Elongation Factor Tus Nucleotide Binding Is Governed by a Thermodynamic Landscape Unique among Bacterial Translation Factors
Insights into the apo Conformation of EF-Tu The first structural insight into the apo conformation of EF-Tu was gleaned from the crystal structure of the EF-Tu•EF-Ts complex 1 . In this structure, switch I of EF-Tu is disordered and the domain arrangement is similar to the GDP conformation. Subsequently, Thirup and co-workers were able to crystalize an EF-Tu•EF-Ts complex with EF-Tu in the closed GTP conformation, indicating that the EF-Tu•EF-Ts complex exhibits conformational flexibility 2 .
... l flexibility 2 . Our kinetic data reported here provides additional insight in the apo state of EF-Tu and suggests that the free apo state of EF-Tu adopts a unique conformation that is able to recognize GTP and GDP similarly. Similar association and different dissociation activation barriers support that nucleotide binding to this non-discriminatory conformation of EF-Tu is followed by conformational changes of EF-Tu that depend on the presence and absence of the gamma-phosphate in GTP. Therefore, the nucleotide release mechanism is likely the inverse in which a conformational change has to occur prior to nucleotide dissociation. This process would require a unique conformation for EF-Tu•apo and is constant with the reported conformational flexibility observed by Johansen et al and Kavaliauskas et al   . Considering the entropic landscape of EF-Tu, the EF-Tu•apo conformation is less stable than both EF-Tu•GTP and EF-Tu•GDP (Fig 4) . Since EF-Tu employs water coordination to entropically stabilize the GTP conformation it is likely that water coordination stabilizes EF-Tu•GTP and EF-Tu•GDP compared to EF-Tu•apo. If GDP is removed from the structure of EF-Tu•GDP the SASA increases by 109 Å 2 . The difference in SASA between EF-Tu•GTP and EF-Tu•GDP is 1041 Å 2 ( Fig S6B) and since the entropy gap between EF-Tu•GDP and EF-Tu•GTP is similar to the entropy gap between EF-Tu•apo and EF-Tu•GDP ( Fig 4C) it is likely that SASA alone does not explain the entropy of the EF-Tu•apo conformation. Therefore, EF-Tu•apo cannot merely be a similar conformation to EF-Tu•GDP without nucleotide, but has to be a unique conformation. Another possible explanation for the entropy of EF-Tu•apo is that this state is less flexible. This is unlikely as switch I is disordered in the EF-Tu•EF-Ts crystalized complex from Kawashima et al 1 . If this is the conformation that switch I adopts in the apo state then it is likely going to be more flexible. However, this does agree with the observation that EF-Tu•apo coordinates more water molecules, as a disordered switch I would have a larger SASA. Since EF-Tu•apo is less entropically stable compared to EF-Tu•GTP or EF-Tu•GDP which cannot be explained simply by the loss of nucleotide or EF-Tu•apo being less flexible then EF-Tu•apo must adopt a unique conformation. To directly compare the thermodynamic contributions of each nucleotide bound state relative to each other we can use the law of mass-action. However, since GTP contains an additional phosphate compared to GDP mass is not conserved in the kinetic mechanism of EF-Tu nucleotide binding, preventing the implementation of mass action (Fig 1) . The fact that there is no difference in the ΔH ‡ a or TΔS ‡ a and that the mass is conserved in the respective halves of the nucleotide dissociation mechanism (k-1 and k-2), which defines the thermodynamic landscape of nucleotide binding, indicates that the mass of Pi has little to no influence The Enthalpic Stability of EF-Tu•GDP is Targeted by EF-Ts for Nucleotide Dissociation Since EF-Tu and EF-Ts have co-evolved, the residues involved in stabilizing the GDP conformation of EF-Tu are likely the same residues targeted by EF-Ts to help mediate GDP release. The current understanding is that EF-Ts stimulates GDP dissociation from EF-Tu through 3 factors: (1) destabilization of the Mg 2+ coordination, (2) flipping of the P-loop and (3) destabilizing the nucleotide-ribose binding site 1, 5-6 . Previously, the interactions between EF-Ts and helix B (switch IIamino acids 84-92), as well as helix D (amino acids 139-144) have been studied as EF-Ts makes direct interactions with these regions of EF-Tu 5-9 . Crystal structures of the EF-Tu•EF-Ts complex show that residues in helix A of EF-Tu contact the C-terminal end of EF-Ts 1 . Additionally, our group has previously shown that these interactions destabilize helices A and F and increase the rate of nucleotide release 10-fold 10 . Since helix A is located in proximity to a number of interactions that stabilize the GDP conformation, it is likely that EF-Ts specifically disrupts these hydrogen bonds. Two possible mechanisms may explain how the C-terminus of EF-Ts helps to stimulate GDP dissociation: (1) it helps position EF-Ts properly onto EF-Tu in order for F81 to insert between H118 and H84 or (2) EF-Ts destabilizes the hydrogen bonding potential of helix A and as a consequence lowers the ∆H ‡ d barrier favoring dissociation. Our data supports the latter hypothesis and is consistent with the crystal structure of the EF-Tu•GDPNP•EF-Ts complex where EF-Ts engages with EF-Tu in a conformation where the C-terminus does not pack against helix A 2 . This structural model suggests that these interactions are not required for EF-Ts binding but are, instead, involved in promoting efficient nucleotide dissociation 2 .