 1 mm b-ME. Aliqu- ots of collected fractions were denatured for SDS ⁄ PAGE as described above. For each condition two controls were also performed: (a) the His6-GluRS solution was substituted by the same volume of buffer; (b) the cell extract was substituted by a crude extract of E. coli BL21 (DE3, pET23b, p20) cells. The gels were stained with Coo- massie Brilliant Blue, destained and the images were scanned with a ImageScanner (GE Healthcare). GluTR and GluRS proteins in the various fractions were quanti- fied using GluRS and GluTR standards as the reference. Bands corresponding to GluTR were identified by western blotting and immunodecoration with rabbit anti-Mt- GluTR IgG prepared for us by PRIMM srl (Milan) using samples of electrophoretically homogeneous Mt-GluTR– His6. The latter was prepared by Ni-nitrilotriacetic acid– denaturing chromatograhy Sepharose conditions (6 m urea) from E. coli BL21 (DE3) cells trans- formed with pGTRHis and grown at 25 (cid:2)C with the addi- tion of IPTG when D600 was 1 and harvested after 15 h. A similar experimental scheme was followed by using extracts of cells that had overproduced GluTR–His and purified GluRS.

the amount of

GluRS ⁄ GluTR interaction

ampicillin medium at

0.1 mgÆmL)1

The method described by Baick et al. [25] to establish the toxicity of the expression of B. subtilis nondiscriminating GluRS in E. coli was adapted. E. coli BL21 (DE3, pET- GTS1) cells were grown in Luria–Bertani medium contain- 30 (cid:2)C and ing

Rv0509 encoding the putative Mt-GluTR was amplified from cosmid MTCY20G9 (Institut Pasteur) following a scheme similar to that described for the construction of pETGTS1. The insert of the resulting plasmid (pGTR) was also reamplified in order to remove the stop codon and engineer a XhoI site at the 3¢-end of the ORF to allow for cloning in pET23b digested with NdeI and XhoI. The resulting plasmid (pGTRHis) encodes a GluTR species car- rying a C-terminal His6 tag (Mt-GluTR–His). With both soluble protein increased by plasmids co-transforming E. coli BL21(DE3) cells with pGTR6 or pGTRHis and p20 [28]. The latter plasmid encodes the E. coli chaperons DnaJ, DnaK and GrpE. Transformed E. coli cells were grown in Luria–Bertani medium supple- mented with 100 lgÆmL)1 ampicillin, 25 lgÆmL)1 chloram- phenicol until the D600 of the culture reached a value of (cid:3) 1. The culture was transferred at 15 (cid:2)C and induction of

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1414

Testing the toxicity of Mt-GluRS in E. coli BL21(DE3) cells

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

References

1 Jain A & Mondal R (2008) Extensively drug-resistant tuberculosis: current challenges and threats. FEMS Immunol Med Microbiol 53, 145–150.

0.1 mm IPTG or

2 Ibba M & Soll D (2000) Aminoacyl–tRNA synthesis.

Annu Rev Biochem 69, 617–650.

or

plates

ampicillin

3 Schimmel P, Tao J & Hill J (1998) Aminoacyl tRNA

synthetases as targets for new anti-infectives. FASEB J 12, 1599–1609.

4 Ibba M, Losey HC, Kawarabayasi Y, Kikuchi H,

Bunjun S & Soll D (1999) Substrate recognition by class I lysyl–tRNA synthetases: a molecular basis for gene displacement. Proc Natl Acad Sci USA 96, 418–423. 5 Sekine S, Nureki O, Dubois DY, Bernier S, Chenevert R, Lapointe J, Vassylyev DG & Yokoyama S (2003) ATP binding by glutamyl–tRNA synthetase is switched to the productive mode by tRNA binding. EMBO J 22, 676–688.

6 Salazar JC, Zuniga R, Raczniak G, Becker H, Soll D & Orellana O (2001) A dual-specific Glu–tRNA(Gln) and Asp–tRNA(Asn) amidotransferase is involved in decod- ing glutamine and asparagine codons in Acidithiobacil- lus ferrooxidans. FEBS Lett 500, 129–131.

7 Curnow AW, Hong K, Yuan R, Kim S, Martins O,

220 r.p.m. until the culture reached an D600 of 1.0. Serial dilutions (to 10)5 and 2 · 10)6) were done in Luria– Bertani broth. Aliquots (200 lL) of the 10)5 dilution were added to 3 mL 0.7% top agar made with Luria–Bertani IPTG and medium containing 0.1 mgÆmL)1 ampicillin. The top agar was poured onto Luria–Bertani with without 0.1 mgÆmL)1 ampicillin, respectively. Aliquots (200 lL) of the 2 · 10)6 diluted culture were mixed with top agar with 0 or 0.1 mgÆmL)1 ampicillin, and the top agar was poured on Luria–Bertani or Luria–Bertani medium containing 0.1 mgÆmL)1 ampicillin plates, respectively. The plates were incubated for up to 48 h at 37 or 25 (cid:2)C (the latter to mim- ick large scale growth conditions). The effect of l-Glu or l-Gln (2.5 and 25 mm) in the Luria–Bertani medium was also tested. In separate experiments, the toxicity caused by the expression of Mt-GluRS in E. coli cells was determined in M9 minimal medium in the presence of different l-Gln or l-Glu concentrations. For these experiments E. coli BL21 (DE3, pETGTS1) was grown in Luria–Bertani med- ium containing 0.1 mgÆmL)1 ampicillin serially diluted in 0.9% NaCl to 10)5 and 2 · 10)6. Aliquots (200 lL) were mixed with top agar and then poured on plates as described above except for the fact that M9 medium was used and that three series of samples were prepared con- taining 0, 2.5 and 25 mm l-Gln or l-Glu in both top agar and plates. After incubation at 37 or 25 (cid:2)C for up to 48 h, the formed colonies were counted.

Winkler W, Henkin TM & Soll D (1997) Glu–tRNAGln amidotransferase: a novel heterotrimeric enzyme required for correct decoding of glutamine codons dur- ing translation. Proc Natl Acad Sci USA 94, 11819– 11826.

8 Heinemann IU, Jahn M & Jahn D (2008) The biochem- istry of heme biosynthesis. Arch Biochem Biophys 474, 238–251.

UV–Vis absorbance measurements were done on HP8453 (Agilent Technologies, Santa Clara, CA, USA), Cary219 or Cary100 (Varian, Palo Alto, CA, USA) spectrophotometers connected to water baths.

9 Jahn D, Verkamp E & Soll D (1992) Glutamyl-transfer RNA: a precursor of heme and chlorophyll biosynthe- sis. Trends Biochem Sci 17, 215–218.

Miscellaneous techniques

Acknowledgements

10 Levican G, Katz A, Valenzuela P, Soll D & Orellana O (2005) A tRNA(Glu) that uncouples protein and tetra- pyrrole biosynthesis. FEBS Lett 579, 6383–6387. 11 Jahn D (1992) Complex formation between glutamyl–

tRNA synthetase and glutamyl–tRNA reductase during the tRNA-dependent synthesis of 5-aminolevulinic acid in Chlamydomonas reinhardtii. FEBS Lett 314, 77–80.

12 Sekine S, Nureki O, Shimada A, Vassylyev DG &

Yokoyama S (2001) Structural basis for anticodon rec- ognition by discriminating glutamyl–tRNA synthetase. Nat Struct Biol 8, 203–206.

13 Sekine S, Shichiri M, Bernier S, Chenevert R, Lapo- inte J & Yokoyama S (2006) Structural bases of transfer RNA-dependent amino acid recognition and activation by glutamyl–tRNA synthetase. Structure 14, 1791–1799.

14 Nureki O, Fukai S, Sekine S, Shimada A, Terada T,

Nakama T, Shirouzu M, Vassylyev DG & Yokoyama S (2001) Structural basis for amino acid and tRNA

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1415

This work was carried out thanks to funds from the Ministero dell’Istruzione, Universita’ e Ricerca MIUR- PRIN2003 (Rome, Italy), the European Union Con- tract QLK2-CT-2000-01761 to BC and Fondazione Cariplo (Milano, Italy) Contract 2004-1580 to MAV. G. Riccardi, E. De Rossi and A. Aliverti are thanked for the initial cloning of Rv0909. We are grateful to J. Lapointe for the gift of pERS, to Dr Rizzi for the gift of decoyinine and of the pyrophosphate analogs tested, to A. Mattevi and M. Nardini for carrying out crystallization trials, to G. Tedeschi and A. Negri for performing MALDI-TOF analyses and N-terminal sequencing, and to G. Deho` for helpful discussions. PVK, MVP and DIS acknowledge support from the EU design study SAXIER (contract RIDS No. 011934).

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

the testing of chaperone-assisted folding of proteins overexpressed in Escherichia coli. Anal Biochem 254, 150–152.

recognition by class I aminoacyl–tRNA synthetases. Cold Spring Harbor Symp Quant Biol 66, 167–173. 15 Schulze JO, Masoumi A, Nickel D, Jahn M, Jahn D,

29 Moser J, Lorenz S, Hubschwerlen C, Rompf A & Jahn D (1999) Methanopyrus kandleri glutamyl–tRNA reduc- tase. J Biol Chem 274, 30679–30685.

Schubert WD & Heinz DW (2006) Crystal structure of a non-discriminating glutamyl–tRNA synthetase. J Mol Biol 361, 888–897.

16 Sassetti CM, Boyd DH & Rubin EJ (2003) Genes

required for mycobacterial growth defined by high den- sity mutagenesis. Mol Microbiol 48, 77–84.

17 Lin SX, Brisson A, Liu J, Roy PH & Lapointe J (1992) Higher specific activity of the Escherichia coli glutamyl– tRNA synthetase purified to homogeneity by a six-hour procedure. Protein Expr Purif 3, 71–74.

18 Lapointe J, Levasseur S & Kern D (1985) Glutamyl–

30 Moser J, Schubert WD, Beier V, Bringemeier I, Jahn D & Heinz DW (2001) V-shaped structure of glutamyl– tRNA reductase, the first enzyme of tRNA-dependent tetrapyrrole biosynthesis. EMBO J 20, 6583–6590. 31 Moser J, Schubert WD, Heinz DW & Jahn D (2002) Structure and function of glutamyl–tRNA reductase involved in 5-aminolaevulinic acid formation. Biochem Soc Trans 30, 579–584.

tRNA synthetase from Escherichia coli. Methods Enzy- mol 113, 42–49.

19 Nakamura J & Lou L (1995) Biochemical characteriza- tion of human GMP synthetase. J Biol Chem 270, 7347–7353.

32 Schauer S, Chaturvedi S, Randau L, Moser J, Kitaba- take M, Lorenz S, Verkamp E, Schubert WD, Nakaya- shiki T, Murai M et al. (2002) Escherichia coli glutamyl–tRNA reductase. Trapping the thioester inter- mediate. J Biol Chem 277, 48657–48663.

33 Kern D, Potier S, Boulanger Y & Lapointe J (1979)

20 Kern D & Lapointe J (1981) The catalytic mechanism of glutamyl–tRNA synthetase of Escherichia coli. A steady- state kinetic investigation. Eur J Biochem 115, 29–38. 21 Desjardins M, Garneau S, Desgagnes J, Lacoste L,

Yang F, Lapointe J & Chenevert R (1998) Glutamyl adenylate analogues are inhibitors of glutamyl–tRNA synthetase. Bioorg Chem 26, 1–13.

22 Kern D & Lapointe J (1980) Catalytic mechanism of

The monomeric glutamyl–tRNA synthetase of Escheri- chia coli. Purification and relation between its structural and catalytic properties. J Biol Chem 254, 518–524. 34 Balg C, Blais SP, Bernier S, Huot JL, Couture M, Lapointe J & Chenevert R (2007) Synthesis of b-ketophosphonate analogs of glutamyl and glutaminyl adenylate, and selective inhibition of the corresponding bacterial aminoacyl–tRNA synthetases. Bioorg Med Chem 15, 295–304.

glutamyl–tRNA synthetase from Escherichia coli. Reac- tion pathway in the aminoacylation of tRNAGlu. Bio- chemistry 19, 3060–3068.

23 Kern D & Lapointe J (1980) The catalytic mechanism

of the glutamyl–tRNA synthetase from Escherichia coli. Detection of an intermediate complex in which gluta- mate is activated. J Biol Chem 255, 1956–1961.

35 Bernier S, Dubois DY, Habegger-Polomat C, Gagnon LP, Lapointe J & Chenevert R (2005) Glutamylsulfa- moyladenosine and pyroglutamylsulfamoyladenosine are competitive inhibitors of E. coli glutamyl–tRNA synthetase. J Enzyme Inhib Med Chem 20, 61–67. 36 Yu M, Magalhaes ML, Cook PF & Blanchard JS

(2006) Bisubstrate inhibition: theory and application to N-acetyltransferases. Biochemistry 45, 14788–14794. 37 Lee J & Hendrickson TL (2004) Divergent anticodon

24 Kern D & Lapointe J (1980) The catalytic mechanism of glutamyl–tRNA synthetase of Escherichia coli. Evi- dence for a two-step aminoacylation pathway, and study of the reactivity of the intermediate complex. Eur J Biochem 106, 137–150.

25 Baick JW, Yoon JH, Namgoong S, Soll D, Kim SI,

recognition in contrasting glutamyl–tRNA synthetases. J Mol Biol 344, 1167–1174.

38 Srivastava A & Beale SI (2005) Glutamyl–tRNA reduc- tase of Chlorobium vibrioforme is a dissociable homodi- mer that contains one tightly bound heme per subunit. J Bacteriol 187, 4444–4450.

Eom SH & Hong KW (2004) Growth inhibition of Esc- herichia coli during heterologous expression of Bacil- lus subtilis glutamyl–tRNA synthetase that catalyzes the formation of mischarged glutamyl–tRNA1 Gln. J Microbiol 42, 111–116.

26 Svergun DI, Barberato C & Koch MHJ (1995) CRY-

SOL – a program to evaluate X-ray solution scattering of biological macromolecules from atomic coordinates. J Appl Crystallogr 28, 768–773.

39 Srivastava A, Lake V, Nogaj LA, Mayer SM, Willows RD & Beale SI (2005) The Chlamydomonas reinhardtii gtr gene encoding the tetrapyrrole biosynthetic enzyme glutamyl–tRNA reductase: structure of the gene and properties of the expressed enzyme. Plant Mol Biol 58, 643–658.

40 Gornall AG, Bardewill CJ & David MM (1949) Deter- mination of serum proteins by means of the biuret reac- tion. J Biol Chem 177, 751–766.

27 Konarev PV, Volkov VV, Sokolova AV, Koch MHJ & Svergun DI (2003) PRIMUS – a Windows-PC based system for small-angle scattering data analysis. J Appl Crystallogr 36, 1277–1282.

28 Castanie MP, Berges H, Oreglia J, Prere MF & Fayet

41 Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein

O (1997) A set of pBR322-compatible plasmids allowing

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1416

S. Paravisi et al.

M. tuberculosis glutamyl–tRNA synthetase

utilizing the principle of protein–dye binding. Anal Bio- chem 72, 248–254.

42 Laemmli UK (1970) Cleavage of structural proteins

during the assembly of the head of bacteriophage T4. Nature 227, 680–685.

separation of Mt-GluRS

43 Matsudaira P (1987) Sequence from picomole quantities of proteins electroblotted onto polyvinylidene difluoride membranes. J Biol Chem 262, 10035–10038. 44 Cantor CR & Schimmel PR (1980) Biophysical Chemistry. Freeman, San Francisco, CA.

45 Segel IH (1975) Enzyme Kinetics. Wiley, New York,

NY.

46 Orr GA & Blanchard JS (1984) High-performance ion- exchange separation of oxidized and reduced nicotin- amide adenine dinucleotides. Anal Biochem 142, 232–234.

47 Ogawa T & Okazaki T (1979) RNA-linked nascent

DNA pieces in phage T7-infected Escherichia coli. III. Detection of intact primer RNA. Nucleic Acids Res 7, 1621–1633.

48 Marini F & Wood RD (2002) A human DNA helicase homologous to the DNA cross-link sensitivity protein Mus308. J Biol Chem 277, 8716–8723.

49 Roessle MW, Klaering R, Ristau U, Robrahn B, Jahn D, Gehrmann T, Konarev PV, Round A, Fiedler S, Hermes C et al. (2007) Upgrade of the small-angle X-ray scattering beamline X33 at the European Molecu- lar Biology Laboratory, Hamburg. J Appl Crystallogr 40, s190–s194.

the kinetics of proteolysis of

50 Svergun DI (1992) Determination of the regularization parameter in indirect-transform methods using percep- tual criteria. J Appl Crystallogr 25, 495–503.

Fig. S4. Inhibition of Mt-GluRS by GoA and pyro- phosphate. Fig. S5. Chromatographic reaction components. Fig. S6. PPi ⁄ ATP exchange reaction of Mt-GluRS. Fig. S7. Alignment of selected GluRS sequences and identification of the limited chymotryptic cleavage sites. Fig. S8. Analysis of Mt- and Ec-GluRS. Fig. S9. Minimal models of the proteolytic events lead- ing to fragments M1–M5 of Mt-GluRS and E1-E3 of Ec-GluRS. Fig. S10. DLS analysis of GluRS aggregation state. Fig. S11. Spectral properties of Mt-GluTR isolated by using immobilized His6–GluRS and aggregation state. Table S1. Alternate substrates and inhibitors of Mt- GluRS: l-Glu, l-Gln and 2-oxoglutarate. Table S2. Alternate substrates and inhibitors of Mt- GluRS: ATP, AMP, pyrophosphate and their analogs. ions on the Mt-GluRS Table S3. Effect of metal activity. Table S4. Testing the toxicity of Mt-GluRS in E. coli BL21(DE3) cells. Table S5. Summary of the properties of the fragments cleavage of obtained during limited chymotryptic Mt- and Ec-GluRS as deduced from the gels shown in Fig. 5, main text. Table S6. Summary of molecular parameters deduced by SAXS. This supplementary material can be found in the online version of this article.

Supporting information

FEBS Journal 276 (2009) 1398–1417 ª 2009 The Authors Journal compilation ª 2009 FEBS

1417

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corre- sponding author for the article. steady-state kinetic the The following supplementary material is available: Fig. S1. Mt-GluRS production and purification. Fig. S2. tRNA aminoacylation activity of Mt-GluRS. Fig. S3. Determination of parameters of the Mt-GluRS reaction.