SufC hydrolyzes ATP and interacts with SufB from Thermotoga maritima.

Genetic experiments in bacteria have shown the suf operon is involved in iron homeostasis and the oxidative stress response. The sufB and sufC genes that always occur together in bacteria are also found in plants, and even the malaria parasite, associated with the plastid organelle. Although the suf operon is believed to encode an iron-dependent ABC-transporter there is no direct evidence. By immunolocalization we show here that SufB and SufC are associated with the membrane of Escherichia coli. We also present kinetic studies with a recombinant version of SufC from Thermotoga maritima that shows it is an ATPase and that it interacts with SufB in vitro.

Chemistry of sulforhodamine--amine conjugates.

Commercially-available sulforhodamine sulfonyl chlorides contain two isomeric monosulfonyl chlorides. Conjugates of these isomers with amines have different properties because the sulfonamide formed from one isomer can undergo ring-closure to a colorless sultam. This chemistry has been examined for a model conjugate with methylamine and for a bioconjugate with 2'(3')-O-[N-(2-aminoethyl)carbamoyl]ATP. The interaction of each isomer of the latter conjugates with myosin subfragment 1 has been characterized. Significant differences between the two isomers are observed in these interactions.

Comparative single-molecule and ensemble myosin enzymology: sulfoindocyanine ATP and ADP derivatives.

Single-molecule and macroscopic reactions of fluorescent nucleotides with myosin have been compared. The single-molecule studies serve as paradigms for enzyme-catalyzed reactions and ligand-receptor interactions analyzed as individual stochastic processes. Fluorescent nucleotides, called Cy3-EDA-ATP and Cy5-EDA-ATP, were derived by coupling the dyes Cy3.29.OH and Cy5.29.OH (compounds XI and XIV, respectively, in, Bioconjug. Chem. 4:105-111)) with 2'(3')-O-[N-(2-aminoethyl)carbamoyl]ATP (EDA-ATP). The ATP(ADP) analogs were separated into their respective 2'- and 3'-O-isomers, the interconversion rate of which was 30[OH(-)] s(-1) (0.016 h(-1) at pH 7.1) at 22 degrees C. Macroscopic studies showed that 2'(3')-O-substituted nucleotides had properties similar to those of ATP and ADP in their interactions with myosin, actomyosin, and muscle fibers, although the ATP analogs did not relax muscle as well as ATP did. Significant differences in the fluorescence intensity of Cy3-nucleotide 2'- and 3'-O-isomers in free solution and when they interacted with myosin were evident. Single-molecule studies using total internal reflection fluorescence microscopy showed that reciprocal mean lifetimes of the nucleotide analogs interacting with myosin filaments were one- to severalfold greater than predicted from macroscopic data. Kinetic and equilibrium data of nucleotide-(acto)myosin interactions derived from single-molecule microscopy now have a biochemical and physiological framework. This is important for single-molecule mechanical studies of motor proteins.

Distances between DNA and ATP binding sites in the TyrR-DNA complex.

The Escherichia coli regulatory protein TyrR controls the expression of eight transcription units that encode proteins involved in the biosynthesis and transport of aromatic amino acids. It binds to DNA as a homodimer with a subunit molecular mass of 57 640 Da, each of which has a single site for the binding of ATP within a central structural domain. This paper reports distances between four sites on the DNA and the ATP binding site as determined by fluorescence resonance energy transfer. The DNA was a 30mer containing a centrally located binding site for TyrR. Replacement of a thymidine residue with an aminouridine residue at positions -9, -7, -3, and 2 of the palindromic oligonucleotide sequence enabled the placement of a single fluorescein group along the major groove of the DNA. The energy transfer acceptor was ATP labeled with a rhodamine group through positions 2' and 3' of the ribose, positions that are known to cause minimal interference with the binding of ATP to protein. The dissociation constant for the binding of rhodamine-ATP to TyrR was 300 nM as determined by steady-state fluorescence anisotropy titrations. The energy transfer efficiencies were determined by measuring the level of quenching of donor fluorescence on binding rhodamine-ATP to the TyrR-DNA complex. The experimental transfer efficiencies were compared to theoretical values calculated for a model of the DNA-TyrR complex in which the position of the ATP binding site was allowed to vary over the surface of the monomer unit. Theory was written to account for the transfer from one donor to two acceptors, one on each monomer unit of the TyrR dimer. The results indicate that the ATP binding site is about 40-45 A from the nearest point on the DNA and distant from the DNA helix-turn-helix binding domain. The effects of ATP binding of (i) increasing the TyrR binding affinity by a factor of 4-5 and (ii) permitting the binding of the tyrosine corepressor must therefore occur because of a significant allosteric change in the conformation of the protein.

Crystal structure and functional analysis of Ras binding to its effector phosphoinositide 3-kinase gamma.

Ras activation of phosphoinositide 3-kinase (PI3K) is important for survival of transformed cells. We find that PI3Kgamma is strongly and directly activated by H-Ras G12V in vivo or by GTPgammaS-loaded H-Ras in vitro. We have determined a crystal structure of a PI3Kgamma/Ras.GMPPNP complex. A critical loop in the Ras binding domain positions Ras so that it uses its switch I and switch II regions to bind PI3Kgamma. Mutagenesis shows that interactions with both regions are essential for binding PI3Kgamma. Ras also forms a direct contact with the PI3Kgamma catalytic domain. These unique Ras/PI3Kgamma interactions are likely to be shared by PI3Kalpha. The complex with Ras shows a change in the PI3K conformation that may represent an allosteric component of Ras activation.

Magnesium fluoride-dependent binding of small G proteins to their GTPase-activating proteins.

GTPase-activating proteins (GAPs) enhance the intrinsic GTPase activity of small G proteins, such as Ras and Rho, by contributing a catalytic arginine to the active site. An intramolecular arginine plays a similar role in heterotrimeric G proteins. Aluminum fluoride activates the GDP form of heterotrimeric G proteins, and enhances binding of the GDP form of small G proteins to their GAPs. The resultant complexes have been interpreted as analogues of the transition state of the hydrolytic reaction. Here, equilibrium binding has been measured using scintillation proximity assays to provide quantitative information on the fluoride-mediated interaction of Ras and Rho proteins with their respective GAPs, neurofibromin (NF1) and RhoGAP. High-affinity fluoride-mediated complex formation between Rho.GDP and RhoGAP occurred in the absence of aluminum; however, under these conditions, magnesium was required. Additionally, the novel observation was made of magnesium-dependent, fluoride-mediated binding of Ras.GDP to NF1 in the absence of aluminum. Aluminum was required for complex formation when the concentration of magnesium was low. Thus, either aluminum fluoride or magnesium fluoride can mediate the high-affinity binding of Rho. GDP or Ras.GDP to GAPs. It has been reported that magnesium fluoride can activate heterotrimeric G proteins. Thus, magnesium-dependent fluoride effects might be a general phenomenon with G proteins. Moreover, these data suggest that some protein.nucleotide complexes previously reported to contain aluminum fluoride may in fact contain magnesium fluoride.

The conserved arginine in rho-GTPase-activating protein is essential for efficient catalysis but not for complex formation with Rho.GDP and aluminum fluoride.

The Rho family of small GTP-binding proteins are downregulated by an intrinsic GTPase, which is enhanced by GTPase-activating proteins (GAPs). RhoGAPs contain a single conserved arginine residue that has been proposed to be involved in catalysis. Here, the role of this arginine has been elucidated by mutagenesis followed by determination of catalytic and equilibrium binding constants using single-turnover kinetics, isothermal titration calorimetry, and scintillation proximity assays. The turnover numbers for wild-type, R282A, and R282K RhoGAPs were 5.4, 0.023, and 0.010 s-1, respectively. Thus, the function of this arginine could not be replaced by lysine or alanine. Nevertheless, the R282A mutation had a minimal effect on the binding affinity of RhoGAP for either Rho. GTP or Rho.GMPPNP, which confirms the importance of the arginine residue for catalysis as opposed to formation of the protein-protein complex. The R282A mutant RhoGAP still increased the hydrolysis rate of Rho.GTP by 160-fold, whereas the wild-type enzyme increased it by 38000-fold. We conclude that this arginine contributes half of the total reduction of activation energy of catalysis. In the presence of aluminum fluoride, the R282A mutant RhoGAP binds almost as well as the wild type to Rho.GDP, demonstrating that the conserved arginine is not required for this interaction. The affinity of wild-type RhoGAP for the triphosphate form of Rho is similar to that for Rho.GDP with aluminum fluoride. These last two observations show that this complex is not associated with the free energy changes expected for the transition state, although the Rho.GDP.AlF4-.RhoGAP complex might well be a close structural approximation.

The importance of two conserved arginine residues for catalysis by the ras GTPase-activating protein, neurofibromin.

Ras proteins are guanine-nucleotide binding proteins that have a low intrinsic GTPase activity that is enhanced 10(5)-fold by the GTPase-activating proteins (GAPs) p120-GAP and neurofibromin. Comparison of the primary sequences of RasGAPs shows two invariant arginine residues (Arg1276 and Arg1391 of neurofibromin). In this study, site-directed mutagenesis was used to change each of these residues in the catalytic domain of neurofibromin (NF1-334) to alanine. The ability of the mutant proteins to bind to Ras.GTP and to stimulate their intrinsic GTPase rate was then determined by kinetic methods under single turnover conditions using a fluorescent analogue of GTP. The separate contributions of each of these residues to catalysis and binding affinity to Ras were measured. Both the R1276A and the R1391A mutant NF1-334 proteins were 1000-fold less active than wild-type NF1-334 in activating the GTPase when measured at saturating concentrations. In contrast, there was only a minor effect of either mutation on NF1-334 affinity for wild-type Ha-Ras. These data are consistent with both arginines being required for efficient catalysis. Neither arginine is absolutely essential, because the mutant NF1-334 proteins increase the intrinsic Ras.GTPase by at least 100-fold. The roles of Arg1276 and Arg1391 in neurofibromin are consistent with proposals based on the recently published x-ray structure of p120-GAP complexed with Ras.

Mechanism of inhibition by arachidonic acid of the catalytic activity of Ras GTPase-activating proteins.

Ras is a guanine nucleotide-binding protein that acts as a molecular switch controlling cell growth. The Ras GTPase-activating proteins (GAPs) p120-GAP and neurofibromin are candidates as Ras effectors. The GTPase-activating activity of both proteins is inhibited by mitogenic lipids, such as arachidonic acid and phosphatidic acid, and differential inhibition of the two GAPs led to the hypothesis that both were effectors in a Ras-controlled mitogenic pathway (Bollag, G., and McCormick, F. (1991) Nature 351, 576-579). We have studied the mechanism of inhibition by arachidonic acid in three ways: first, by measurements of catalytic activity under multiple turnover conditions; second, using p-((6-phenyl)-1,3,5-hexatrienyl)benzoic acid as a fluorescent probe for ligands binding to GAPs; and third, by using a scintillation proximity assay to measure direct binding of Ras to neurofibromin. We found no significant differential inhibition between p120-GAP and neurofibromin by arachidonic acid. The inhibition by arachidonic acid included a major component that is competitive with Ras GTP. These data suggest that insomuch as the mitogenic effects of lipids are mediated via inhibition of GAPs, GAPs are not Ras effector proteins. Additionally, lipids can exert a non-competitive type effect, consistent with a protein denaturing activity, making difficult extrapolations from in vitro data to the situation within cells, and possibly explaining the variability of literature data on inhibition by lipids.

Kinetics of interaction between normal and proline 12 Ras and the GTPase-activating proteins, p120-GAP and neurofibromin. The significance of the intrinsic GTPase rate in determining the transforming ability of ras.

Single turnover and equilibrium binding measurements on the interaction of Gly-12 and Pro-12 Ras.GTP with the catalytic domains of the GTPase-activating proteins, p120-GAP and neurofibromin, have been made utilizing fluorescent 2'(3')O-(N-methylanthraniloyl)-nucleotides. These have enabled the equilibrium dissociation constants (Kd) for their initial binding and the rate constants of the hydrolysis step to be measured. p120-GAP binds to both Ras proteins with a Kd of 17 microM, whereas neurofibromin binds to both Ras proteins with a Kd of 1 microM. Both p120-GAP and neurofibromin increased the rate constant of the GTP hydrolysis step of Pro-12 Ras, but the maximal activation at 30 degrees C was 120-fold and 560-fold, as compared with 70,000- and 52,000-fold, with Gly-12 Ras. The affinity with which p120-GAP and neurofibromin binds to either Gly-12 or Pro-12 Ras protein was decreased dramatically by increasing ionic strength caused by addition of NaCl. The rate constant of the cleavage step of hydrolysis catalyzed by neurofibromin increases with increasing ionic strength, whereas that catalyzed by p120-GAP appears to be unaffected. The high ionic strength within the cell might result in a much lower overall GTPase-activating protein activity than is measured under conditions of low ionic strength in vitro, with p120-GAP being more severely inhibited. The GTP hydrolysis rate of Pro-12 Ras is 2-fold faster than that of normal Ras. The low oncogenicity of Pro-12 ras is explained by a model in which the intrinsic rates of hydrolysis and exchange, as well as GTPase-activating protein- and exchange factor-stimulated rates, are determinants of the biological activity of Ras proteins in fibroblasts.

Solution dynamics of p21ras proteins bound with fluorescent nucleotides: a time-resolved fluorescence study.

The solution dynamics of normal and transforming p21ras proteins in both the GTP- and GDP-bound forms were examined with time-resolved fluorescence spectroscopy. The fluorescent 2'(3')-O-(N-methylanthraniloyl) derivatives (mant derivatives) of GTP, dGTP, and GDP and the aminocoumarin and fluorescein derivatives of GTP and GDP were synthesized and used as reporter groups. The fluorescence lifetimes at 5 degrees C of the mant nucleotide derivatives increased from approximately 4 ns in solution to approximately 9 ns when bound to p21ras. At 30 degrees C, there was a 7.8% difference in lifetime between normal p21ras.mantGTP and p21ras.mantGDP, but no difference between similar complexes of the [Asp-12]p21ras protein. These data are consistent with steady-state fluorescence intensity differences among p21ras.mantGTP, p21ras.mantGDP, and the free nucleotides. Rotational correlation times for the mantGTP- and mantGDP-bound p21 proteins, N-ras, K-ras, and H-ras, were similar at 26 ns (5 degrees C), which is significantly longer than the 15-ns rotational correlation time predicted for a globular 21,000-Da protein. The p21-bound fluorescein and aminocoumarin nucleotide derivatives reported correlation times of 19 and 29 ns, respectively. Global analysis of the three fluorophore.p21 complexes with linked protein rotational correlation functions were best fit with a common rotational correlation time of 28 ns. Gel permeation chromatography of the GDP and mantGDP complexes of normal p21N-ras also showed greater apparent molecular weights than were expected in both cases, demonstrating that the high rotational correlation times obtained from time-resolved fluorescence measurements were not a result of the introduction of the fluorophore.(ABSTRACT TRUNCATED AT 250 WORDS)

Mechanism of GTP hydrolysis by p21N-ras catalyzed by GAP: studies with a fluorescent GTP analogue.

The mechanism of the hydrolysis of GTP by p21N-ras and its activation by the catalytic domain of p120 GTPase activating protein (GAP) have been studied using a combination of chemical and fluorescence measurements with the fluorescent GTP analogue, 2'(3')-O-(N-methylanthraniloyl)GTP (mantGTP). Since the concentration of active p21 is important in these measurements, various assays for both total protein and active p21 were investigated. All assays gave good agreement except the filter binding assay of [3H]-GDP bound to p21, which gave values of 35-40% compared to the other methods. Concentrations of p21 were thus based on the absorbance of the mant-chromophore of the p21-mant-nucleotide complexes. The rate constants of the elementary steps of the p21 intrinsic GTPase activity and the GAP activated activity were similar between GTP and mantGTP. Incubation of a stoichiometric complex of p21.mantGTP results in a biphasic decrease in fluorescence. The second phase occurs with the same rate constant as the cleavage step and is accelerated by GAP. No other steps of the mechanism are affected by GAP. Incubation of a stoichiometric complex of p21.mantGpp[NH]p also results in a biphasic decrease in fluorescence even though cleavage does not occur. This is interpreted that the cleavage step of p21.GTP is preceded by and controlled by an isomerization of the p21.GTP complex. GAP accelerates the rate constant of the second fluorescence phase occurring with p21.mantGpp[NH]p. This result shows that GAP accelerates the proposed isomerization which limits GTP cleavage rather than the cleavage step itself.

The kinetic mechanism of the GAP-activated GTPase of p21 ras.

Guanine nucleotides modified by acetylation of the ribose moiety with the small fluorophore N-methylanthranilic acid(mant) have been shown to bind to p21 ras with similar equilibrium and kinetic rate constants as the parent nucleotides. Hydrolysis of p21.mantGTP to p21.mantGDP results in a 10% decrease in fluorescence intensity occurring at the same rate as the cleavage step. A similar process occurs with the non-hydrolysable analogue mantGMP.PNP, and this has led to the proposal that a conformational change of p21.mantGTP precedes and controls the rate of the cleavage step. The fluorescence change with p21.mantGMP.PNP is accelerated in the presence of the C-terminal catalytic domain of GAP, which is consistent with this mechanism. The same conformational change does not occur with oncogenic mutants of p21 ras, Asp-12 and Val-12, but does occur with the weakly oncogenic Pro-12 mutant. Stopped flow measurements of the interaction of GAP with p21.mantGTP show an exponential decrease in fluorescence, the rate of which does not vary linearly with GAP concentration. These data imply a rapidly reversible formation of the p21.mantGTP complex with GAP followed by the isomerization of this complex. This is at least 10(5)-fold faster than the same process in the absence of GAP.

Fluorescence approaches to the study of the p21ras GTPase mechanism.

The use of ribose-modified guanine nucleotides and tryptophan mutants of p21ras, neither of which have significant effect on the kinetic mechanism of the p21ras GTPase and the GAP-activated p21ras GTPase, will now allow a detailed kinetic study of how GAP and other regulatory proteins interact with p21ras. This will lead to a better understanding of how the relative concentrations of 'active' p21ras. GTP and 'inactive' p21ras. GDP are regulated in the cell.

Kinetics of the interaction of 2'(3')-O-(N-methylanthraniloyl)-ATP with myosin subfragment 1 and actomyosin subfragment 1: characterization of two acto-S1-ADP complexes.

We have used a fluorescent analogue of ATP, mantATP [2'(3')-O-(N-methylanthraniloyl)-adenosine 5'-triphosphate; Hiratsuka T. (1983) Biochim. Biophys. Acta 742, 496-508], and made a detailed kinetic study of the interaction of mantATP and mantADP with S1 and acto-S1. We have shown that these analogues behave like ATP and ADP, respectively. In addition, we have demonstrated that this analogue can distinguish between two acto-S1 complexes, the A-M.N (attached) and A.M.N (rigor-like) states [Geeves, M. A., Good, R. S., & Gutfreund, H. (1984) J. Muscle Res. Cell Motil. 5, 351-361]. Previously, these two states were observed with a pyrene label on Cys 374 of actin. This isomerization can now be monitored at two spatially distinct sites on the ternary complex, indicative of a major conformational change in the ternary complex. Also, we have measured the rate of ADP dissociation from both A-M.N and A.M.N directly and shown these to differ by a factor of 1000. Thus the results presented here support the model of Geeves et al. and are consistent with the A-M.N to A.M.N transition being coupled to the force-generating event of the crossbridge cycle.

Hydrolysis of GTP by p21NRAS, the NRAS protooncogene product, is accompanied by a conformational change in the wild-type protein: use of a single fluorescent probe at the catalytic site.

2'(3')-O-(N-Methyl)anthraniloylguanosine 5'-triphosphate (mantGTP) is a fluorescent analogue of GTP that has similar properties to the physiological substrate in terms of its binding constant and the kinetics of its interactions with p21NRAS, the NRAS protooncogene product. There is a 3-fold increase in fluorescence intensity when mantGTP binds to p21NRAS. The rate constant for the cleavage of mantGTP complexed with the protein is similar to that of GTP, and cleavage is accompanied by a fluorescence intensity change in the wild-type protein complex. A two-phase fluorescence change also occurs when the nonhydrolyzable analogue 2'(3')-O-(N-methyl)anthraniloylguanosine 5'-[beta, gamma-imido]triphosphate (mantp[NH]ppG) binds to wild-type p21NRAS. The second phase occurs at the same rate as the second phase observed after mantGTP binding. Thus this second phase is probably a conformation change of the p21NRAS nucleotiside triphosphate complex and that the change controls the rate of GTP hydrolysis on the protein. With a transforming mutant, [Asp12]-p21NRAS, there is no second phase of the fluorescence change after mantGTP or mantp[NH]ppG binding, even though mantGTP is hydrolyzed. This shows that an equivalent conformational change does not occur and thus the mutant may stay in a "GTP-like" conformation throughout the GTPase cycle. These results are discussed in terms of the proposed role of p21NRAS in signal transduction and the transforming properties of the mutant.

Photoreceptor channel activation by nucleotide derivatives.

Cyclic nucleotide activated sodium currents were recorded from photoreceptor outer segment membrane patches. The concentration of cGMP and structurally similar nucleotide derivatives was varied at the cytoplasmic membrane face; currents were generated at each concentration by the application of a voltage ramp. Nucleotide-activated currents were analyzed as a function of both concentration and membrane potential. For cGMP, the average K0.5 at 0 mV was 24 microM, and the activation was cooperative with an average Hill coefficient of 2.3. Of the nucleotide derivatives examined, only 8-[[(fluorescein-5-yl-carbamoyl)methyl]thio]-cGMP (8-Fl-cGMP) activated the channel at lower concentrations than cGMP with a K0.5 of 0.85 microM. The next most active derivative was 2-amino-6-mercaptopurine riboside 3',5'-monophosphate (6-SH-cGMP) which had a K0.5 of 81 microM. cIMP and cAMP had very high K0.5 values of approximately 1.2 mM and greater than 1.5 mM, respectively. All nucleotides displayed cooperativity in their response and were rapidly reversible. Maximal current for each derivative was compared to the current produced at 200 microM cGMP; only 8-Fl-cGMP produced an identical current. The partial agonists 6-SH-cGMP, cIMP, and cAMP activated currents which were approximately 90%, 80%, and 25% of the cGMP response, respectively. 5'-GMP, 2-aminopurine riboside 3',5'-monophosphate, and 2'-deoxy-cGMP produced no detectable current. The K0.5 values for cGMP activation, examined from -90 to +90 mV, displayed a weak voltage dependence of approximately 400 mV/e-fold; the index of cooperativity was independent of the applied field.(ABSTRACT TRUNCATED AT 250 WORDS)

Kinetic analysis of the hydrolysis of GTP by p21N-ras. The basal GTPase mechanism.

The rate constants have been determined for elementary steps in the basal GTPase mechanism of normal p21N-ras (Gly-12) and an oncogenic mutant (Asp-12): namely GTP binding, hydrolysis, phosphate release, and GDP release. By extrapolation from data at lower temperatures, the GTP association rate constant at 37 degrees C is 1.4 x 10(8) M-1 s-1 for the normal protein and 4.8 x 10(8) M-1 s-1 for the mutant. Other rate constants were measured directly at 37 degrees C, and three processes have similar slow values. GTP dissociation is at 1.0 x 10(-4) s-1 (normal) and 5.0 x 10(-4) s-1 (mutant). The hydrolysis step is at 3.4 x 10(-4) s-1 (normal) and 1.5 x 10(-4) s-1 (mutant). GDP dissociates at 4.2 x 10(-4) s-1 (normal) and 2.0 x 10(-4) s-1 (mutant). GDP association rate constants are similar to those for GTP, 0.5 x 10(8) M-1 s-1 for normal and 0.7 x 10(8) M-1 s-1 for mutant. Both hydrolysis and GDP release therefore contribute to rate limitation of the basal GTPase activity. There are distinct differences (up to 5-fold) between rate constants for the normal and mutant proteins at a number of steps. The values are consistent with the reduced GTPase activity for this mutant and suggest little difference between normal and mutant proteins in the relative steady-state concentrations of GTP and GDP complexes that may represent active and inactive states. The results are discussed in terms of the likely role of p21ras in transmembrane signalling.

Characterization of the GTPase reaction of elongation factor Tu. Determination of the stereochemical course in the presence of antibiotic X5108.

The stereochemical course of the GTPase of elongation factor Tu from Escherichia coli has been determined by making use of the reaction dependent on antibiotic X5108 (an N-methylated derivative of kirromycin). Guanosine 5'-(gamma-thio)triphosphate stereospecifically labeled with 17O and 18O in the gamma-position was hydrolyzed in the presence of elongation factor Tu and X5108. The configuration of the product, inorganic [16O, 17O, 18O]thiophosphate was analyzed by 31P NMR after its stereospecific incorporation into adenosine 5'-(beta-thio)triphosphate. The analysis showed that the hydrolysis proceeds with inversion of configuration at the transferred phosphorus, implying that there is not a phosphoenzyme intermediate. Incubation of [beta gamma-18O, gamma-18O3]GTP with elongation factor Tu leaves the 18O-labeling unaltered, as shown by 31P NMR. No exchange of oxygens with water nor beta gamma-beta positional isotope exchange occurs, implying that not even transient cleavage can occur with the elongation factor alone. Only on interaction with X5108, kirromycin, or ribosomes does the cleavage occur, most likely by a single step, in-line transfer of the terminal phosphorus from GDP to a water oxygen. These properties of the GTP hydrolysis mechanism of elongation factor Tu are similar to those of elongation factor G.