Chester Clinic

Peculiar inhibition of human mitochondrial aspartyl-trna synthetaseby adenylate analogs

Biochimie 91 (2009) 596–603 Contents lists available at Peculiar inhibition of human mitochondrial aspartyl-tRNA synthetaseby adenylate analogs Marie Messmer Se´bastien P. Blais , Christian Balg Robert Cheˆnevert Luc Grenier , Patrick Lagu¨e ,Claude Sauter Marie Sissler Richard Giege´ Jacques Lapointe , Catherine Florentz a Architecture et Re´activite´ de l'ARN, Universite´ Louis Pasteur, CNRS, IBMC 15 rue Rene´ Descartes, 67084 Strasbourg Cedex, Franceb De´partement de Biochimie et de Microbiologie, Centre de recherche sur la fonction, la structure et l'inge´nierie des prote´ines (CREFSIP), Faculte´ des sciences et de ge´nie, Universite´Laval, Que´bec, Canada G1V 0A6c De´partement de Chimie, Centre de recherche sur la fonction, la structure et l'inge´nierie des prote´ines (CREFSIP), Faculte´ des sciences et de ge´nie, Universite´ Laval, Que´bec,Canada G1V 0A6 Human mitochondrial aminoacyl-tRNA synthetases (mt-aaRSs), the enzymes which esterify tRNAs with Received 16 December 2008 the cognate specific amino acid, form mainly a different set of proteins than those involved in the Accepted 18 February 2009 cytosolic translation machinery. Many of the mt-aaRSs are of bacterial-type in regard of sequence and Available online 28 February 2009 modular structural organization. However, the few enzymes investigated so far do have peculiarbiochemical and enzymological properties such as decreased solubility, decreased specific activity and enlarged spectra of substrate tRNAs (of same specificity but from various organisms and kingdoms), as Human mitochondria compared to bacterial aaRSs. Here the sensitivity of human mitochondrial aspartyl-tRNA synthetase Aspartyl-tRNA synthetase (AspRS) to small substrate analogs (non-hydrolysable adenylates) known as inhibitors of Escherichia coli and Pseudomonas aeruginosa AspRSs is evaluated and compared to the sensitivity of eukaryal cytosolic Molecular docking human and bovine AspRSs. L-aspartol-adenylate (aspartol-AMP) is a competitive inhibitor of aspartyla-tion by mitochondrial as well as cytosolic mammalian AspRSs, with Ki values in the micromolar range (4–27 mM for human mt- and mammalian cyt-AspRSs). 50-O-[N-(L-aspartyl)sulfamoyl]adenosine (Asp-AMS)is a 500-fold stronger competitive inhibitor of the mitochondrial enzyme than aspartol-AMP (10 nM) anda 35-fold lower competitor of human and bovine cyt-AspRSs (300 nM). The higher sensitivity of humanmt-AspRS for both inhibitors as compared to either bacterial or mammalian cytosolic enzymes, is notcorrelated with clear-cut structural features in the catalytic site as deduced from docking experiments,but may result from dynamic events. In the scope of new antibacterial strategies directed against aaRSs,possible side effects of such drugs on the mitochondrial human aaRSs should thus be considered.
Ó 2009 Elsevier Masson SAS. All rights reserved.
within the cytosol and imported . They include the sets of ami-noacyl-tRNA synthetases (aaRSs), ribosomal proteins, translation Two distinct translational machineries coexist in mammalian factors and tRNA maturation and modification enzymes. AaRSs are cells. The mitochondrial machinery is still in the process of char- the enzymes which catalyze specific esterification of their cognate acterization. While its 22 tRNAs, 11 mRNAs (2 are polycistronic) that tRNAs by the corresponding amino acids. Most of the genes code for 13 proteins, and 2 rRNAs are encoded by the mitochondrial encoding the human aaRSs have been annotated, demonstrating (mt) genome, all other macromolecules needed for protein their distribution into two distinct sets Except for GlyRS synthesis are coded by the nuclear chromosome, synthesized and LysRS , mt- and cytosolic-aaRSs (cyt-aaRSs) are encoded bydistinct genes.
In agreement with the endosymbiotic hypothesis for the origin of mitochondria sequence features and modular organizationof many mt-aaRSs are of bacterial-type and thus differ from the Abbreviations: aaRS, aminoacyl-tRNA synthetase, with aa for the amino acid in eukaryotic-type corresponding cyt-aaRSs Biochemical and three-letter abbreviation; mt, mitochondrial; cyt, cytosolic; aspartol-AMP, aspartol- enzymatic characterization of an initial set of human bacterial-type mt-aaRSs revealed however unexpected properties making these * Corresponding author. Tel.: þ33 3 88 41 70 59; fax: þ33 3 88 60 22 18.
E-mail address: (C. Florentz).
enzymes functionally distinct from their bacterial counterparts. As 0300-9084/$ – see front matter Ó 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.biochi.2009.02.005 M. Messmer et al. / Biochimie 91 (2009) 596–603 an example, human mt-AspRS and mt-TyrRS share about 40% 2. Materials and methods sequence identity with the corresponding Escherichia coli enzymes(including strongly conserved functional amino acids) and present 2.1. Materials and enzymes the same modular organization . The crystallographic structureof mt-TyrRS reveals a three-dimensional fold very similar to that of Total E. coli tRNA was purchased from Roche Diagnostics and total Bacillus subtilis and E. coli TyrRSs . However, both mt-aaRSs calf liver tRNA from Novagen. L-[2,3-3H]aspartic acid (specific aminoacylate their substrates with 10- to 40-fold less efficiency activity 34 Ci/mmol) was from GE Healthcare. Aspartol-AMP and than the corresponding E. coli aaRSs . These mt-enzymes require Asp-AMS were synthesized as reported Ni-NTA resin was from restricted sets of identity elements within their cognate tRNAs Qiagen Inc. Human (Homo sapiens) mt-AspRS was previously cloned compared to bacterial AspRSs or TyrRSs . Finally, both mt- into pQE70 vector that introduces a poly-His tag to the C-terminus of AspRS and mt-TyrRS likely have an enlarged spectrum of possible the expressed protein. Overproduction and purification steps were tRNA substrates, as first observed after comparing E. coli aaRSs with conducted as described Human cyt-AspRS was a kind gift of homologous bovine enzymes . Indeed mitochondrial enzymes M. Frugier (Stasbourg). Bovine (Bos taurus) cyt-AspRS was purchased aminoacylate tRNAs of same specificity from a large range of from Bio S&T Inc. (Montreal, Canada) as a mixture of different ami- organisms, while most bacterial enzymes recognize and amino- acylate only their own tRNA.
Aminoacyl-tRNA synthetases have been subjected to signifi- 2.2. Aminoacylation and inhibition assays cant evolutionary divergence, so that selective inhibition ofbacterial enzymes appears as a valuable strategy for the Aspartylation assays in the presence of human mt-AspRS were production of new antibiotics (reviewed in Refs. Such carried out in 50 mM HEPES-NaOH pH 7.5, 2.5 mM ATP, 12 mM antibiotics are expected to have strong negative effects on MgCl2, 25 mM KCl, 0.2 mg/ml BSA, 1 mM spermine and 40 mM total pathogenic bacteria, but should not affect the human host.
E. coli tRNA. For establishing the Km for aspartate, this substrate was Pseudomonic acid (mupirocin) is the first known effective anti- added at concentrations ranging from 0.7 to 40 mM. The reaction biotic of this type and inhibits IleRSs from Gram positive (e.g.
was initiated by adding the pre-warmed enzyme at 37 C to a final Staphylococcus aureus) and Gram negative (e.g. Neisseria menin- concentration of 62.5 nM. The amount of aspartyl-tRNA formed gitidis) bacteria with a 8000-fold higher affinity than for was determined by the radioactivity present in 5% trichloroacetic mammalian cyt-IleRS . This natural product is a stable acid precipitates of reaction mixture aliquots, as previously adenylate analog and is in clinical use Beside IleRS, many described Initial reaction rates were determined by measuring other aaRSs are inhibited by adenylate derivatives, of synthetic [3H]aspartyl-tRNA formed in 5 ml aliquots (from a total volume of and in a few cases of natural origin, that can be considered as 50 ml) taken at 1 min intervals over 6 min. Inhibition constants (Ki) potential drugs targeting aaRSs .
were determined at the aspartate concentration corresponding to We have previously synthesized aspartyl-adenylate analogs the Km value and the inhibitors aspartol-AMP and Asp-AMS were and established that they have inhibitory effects on bacterial added at various concentrations from 0.5 to 100 mM and from 1 to aaRSs as tested on E. coli and Pseudomonas aeruginosa AspRSs 50 nM, respectively, to reaction media pre-heated to 37 C, 2 min Here, the effect of Asp-AMS and aspartol-AMP () is explored on before addition of the synthetase. The error range was 15% for human mt-AspRS as well as on mammalian cytosolic AspRSs (human and bovine). The inhibition produced by Asp-AMS and Aspartylation assays performed in the presence of human cyt- aspartol-AMP on the activity of the three enzymes was investigated AspRS were carried out as for human mt-AspRS but with 80 mM total and compared with the effect produced on bacterial AspRSs.
calf tRNA and 0.5 nM enzyme. To determine the Km for aspartate for Functional studies were completed by computer-assisted docking the human cyt-AspRS, this substrate was added to final concentra- of the adenylates in the catalytic site of the diverse AspRSs. Data tions ranking from 2.5 to 120 mM. Aliquots of 5 ml were taken from reveal differences between the three types of enzymes (bacterial, a total volume of 50 ml, at 2 min intervals over 12 min and treated as mt- and cyt-eukaryal) and strikingly highest sensitivity of the described above. Establishment of Ki for aspartol-AMP and Asp-AMS mitochondrial enzyme to both inhibitors. They further support was set to 24 mM of aspartate (Km value) and was done as described functional differences between the bacterial-type human mt- above but with inhibitor concentrations varying from 5 mM to AspRS and bacterial AspRSs. These functional peculiarities are not 100 mM and from 50 nM to 1 mM, respectively. The error range was due to striking structural idiosyncrasies in the catalytic domain of 10% for triplicate experiments.
the AspRSs, in particular of human mt-AspRS, as suggested bydocking of the adenylate in the catalytic sites. Structure–function 2.3. Determination of inhibition type and constant (Ki) relationships and the implications for medical research dedicatedto the discovery of new antibiotics using aaRSs as targets will be The Km values of human cyt-AspRS and human mt-AspRS for the aspartate substrate were calculated from Michaelis–Menten plots.
Fig. 1. Comparison of the chemical structures of aspartyl-adenylate and its two analogs, L-aspartol-adenylate and 50-O-[N-(L-aspartyl)sulfamoyl]adenosine.

M. Messmer et al. / Biochimie 91 (2009) 596–603 The rate ‘vi' of the aminoacylation reaction in the presence of Docking of the natural aspartyl-adenylate and of two analogs inhibitor at various concentrations [I] is given by the following (Asp-AMS and aspartol-AMP) was performed using AUTODOCK 3.0.5 . Hydrogen atoms were added to proteins and ligandsusing Gast-Eigen charges were computed for the ligand partial atomic ½S þ Kmð1 þ ð½I=KiÞÞ Three-dimensional grids of interaction energies based on the When reactions are conducted at the amino acid concentration macromolecular target using the AMBER force field were calculated corresponding to the Km value and in the presence of saturating using AutoGrid. The cubic grid box of 60 Å size (x, y, z) with concentrations of ATP and tRNA, Eq. can be rearranged as Eq. a spacing of 0.375 Å and grid maps were centered on the respective with the ratio vi/v0 (where v0 is the rate in the absence of inhibitor AspRS active sites. Automated docking studies were carried out to under the same substrate concentrations) expressed as a function evaluate the binding free energy of the inhibitors within the of [I]. This equation illustrates competitive inhibition with one macromolecules. The GA-LS search algorithm (algorithm with local binding site for the inhibitor and can be simplified in Eq. search) was chosen to search for the best conformers. The param- eters were set using the software ADT. For all docking parameters, default values were used with 30 independent docking runs for each docking case. Each three-dimensional model was used as ½S þ Kmð1 þ ð½I=KiÞÞ a rigid scaffold, but the three ligands benefited from a full freedomwith respect to their flexibility to be able to adapt to the catalytic groove. A theoretical Ki value was derived from the calculated binding free energy [Ki ¼ exp(DGbinding/RT)]. An average pKi value Curve-fitting of the data was made with the Sigmaplot software (pKi ¼ log Ki) is given for all trials falling within 2 Å rmsd from the (SPSS Inc) and was used to identify the type of inhibition and to position of the natural adenylate in crystallographic structure of E. coli ternary complex. The figure presenting docking results was i values. Inhibition type of Asp-AMS with respect to aspartate for both human AspRSs was further confirmed by prepared with PyMOL (DeLano Scientific LLC, CA).
the determination of the apparent Km for aspartate in thepresence of several concentrations of this inhibitor (from 0 to 40 nM for human mt-AspRS and from 0 to 1.2 mM for human cyt-AspRS) and under fixed and saturating concentrations of the two 3.1. Activity of human and bovine AspRSs inhibited by aspartyl- other substrates (2.5 mM ATP and 40 mM E. coli total tRNA for adenylate analogs human mt-AspRS or 80 mM total calf tRNA for human cyt-AspRS).
As a prerequisite to the search of the inhibition of adenylates on AspRS activity (see ) we determined the 2.4. Docking of adenylate and analogs Km for aspartate of human cyt- and mt-AspRSs and for bovine cyt-AspRS. Initial rates of aspartylation, established under saturating Three-dimensional models of candidate AspRSs were derived tRNA and ATP concentrations were obtained for aspartate from crystallographic structures of close relatives in complex with concentrations ranging from 1 to 40 mM (mt-AspRS) and 2.5 to tRNA. Yeast binary complex (1ASY.pdb – Ref. ) and the E. coli 120 mM (cyt-AspRSs) , top). Analysis of the Lineweaver–Burk ternary complex (1CA0.pdb – Ref. were used to model bovine plots (, bottom) yielded similar Km values for the two cytosolic and human cyt-AspRSs and bacterial-type enzymes (P. aeruginosa AspRSs (24 mM and 37 mM for the human and bovine enzymes, and human mitochondria, respectively). The three-dimensional respectively) and a strikingly lower value (1.5 mM) for the human model of the mt-AspRS was built using modeler as described previously whereas others were generated using the web- The inhibition constants (Ki) of Asp-AMS and aspartol-AMP have based SWISS-MODEL workspace .
been determined with respect to aspartate for the three AspRSs at Fig. 2. Determination of the apparent Km values for aspartate for H. sapiens mt-AspRS (A), H. sapiens cyt-AspRS (B) and for B. taurus cyt-AspRS (C).

M. Messmer et al. / Biochimie 91 (2009) 596–603 fixed and saturating concentrations of ATP and tRNA and at Km coordinates from E. coli complex used as a template for homology concentration for the aspartic acid. Kinetic data were displayed as modeling did contain this ligand. The difference of 3 theoretical pKi normalized initial rates of tRNA aminoacylation (vi/v0,) as a func- units between bacterial- and eukaryotic-type systems may also be tion of inhibitor concentration Ii (with vi being the initial rates in linked to the resolution of the original template, the E. coli structure the presence of inhibitor and v0 the rate in the absence of inhibitor).
determined with a higher accuracy (2.3 vs 2.9 Å resolution) giving If inhibitions are competitive, as can be anticipated for adenylate the highest docking scores. Beside these technical aspects, the most analogs, experimental data should fit on sigmoidal curves (see striking feature is that pKi values obtained for a given AspRS do not dramatically vary from one ligand to the others. Moreover, no computed for aspartol-AMP fit perfectly with the experimental significant behavior difference is detected between AspRSs points indicating that the inhibition is indeed competitive belonging to the same group.
for the three enzymes with Ki values ranging from 4.6 to 27 mM As an illustration, shows the locations of aspartyl-adeny- ). In the experiments done in the presence of Asp-AMS, the late, aspartol-AMP and Asp-AMS superimposed in the active site fit is less perfect, especially in the case of the human mt-AspRS.
human mt-AspRS (in the homology model of E. coli AspRS). The Assuming inhibitions are competitive, extracted Ki values are quite docking suggests excellent superimpositions of the adenine similar for the two mammalian cyt-AspRSs (390 and 280 nM for the (A, right) and aspartate , left) moieties at the distal human and bovine enzymes) and 9.8 nM for the mt-AspRS. These extremities of the adenylate molecules and slight changes in the values are 2–3 orders of magnitude below those measured for orientation of the ribo-phosphate and ribo-sulfamoyl groups in the central part of these molecules.
To verify whether the deviations from the theoretical curve with Asp-AMS originate from experimental errors or result from a more complex kinetic behavior, we undertook a classical Lineweaver–Burk analysis for the human mitochondrial and cytosolic enzymes 4.1. General considerations Apparent Km values for aspartate have been determinedunder large ranges of inhibitor concentrations. All lines cross the y- Aminoacyl-tRNA synthetases catalyze the esterification of their axis into a single point that corresponds to 1/Vmax, conclusively cognate tRNA with the specific amino acid in a two-step process. In demonstrating that inhibition by Asp-AMS is of competitive type the first step, the amino acid is recognized by the enzyme and with respect to aspartate both for human cyt-AspRS and mt-AspRS.
reacts with ATP to form an enzyme-bound mixed anhydride (aa- Kinetic parameters in these experiments are very close to those AMP or aminoacyl-adenylate) with release of pyrophosphate .
In this intermediate, the high-energy anhydride bond activates thecarboxyl group of the amino acid. In the second step, the activated 3.2. Molecular docking of aspartyl adenylate and analogs in the amino acid is transferred to the 30-terminal adenosine of the cor- active site of AspRSs responding tRNA to form aminoacyl-tRNA and AMP (reviewed inRefs. This overall mechanism applies to both class I and Three-dimensional models of four AspRSs (from P. aeruginosa, class II aaRSs (reviewed in Refs. ). While the overall func- human mitochondria, human and bovine cytoplasms) were tioning of aaRSs is essentially conserved in evolution, one notes generated to study the binding of aspartyl-adenylate and its idiosyncrasies when comparing properties of aaRSs from phylo- analogs (Asp-AMS and aspartol-AMP) by molecular docking. These genetically distant species or organelles and this opens the three-dimensional models were based on X-ray structures from possibility to find or design species-selective inhibitors of aaRSs.
E. coli and yeast AspRS:tRNAAsp complexes (see Here we focus on human mt-AspRS for a better understanding of ). Individual monomers were considered (i.e. one active its functional and structural idiosyncrasies, especially in regard of site) in the absence of tRNA. In order to get comparative scores, inhibition by small substrate analogs targeting its catalytic site. For docking trials were performed for each ligand, both on the original comparative purposes, two novel mammalian cyt-AspRSs (human X-ray structures and on the four models (with protein backbone and bovine) were studied for their behavior to interact with aspartic and side chain orientation maintained as in the reference X-ray acid and two adenylate analogs. reports the Km and Ki values structures). The results are presented in In the case of for the three mammalian enzymes and compares these values with bacterial-type enzymes, the natural adenylate systematically gives those previously determined for E. coli and P. aeruginosa AspRSs .
a slightly better score, probably due to a structural bias: the X-ray Remarkable variations are observed that are best visualized in the Fig. 3. Inhibition kinetics with Asp-AMS (left) and aspartol-AMP (right) of H. sapiens mt-AspRS, H. sapiens cyt-AspRS and B. taurus cyt-AspRS. Abbreviations used: Hs, Homo sapiens;Bt, Bos taurus.

M. Messmer et al. / Biochimie 91 (2009) 596–603 27 mM) and for Asp-AMS (390 and 280 nM). Over the 5 enzymes Kinetic parameters Km of aspartyl-adenylate and Ki of non-hydrolysable analogs for considered, human mt-AspRS is the most sensitive enzyme towards mammalian and bacterial AspRSs. Abbreviations used: Pa, Pseudomonas aeruginosa; each inhibitor with a Ki of 4.6 mM for aspartol-AMP and of 9.8 nM Ec, Escherichia coli; Hs, Homo sapiens; Bt, Bos taurus. n.d. stands for non-determined.
for Asp-AMS ().
Aspartate Km (mM) Aspartol-AMP Ki (mM) 4.3. H. sapiens mitochondrial AspRS is more sensitive to adenylates than bacterial AspRSs The data of and highlight distinct behaviors for the three families of enzymes considered (bacterial, bacterial-type, eukaryal). Considering either K Experimental data taken from Ref. m values for the natural substrate aspartate or Ki values of the inhibitors, the mitochondrial enzymebehaves apart from the four other AspRSs discussed here. Not only histogram comparing the inverse of the Km and Ki values ().
is aspartic acid retained with the best relative affinity for this Human mt-AspRS presents the most atypical functional behavior enzyme (assuming that the inverse of Km values is representative of deviating significantly from what observed with other AspRSs. It is the affinity) but also the adenylate analogs do present the highest common sense to believe that the functional differences are due to relative inhibitory properties.
structural idiosyncrasies of the different AspRSs, but as shown in Mt-AspRS displays a much higher affinity for aspartic acid than other tRNA aminoacylation systems, large functional differences the two cyt-AspRSs (16–25-fold) and than the two bacterial AspRSs could originate from faint structural effects . For interpreta- (60–70-fold). This markedly small Km value is in support of tion of the present data it should be kept in mind that all the above distinct kinetic properties for the enzyme families considered and results were obtained by kinetic analyses conducted in the presence especially of the mitochondrial bacterial-type enzyme as compared to of tRNA, and that former experiments with yeast AspRS have shown the two other families. Bacterial synthetases (E. coli and P. aeruginosa) that tRNAAsp significantly increases the affinity of aspartyl-adenylate present the poorest affinity for their amino acid substrate, while for the synthetase .
the two eukaryal enzymes present a 3–4-fold better affinity.
In regard to inhibitors, aspartol-AMP presents a 2–6-fold lower Ki 4.2. Aspartol-AMP and Asp-AMS are competitive inhibitors of for mt-AspRS than for the human or bovine cyt-AspRSs and about H. sapiens and B. taurus AspRSs 10-fold lower than the bacterial AspRSs. Asp-AMS has also thehighest inhibitory effect for the mitochondrial enzyme, but it Aspartol-AMP and Asp-AMS are analogs of aspartyl-adenylate, remains close to those measured for E. coli AspRS but distinguishes the natural derivative formed by AspRS in the presence of aspartic strongly from the cyt-enzymes. In summary, whatever the inhibitor, acid and ATP, during the first step of the aminoacylation reaction.
it has the strongest effect on the mitochondrial enzyme. This Aspartol-AMP differs from aspartyl-adenylate in converting an enzyme distinguishes thus from both other families of considered aminoacyl-adenylate into an aminoalcohol-adenylate, while Asp- AspRSs, namely bacterial and eukaryal cytosolic AspRSs. Note that AMS has a sulfamoyl function In a previous work, it was both these families present an inverted reactivity, with inhibition by shown as anticipated, that both molecules are competitive inhibi- aspartol-AMP more important with the eukaryal enzymes and tors of aspartate in bacterial AspRSs (E. coli and P. aeruginosa) inhibition by Asp-AMS more important with the prokaryal enzymes.
. While aspartol-AMP is a weak inhibitor for these twoAspRSs with Ki values in the micromolar (mM) range, Asp-AMS is 4.4. Search for structure–function relationships a strong inhibitor with Ki in the nanomolar (nM) range ). Inthe present work, that extends the analysis to eukaryal cyt-AspRSs Two aspects have to be considered here. First, the functional (human and bovine) and to human mt-AspRS, both adenylate difference between the human mt-AspRS and the two bacterial analogs behave also as competitive inhibitors. Inhibition constants AspRSs from E. coli and P. aeruginosa, given the fact that the human (Ki) of aspartol-AMP remain in the mM range and those of Asp-AMS enzyme is of bacterial-type and thus structurally similar to remain in the nM range, as was the case for the bacterial enzymes bacterial AspRSs. Second, the strong inhibitory effect produced by . Interestingly, both cytosolic mammalian AspRSs (human Asp-AMS with an affinity about three orders of magnitude higher and bovine) have about the same Ki for aspartol-AMP (10 and than the closely related aspartol-AMP ). In the absence of Fig. 4. Inhibition with Asp-AMP of the aminoacylation activity of H. sapiens mt-AspRS (A) and cyt-AspRS (B). Experiments have been performed in the presence of different fixedconcentrations of Asp-AMS.

M. Messmer et al. / Biochimie 91 (2009) 596–603 Fig. 5. Docking of aspartyl-adenylate and of two analogs in the active site of AspRSs. (A) Example of the best docking solution (i.e. highest binding energy) for the natural adenylate(or AMO with carbon atoms in medium blue), aspartol-AMP (or AOA, in green) and Asp-AMS (or AMS, in yellow). The protein backbone (with the antiparallel b-sheet characteristicof class II aaRSs) is represented in blue. Small variations are observed at the connection of the two adenylate moieties, whereas the position of the aspartate side chains and of theadenine ring is almost conserved. (B) Average docking scores obtained for the three substrates (same color code) in the active site of five AspRSs. Scores are indicated in theoreticalpKi values based on computed binding energies (see ). (For interpretation of color in this figure, the reader is referred to the web version of this article.) crystallographic structures of human mt-AspRS in its apo and catalysis (see legend to ) and only 3 differ in the two enzymes, liganded versions, as well as of the other AspRSs investigated in this namely Ile536, Ile581 and Leu583 in mt-AspRS, replaced respec- work, except the E. coli enzyme , a structural analysis of AspRSs tively by Phe533, Val483 and Leu531 in E. coli AspRS. These amino completed by docking studies of the adenylates in the active site of acids are not predicted to make energetically favorable bounds AspRSs can be useful.
with the adenylate, and in addition were not identified by crystal- AspRSs are modular proteins that belong to class IIb aaRSs. Their lography to contribute to adenylate binding in E. coli AspRS. This catalytic core encompasses a seven-stranded antiparallel b-sheet, suggests that the functional differences between the two AspRSs surrounded by a-helices that encompass the three class II signature are due to subtle structural effects and are not accounted by sole motifs. This fold is common to all class II aaRSs and differs from the thermodynamic binding features but are also kinetically driven Rossmann-fold of class I aaRSs The b-sheet offers a platform with induced fit and indirect effects. Such an interpretation finds where adenylates are formed. The overall structure and the cata- support from a mutational analysis of the active site of yeast AspRS lytic core of AspRSs are roughly conserved in evolution, but present that identified 23 functionally important amino acids by a genetic idiosyncrasies specific to phylogenic kingdoms (lower and higher selection method. Among these amino acids located around the eukarya, bacteria, archaea and organelles) (reviewed in Ref. ATP binding site, 10 act indirectly and were not identifiable by Interestingly, crystallographic structures tell us that the inter- crystallography .
action of aspartyl-adenylate with AspRSs is essentially the same In conclusion this analysis suggests that the Km and Ki variations than with free aspartate and the adenosine moiety of ATP. Indeed, observed in the test tube are not only the consequence of the as found with E. coli AspRS (), the AMP moiety of aspartyl- architecture of the active site itself and of the direct atomic envi- adenylate is positioned in a class II conserved manner, with ronment of the ligands, but also rely on the dynamics of ligand interactions of the a-phosphate with conserved Arg217 (Arg266 in binding, tRNA and small substrates, and the associated conforma- mt-AspRS) from motif 2. Further, recognition of the a-carboxyl and tional changes. In this process, adaptability of the flexible tRNA a-amino groups of aspartate by conserved AspRS residues is also molecule on the protein will likely be crucial . Deciphering class II characteristic, but with a system-specific interaction of the these subtleties of human mt-AspRS will require more functional side chain carboxylic group with Lys198, Arg489 (Lys247 and and structural work. In regard to functional investigation, it should Arg542 in mt-AspRS) whose basic side chains are stabilized by saltbridges with Asp233, Glu235 (would be Asp282 and Glu284 in mt-AspRS). Comparison between the E. coli AspRS structure containingaspartyl-adenylate and the apo structure shows that there is noconformational change whatsoever of the four conserved residuesLys198, Asp233, Glu235 and Arg489 (correspond to Lys247, Asp282,Glu284, Arg542 in mt-AspRS) from the catalytic domain uponaspartate binding . This implies a ‘‘lock-and-key'' recognition ofthe preformed adenylate analogs that is also found in archaealAspRS from Pyrococcus kodakaraensis and contrasts with theinduced fit occurring upon recognition of ATP with conformationalchanges in the active site of the E. coli enzyme, in particular atArg217.
On the other hand, sequence analysis of human mt-AspRS together with modeling of its three-dimensional structure (basedon the structure of E. coli AspRS:tRNAAsp:adenylate ternarycomplex, see and docking of the adeny-lates, indicates similar interaction patterns of the adenylates (see). The amino acids in human mt-AspRS and E. coli AspRSidentified by the docking procedure to make energetically favorable Fig. 6. Histogram comparing the (1/Km) values of aspartate and the (1/Ki) values of thetwo non-hydrolysable aminoalcohol (aspartol-AMP) and sulfamoyl (Asp-AMS) deriv- bonds with the adenylates or to be in vicinity of the adenylate in the atives for five different AspRSs. Abbreviations used: Pa, Pseudomonas aeruginosa; Ec, active cavity of the AspRSs are shown in Among these 24 Escherichia coli; Hs, Homo sapiens; Bt, Bos taurus. Experimental data for Pa and Ec amino acids, 10 were identified by crystallography to play a role in AspRSs are taken from Ref.
M. Messmer et al. / Biochimie 91 (2009) 596–603 Fig. 7. Schematized comparison of the active sites of H. sapiens mt-AspRS and E. coli AspRS in interaction with aspartyl-adenylate and its two analogs, Asp-AMS and aspartol-AMP.
The figure shows the chemical structure of aspartyl-adenylate and the structural changes in Asp-AMS (in red) and aspartol-AMP (in magenta). The dashed line is the computedproximity contour of the adenylates in the binding cavity of AspRSs. The diameter of the shadowed blue circles is proportional to exposure of adenylate atoms to the solvent. Theamino acids forming the catalytic cavity are shown circled in three-letter abbreviations and numbering as in E. coli and human mitochondrial AspRSs (colored black and blue,respectively). The 10 amino acids in E. coli AspRS that play a role in catalysis , and present in human mt-AspRS, are displayed on a green background. Notice that Lys198 andArg489 were also predicted by free energy simulations to be the main contributors for the specificity of aspartate recognition by AspRSs . Conserved residues in all AspRSs are inbold, the other amino acids being semi-conserved and characteristic of bacterial- and mt-AspRSs. The green and blue arrows show respectively the amino acid side chains orbackbones predicted by the docking simulations to hydrogen bond with atoms from the adenylates. Note that several putative bonds are not possible with the analogs. (Forinterpretation of color in this figure, the reader is referred to the web version of this article.) be kept in mind that the aminoacylation reaction is a two-step a stronger effect on bacterial AspRSs (E. coli and P. aeruginosa) than process including (i) formation of the aminoacyl-adenylate and (ii) on human cytosolic AspRS. Here, for the first time, a very strong transfer of the amino acid to the tRNA. Accordingly, the Ki value is inhibition by Asp-AMS of a human mitochondrial synthetase has not necessarily equal to the dissociation constant of the inhibitor been measured. These data suggest that medical applications of for the enzyme/inhibitor complex. Different relative rates of the aaRS substrate analogs as inhibitors of pathogens could potentially two steps of the reaction may account at least in part for the affect the host mitochondrial enzymes. Since mitochondria are the difference between the Ki values observed with mt-AspRS and cyt- powerhouse of eukaryal cells, side effects can indeed not be ruled AspRS. In regard to further structural investigations, we noticed out. However, the toxicity of adenylate analogs in vivo is difficult to that the strong binding of Asp-AMS, that differentiates the mt- predict precisely since it will depend also on the ability of the drug enzyme from other AspRSs, decreases its propensity to aggregate to cross the mitochondrial membranes and further on the intra and increases its solubility (not shown). Such a property, also found mitochondrial concentration of free amino acid competing with the in the case of human mt-TyrRS , becomes a positive hint drug for the active site of the synthetase. Additional investigations towards successful crystallization assays.
need to be performed to understand the contribution of theseparameters in detail.
The dramatic adaptation of pathogens to antibiotics calls for new target macromolecules and new types of inhibitors. Along M. Frugier is acknowledged for the generous gift of human cyt- evolution, aaRSs acquired subtle differences in their active site, AspRS. This work was supported by a collaborative France-Que´bec making this family of macromolecules attractive targets in such grant (project #61.103). We thank also Centre National de la a strategy . Efficient inhibition by adenylate analogs has Recherche Scientifique (CNRS) including a PICS project, Universite´ already been obtained for bacterial AspRS IleRS Louis Pasteur Strasbourg (ULP), Association Française contre les MetRS , GluRS and GlnRS Our present data confirm Myopathies (AFM) in France and the Fonds Que´be´cois de Recherche and extend a differential sensitivity of AspRSs from various sur la Nature et les Technologies (FQRNT) in Canada for financial organisms to aspartyl-adenylate analogs. Asp-AMS is the most support. M.M. received a fellowship from the French Ministe re de active inhibitor with Ki values in the nanomolar range, with l'Enseignement Supe´rieur et de la Recherche.
M. Messmer et al. / Biochimie 91 (2009) 596–603 K. Arnold, L. Bordoli, J. Kopp, T. Schwede, The SWISS-MODEL workspace:a web-based environment for protein structure homology modelling, Bio-informatics 22 (2006) 195–201.
I.E. Scheffler, Mitochondria, second ed. Willey-Liss, 2007.
R.J. Rosenfeld, D.S. Goodsell, R.A. Musah, G.M. Morris, D.B. Goodin, A.J. Olson, L. Bonnefond, A. Fender, J. Rudinger-Thirion, R. Giege´, C. Florentz, M. Sissler, Automated docking of ligands to an artificial active site: augmenting crys- Towards the full set of human mitochondrial aminoacyl-tRNA synthetases: tallographic analysis with computer modeling, J. Comput. Aided Mol. Des. 17 characterization of AspRS and TyrRS, Biochemistry 44 (2005) 4805–4816.
(2003) 525–536.
M. Sissler, B. Lorber, M. Messmer, A. Schaller, J. Pu¨tz, C. Florentz, Handling P. Berg, E.J. Offengand, An enzymatic mechanism for linking amino acids to mammalian mitochondrial tRNAs and aminoacyl-tRNA synthetases for RNA, Proc. Nat. Acad. Sci. U.S.A. 44 (1958) 78–86.
functional and structural characterization, Methods 44 (2008) 176–189.
J. Lapointe, R. Giege´, Transfer RNAs and aminoacyl-tRNA synthetases, in: K. Shiba, P. Schimmel, H. Motegi, T. Noda, Human glycyl-tRNA synthetase.
H. Trachsel (Ed.), Translation in Eukaryotes, CRC Presss Inc, Boca Raton, FL, Wide divergence of primary structure from bacterial counterpart and 1991, pp. 35–69.
species-specific aminoacylation, J. Biol. Chem. 269 (1994) 30049–30055.
R. Giege´, The early history of tRNA recognition by aminoacyl-tRNA synthe- S.J. Mudge, J.H. Williams, H.J. Eyre, G.R. Sutherland, P.J. Cowan, D.A. Power, tases, J. Biosci. 31 (2006) 477–488.
Complex organisation of the 50-end of the human glycine tRNA synthetase M. Ibba, D. So¨ll, Aminoacyl-tRNA synthesis, Annu. Rev. Biochem. 69 (2000) gene, Gene 209 (1998) 45–50.
E. Tolkunova, H. Park, J. Xia, M.P. King, E. Davidson, The human lysyl-tRNA M. Ibba, C. Francklyn, S. Cusack (Eds.), The Aminoacyl-tRNA Synthetases, synthetase gene encodes both the cytoplasmic and mitochondrial enzymes Landes Bioscience, 2005.
by means of an unusual splicing of the primary transcript, J. Biol. Chem. 275 K.B. Jacobson, Reaction of aminoacyl-tRNA synthetases with heterologous tRNA's, Prog. Nucl. Acid Res. Mol. Biol. 11 (1971) 461–488.
L. Margulis, Origin of Eukaryotic Cells, New Haven Yale University Press, 1970.
A.R. Fersht, J.P. Shi, J. Knill-Jones, D.M. Lowe, A.J. Wilkinson, D.M. Blow, M.W. Gray, G. Burger, B.F. Lang, The origin and early evolution of mito- P. Brick, P. Carter, M.M. Waye, G. Winter, Hydrogen bonding and biological chondria, Genome Biol. 2 (2001) 1018.1–1018.5.
specificity analysed by protein engineering, Nature 314 (1985) 235–238.
B. Brindefalk, J. Viklund, D. Larsson, M. Thollesson, S.G. Andersson, Origin and J. Pu¨tz, J.D. Puglisi, C. Florentz, R. Giege´, Additive, cooperative and anti- evolution of the mitochondrial aminoacyl-tRNA synthetases, Mol. Biol. Evol.
cooperative effects between identity nucleotides of a tRNA, EMBO J. 12 24 (2007) 743–756.
(1993) 2949–2957.
L. Bonnefond, M. Frugier, E. Touze´, B. Lorber, C. Florentz, R. Giege´, C. Sauter, B. Lorber, D. Kern, R. Giege´, J.-P. Ebel, Covalent attachment of aspartic acid to J. Rudinger-Thirion, Crystal structure of human mitochondrial tyrosyl-tRNA yeast aspartyl-tRNA synthetase, FEBS Lett. 146 (1982) 59–64.
synthetase reveals common and idiosyncratic features, Structure 15 (2007) D. Kern, B. Lorber, Y. Boulanger, R. Giege´, A peculiar property of aspartyl- tRNA synthetase from bakers yeast: chemical modification of the protein by A. Fender, C. Sauter, M. Messmer, J. Pu¨tz, R. Giege´, C. Florentz, M. Sissler, Loss the enzymatically synthesized aminoacyl adenylate, Biochemistry 24 (1985) of a primordial identity element for a mammalian mitochondrial amino- acylation system, J. Biol. Chem. 281 (2006) 15980–15986.
D. Bernard, P.M. Akochy, S. Bernier, O. Fisette, O.C. Brousseau, R. Cheˆnevert, L. Bonnefond, M. Frugier, R. Giege´, J. Rudinger-Thirion, Human mitochondrial P.H. Roy, J. Lapointe, Inhibition by TyrRS disobeys the tyrosine identity rules, RNA 11 (2005) 558–562.
L-aspartol adenylate of a nondiscriminating aspartyl-tRNA synthetase reveals differences between the interactions of its Y. Kumazawa, H. Himeno, K.-I. Miura, K. Watanabe, Unilateral aminoacylation active site with tRNA(Asp) and tRNA(Asn), J. Enzyme Inhib. Med. Chem. 22 specificity between bovine mitochondria and eubacteria, J. Biochem. 109 (2007) 77–82.
(1991) 421–427.
S. Cusack, C. Berthet-Colominas, M. Hartlein, N. Nassar, R. Leberman, A J. Tao, P. Schimmel, Inhibitors of aminoacyl-tRNA synthetases as novel anti- second class of synthetase structure revealed by X-ray analysis of Escherichia infectives, Exp. Opin. Invest. Drugs 9 (2000) 1767–1775.
coli seryl-tRNA synthetase at 2.5 Å, Nature 347 (1990) 249–255.
R. Cheˆnevert, S. Bernier, J. Lapointe, Inhibitors of aminoacyl-tRNA synthetases G. Eriani, M. Delarue, O. Poch, J. Gangloff, D. Moras, Partition of tRNA as antibiotics and tools for structural and mechanistic studies, in: J. Lapointe, synthetases into two classes based on mutually exclusive sets of sequence L. Brakier-Gringas (Eds.), Translation Mechanisms, Landes Bioscience, 2003, motifs, Nature 347 (1990) 203–206.
pp. 416–428.
R. Giege´, B. Rees, Aspartyl-tRNA synthetases, in: M. Ibba, C. Francklyn, S. Kim, C. Lee, E.C. Choi, S.Y. Choi, Aminoacyl-tRNA synthetases and their S. Cusack (Eds.), Aminoacyl-tRNA synthetases, Landes Biosciences, 2005, pp.
inhibitors as a novel family of antibiotics, Appl. Microbiol. Biotechnol. 61 (2003) 278–288.
E. Schmitt, L. Moulinier, S. Fujiwara, T. Imanaka, J.-C. Thierry, D. Moras, J. Pohlmann, H. Brotz-Oesterhelt, New aminoacyl-tRNA synthetase inhibitors Crystal structure of aspartyl-tRNA synthetase from Pyrococcus kodakaraensis as antibacterial agents, Curr. Drug Targets Infect. Disord. 4 (2004) 261–272.
KOD: archeon specificity and catalytic mechanism of adenylate formation, M. Torchala, M. Hoffmann, IA, Database of known ligands of aminoacyl-tRNA EMBO J. 17 (1998) 5227–5237.
synthetases, J. Comput. Aided Mol. Des. 21 (2007) 523–525.
L. Ador, A. Camasses, P. Erbs, J. Cavarelli, D. Moras, J. Gangloff, G. Eriani, A. Fuller, Pseudomonic acid: an antibiotic produced by Pseudomonas Active site mapping of yeast aspartyl-tRNA synthetase by in vivo selection fluorescens, Nature 234 (1971) 416–417.
of enzyme mutations lethal for cell growth, J. Mol. Biol. 288 (1999) J. Hughes, G. Mellows, Inhibition of isoleucyl-transfer ribonucleic acid synthe- tase in Escherichia coli by pseudomonic acid, Biochem. J. 176 (1978) 305–318.
R. Giege´, Toward a more complete view of tRNA biology, Nat. Struct. Mol.
R.O. Nicholas, V. Berry, P.A. Hunter, J.A. Kelly, The antifungal activity of Biol. 15 (2008) 1007–1014.
mupirocin, J. Antimicrob. Chemother. 43 (1999) 579–582.
M. Sissler, J. Pu¨tz, F. Fasiolo, C. Florentz, Mitochondrial aminoacyl-tRNA S. Bernier, P.M. Akochy, J. Lapointe, R. Cheˆnevert, Synthesis and aminoacyl- synthetases, in: M. Ibba, C. Francklyn, S. Cusack (Eds.), Aminoacyl-tRNA tRNA synthetase inhibitory activity of aspartyl adenylate analogs, Bioorg.
synthetases, Landes Biosciences, 2005, pp. 271–284.
Med. Chem. 13 (2005) 69–75.
A.J. Pope, M. McVey, K. Fantom, K.J. Moore, Effects of substrate and inhibitor J. Lapointe, S. Levasseur, D. Kern, Glutamyl-tRNA synthetase from Escherichia binding on proteolysis of isoleucyl-tRNA synthetase from Staphylococcus coli, Methods Enzymol. 113 (1985) 42–49.
aureus, J. Biol. Chem. 273 (1998) 31702–31706.
C. Balg, S.P. Blais, S. Bernier, J.L. Huot, M. Couture, J. Lapointe, R. Cheˆnevert, C.F. Crasto, A.K. Forrest, T. Karoli, D.R. March, L. Mensah, P.J. O'Hanlon, Synthesis of beta-ketophosphonate analogs of glutamyl and glutaminyl M.R. Nairn, M.D. Oldham, W. Yue, M.G. Banwell, C.J. Easton, Synthesis and adenylate, and selective inhibition of the corresponding bacterial aminoacyl- activity of analogues of the isoleucyl tRNA synthetase inhibitor SB-203207, tRNA synthetases, Bioorg. Med. Chem. 15 (2007) 295–304.
Bioorg. Med. Chem. 11 (2003) 2687–2694.
M. Ruff, S. Krishnaswamy, M. Boeglin, A. Poterszman, A. Mitschler, J. Lee, S.E. Kim, J.Y. Lee, S.Y. Kim, S.U. Kang, S.H. Seo, M.W. Chun, T. Kang, A. Podjarny, B. Rees, J.-C. Thierry, D. Moras, Class II aminoacyl transfer RNA S.Y. Choi, H.O. Kim, N-Alkoxysulfamide, N-hydroxysulfamide, and sulfamate synthetases: crystal structure of yeast aspartyl-tRNA synthetase complexed analogues of methionyl and isoleucyl adenylates as inhibitors of methionyl- with tRNAAsp, Science 252 (1991) 1682–1689.
tRNA and isoleucyl-tRNA synthetases, Bioorg. Med. Chem. Lett. 13 (2003) S. Eiler, A.C. Dock-Bregeon, L. Moulinier, J.-C. Thierry, D. Moras, Synthesis of aspartyl-tRNAAsp in Escherichia coli – a snapshot of the second step, EMBO J.
G. Archontis, T. Simonson, D. Moras, M. Karplus, Specific amino acid recog- 18 (1999) 6532–6541.
nition by aspartyl-tRNA synthetase studied by free energy simulations, J. Mol.
A. Sali, T.L. Blundell, Comparative protein modelling by satisfaction of spatial Biol. 275 (1998) 823–846.
restraints, J. Mol. Biol. 234 (1993) 779–815.

Source: http://cj.sauter.free.fr/xtal/pdf/messmer-biochimie09.pdf