Nuclear overhauser effect studies of the conformations of Mg(alpha, beta-methylene)ATP bound to E. coli isoleucyl-tRNA synthetase

Internuclear distances obtained from transferred nuclear Overhauser effects were used in combination with distance geometry calculations to define the E. coli isoleucyl-tRNA synthetase bound conformation of Mg(alpha, beta-methylene)ATP both in the absence and in the presence of the cognate and noncognate amino acids L-isoleucine and L-valine, respectively. A single nucleotide structure having an anti adenine-ribose glycosidic torsional angle of -114 degrees was found to satisfy the experimental distance constraints. The nearly identical anti glycosidic torsional angles observed in all three complexes demonstrate that the conformation of the adenosine moiety of the enzyme-bound nucleotide is not sensitive to the presence or to the nature of the amino acid bound at the aminoacyladenylate site. In addition, the acceptable range of Mg(alpha, beta-methylene)ATP conformations bound to the E. coli isoleucyl-tRNA synthetase was found to be nearly identical to that previously determined for the E. coli methionyl-tRNA synthetase (Williams and Rosevear (1991) J. Biol. Chem. 266, 2089-2098). Thus, the predicted structural homology between the isoleucyl- and methionyl-tRNA synthetases, both members of the same class of synthetases on the basis of common consensus sequences, is further supported by consensus enzyme-bound nucleotide conformations.     

Nuclear Overhauser effect studies of the conformation of Co(NH3)4ATP bound to kidney Na,K-ATPase

Transferred nuclear Overhauser effect measurements (in the two-dimensional mode) have been used to determine the three-dimensional conformation of an ATP analogue, Co(NH3)4ATP, at the active site of sheep kidney Na,K-ATPase. Previous studies have shown that Co(NH3)4ATP is a competitive inhibitor with respect to MnATP for the Na,K-ATPase [Klevickis, C., & Grisham, C.M. (1982) Biochemistry 21, 6979. Gantzer, M.L., et al. (1982) Biochemistry 21, 4083]. Nine unique proton-proton distances on ATPase-bound Co(NH3)4ATP were determined from the initial build-up rates of the cross-peaks of the 2D-TRNOE data sets. These distances, taken together with previous 31P and 1H relaxation measurements with paramagnetic probes, are consistent with a single nucleotide conformation at the active site. The bound Co(NH3)4ATP) adopts an anti conformation, with a glycosidic torsion angle of 35 degrees, and the conformation of the ribose ring is slightly N-type (C2'-exo, C3'-endo). The delta and gamma torsional angles in this conformation are 100 degrees and 178 degrees, respectively. The nucleotide adopts a bent configuration, in which the triphosphate chain lies nearly parallel to the adenine moiety. Mn2+ bound to a single, high-affinity site on the ATPase lies above and in the plane of the adenine ring. The distances from enzyme-bound Mn2+ to N6 and N7 are too large for first coordination sphere complexes, but are appropriate for second-sphere complexes involving, for example, intervening hydrogen-bonded water molecules. The NMR data also indicate that the structure of the bound ATP analogue is independent of the conformational state of the enzyme.(ABSTRACT TRUNCATED AT 250 WORDS)     

Nuclear Overhauser effect studies of the conformations of MgATP bound to the active and secondary sites of muscle pyruvate kinase

MgATP binds both at the active site (site 1) and at a secondary site (site 2) on each monomer of muscle pyruvate kinase as previously found by binding studies and by X-ray analysis. Interproton distances on MgATP bound at each site have been measured by the time-dependent nuclear Overhauser effect in the absence and presence of phosphoenolpyruvate (P-enolpyruvate), which blocks ATP binding at site 1. Interproton distances at site 2 are consistent with a single conformation of bound ATP with a high antiglycosidic torsional angle (chi = 68 +/- 10 degrees) and a C3'-endo ribose pucker (delta = 90 +/- 10 degrees). Interproton distances at site 1, determined in the absence of P-enolpyruvate by assuming the averaging of distances at both sites, cannot be fit by a single adenine-ribose conformation but require the contribution of at least three low-energy structures: 62 +/- 10% low anti (chi = 30 degrees), C3'-endo; 20 +/- 8% high anti (chi = 55 degrees), O1'-endo; and 18 +/- 8% syn (chi = 217 degrees), C2'-endo. Although a different set of ATP conformations might also have fit the interproton distances, the mixture of conformations used also fits previously determined distances from Mn2+ to the protons of ATP bound at site 1 [Sloan, D. L., & Mildvan, A. S. (1976) J. Biol. Chem. 251, 2412] and is similar to the adenine-ribose portion of free Co(NH3)4ATP, which consists of 35% low anti, 51% high anti, and 14% syn [Rosevear, P. R., Bramson, H. N., O'Brian, C., Kaiser, E. T., & Mildvan, A. S. (1983) Biochemistry 22, 3439].(ABSTRACT TRUNCATED AT 250 WORDS)     

Adenosine conformations of nucleotides bound to methionyl tRNA synthetase by transferred nuclear Overhauser effect spectroscopy.

The conformations of MgATP and AMP bound to a monomeric tryptic fragment of methionyl tRNA synthetase have been investigated by two-dimensional proton transferred nuclear Overhauser effect spectroscopy (TRNOESY). The sample protocol was chosen to minimize contributions from adventitious binding of the nucleotides to the observed NOE. The experiments were performed at 500 MHz on three different complexes, E.MgATP, E.MgATP.L-methioninol, and E.AMP.L-methioninol. A starter set of distances obtained by fitting NOE build-up curves (not involving H5' and H5") were used to determine a CHARMm energy-minimized structure. The positioning of the H5' and H5" protons was determined on the basis of a conformational search of the torsion angle to obtain the best fit with the observed NOEs for their superposed resonance. Using this structure, a relaxation matrix was set up to calculate theoretical build-up curves for all of the NOEs and compare them with the observed curves. The final structures deduced for the adenosine moieties in the three complexes are very similar, and are described by a glycosidic torsion angle (chi) of 56 degrees +/- 5 degrees and a phase angle of pseudorotation (P) in the range of 47 degrees to 52 degrees, describing a 3(4)T-4E sugar pucker. The glycosidic torsion angle, chi, deduced here for this adenylyl transfer enzyme and those determined previously for three phosphoryl transfer enzymes (creatine kinase, arginine kinase, and pyruvate kinase), and one pyrophosphoryl enzyme (PRibPP synthetase), are all in the range 52 degrees +/- 8 degrees. The narrow range of values suggests a possible common motif for the recognition and binding of the adenosine moiety at the active sites of ATP-utilizing enzymes, irrespective of the point of cleavage on the phosphate chain.     

Cobalt(III) labeling of methionyl-tRNA synthetase from Escherichia coli.

Native and trypsin-modified methionyl-tRNA synthetases from Escherichia coli were found to be inactivated by incubation in the presence of Co(III) complexes of ATP, stabilized either by imidazole or phenanthroline, or by oxidation in situ to Co(III) of the substrate ATP-Co(II). It has been shown that the inactivation proceeds by specific labeling of the catalytic ATP-Mg(II) site of the synthetases. The enzymes are completely inactivated by the incorporation of one cobalt atom and one ATP molecule per active site. The inactivated enzymes may be stored for a long period without significant reactivation or removal of the cobalt label. In the presence of dithiothreitol or 2-mercaptoethanol, the labeled enzymes recover full activity with concomittant release of the bound label molecules.     

Binding of the anticodon domain of tRNA(fMet) to Escherichia coli methionyl-tRNA synthetase

A stem and loop RNA domain carrying the methionine anticodon (CAU) was designed from the tRNA(fMet) sequence and produced in vitro. This domain makes a complex with methionyl-tRNA synthetase (Kd = 38(+/- 5) microM; 25 degrees C, pH 7.6, 7 mM-MgCl2). The formation of this complex is dependent on the presence of the cognate CAU anticodon sequence. Recognition of this RNA domain is abolished by a methionyl-tRNA synthetase mutation known to alter the binding of tRNA(Met).     

NMR analyses of the conformations of L-isoleucine and L-valine bound to Escherichia coli isoleucyl-tRNA synthetase

The 400-MHz 1H NMR spectra of L-isoleucine and L-valine were measured in the presence of Escherichia coli isoleucyl-tRNA synthetase (IleRS). Because of chemical exchange of L-isoleucine or L-valine between the free state and the IleRS-bound state, a transferred nuclear Overhauser effect (TRNOE) was observed among proton resonances of L-isoleucine or L-valine. However, in the presence of isoleucyl adenylate tightly bound to the amino acid activation site of IleRS, no TRNOE for L-isoleucine or L-valine was observed. This indicates that the observed TRNOE is due to the interaction of L-isoleucine or L-valine with the amino acid activation site of IleRS. The conformations of these amino acids in the amino acid activation site of IleRS were determined by the analyses of time dependences of TRNOEs and TRNOE action spectra. The IleRS-bound L-isoleucine takes the gauche+ form about the C alpha-C beta bond and the trans form about the C beta-C gamma 1 bond. The IleRS-bound L-valine takes the gauche- form about the C alpha-C beta bond. Thus, the conformation of IleRS-bound L-valine is the same as that of IleRS-bound L-isoleucine except for the delta-methyl group. The side chain of L-isoleucine or L-valine lies in an aliphatic hydrophobic pocket of the active site of IleRS. Such hydrophobic interaction with IleRS is more significant for L-isoleucine than for L-valine. The TRNOE analysis is useful for studying the amino acid discrimination mechanism of aminoacyl-tRNA synthetases.     

Crystallographic study at 2.5 A resolution of the interaction of methionyl-tRNA synthetase from Escherichia coli with ATP

The crystal structure of the tryptic fragment of the methionyl-tRNA synthetase from Escherichia coli, complexed with ATP, has been refined to a crystallographic R-factor of 0.220, at 2.5 A resolution (for 4433 protein atoms). In the last stages of the refinement, the simulated annealing refinement method was fully applied, contributing to a drastic improvement of the model and the identification of the missing atoms. In the final model, the root-mean-square deviation from ideality for bond distances is 0.021 A and for angle distances is 0.054 A. The position of the zinc ion has been confirmed and is located near the active site. The tryptic fragment is composed of two globular domains. The first domain, from the N terminus to Thr360, contains a nucleotide-binding fold into which two long polypeptides of 101 and 70 residues are inserted. The nucleotide-binding fold is strengthened by the presence of the zinc ion in the vicinity of the active site. The second domain, up to Pro526, is mainly alpha-helical. The C-terminal polypeptide, Phe527 to Lys551, folds back towards the first domain, making a link between the two domains. The heptapeptide 528-534 partly shapes a deep cavity that plunges into the central core of the nucleotide-binding fold, where the ATP molecule is located. The adenine ring, deeply buried in the bottom of the cleft, is blocked between the first helix HA, and the strands A and D of the beta-sheet and makes no polar interaction with the enzyme. The 2' and 3' hydroxyl groups of the ribose, whose conformation is C2' endo, interact with the main-chain carbonyl oxygen atoms of Ile231 and Glu241, respectively. The side-chain nitrogen atom of Lys142 is at hydrogen-bonding distance from the ring oxygen O-4' of the ribose. One of the alpha-phosphate oxygen atoms and one of the gamma-phosphate oxygen atoms interact with the imidazole ring of His21, which is well conserved in many of the known synthetases; this indicates a possible crucial role for this residue in binding ATP. The beta-phosphate group is linked to the main-chain carbonyl oxygen atom of Tyr15 through an intermediate water molecule. The gamma-phosphate group interacts with the carbonyl oxygen atom and the side-chain of Asn17.(ABSTRACT TRUNCATED AT 400 WORDS)     

Nuclear Overhauser effect studies of the conformations and binding site environments of deoxynucleoside triphosphate substrates bound to DNA polymerase I and its large fragment

The conformations and binding site environments of Mg2+TTP and Mg2+dATP bound to Escherichia coli DNA polymerase I and its large (Klenow) fragment have been investigated by proton NMR. The effect of the large fragment of Pol I on the NMR line widths of the protons of Mg2+TTP detected one binding site for this substrate with a dissociation constant of 300 +/- 100 microM and established simple competitive binding of deoxynucleoside triphosphates at this site in accord with previous equilibrium dialysis experiments with whole Pol I [Englund, P. T., Huberman, J.A., Jovin, T.M., & Kornberg, A. (1969) J. Biol. Chem. 244, 3038]. Primary negative nuclear Overhauser effects were used to calculate interproton distances on enzyme-bound Mg2+dATP and Mg2+TTP. These distances established that each substrate was bound with an anti-glycosidic torsional angle (chi) of 50 +/- 10 degrees for Mg2+dATP and 40 +/- 10 degrees for Mg2+TTP. The sugar pucker of both substrates was predominantly O1'-endo, with a C5'-C4'-C3'-O3' exocyclic torsional angle (delta) of 95 +/- 10 degrees for Mg2+dATP and 100 +/- 10 degrees for Mg2+TTP. The consistency of these conformations with those previously proposed, on the basis of distances from Mn2+ at the active site [Sloan, D. L., Loeb, L. A., Mildvan, A.S., & Feldman, R.J. (1975) J. Biol. Chem. 250, 8913], indicates a unique conformation for each bound nucleotide. The chi and delta values of the bound substrates are appropriate for nucleotide units of B DNA.(ABSTRACT TRUNCATED AT 250 WORDS)     

Structural characterization of adenine nucleotides bound to Escherichia coli adenylate kinase. 1. Adenosine conformations by proton two-dimensional transferred nuclear Overhauser effect spectroscopy

Adenosine conformations of adenosine 5'-triphosphate (ATP) and adenosine 5'-monophosphate (AMP), and of an ATP analogue, adenylyl imidodiphosphate (AMPPNP), bound to Escherichia coliadenylate kinase (AKe) in the complexes of AKe.Mg(II)ATP, AKe.AMP.Mg(II)GDP, AKe. AMPPNP, and AKe.Mg(II)AMPPNP were determined by transferred two-dimensional nuclear Overhauser effect spectroscopy (TRNOESY) measurements and molecular dynamics simulations. The glycosidic torsion angles, chi, deduced for the adenine nucleotides in these complexes are 51 degrees, 37 degrees, 49 degrees, and 47 degrees, respectively, with an experimental error of about +/-5 degrees. These values are in general agreement with those previously measured for other ATP-utilizing enzymes, suggesting a possible common motif for adenosine recognition and binding. The pseudorotational phase angle, P, of the sugar puckers for the bound nucleotides varied between 50 degrees and 103 degrees. These solution-state conformations are significantly different from those in published data from X-ray crystallography. A computation of the ligand NOEs, made by using the program CORCEMA [Moseley, H. N. B., Curto, E. V., and Krishna, N. R. (1995) J. Magn. Reson. B108, 243-261] with the protein protons in the vicinity of nucleotide included, on the basis of the X-ray structure of the AKe.AMP.AMPPNP complex [Berry, M. B., Meador, B., Bilderback, T., Liang, P., Glaser, M., and Philips, G. N. , Jr. (1994) Proteins: Struct., Funct., Genet. 19, 183-198], showed that polarization transfer to the protein protons does not produce significant errors in the structures determined by considering the ligand NOEs alone.     

Identification of the tRNA anticodon recognition site of Escherichia coli methionyl-tRNA synthetase

We have previously shown that the anticodon of methionine tRNAs contains most, if not all, of the nucleotides required for specific recognition of tRNA substrates by Escherichia coli methionyl-tRNA synthetase [Schulman, L. H., & Pelka, H. (1988) Science 242, 765-768]. Previous cross-linking experiments have also identified a site in the synthetase that lies within 14 A of the anticodon binding domain [Leon, O., & Schulman, L. H. (1987) Biochemistry 26, 5416-5422]. In the present work, we have carried out site-directed mutagenesis of this domain, creating conservative amino acid changes at residues that contain side chains having potential hydrogen-bond donors or acceptors. Only one of these changes, converting Trp461----Phe, had a significant effect on aminoacylation. The mutant enzyme showed an approximately 60-100-fold increase in Km for methionine tRNAs, with little or no change in the Km for methionine or ATP or in the maximal velocity of the aminoacylation reaction. Conversion of the adjacent Pro460 to Leu resulted in a smaller increase in Km for tRNA(Mets), with no change in the other kinetic parameters. Examination of the interaction of the mutant enzymes with a series of tRNA(Met) derivatives containing base substitutions in the anticodon revealed sequence-specific interactions between the Phe461 mutant and different anticodons. Km values were highest for tRNA(mMet) derivatives containing the normal anticodon wobble base C. Base substitutions at this site decreased the Km for aminoacylation by the Phe461 mutant, while increasing the Km for the wild-type enzyme and for the Leu460 mutant to values greater than 100 microM.(ABSTRACT TRUNCATED AT 250 WORDS)     

Conformational change of Escherichia coli initiator methionyl-tRNA(fMet) upon binding to methionyl-tRNA formyl transferase

The specific formylation of initiator methionyl-tRNA (Met-tRNA) by methionyl-tRNA formyltransferase (MTF) is important for the initiation of protein synthesis in Escherichia coli. The determinants for formylation are located in the acceptor stem and in the dihydrouridine (D) stem of the initiator tRNA (tRNA^fMet). Here, we have used ethylation interference analysis to study the interactions between the Met-tRNA^fMet and MTF in solution. We have identified three clusters of phosphates in the tRNA that, when ethylated, interfere with binding of MTF. Interference due to ethylation of phosphates in the acceptor stem and in the D stem is most likely due to the close proximity of the protein as seen in the crystal structure of the MTF.fMet-tRNA^fMet complex. The third cluster of phosphates, whose ethylation interferes with binding of MTF, is dispersed along the anticodon stem, which is distal to the sites of tRNA protein contacts. Interestingly, these latter positions correspond to sites of increased cleavages by RNase V1 in RNA footprinting experiments. Together, these results suggest that in addition to the protein, which binds to the substrate tRNA in an induced fit mechanism, the tRNA also undergoes induced structural changes during its binding to MTF.     

Antico-operative binding of bacterial and mammalian initiator tRNAMet to methionyl-tRNA synthetase from Escherichia coli

The reaction scheme of methionyl-tRNA synthetase from Escherichia coli with the initiator tRNA_s^Met from E. coli and rabbit liver, respectively, has been resolved. The statistical rate constants for the formation, k_R, and for the dissociation, k_D, of the 1:1 complex of these tRNAs with the dimeric enzyme have been calculated. Identical k_R values of 250 μm^?1 s^?1 reflect similar behaviour for antico-operative binding of both tRNA_s^Met to native methionyl-tRNA synthetase. Advantage was taken of the difference in extent of tryptophan fluorescence-quenching induced by the bacterial and mammalian initiator tRNA_s^Met to measure the mode of exchange of these tRNAs antico-operatively bound to the enzyme. Analysis of the results reveals that antico-operativity does not arise from structural asymmetric assembly of the enzyme subunits. Indeed, both subunits can potentially bind a tRNA molecule. Exchange between tRNA molecules can occur via a transient complex in which both sites are occupied. Either strong and weak sites reciprocate between subunits on the transient complex or occupation of the weak site induces symmetry of this complex. While in the present case, these two alternatives are kinetically indistinguishable, they do account for the observation that, upon increasing the concentration of the competing mammalian tRNA, the rate of exchange of the E. coli initiator tRNA^Met is enhanced, due to its faster rate of dissociation from the transient complex. Finally, it has been verified that in the case of the trypsin-modified methionyl-tRNA synthetase which cannot provide more than one binding site for tRNA, exchange of enzymebound bacterial tRNA by mammalian tRNA does proceed to a limiting rate independent of the mammalian tRNA concentration present in the solution.     

Molecular cloning and primary structure of the Escherichia coli methionyl-tRNA synthetase gene.

The intact metG gene was cloned in plasmid pBR322 from an F32 episomal gene library by complementation of a structural mutant, metG83. The Escherichia coli strain transformed with this plasmid (pX1) overproduced methionyl-tRNA synthetase 40-fold. Maxicell analysis showed that three major polypeptides with MrS of 76,000, 37,000, and 29,000 were expressed from pX1. The polypeptide with an Mr of 76,000 was identified as the product of metG on the basis of immunological studies and was indistinguishable from purified methionyl-tRNA synthetase. In addition, DNA-DNA hybridization studies demonstrated that the metG regions were homologous on the E. coli chromosome and on the F32 episome. DNA sequencing of 642 nucleotides was performed. It completes the partial metG sequence already published (D. G. Barker, J. P. Ebel, R. Jakes, and C. J. Bruton, Eur. J. Biochem. 127:449-451, 1982). Examination of the deduced primary structure of methionyl-tRNA synthetase excludes the occurrence of any significant repeated sequences. Finally, mapping of mutation metG83 by complementation experiments strongly suggests that the central part of methionyl-tRNA synthetase is involved in methionine recognition. This observation is discussed in the light of the known three-dimensional crystallographic structure.     

Identification of peptide sequences at the tRNA binding site of Escherichia coli methionyl-tRNA synthetase

Four different structural regions of Escherichia coli tRNAfMet have been covalently coupled to E. coli methionyl-tRNA synthetase (MetRS) by using a tRNA derivative carrying a lysine-reactive cross-linker. We have previously shown that this cross-linking occurs at the tRNA binding site of the enzyme and involves reaction of only a small number of the potentially available lysine residues in the protein [Schulman, L. H., Valenzuela, D., & Pelka, H. (1981) Biochemistry 20, 6018-6023; Valenzuela, D., Leon, O., & Schulman, L. H. (1984) Biochem. Biophys. Res. Commun. 119, 677-684]. In this work, four of the cross-linked peptides have been identified. The tRNA-protein cross-linked complex was digested with trypsin, and the peptides attached to the tRNA were separated from the bulk of the tryptic peptides by anion-exchange chromatography. The tRNA-bound peptides were released by cleavage of the disulfide bond of the cross-linker and separated by reverse-phase high-pressure liquid chromatography, yielding five major peaks. Amino acid analysis indicated that four of these peaks contained single peptides. Sequence analysis showed that the peptides were cross-linked to tRNAfMet through lysine residues 402, 439, 465, and 640 in the primary sequence of MetRS. Binding of the tRNA therefore involves interactions with the carboxyl-terminal half of MetRS, while X-ray crystallographic data have shown the ATP binding site to be located in the N-terminal domain of the protein [Zelwer, C., Risler, J. L., & Brunie, S. (1982) J. Mol. Biol. 155, 63-81].(ABSTRACT TRUNCATED AT 250 WORDS)     

Identification of residues involved in the binding of methionine by Escherichia coli methionyl-tRNA synthetase

Comparison of the amino-acid sequences of several methionyl-tRNA synthetases indicates the occurrence of a few conserved motifs, having a possible functional significance. The role of one of these motifs, centered at position 300 in the E. coli enzyme sequence, was assayed by the use of site-directed mutagenesis. Substitution of the His301 or Trp305 residues by Ala resulted in a large decrease in methionine affinity, whereas the change of Val298 into Ala had only a moderate effect. The catalytic rate of the enzyme was unimpaired by these substitutions. It is concluded that the above conserved amino-acid region is located at or close to the amino-acid binding pocket of methionyl-tRNA synthetase.     

Affinity chromatography on agarose-hexyl-adenosine-5'-phosphate of methionyl-tRNA synthetase from Escherichia coli. Application of the couplings between the methionine and ATP sites.

Recent studies by us [Biochemistry (1977) 16, 2570-2579] have shown that L-methioninol, a methionine analog lacking the carboxylate negative charge, enhances the affinity of AMP for methionyl-tRNA synthetase while L-methionine antagonizes the nucleotide binding. Such couplings between ligands of the enzyme have now been applied to affinity chromatography of methionyl-tRNA synthetase on an agarose-hexyl-adenosine-5'-phosphate gel (the spacer is attached to AMP at the adenine C-8 position). Retention of the enzyme on this gel column was shown to be dependent on the presence of appropriate concentrations of magnesium and of L-methioninol in the equilibration buffer. The enzyme was then specifically recovered from the column by omitting the amino alcohol or by adding an excess of L-methionine which antagonizes the cooperative effect of L-methioninol. This approach has provided the basis for a new purification procedure of methionyl-tRNA synthetase which leads to a 200-fold purification in a single chromatographic step. In this manner, after 30-50% ammonium sulfate fractionation of extracts of Escherichia coli EM 20031 (carrying the F32 episome), 0.25 mg X methionyl-tRNA synthetase was obtained at 90% purity per ml of agarose-hexyl-adenosine-5'-phosphate gel.     

Interconversion of ATP binding and conformational free energies by tryptophanyl-tRNA synthetase: structures of ATP bound to open and closed,pre-transition-state conformations

Binding ATP to tryptophanyl-tRNA synthetase (TrpRS) in a catalytically competent configuration for amino acid activation destabilizes the enzyme structure prior to forming the transition state. This conclusion follows from monitoring the titration of TrpRS with ATP by small angle solution X-ray scattering, enzyme activity, and crystal structures. ATP induces a significantly smaller radius of gyration at pH=7 with a transition midpoint at approximately 8mM. A non-reciprocal dependence of Trp and ATP dissociation constants on concentrations of the second substrate show that Trp binding enhances affinity for ATP, while the affinity for Trp falls with the square of the [ATP] over the same concentration range ( approximately 5mM) that induces the more compact conformation. Two distinct TrpRS:ATP structures have been solved, a high-affinity complex grown with 1mM ATP and a low-affinity complex grown at 10mM ATP. The former is isomorphous with unliganded TrpRS and the Trp complex from monoclinic crystals. Reacting groups of the two individually-bound substrates are separated by 6.7A. Although it lacks tryptophan, the low-affinity complex has a closed conformation similar to that observed in the presence of both ATP and Trp analogs such as indolmycin, and resembles a complex previously postulated to form in the closely-related TyrRS upon induced-fit active-site assembly, just prior to catalysis. Titration of TrpRS with ATP therefore successively produces structurally distinct high- and low-affinity ATP-bound states. The higher quality X-ray data for the closed ATP complex (2.2A) provide new structural details likely related to catalysis, including an extension of the KMSKS loop that engages the second lysine and serine residues, K195 and S196, with the alpha and gamma-phosphates; interactions of the K111 side-chain with the gamma-phosphate; and a water molecule bridging the consensus sequence residue T15 to the beta-phosphate. Induced-fit therefore strengthens active-site interactions with ATP, substantially intensifying the interaction of the KMSKS loop with the leaving PP(i) group. Formation of this conformation in the absence of a Trp analog implies that ATP is a key allosteric effector for TrpRS. The paradoxical requirement for high [ATP] implies that Gibbs binding free energy is stored in an unfavorable protein conformation and can then be recovered for useful purposes, including catalysis in the case of TrpRS.