Technical Comments

Response to Comment on “The Mechanism for Activation of GTP Hydrolysis on the Ribosome”

See allHide authors and affiliations

Science  01 Jul 2011:
Vol. 333, Issue 6038, pp. 37
DOI: 10.1126/science.1202532


Our report of the crystal structure of elongation factor Tu (EF-Tu) and aminoacyl–transfer RNA bound to the ribosome with a guanosine triphosphate (GTP) analog included a proposed mechanism of GTP hydrolysis by EF-Tu involving histidine-84. Liljas et al. summarize experimental evidence against this mechanism and propose a substrate-assisted catalytic model. However, these experiments and the model are also problematic. Further study is required to definitively determine the mechanism of GTP hydrolysis by EF-Tu.

Our study (1) described the structure of a cognate ternary complex of elongation factor Tu (EF-Tu), transfer RNA (tRNA), and a guanosine triphosphate (GTP) analog (β-γ-methyleneguanosine 5′-triphosphate, GDPCP) bound to the ribosome, thus mimicking the state of decoding just before GTP hydrolysis by EF-Tu. The structure suggests how codon recognition in the decoding center results in conformational rearrangements that activate GTP hydrolysis by EF-Tu. Furthermore, the structure identified a critical interaction between 23S ribosomal RNA (rRNA) residue A2662 and the conserved His84 of EF-Tu that reorients His84 into the GTPase center to catalyze GTP hydrolysis. The discussion section of our paper included a tentative model for the catalytic mechanism of the GTP hydrolysis reaction itself. The proposal had its basis in the conformation of the GTPase center and included His84 acting as a general base.

Liljas et al. (2) suggest that this mechanism is unlikely for a variety of reasons. They argue that the local environment of His84 between two charged phosphates (A2662 of the 23S rRNA and the γ-phosphate of GTP) would elevate its pKa (where Ka is the acid dissociation constant) such that it is protonated and unable to act as the general base. In support, they cite a study that concludes that the rate of GTP hydrolysis for EF-Tu is independent of pH between 6.5 and 8.5 (3). However, these experiments used 2′(3′)-O-(N-Methylanthraniloyl)-guanosine 5′-[γ-thio]triphosphate (mant-GTP-γ-S) in place of GTP, which alters a functional group directly involved in the reaction and has a hydrolysis rate that is a factor of ~60,000 slower than GTP (3, 4). GTP-γ-S is known to alter the arrangement of enzyme active sites (5), and the pH independence observed may not necessarily hold true for the native GTP. Furthermore, the interaction of His84 with the phosphate of A2662 and the possible perturbation of its pKa would presumably not occur until His84 is repositioned into the GTPase center during GTPase activation. At this point, an increase in the pKa of His84 would only make it a better base and thus more efficient at abstracting a proton from the catalytic water as we had proposed. Liljas et al. also point out that, although the substitution of Ala for His84 reduces the rate by six orders of magnitude (3), substitution with Gln (H84Q) results in only a modest impairment (68). However, the magnitude of the H84Q impairment is unclear: Previous reports have investigated either the rates of poly-Phe synthesis (7), where GTP hydrolysis is not normally rate-limiting, or the extremely slow rates of GTP hydrolysis catalyzed by Thermus thermophilus EF-Tu on Escherichia coli ribosomes (8). Understanding the catalytic role of a Gln84 substitution in EF-Tu will require a more direct measurement of the catalytic rate, preferably in conjunction with a structure of the mutant complex.

Next, Liljas et al. point out that the hydrolysis reaction in cellular GTPases, such as Ras, proceeds through a substrate-assisted catalytic mechanism (9). It is possible that translational GTPases use a similar strategy, but the proposed analogy with Ras has several problems. For example, Liljas et al. argue that His84 must be protonated and donating a hydrogen bond to the nucleophilic water, in part because the water hydrogens are interacting with the carbonyl oxygen of Thr62 and the γ-phosphate of GTP. However, these interacting partners result in a highly unfavorable geometry for the water molecule (1), with the caveats that the bond angles of GDPCP differ from those of GTP and that the structure was determined to relatively modest resolution (3.2 Å). Furthermore, in cellular GTPases such as Ras or Ran, the organization of the active site is extremely similar, with the exception that His84 is replaced by the unprotonated carbonyl oxygen of Gln61 (Ras numbering) (10). If His84 were playing the role of Gln61, it would, at least in the ground state, be required to function as a hydrogen bond acceptor and therefore be unprotonated.

Although a common mechanism with cellular GTPases may be aesthetically pleasing, it is not clear that the His84 in EF-Tu is directly analogous to the Gln61 in Ras/Ran. For example, His84 cannot possibly perform the dual hydrogen bond donor and acceptor role of Gln61 in the mechanism cited by Liljas et al. for Ras (9). Indeed, a Q61H mutation in Ras decreases catalysis rates by six orders of magnitude (9), suggesting that the roles of His and Gln are not interchangeable. The absolute conservation of His84 in translational GTPases, which can differ radically outside the G domain (11), suggests that this histidine is catalytically important in this particular class of GTPases. Accordingly, in some other GTPases that contain residues other then Gln at this position (12), such as the Rap and SRP proteins, GTP hydrolysis is catalyzed by entirely different mechanisms (13, 14).

Lastly, Liljas et al. suggest that His84 plays the role of both Gln61 and the arginine finger in Ras/Ran, which is not present in all GTPases (10). However, His84 is over 6 Å away from the bridging phosphate oxygen, where the arginine finger localizes in Ras/Ran. Arginine fingers can also interact with a nonbridging oxygen of the γ-phosphate, but His84 would have to change conformation to accomplish this, which would likely disrupt its interaction with the nucleophilic water and weaken the interaction between His84 and the phosphate of A2662 of the sarcin-ricin loop.

Instead of the roles proposed by Liljas et al., it is possible that the contribution of His84 to catalysis is entirely indirect, similar to the role suggested for Gln61 in some cellular GTPases (15).

Thus although Liljas et al. have raised legitimate concerns, we continue to feel that the precise mechanism of catalysis and the contribution of His84 will need, as we stated, “to be assessed by [additional] biochemical experiments” (1) and further structural studies. However, the main focus of our paper, the role of the ribosome in organizing the GTPase center into a catalytically competent arrangement, in particular the reorientation of His84 by A2662 of the sarcin-ricin loop of 23S rRNA, is still likely to be a universal feature for all translational GTPases.


View Abstract

Navigate This Article