Mg2+ and a Key Lysine Modulate Exchange Activity of Eukaryotic Translation Elongation Factor 1Bα*

To sustain efficient translation, eukaryotic elongation factor Bα (eEF1Bα) functions as the guanine nucleotide exchange factor for eEF1A. Stopped-flow kinetics using 2′-(or 3′)-O-N-methylanthraniloyl (mant)-GDP showed spontaneous release of nucleotide from eEF1A is extremely slow and accelerated 700-fold by eEF1Bα. The eEF1Bα-stimulated reaction was inhibited by Mg2+ with a K½ of 3.8 mm. Previous structural studies predicted the Lys-205 residue of eEF1Bα plays an important role in promoting nucleotide exchange by disrupting the Mg2+ binding site. Co-crystal structures of the lethal K205A mutant in the catalytic C terminus of eEF1Bα with eEF1A and eEF1A·GDP established that the lethality was not due to a structural defect. Instead, the K205A mutant drastically reduced the nucleotide exchange activity even at very low concentrations of Mg2+.A K205R eEF1Bα mutant on the other hand was functional in vivo and showed nearly wild-type nucleotide dissociation rates but almost no sensitivity to Mg2+. These results indicate the significant role of Mg2+ in the nucleotide exchange reaction by eEF1Bα and establish the catalytic function of Lys-205 in displacing Mg2+ from its binding site.

Successive rounds of translation elongation are required for the stepwise addition of amino acids onto the growing polypeptide chain (reviewed in Ref. 1). Several elongation factors are required to facilitate efficient protein synthesis. The eukaryotic translation elongation factor 1 (eEF1) 4 complex consists of three or four subunits: eEF1A, eEF1B␣, eEF1B␥, and eEF1B␤. The eEF1A subunit, which is homologous to prokaryotic EF-Tu, facilitates the transport of aminoacyl-tRNA (aa-tRNA) to the A-site of the elongating ribosome. Following formation of the proper codon-anticodon pair and the peptidyl transfer reaction, eEF2 catalyzes translocation. This movement of the peptidyl tRNA to the P-site, the deacylated tRNA to the E-site, and the mRNA by three nucleotides allows for another elongation cycle. The fungal-specific eEF3 protein promotes release of the deacylated tRNA from the E-site (2,3). All three of these elongation factors are essential nucleotide-binding proteins.
After the delivery of aa-tRNA to the ribosome, GTP bound to eEF1A is hydrolyzed to GDP. The guanine nucleotide exchange factor (GEF) complex eEF1B recycles eEF1A back to the active GTP form by stimulating GDP release, allowing rebinding of GTP and subsequently aa-tRNA for further rounds of peptide chain elongation (4). The yeast Saccharomyces cerevisiae eEF1B complex contains the eEF1B␣ and eEF1B␥ subunits, whereas mammalian systems have an additional eEF1B␤ subunit (reviewed in Ref. 1). The eEF1B␣ subunit is catalytic (5), whereas eEF1B␥ binds to and stimulates the nucleotide exchange activity of eEF1B␣ in vitro (6). Co-crystal structure of the eEF1A⅐C terminus of eEF1B␣ revealed that eEF1A has three defined domains, and eEF1B␣ spans domains I and II of eEF1A to create two interfaces. Due to the proposed overlap of the aa-tRNA recognition site with that of eEF1B␣ in domain II of eEF1A, the eEF1A⅐GTP⅐eEF1B␣ intermediate complex is also proposed to interact with aa-tRNA before the dissociation of eEF1B␣ (7). eEF1A is also an actin-binding and -bundling protein (8). Actin and aa-tRNA binding, however, are mutually exclusive (9). Thus, eEF1B␣ may modulate access of these two substrates to eEF1A. The third eEF1B␤ subunit of the mammalian system has sequence similarity to eEF1B␣ and can function as a GEF. eEF1B␤ may play a role in the organization of the eEF1B complex with aa-tRNA synthetases, particularly valyl tRNA synthetase (10). eEF1B␣ is essential in vivo (11). However, its requirement for viability can be bypassed by an extra copy of the TEF1 or TEF2 genes encoding eEF1A (12), which is referred to as an eEF1B␣-deficient strain. Two eEF1B␥ isoforms encoded by the TEF3 and TEF4 genes are expressed in the cell, but both genes can be deleted with no adverse effect on cell growth (13). The loss of either eEF1B subunit promotes a high level of resistance to oxidative stress, indicating a potential role in a post-transcriptional control mechanism under stress conditions (14).
Recently, a structural model of the mechanism of eEF1B␣-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The atomic coordinates and structure factors ( catalyzed guanine nucleotide exchange was proposed (7,15). The structure predicts that Lys-205 of eEF1B␣ is critical for the nucleotide exchange mechanism by displacing the Mg 2ϩ ion associated with the bound GDP, consequently releasing GDP and stabilizing the nucleotide free state of eEF1A. The essential function of this residue is also supported by the lethality of the K205A mutation (15). Current models suggest that the action of eEF1B␣ on eEF1A⅐GTP regeneration is the rate-limiting step in translation elongation (4). This is important because the spontaneous rate of GDP dissociation of eEF1A is very slow (4), and the protein has no preference for the type of nucleotide as the equilibrium dissociation constants for GDP and GTP are 10 ϫ 10 Ϫ7 and 7 ϫ 10 Ϫ7 M, respectively (16). Mg 2ϩ is important for high affinity binding of nucleotides to proteins as well as nucleotide hydrolysis. Therefore, it was suggested that the function of a GEF protein in vivo is to overcome the possible inhibition of nucleotide exchange caused by the physiological levels of Mg 2ϩ (reviewed in Ref. 17). The mechanism of nucleotide dissociation by the eukaryotic elongation GEF has not yet been determined. Prior kinetic analysis demonstrated that EF-Ts, the GEF for EF-Tu, accelerates the dissociation of GDP by 6 ϫ 10 4 (18). It was also shown that the removal of Mg 2ϩ from the EF-Tu⅐GDP complex accelerates nucleotide dissociation by a factor of 150 -300 in the absence of EF-Ts (18).
In this study, the interaction between GDP and wild-type and mutants of Lys-205 of the eukaryotic GEF eEF1B␣ was determined at varying Mg 2ϩ concentrations. The x-ray crystal structure of a complex between the C-terminal fragment of the K205A eEF1B␣ mutant and eEF1A, with and without GDP, revealed that there are no large structural changes as compared with the wild-type structure. Using a fluorescent guanine nucleotide analog (2Ј-(or 3Ј)-O-N-methylanthraniloyl-GDP (mant-GDP)) and stopped-flow kinetics methods, we show that the guanine nucleotide exchange of eEF1A is accelerated 700fold by wild-type eEF1B␣. The GDP dissociation rate of the eEF1A⅐eEF1B␣ complex is highly Mg 2ϩ -dependent, which unlike EF-Tu is not a property of spontaneous GDP dissociation of eEF1A in the absence of its GEF. On the other hand, the K205A eEF1B␣ mutant reduces the rate of GDP dissociation, and increasing the concentration of Mg 2ϩ does not influence the GDP dissociation rates due to the inability of K205A to disrupt the Mg 2ϩ binding pocket. Interestingly, K205A eEF1B␣ also dissociates slowly from eEF1A compared with the wildtype protein. In vivo analysis of the lethal K205A mutation was performed using an eEF1B␣-deficient strain supported by excess eEF1A (12). The K205A mutant suppressed the growth and translational defects of a strain modified to survive without eEF1B␣, indicating that it is not a complete loss of function mutant. Substituting Lys-205 by arginine, however, results in a protein sufficient for viability in vivo with a slight growth defect compared with the wild-type strain. The K205R protein, however, displays a slightly reduced GDP dissociation rate constant that is independent of the Mg 2ϩ concentration, likely due to the occlusion of the Mg 2ϩ binding site. Taken together, these results demonstrate the kinetic effects of a eukaryotic guanine nucleotide exchange factor and the key role of Mg 2ϩ .

EXPERIMENTAL PROCEDURES
Yeast Techniques and Mutant Preparation-S. cerevisiae strains used are listed in Table 1. Standard yeast genetic methods were employed (19). Yeast cells were grown in either YEPD (1% Bacto yeast extract, 2% peptone, 2% dextrose) or defined synthetic complete media (C or CϪ) supplemented with 2% dextrose as a carbon source. Yeast were transformed by the lithium acetate method (20). The K205A and K205R eEF1B␣ mutants were prepared in pTKB500 (TEF5 LEU2) by PCR mutagenesis using the QuikChange method (Stratagene) with oligonucleotides K205A 5Ј-CCGATATTGCTGCTATGCAA-GCTTTATAAAAGGC-3Ј or K205R 5Ј-CCGATATTGCTGT-ATGCAAAGATTATAAAAAGCTTTTTTATAAAC-3Ј. The resulting plasmids pTKB526 (tef 5 K205A) and pTKB590 (tef 5 K205R) as well as the wild-type eEF1B␣ plasmid and an empty vector pRS315 (LEU2) were transformed separately into the eEF1B␣-deficient strain, TKY298. The plasmids in each strain were maintained by growth on C-Ura-Leu media. pTKB590 was also transformed into the strain TKY406 (tef5⌬TRP1 pTEF5 URA3), and the ability of the mutant to function in place of wild-type eEF1B␣ was monitored by growth on 5-fluoroorotic acid, resulting in the production of strain TKY925 expressing only K205R eEF1B␣.
Growth and Drug Assays-Growth of strains at the permissive and non-permissive temperatures was assayed by spotting equal amounts of 1/10 serial dilutions of cells on YEPD or C-Ura-Leu media at 13, 24, 30, and 37°C for 2-7 days. The doubling times were determined by measuring the growth of cell cultures in liquid YEPD or C-Ura-Leu media. Strains grown for 1 day at 30°C were diluted to an A 600 of ϳ0.07 in the appropriate media and incubated at 30°C with vigorous shaking. The optical density (A 600 ) was assayed approximately every hour, and doubling times were determined by continually maintaining the cells in the log phase by dilution when the A 600 reached mid-log phase (0.4 -0.6 A 600 units). Drug sensitivity assays were performed on cultures of each strain grown at 30°C in YEPD or C-Ura-Leu media to mid-log phase, 0.3 ml was then spread plated onto the corresponding solid media, and 10 l of each drug was pipetted onto a sterile BBL 1/4-inch diameter paper disc. The concentrations of drugs used were 10 mM cycloheximide, 25 mM hygromycin B, and 406 mM paromomycin. The plates were incubated for 2 days at 30°C. Sensitivity to each drug was measured by the radius of the zone of growth inhibition in millimeters around each disc. Protein Purification and Crystallization-eEF1A was purified as described (21) with the following modifications. TKY368 was grown in YEPD media and centrifuged, and the cell pellet was frozen and stored in liquid nitrogen. Two ml/g of lysis buffer (60 mM Tris, pH 7.5, 50 mM NH 4 Cl, 5 mM MgCl 2 , 0.1 mM EDTA, pH 8.0, 10% glycerol, 1 mM dithiothreitol (DTT), and 0.2 mM phenylmethylsulfonyl fluoride) was added to the thawed cell pellet, and lysed by three passages through a M110Y Microfluidizer Processor System (Microfluidics). Cellular debris was removed by two rounds of centrifugation for 30 min at 8,000 rpm and 90 min at 50,000 rpm. The supernatant was passed through a 0.22-m pore-size filter, glycerol was added to a final concentration of 25% and the protein solution applied to DE52 (Whatman) pre-equilibrated with buffer 1 (20 mM Tris, pH 7.5, 0.1 mM EDTA, pH 8.0, 25% glycerol, 1 mM DTT, and 0.2 mM phenylmethylsulfonyl fluoride) with 100 mM KCl. The unbound material was added to SP-Sepharose Fast Flow (Amersham Biosciences) pre-equilibrated with buffer 1 with 100 mM KCl. The material eluted with 500 mM KCl was dialyzed overnight against buffer 1 with 50 mM KCl, centrifuged for 20 min at 20,000 rpm, filtered through a 0.22-m filter, and loaded onto a Source 15S PE 4.6/100 column (Amersham Biosciences) pre-equilibrated with buffer 1 with 50 mM KCl using an AKTA fast-protein liquid chromatography system (Amersham Biosciences). The column was washed with buffer 1 with 50 mM KCl at a flow rate of 1 ml/min and eluted using a linear salt gradient of buffer 1 with 150 to 300 mM KCl. The appropriate fractions were dialyzed overnight against 10 volumes of buffer 1 with 100 mM KCl, and the protein concentration was determined by the Bradford assay (Bio-Rad).
A pET11d vector expressing the N-terminally His 6 -tagged eEF1B␣ fusion protein (TKB448) was used to prepare the K205A and K205R mutants using the QuikChange method (Stratagene). The resulting plasmids were pTKB529 and pTKB897, respectively. One liter of Escherichia coli strain BL21 containing one of the expression plasmids was grown to an A 600 of 0.6 in Luria broth with 100 g/ml ampicillin medium. Protein expression was induced with 1 mM isopropyl-␤-D-thiogalactopyranoside at 37°C for 2-3 h. Cells were harvested by centrifugation and lysed by sonication, and the recombinant protein was purified in accordance with the QIAexpressionist protocol for His 6 -tagged proteins under native conditions. Protein-containing fractions were dialyzed with 20 mM Tris, pH 7.5, 1 mM DTT, 0.1 mM EDTA, pH 8.0, 100 mM KCl, and 10% glycerol.
Expression and purification of the yeast eEF1B␣ K205A mutant encoded in a pET11d expression vector for crystallography and subsequent complex formation with yeast eEF1A was essentially performed as for the native protein complex (22). Crystallization was based on the conditions used for the native complex but occurred at a slightly lower pH and polyethylene glycol 2000 monomethyl ether concentration.
Data Collection and Structure Refinement-The reservoir solution of the crystallization tray was substituted with cryobuffer 1 (12% polyethylene glycol 2k monomethyl ether, 20% glycerol, 100 mM Tris-HCl, pH 8, 100 mM KCl, and 3 mM DTT) 20 h prior to data collection. The crystal was then transferred to cryobuffer 1 for 10 min prior to flash freezing in the nitrogen stream and collecting data. A second crystal had its reservoir substituted with cryobuffer 2 (37.5% polyethylene glycol 750 monomethyl ether, 100 mM Tris-HCl, pH 7.6, 100 mM KCl, 3 mM DTT, 2 mM GDP, and 5 mM MgCl 2 ) 4 h before data collection. The crystal was transferred 10 min before data collection to cryobuffer 2 and frozen as the first crystal, and data were collected. Data were collected at the XRD1 beamline at Elettra in Trieste, Italy. A high and low resolution data set was collected for the apo complex and a single data set for the nucleotide soaked crystal. All data were processed with the HKL package (23) ( Table 2). The native structure (PDB ID code 1F60) was used for solving both structures by rigid body refinement in CNS (24). The search model was divided into the three structural domains of eEF1A (2-240, 241-333, and 334 -441) and the C-terminal domain of eEF1B␣ (117-204). Manual model rebuilding was performed in O (25). The electron density for residue 205 in eEF1B␣ corresponded to the expected alanine ( Fig. 1A), confirming the mutation. The model for the eEF1A-eEF1B␣-GDP complex was initially refined without GDP. After two rounds of refinement of the search model against the data for the GDP-and Mg 2ϩ -soaked crystal, a GDP molecule was fitted into the density of a F o Ϫ F c map.
where F obs and F calc are the observed and calculated structure factors, respectively, and k is a scaling factor. d R free is identical to R on a subset of test reflections not used in refinement.
The Mg 2ϩ ion present in cryobuffer 2 could not be modeled. After iterative cycles of model refinement in O and CNS, and addition of water molecules, the quality of both structures was checked with PROCHECK (26) and Oops2. 5 A few minor refinements were performed, and the structures were refined in CNS a final time (Table 2). For the nucleotide-free complex residues 5-441 of eEF1A, 117-206 of eEF1B␣, and 539 water molecules could be modeled. For the GDP-containing complex these numbers where 2-441, 117-206 and 268, respectively. The nucleotide-free and GDP-containing structures have been deposited at the Protein Data Bank with ID codes 2B7C and 2B7B, respectively. Fig. 1 was made using PyMOL. 6 mant-GDP Binding Assay-mant-GDP was purchased from Molecular Probes. The 400-l mant-GDP solution was diluted with 5 ml of dH 2 O and loaded onto a 1-ml Hi-Trap DEAE-Sepharose Fast Flow column (Amersham Biosciences) preequilibrated with 20 mM triethylammonium bicarbonate, pH 7.8, using an AKTA fast-protein liquid chromatography system (Amersham Biosciences). The column was washed with 2 column volumes of 20 mM triethylammonium bicarbonate and eluted using a linear gradient from 0.02 to 1 M triethylammonium bicarbonate over 20 column volumes. All nucleotide-containing fractions were combined, lyophilized using a speed vacuum, resuspended in 100 l of dH 2 O, and stored at Ϫ20°C. The purity of the fluorescence-labeled nucleotide was analyzed by thin layer chromatography, and the concentration was determined by reading the absorbance at 356 nm and using the absorption coefficient of the N-methylanthraniloyl group ⑀356 ϭ 5800 M Ϫ1 cm Ϫ1 .
The affinity for mant-GDP at 25°C was determined by a fluorometric titration assay as previously described with minor modifications (29). Using a FluoroMax-2 spectrofluorometer (Jobin Yvon, Spex Instruments S.A., Inc.), 1 M eEF1A in 2.5 ml of binding buffer (10% glycerol, 50 mM Tris-Cl, pH 8.0, 50 mM KCl, and 5 mM MgCl 2 ) was placed in a 10-ϫ 10-ϫ 40-mm quartz cuvette with a magnetic stirring bar. Aliquots of mant-GDP were added with continuous stirring for 3 min before measuring the fluorescence (F obs ) with fluorescence resonance energy transfer via excitation at 280 nm and emission of 440 nm for the mant moiety of the nucleotide with both slit widths of 1.05 nm. The protein and nucleotide complex-dependent fluorescence values (F em ) were obtained by correcting for titration volume and inner filter effect using Equation 1, plotted against mant-GDP concentration and fit to Equations 2 and 3, where V f is the final volume, V o is the initial volume, Ab ex (280) is the excitation absorbance of mant-GDP, and Ab em (440) is the emission absorbance of mant-GDP, C is background fluorescence, ƒ M is the fluorescence coefficient of free mant-GDP, ƒ Eb is the fluorescence coefficient of mant-GDP bound to eEF1A, E t is the total eEF1A protein concentration, M t is the total concentration of mant-GDP, the concentration of eEF1A bound to mant-GDP is E b , and K d is the mant-GDP dissociation constant.
Stopped-flow Kinetics Measurements-mant-nucleotide fluorescence was monitored via fluorescence resonance energy transfer from endogenous tryptophans and tyrosines of eEF1A excited at 280 nm, and light was measured after passing through a KV399 filter of a stopped-flow fluorometer (KinTek). The maximum emission of mant-GDP is 440 nm. All experiments were performed in reaction buffer (10% glycerol, 50 mM Tris-Cl, pH 8.0, and 50 mM KCl) at 37°C. MgCl 2 was added to various concentrations, and reaction buffer without MgCl 2 contained 5 mM EDTA, pH 8.0. Reaction buffer containing 1 M eEF1A and 1 M mant-GDP was rapidly mixed with 45 M GDP and eEF1B␣ in reaction buffer. Dissociation of eEF1A⅐mant-GDP complexes was initiated by excess non-labeled GDP and catalyzed by eEF1B␣. No fluorescence change was observed when mant-GDP was mixed with the reaction buffer alone. At least seven fluorescence trace curves were averaged for each condition, and the resulting averaged curves were fitted to single or double exponential decay (Equation 4) to calculate the dissociation rate constants and amplitudes, where A is the amplitude of each phase, k is the dissociation rate constant of each phase, t is time, F o is the final fluorescence value, and F is fluorescence. All data fitting and residuals were performed using Sigma Plot version 9 (Systat Software, Inc.).

Inhibition of eEF1B␣ Function by the K205A Mutant Is Not
Due to a Structural Defect-Prior analysis demonstrated that a K205A mutation of eEF1B␣ is not viable as the only form of the protein, although it maintains the ability to bind to eEF1A as determined by a pull-down assay (15). The apo structure of the wild-type eEF1A⅐eEF1B␣ C-terminal catalytic fragment shows that Lys-205 inserts into the GDP/GTP-Mg 2ϩ binding pocket of eEF1A (7). Based on the ability to maintain substrate binding, co-crystal structures were resolved of the C-terminal fragment of amino acids 117-206 of eEF1B␣ K205A with eEF1A and eEF1A⅐GDP at resolutions of 1.8 and 2.6 Å, respectively. Mg 2ϩ was included during crystallization for the GDP structure, but Mg 2ϩ ions were not detectable in the density. Comparison to the structure of the wild-type eEF1A⅐eEF1B␣ C terminus complex indicates that there are no major changes in the K205A mutant structure except the lack of the Lys-205 side chain, supporting that this residue of eEF1B␣ has a functional requirement for catalytic activity (Fig. 1A).
The electron density for residue 205 of eEF1B␣ clearly corresponds to an alanine in both structures (Fig. 1A). In previous structures of nucleotides bound to the yeast eEF1A⅐eEF1B␣ C terminus complex, the ␤and ␥-phosphates appeared mobile. The structure of the mutant complex with GDP revealed the position of the ␤-phosphate for the bound nucleotide, but the Mg 2ϩ ion present in the cryo solution could not be modeled (Fig. 1B). The ␣-phosphate of GDP in the structure presented here interacts with the backbone amide group of T22 of eEF1A and two water molecules. The ␤-phosphate is positioned between the N terminus of helix A and the P-loop preceding it. The oxygen atoms of this phosphate interact with the backbone amide groups of Gly-19, Lys-20, and Ser-21 of eEF1A. The positive dipole of helix A contributes to the electrostatic binding of both the ␣and ␤-phosphates. One of the ␤-phosphate oxygens also interacts with a water molecule 2.53 Å away. This water molecule lies approximately at the position of the N atom from Lys-205 in the native nucleotide-containing complexes of eEF1A and the C-terminal fragment eEF1B␣. Asp-17 from the P-loop bends slightly toward the ␤-phosphate, but no other changes are seen in this region. Another minor difference between the six yeast complex structures is the orientation of Trp-194. This residue lines the GDP/GTP binding site in a region involved in binding the base of the nucleotide. In the native nucleotide-free and GDPcontaining structures, as well as the two structures presented here, this side chain points in one direction, whereas in the GDPNP structure and the GDP-Mg 2ϩ structure it is flipped ϳ180°around the C ␤ -C ␥ bond and rotated slightly about the C ␣ -C ␤ bond. However, the indole side chain lies in the same plane in all of the structures.
When comparing the phosphate-binding site of Sulfolobus solfataricus archeal EF1A⅐GDP (30) with all of the yeast complex structures, the peptide bond between Val-15 and Asp-16, equivalent to Val-16 and Asp-17 in yeast eEF1A, is flipped almost 180°. The orientation of the bond in the yeast structures disrupts the hydrogen-bonding interaction between the backbone amide of Asp-17 and one of the ␤-phosphate oxygens of GDP. The carbonyl oxygen in this peptide bond is positioned close to two ␤-phosphate oxygens, 2.61 and 2.81 Å away, respectively. This is energetically unfavorable, and it favors the exclusion of the phosphate moiety from the P-loop, in accordance with the mechanism suggested for GDP release in EF-Tu (31). Although the peptide was flipped in the yeast structure, the position of the ␤-phosphate was very similar to the one in archeal EF1A⅐GDP. The fact that there was density for the ␤-phosphate in the structure presented here argues strongly for the importance of Lys-205 in the nucleotide exchange reaction in yeast. When comparing the eEF1B␣ K205A mutant GDP structure with the native structure, it was clear that Lys-205 does not superimpose with the ␤-phosphate of GDP (Fig. 1C). It therefore seems likely that the function of Lys-205 is to disrupt the Mg 2ϩ and phosphate binding sites.
The Lethal Mutation eEF1B␣ K205A Suppresses an eEF1B␣deficient Strain-To further investigate the lethal K205A eEF1B␣ mutant, we analyzed its in vivo effects using an S. cerevisiae strain where the essential TEF5 gene encoding eEF1B␣ gene was deleted but the cells were supplemented with an extra copy of the TEF2 gene encoding eEF1A gene on a plasmid. This eEF1B␣-deficient strain ( Fig. 2A, pRS315), has two chromosomal and one plasmid-borne eEF1A-encoding genes, and thus increased eEF1A protein levels compensate for the loss of eEF1B␣. This strain exhibits cold-and temperature-sensitive growth defects (12). The addition of wild-type eEF1B␣ in this strain results in enhanced growth compared with the eEF1B␣deficient strain ( Fig. 2A, WT). Surprisingly, this increased growth was equivalent when the lethal K205A eEF1B␣ mutant  and eEF1B␣ in violet). B, stereo view of the A -weighted 2F o Ϫ F c electron density map contoured at 1.5 around GDP after the final refinement. The GDP molecule omitted from the map calculation has been superimposed on the density and shown in yellow (carbon), blue (nitrogen), red (oxygen), and green (phosphorus). C, superposition of the native apo complex with that of the K205A GDP-containing complex showing the relative position of Lys-205 to GDP. The native structure is shown in wheat and slate for eEF1A and eEF1B␣, respectively, and the mutant complex is colored as above.
was added to the eEF1B␣-deficient strain ( Fig. 2A, K205A). Doubling time assay results at the permissive temperature confirmed the enhanced growth rate of an eEF1B␣-deficient strain in the presence of wild-type eEF1B␣ (4.2 Ϯ 0.03 h) or the K205A mutant form of eEF1B␣ (4.4 Ϯ 0.14 h) compared with a rate of 7.15 Ϯ 0.41 h for the eEF1B␣-deficient strain. The K205A mutant protein was stably expressed in the eEF1B␣-deficient strain as confirmed by Western blot analysis (data not shown).
Because an eEF1B␣-deficient strain was also previously demonstrated to have hypersensitivity to inhibitors of translation elongation (12), we determined whether the addition of wild-type or K205A mutant eEF1B␣ had the ability to suppress the translational defects conferred by the eEF1B␣deficient strain. The sensitivity to the translation elongation inhibitor cycloheximide in this strain was suppressed equally in the presence of K205A mutant compared with wild-type eEF1B␣ (data not shown). This supports that suppression is likely due to the presence of residual activity of K205A eEF1B␣.
The eEF1B␣ K205R Mutant Is Functional in Vivo-To investigate the potential of occupying the Mg 2ϩ binding site constitutively, we replaced Lys-205 with Arg, which maintains a positive charge but extends one carbon further toward the Mg 2ϩ binding site. The K205R mutant, unlike K205A, is functional in vivo as the only form of eEF1B␣, and the strain exhibits comparable growth to a wild-type strain on YEPD media at permissive and non-permissive temperatures (Fig. 2B). However, the doubling times at permissive temperature in liquid media reveal that the K205R mutation causes an 18% reduction of growth over a wild-type strain (3.9 Ϯ 0.07 h versus 3.1 Ϯ 0.12 h). To test the possibility that K205R might result in a translational defect, we tested the sensitivity to drugs that inhibit translation elongation. K205R eEF1B␣ does not affect sensitivity to the translational inhibitors hygromycin B, cycloheximide, or paromomycin (data not shown). The presence of the K205R eEF1B␣ mutant also fully suppresses the growth defect (4.2 Ϯ 0.11 h) and sensitivity to cycloheximide observed in an eEF1B␣-deficient strain ( Fig. 2A and data not  shown). These results indicate that the K205R mutant allele is not significantly defective in vivo.
eEF1B␣-catalyzed Nucleotide Exchange on eEF1A-To assess the nucleotide binding and exchange properties of eEF1A, the GDP nucleotide analog, mant-GDP, was used. This nucleotide contains a fluorophore, which increases its fluorescence signal in response to hydrophobicity. The increase in fluorescence indicates the binding of the nucleotide to protein. Upon the addition of mant-GDP to eEF1A, a considerable increase in fluorescence was observed by fluorescence resonance energy transfer from tryptophans or tyrosines of eEF1A to the mant group of GDP using the fluorometric titration assay (Fig. 3A).
To test if mant-GDP and non-fluorescent GDP possess equivalent affinities for eEF1A, the equilibrium dissociation constant (K d ) of eEF1A for mant-GDP in the presence of 5 mM Mg 2ϩ was obtained by using a FluoroMax-2 spectrofluorometer. Both mant-GDP and non-labeled GDP bind to eEF1A with similar affinities. The K d for mant-GDP binding to eEF1A is 1.8 ϫ 10 Ϫ7 M (Fig. 3A), whereas the published K d of GDP for eEF1A is 10 ϫ 10 Ϫ7 M (16). Thus, the change in mant-GDP fluorescence provides an accurate depiction of the eEF1A-GDP interaction.
To study spontaneous GDP dissociation from eEF1A in the absence of its known GEF, we determined the dissociation rate constant of the mant-GDP⅐eEF1A complex using stopped-flow kinetics. eEF1A bound to mant-GDP was rapidly mixed with excess GDP in 5 mM Mg 2ϩ to displace the nucleotide analog causing a decrease in fluorescence. The average dissociation rate constant was 0.17 s Ϫ1 (Fig. 3B). We examined the rate of GDP release catalyzed by eEF1B␣ in 5 mM Mg 2ϩ by monitoring the dissociation of mant-GDP from eEF1A over time in the presence of excess GDP and wild-type eEF1B␣ protein. The dissociation rate of mant-GDP was accelerated 7 ϫ 10 2 -fold. These results support the significance of catalyzing the nucleotide exchange reaction by eEF1B␣, so sufficient GTP-bound eEF1A can be available to sustain the overall rate of elongation. Two exponential phases of the dissociation of mant-GDP from eEF1A were observed in eEF1B␣-catalyzed reactions (Fig. 3C). The faster dissociation rate constant was 70 times more rapid and accounted for ϳ90% of the total amplitude (Fig. 3C). The two phases could be a result of several effects. First, a biphasic effect related to the presence of two isomers in the mant-GDP mixture was previously observed for the dissociation of mant-GDP from Ras complexed with its GEF, Cdc25 (32). Second, the effect could be attributed to subpopulations of eEF1A unable to be activated by eEF1B␣ or different conformations of eEF1B␣ that have differing effects on nucleotide dissociation reaction, thereby causing a much slower reaction.   ): B, k 1 ϭ 0.l7 Ϯ 0.006, A 1 ϭ 1.2 Ϯ 0.002; C, k 1 ϭ 92.4 Ϯ 0.1, A 1 ϭ 1.33 Ϯ  0.006, k 2 ϭ 1.31 Ϯ 0.07, A 2 ϭ 0.17 Ϯ 0.003. Residual plots were prepared to detect experimental error for the fitted data subsets.
The Role of Lys-205 eEF1B␣ in the Nucleotide Exchange Activity of eEF1A-To test the proposed involvement of the Lys-205 residue of eEF1B␣ in the nucleotide exchange mechanism, we measured the rate of mant-GDP dissociation catalyzed by the K205A and K205R forms of eEF1B␣ by stopped-flow kinetics. mant-GDP dissociated from eEF1A at a rate of 7.4 Ϯ 0.03 s Ϫ1 when catalyzed by K205A eEF1B␣ (Fig. 4A). This is a 13-fold reduction compared with the mant-GDP dissociation rate in the presence of wild-type eEF1B␣ (Fig. 3C). Nevertheless, the nucleotide dissociation rate constant was stimulated 43 times by the mutant K205A eEF1B␣. This result indicates that the K205A eEF1B␣ protein is not a complete loss of function mutant but is defective in its catalytic activity to release GDP from eEF1A. The K205R eEF1B␣ protein also displayed a decreased mant-GDP dissociation rate of 49 Ϯ 0.4 s Ϫ1 (Fig. 4B); however, this rate was more comparable with that of the wild-type eEF1B␣, 92 Ϯ 0.1 s Ϫ1 (Fig. 3C). The rate of mant-GDP dissociation from eEF1A catalyzed by the K205R eEF1B␣ protein stimulated the mant-GDP dissociation rate by ϳ7-fold compared with the K205A mutant (Fig. 4). These results support the contribution of Lys-205 eEF1B␣ in catalysis of GDP dissociation from eEF1A.
To determine the maximal dissociation rate constant of the mant-GDP⅐eEF1A complex, experiments were performed at varying concentrations of eEF1B␣. A time course of the fluorescence intensity was monitored for each eEF1B␣ concentration to obtain the dissociation rate constants by fitting the data to a single or double exponential decay curve. The dissociation rate constants were plotted against the eEF1B␣ concentration and fitted to a hyperbolic curve and the apparent equilibrium dissociation constant (K d ) and the maximal dissociation rate constant (k off ) values were obtained. The maximal dissociation rate constant of the mant-GDP⅐eEF1A complex under saturating conditions of wild-type eEF1B␣ was 122 Ϯ 8 s Ϫ1 , and the apparent K d value was 4 Ϯ 0.8 M, whereas the K205A eEF1B␣ mutant caused the maximal rate to be 15-fold slower (8 Ϯ 0.3 s Ϫ1 , Fig. 5, A and B). Previous data demonstrated the critical function of Lys-205 by the lethality of a mutation to alanine in vivo, and this was not due to the loss of the ability to bind its substrate, eEF1A (15). Interestingly, K205A exhibited an extremely low apparent K d value of 0.4 Ϯ 0.1 M compared with the wild-type protein (Fig. 5B). K205R eEF1B␣ caused the maximal nucleotide dissociation rate of eEF1A to be lowered to 68 Ϯ 6.6 s Ϫ1 (Fig. 5C). The apparent K d value of the K205R for the mant-GDP⅐eEF1A complex was 2.4 Ϯ 0.8 M, which was essentially the same as the wild-type eEF1B␣ pro-tein. Therefore, we conclude that the conserved Lys-205 residue contributes to the catalytic activity of eEF1B␣.
Mg 2ϩ Effects on the Guanine Nucleotide Exchange Activity-Mg 2ϩ is modeled to be an inhibitory factor of guanine nucleotide exchange in well studied G-proteins (17,33,34). To determine the effects of Mg 2ϩ on the guanine nucleotide exchange of eEF1A in the absence or presence of a wild-type or K205A and K205R mutant forms of eEF1B␣, we used stopped-flow kinetics at varying concentrations of Mg 2ϩ . eEF1A bound to mant-GDP was rapidly mixed with excess GDP and a saturating amount of eEF1B␣ at different Mg 2ϩ concentrations. Using a single or double exponential equation, the dissociation rate constants and amplitudes were calculated ( Table 3). The Mg 2ϩ concentration and GDP dissociation rate constant for eEF1A with wild-type eEF1B␣ had an inverse relationship (Fig. 6B), however, the spontaneous exchange activity of eEF1A did not show an inhibitory effect with Mg 2ϩ (Fig. 6A). Consequently, we fitted the dissociation rate constants to a hyperbolic decay curve to obtain the apparent K1 ⁄ 2 value that reflects Mg 2ϩ binding to eEF1A and the maximal nucleotide dissociation rate constant in the absence of Mg 2ϩ . In the absence of Mg 2ϩ , the dissociation rate of the eEF1A⅐mant-GDP complex with wild-type eEF1B␣ protein accelerated to 260 Ϯ 6.5 s Ϫ1 compared with the maximal dissociation rate constant, 122 Ϯ 8 s Ϫ1 in the presence of 5 mM Mg 2ϩ (Fig. 6B). These data support the model that physiological levels of Mg 2ϩ inhibit the rate at which GDP is released from the eEF1A⅐eEF1B␣ complex (17). However, the dissociation rate constant of the eEF1A⅐mant-GDP complex without eEF1B␣ in the absence of Mg 2ϩ was 0.17 s Ϫ1 , which was equivalent to the dissociation rate constant in the presence of 5 mM Mg 2ϩ (Fig. 6A). Therefore, cellular Mg 2ϩ levels did not affect the spontaneous release of GDP from eEF1A in the absence of eEF1B␣. The apparent K1 ⁄ 2 for Mg 2ϩ to eEF1A⅐wild-  . A time course of fluorescence intensity was monitored for each eEF1B␣ concentration, and data were fitted to a single or double exponential decay equation to calculate the dissociation rate constants (F). A hyperbolic equation was used to fit the given dissociation rate constants to calculate the K d (micromolar) and k off (s Ϫ1 ) values of K d ϭ 4 Ϯ 0.8 and k off ϭ 122 Ϯ 8 (A), K d ϭ 0.4 Ϯ 0.1 and k off ϭ 8 Ϯ 0.3 (B), and K d ϭ 2.4 Ϯ 0.8 and k off ϭ 68 Ϯ 6.6 (C ). Residual plots were prepared to detect experimental error for the fitted data subsets. type eEF1B␣ complex was 3.8 Ϯ 0.4 mM, and the maximal nucleotide dissociation rate in the absence of Mg 2ϩ was 274 Ϯ 6 s Ϫ1 (Fig. 6B).
The dissociation rates of eEF1A⅐mant-GDP complex in the presence of the K205A eEF1B␣ mutant at varying Mg 2ϩ concentrations were fitted to a hyperbolic decay curve, and the apparent K1 ⁄ 2 value decreased 27-fold (0.14 Ϯ 0.03 mM, Fig. 6C). This result confirms the structural model that the K205A eEF1B␣ mutant is unable to efficiently disrupt the Mg 2ϩ binding pocket, and therefore the release of GDP is slower. This likely accounts for the lethality in vivo. The maximal dissociation rate constant of the K205A mutant in the absence of Mg 2ϩ was 136 Ϯ 9 s Ϫ1 , which was decreased by only 2-fold compared with the wild-type eEF1B␣ protein (Fig. 6C ).
The K205R mutant causes the apparent K1 ⁄ 2 value for Mg 2ϩ to increase considerably to 27 Ϯ 13 mM, suggesting that the Mg 2ϩ ion is not binding to eEF1A as well as in the presence of wildtype eEF1B␣. Therefore, Mg 2ϩ does not affect the catalytic activity of this mutant (Fig. 6D). In the absence of Mg 2ϩ , the maximal GDP dissociation rate constant from the eEF1A⅐K205R eEF1B␣ complex is 62.6 Ϯ 3.4 s Ϫ1 (Fig. 6D), which is equivalent to the maximal dissociation rate constant in 5 mM Mg 2ϩ , 68 Ϯ 6.6 s Ϫ1 (Fig. 5C ). These results support the hypothesis that Arg at position 205 inserts into the Mg 2ϩ binding pocket to reduce Mg 2ϩ binding to eEF1A. Thus, the Lys-205 residue contributes to the catalytic function of eEF1B␣ by displacing Mg 2ϩ from its binding site, allowing the nucleotide exchange activity of eEF1A.

DISCUSSION
A GEF acts as a catalyst to accelerate the rate of GDP release from a G-protein (17). The mechanism of guanine nucleotide exchange has some conserved aspects, such as a high conservation of the nucleotide binding domains of G-proteins and a proposed key role for Mg 2ϩ in the affinity for the nucleotide. This is in contrast to the vast diversity of the size, structure, and sequence of GEF proteins themselves. However, a general model for GEF function is to overcome the inhibitory potential of Mg 2ϩ allowing release of GDP from G-proteins (17). eEF1B␣ is a component of the eEF1 complex and functions as an essential GEF required to sustain sufficient translation elongation in eukaryotes. To gain insight into the guanine nucleotide exchange mechanism of the eEF1A⅐eEF1B␣ complex, we have examined the roles of Mg 2ϩ and a key lysine of eEF1B␣ in catalysis.
Structural data established that the K205A eEF1B␣ C terminus⅐eEF1A complex is essentially the same as the wildtype structure except for the loss of the Lys-205 side chain. This suggests that the lethality of the K205A mutant is caused by an inefficient catalytic GEF function. Stopped-flow kinetics demonstrates that the K205A eEF1B␣ protein shows a reduced guanine nucleotide exchange activity compared with the wild-type protein. Furthermore, Mg 2ϩ is more effective in decreasing the nucleotide exchange rates of the eEF1A⅐K205A eEF1B␣ complex. Thus the half Mg 2ϩ concentration needed to inhibit nucleotide exchange by wild-type eEF1B␣ was much greater than the eEF1B␣ K205A mutant. These data confirm the model that the K205A mutant is likely unable to displace the Mg 2ϩ ion, which is critical for the nucleotide exchange reaction. However, the mutant protein still enhances GDP dissociation from eEF1A significantly over the spontaneous rate, thereby suggesting additional mechanisms for rate enhancement. We speculate that other contacts between eEF1B␣ and eEF1A disrupt interactions between the P-loop and switch regions of eEF1A with GDP to release the nucleotide.
Using an eEF1B␣-deficient strain that survives due to the presence of excess eEF1A (12) allows for in vivo analysis of the normally lethal eEF1B␣ K205A mutant. The addition of the K205A mutant to an eEF1B␣-deficient strain causes suppression of a growth defect of that strain. This is a strong indication that either the residual activity in exchange is above the threshold required for viability or eEF1B␣ might possess other cellular functions. Because K205A eEF1B␣ also suppresses sensitivity to a translation elongation inhibitor, the mutant likely has residual translation activity in vivo. One possible explanation for this finding is that the introduction of the wild-type eEF1B␣ in the eEF1B␣-deficient strain is now an eEF1A overexpression strain. eEF1A overexpression compared with a wildtype strain shows slow growth and actin cytoskeleton disruption phenotypes (35). The addition of the K205A mutant in the eEF1B␣-deficient strain may not behave the same as an eEF1A overexpression strain, due to decreased catalytic GEF activity and higher affinity for eEF1A. This may allow less eEF1A to bind actin (9) and consequently avoid the negative consequences on growth. Thus, the eEF1B␣ K205A mutant may prevent eEF1A from inappropriate association with the actin cytoskeleton and thus avoid the negative consequences seen with eEF1A overexpression (35). Because eEF1A binding to K205A eEF1B␣ is maintained, the mutant may also function in channeling of reactants in protein synthesis, such as the suggested role for eEF1B␣ in aa-tRNA binding (7). Actin binding by eEF1A is mutually exclusive with aa-tRNA binding (9). Conceivably, a fine balance between eEF1A functions in protein synthesis versus actin binding must be achieved for proper cellular homeostasis.