The P-site A76 2′-OH acts as a peptidyl shuttle in a stepwise peptidyl transfer mechanism

Abstract

Notwithstanding various unsolved questions regarding the structure and function of the ribosome, the process of peptide bond formation is of particular importance, being the heart of protein synthesis. Several experimental studies have been carried out on the pre- and post-peptidyl transfer structures in the ribosomal active site. Based on these structures, different reaction mechanisms have been proposed and further investigated using computational techniques. However, the detailed mechanism of peptidyl transfer, as well as the atoms and functional groups involved in this process are still in limbo. Although it was suggested that the A2451 is present in the active site of the ribosome in the previous crystallographic structures, the details of its participation have not been fully investigated. Furthermore, despite the highlighted importance of the P-site A76 2′-OH group in previous studies, its actual role during the process is still unclear. Finally, whether the process of peptidyl transfer is a stepwise mechanism or a concerted one is still under debate. Several computational mechanistic studies have been carried out to investigate the catalytic power of the ribosome, yet, they do not cover all the three concerns mentioned above. Therefore, as well as re-investigating the previously proposed reaction mechanisms with a higher level of theory and basis set (i.e. M06-2X/6-31++G(d,p)//M06-2X/6-311++G(d,p)) in this study, we propose three new reaction mechanisms based on three different pre- and post-peptidyl transfer structures obtained from previous experimental studies. The results of this study highlights the important role played by the P-site A76 2′-OH group in catalyzing the reaction. However, instead of acting as a so-called proton shuttle, this group acts as a polypeptidyl shuttle, transferring the growing polypeptide chain from the leaving 3′-O to the attacking nucleophile through a mechanism known as transesterification. This reaction is aided by the A2451 3′-OH rRNA base, acting as a proton shuttle between the P-site 3′-O and the 2′-O and further stabilizing the transition structure.

---

1. Introduction

The measurements of high resolution crystal structures of the ribosome have made a major contribution towards understanding the different stages of protein synthesis.^1–11 Of particular importance is the elongation cycle, the stage in which the process of peptide bond formation occurs.¹² However, the source of catalysis for such a mechanism is still unknown. The main challenge of understanding this mechanism started when the crystal structure of 50S ribosome was obtained from Haloarcula marismortui (Hma) in high resolution, indicating that the ribosome is a ribozyme.^13,14 This data supported an earlier study on peptide bond formation in “proteinless” ribosome and participation of the 23S-rRNA (ribosomal-RNA) residue on the peptidyl transfer mechanism.¹⁵ It was a major breakthrough towards understanding the catalytic role of the ribosome; that is, the ribosomal proteins do not directly participate in the chemistry of the reaction and the catalytic power of the ribosome is mainly from its RNA residues. Among these residues is the so-called single adenosine A2451 base (E. coli numbering is used throughout the manuscript) which not only was observed in a close proximity of the active site group,^{7,10,11,16–19} but whose mutation was also suggested to be lethal in vivo.¹⁶ However, the juxtaposition of A2451 rRNA base to the active site and its direct participation in the mechanism was^7,19,20 and still is¹² under debate. The main reason for this controversy is the observed decrease in reaction rate upon mutation of the A76 2′-OH group of the P-site tRNA (transfer-RNA) by 100-fold^21–23 compared to the 10-fold¹⁷ decrease upon mutation of the A2451 rRNA base. This proposes a substrate-assisted catalysis mechanism in which the rRNA bases would only facilitate the reaction by positioning the substrates in close proximity to one another and the main chemical catalysis is carried out by the P-site A76 2′-OH group.^10,24–27 However, an only 10-fold difference in reaction rate reduction could be an indication of the fact that the high rate of peptide bond formation is not solely due to the substrate assisted catalysis, nor it is solely due to the ribosomal catalysis, though, it could be due to a combination of both.

Despite these observations, the mechanistic principles of the ribosomal catalysis still remain unsolved. The main challenge in the mechanistic study of peptide bond formation is to figure out the role played by the A76 2′-OH group as well as the A2451 since their activity was clearly observed in many empirical studies.^7,10,24–27 In the computational study carried out by Gindulyte et al., the A76 2′-OH group was suggested to bring the reacting groups of the A-site and P-site amino acids in close proximity without directly participating in the reaction.²⁸ On the other hand, Åqvist proposed that the 2′-OH is directly involved in the reaction, forming a 6-membered transition structure; however, it does not constitute a catalytic effect.^29,30 Later in 2010, they proposed a mechanism in which a water molecule is also involved in the proton shuttle additional to the 2′-OH, forming an 8-membered transition structure based on an earlier high resolution crystallographic data carried out by Steitz which confirmed the presence of water molecules in the active site.^10,11 Steitz however, proposed the formation of a carbon tetrahedral intermediate in a stepwise mechanism where a water molecule neutralizes the oxyanion. Based on the observation of this intermediate, various mechanisms have been proposed and computationally investigated, in most of which, formation of this intermediate is also known to be rate determining,^31–33 while in a few others it appears to occur quite fast relative to the second step, making the latter rate determining.^34,35 Even though these studies share a similar tetrahedral intermediate structure, they go through different types of transition states (TSs) in terms of number of protons “in flight” in both step 1 and step 2 (Scheme 1).

Scheme 1 Previously proposed concerted and stepwise mechanisms, indicated in valence and core of the scheme respectively. (A) Both TS4 and TS6 correspond to the concerted 4- and 6-member mechanisms. TS1-4 → TS2-4 corresponds to a 4-membered mechanism in both steps. TS1-4 → TS2-6 corresponds to a 4-membered mechanism in the first step and a 6-membered mechanism in the second step, and so on. (B) TS8 corresponds to a concerted 8-membered mechanism, while TS1-6 → TS2-8 corresponds to a 6-membered one in the first step and an 8-membered one in the second step.

Except for the study carried out by Świderek et al., and Rangelov et al. in which the amino proton is transferred directly to the P-site carbonyl oxygen without going through the 2′-OH of the P-site ribose (corresponding to a 4-membered transition structure),^31,35 the first step in all stepwise mechanism studies are the same and involve a double proton transfer from an attacking nucleophile to the 2′-OH and from the latter to the carbonyl oxygen (corresponding to a 6-membered transition structure).^32–34

For some unknown reason, regardless of the confirmed direct participation of the P-site 2′-OH group during the reaction and its role in catalyzing the reaction in the experiments,^{10,21,24,26,36} the proposed mechanisms in which the 2′-O is donating its proton to the leaving 3′-O where there is no stabilizing water in the active site occur in a relatively low rate in computational studies. Finding out the reason can lead towards understanding the origin of the catalytic power of the ribosome.

Firstly, it is not known which group (if any) of the A2451 ribosomal base is involved in proton transfer. Secondly, the source of the proton in ester bond dissociation and the destination of the proton from attacking nucleophile are unidentified. Finally, whether these two mechanisms both occur in one step (concerted) or in two different steps (stepwise) is still unknown. To address these problems, we have performed DFT (Density Functional Theory) calculations using the M06-2X/6-31++G(d,p)//M06-2X/6-311++G(d,p) model chemistry on a model system involving all the groups mentioned above which are essential for the reaction. The model system we used covers a larger radius of the active site, i.e. the adenine bases of both A and P-site tRNAs added to the ribose as well as the whole A2451 adenosine base, compared to what have been calculated previously. The initial geometries of both pre- and post-peptidyl transfer states are based on the key interactions in the structures of Steitz,¹⁰ Ramakrishnan,⁷ and Polacek.¹⁷ We then propose three novel reaction mechanisms with well-defined active groups after a re-optimization of the abovementioned structures. Additional to those concerning our own proposed mechanisms, we have also performed calculations of the previously proposed 4, 6, and 8-membered TS with concerted^34,35,37,38 as well as stepwise^32,33 mechanisms in order to make a more accurate comparison. Based on the new structures which include the adenine bases attached to all three ribose sugars, we have proposed a mechanism which provides a new insight into ribosomal catalysis.

2. Methods

All ab initio calculations were performed at the constant temperature of 298.15 K and pressure of 1 atmosphere using the Gaussian09 suite of programs³⁹ on a 64× dual core Intel Itanium2 64-bits Symmetric Multi-Processing machine.

2.1. Molecular modeling

Recent advances in large scale molecular modeling methods made it possible to study the enzymatic reactions by simulating them as a whole or a large part of their active site. That is if the identity and orientation of the enzymatic catalytic arrays are well-defined. For large and complicated enzymes such as ribosome, neither the identity nor the orientation of the catalytic residues are well-defined and different crystal structures of the active site report different pre-peptidyl transfer structures with significantly different orientation of the active site groups. This results in various possible reaction mechanisms where the large scale modeling would only increase the computer time without aiming to a proper direction. However, based on the idea of theozymes or compuzymes from Tantillo et al.,⁴⁰ we are able to determine the most appropriate arrangement of catalytic arrays in the active site during the peptidyl transfer reaction by modeling a smaller part of the active site.

A theozyme is a theoretical model for studying the biological catalysis. Tantillo et al., defines a theozyme as: “An array of functional groups in a geometry predicted by theory to provide TS stabilization”. This is an indication of theoretically predicted geometry and arrangement of functional groups which can quantitatively validate the individual atomic contribution in TS stabilization which is the role played by all enzymes. The 10⁷-fold enhancement in the rate of peptide bond formation in the ribosome compared to that in solution is a very good example of this role.⁴¹ Figuring out the type of enzymatic functional groups and their role in ribosomal catalysis is the main goal of this study.

2.1.1. Constructing a theozyme. The diagram below indicates the different possibilities of constructing a theozyme based on the available crystallographic structures. For the known catalytic arrays and their orientation around the substrates, constructing the model system is rather straight forward and the possible reaction mechanisms are narrowed down to a few ones only. In such cases, one can model a larger part of the enzyme and use methods such as QM/MM-MD to directly observe the reaction. However, not always the orientation of the catalytic arrays is well-defined, in which case, testing more possible reaction mechanisms are required in order to figure out the most plausible one. The latter can then be used as the model system for large scale QM/MM modeling of the enzymatic reaction. In case of ribosomal catalysis, not only the orientations of the active site residues are not known, they are also not well-defined, leaving the catalytic role and power of the ribosome a dilemma and us at the far right side of the diagram. Though, with the high resolution crystal structures and various computational and theoretical studies on this problem, we are slowly moving towards the left side of the diagram. Even though the A2451 is observed in a close proximity of the reacting substrates, whether it can be considered as a catalytic residue is still on debate.^{16,18,19,34,35,37,38}

To construct an ideal theozyme knowing exactly where we are in the diagram, we have used three different crystallographic structures from three different studies.^7,10,17 These three structures have one thing in common, that is, the presence and interaction of the A2451 rRNA base in the active site. The orientation and type of interaction however varies significantly between them. Therefore, we have proposed three different reaction mechanisms with three different orientations of A2451 (suggested catalytic residue by all three studies) around the substrates. According to the definition of the theozyme, the catalytic array which has a more stabilizing effect on the transition structure could bring us closer to the actual enzymatic catalysis.

2.1.2. Reaction mechanisms. Overall, the reaction mechanisms calculated in this study divide into two main groups: (i) those which follow the previously proposed mechanisms where the adenine bases are absent and only the sugar moieties of the A- and P-site tRNAs have been taken into consideration (Scheme 1), and (ii) three original proposed reaction mechanisms where the participating groups in the reaction are defined by optimizing the active site with presence of three adenosine bases (i.e. the A- and P-site A76 of the tRNAs as well as the A2451 of 23S rRNA of the ribosome) (Scheme 2) as well as a mechanism with three ribose sugars representing the A- and P-site tRNAs and the A2451 rRNA base along with two water molecules in the active site (Scheme 3).

Scheme 2 The two originally proposed mechanisms based on the Ramakrishnan (left) and Polacek (right) structures for the pre- and post-peptidyl transfer mechanisms: TS-6 on the left indicates the concerted 6-membered mechanism with the N3 of A2451 interacting with the α-amine, TS1-6 → TS2-4 on the right indicates the stepwise mechanism where there is no interaction between N3 of A2451 and the α-amine.

Scheme 3 The originally proposed water catalyzed peptide bond formation based on Steitz's structure with a 12-membered proton transfer mechanism in step 1, with 5 protons in flight, and a 4-membered transition structure in the second step of the reaction. In this process, the 3′-OH group of the A2451 acts as a peptidyl shuttle to transfer the growing polypeptide chain from P-site leaving group to the A-site amino acid.

In this model which is fully treated quantum mechanically, the phosphate backbones are substituted by hydrogen atoms which signify the boundary between our theozyme and the rest of the RNA molecule. For the water catalyzed reaction mechanism, the adenine bases are excluded since there is no key interaction of these bases with the active site in that proposed mechanism.

In all the stepwise mechanisms, the additional intermediate is treated as the product of the first TS calculation and the reactant of the second TS calculation.

2.2. Model chemistry

The model chemistry used in this study for all calculations is M06-2X/6-31++G(d,p)//M06-2X/6-311++G(d,p) with the M06-2X theory which was introduced by Truhlar's group^42,43 and the split valence double and triple ζ basis sets, introduced by Pople's group.⁴⁴ The meta hybrid GGA DFT functionals in general have been an improvement over the “standard” exchange–correlation (XC) functionals in treating the electron dispersion forces.⁴⁵ Furthermore, based on the case studies carried out by Truhlar's group, the M06-2X provides more reliable energy barriers compared to the well-known and widely studied B3LYP method.⁴³ Though, in calculating the reaction energies, these functionals are known to give large errors for certain integration grids.⁴⁶ Of particular importance is the SG-1 grid which is a pruned version of (50,194). The default grid in Gaussian is FineGrid with a pruned (75,302) which is used for the geometry optimizations in this study in the M06-2X/6-31++G(d,p) model chemistry. To further increase the accuracy of the reaction energies, we have used the UltraFine grid with the pruned version of (99,590) in a single point frequency calculation on the optimized structures using M06-2X/6-311++G(d,p).

Among the appropriate groups of basis sets used in developing, testing and validating such density functionals is the Pople's 6-31++G(d,p) and 6-311++G(d,p) basis sets,⁴² thus, the basis sets of our choice. The basis function used in this study describes the set of d-functions which are added to the heavy atoms and a set of p-functions which are added to the light atoms (indicated as “d,p”). Additional to the function (polarization function) described above, the diffuse function is also included to our basis function, covering the portion of atomic orbitals which are distant from nuclei. The diffuse function we used, further takes into consideration the light atoms (i.e. hydrogen in this study) additional to the heavy atoms.

2.2.1. Medium. The calculations for the previously proposed mechanisms are performed both in vacuum and implicit water. The water solvent effect is calculated using the SMD continuum solvation model (also known as implicit solvation model) introduced by Truhlar's group.⁴⁷ The observable solvation free energy in this model is divided into two main components. One involves the contribution of the electronic and polarization components of the free energy (ΔG_EP) to the total free energy which is achieved by self-consistent reaction field calculation where the latter is defined by solving the nonhomogeneous Poisson equation for electrostatics in terms of the integral equation formalism polarizable continuum model (IEF-PCM). The other one is related to the free energy change associated with solvent cavitation, dispersion and structure which is raised from short range interactions between the solute and the solvent. Unlike the previously developed SM8 model by the same group⁴⁸ in which the solute molecules are represented as partial atomic charges in a cavity using the GB (generalized Born) approximation,⁴⁹ the SMD model is based on the continuous charge density (hence the name with the “D” letter at the end) of the solute. This makes such model to be applicable in a wider range of electronic structure methods compared to SM8, since the SMD model does not rely on the accurate calculation of partial atomic charges. A few recent studies have also mentioned the aptness of this solvation model with the M06-2X level of theory and reported rather reliable free energies of activation for the ribosomal peptide bond formation.^33,34

2.3. Searching for the TS coordinate

The Quadratic Synchronous Transit (QST) global method from Synchronous Transit Quasi-Newton (STQN) algorithm was used to search for the saddle point. As opposed to the local eigenvector following algorithms which require a quite reasonable initial guess for the input transition structure,⁵⁰ the STQN algorithm generates an initial guess for the first order saddle point by moving along the parabola that connects the reactant and the product.^51,52 It then uses the local eigenvector following method to complete the search. In order to confirm the precision of the TS structure, an intrinsic reaction coordinate (IRC) calculation was carried out after finding the transition structure.

Vibrational eigenmodes were used to calculate the energy of activation from the reactants to the TS. Applying it in the Eyring equation (eqn (1)), we then calculated the rate of reaction.

where k is the rate constant, k_B and h are the Boltzmann's and Planck's constants respectively, ΔG^‡ is the Gibbs free energy of activation, calculated from the difference between the free energy of reactants and that of TS, R is the gas constant and T is the temperature.

3. Results and discussion

3.1. Previously proposed mechanisms

The most theoretically studied mechanisms are the concerted 4-, 6-, and 8-membered mechanisms as well as the stepwise mechanisms (Scheme 1). Many studies suggest that the concerted mechanism is the most favorable one,^35,37,38,53 while some other studies are in a better agreement with stepwise mechanism.^31–33 Table 1 illustrates the activation free energy and reaction rate of several previously studied mechanisms as well as those carried out in this study.

Table 1 Indicating the activation energies and reaction rates for proposed 4-member (TS4), 6-member (TS6), 8-member (TS8) and stepwise (SW) reaction mechanisms of peptide bond formation in this study and other studies. The reaction rate reported in the SW section is the overall rate which is the rate determining step^a

Mechanism	Study	Medium		Method	ΔG^‡	k/s
a The energy unit for ΔG^‡ is kcal mol^−1.
TS4	This study	Vacuum		M06-2X/6-31++G**//M06-2X/6-311++G**	44.47	0.2 × 10^−19
		Implicit water		M06-2X/6-31++G** (SMD)//M06-2X/6-311++G** (SMD)	38.84	0.3 × 10^−15
	Świderek et al.^35	Explicit water		M06-2X/6-31+G(d,p)	24.30	1.4 × 10^−5
	Acosta et al.^34	Implicit water		MP2/6-311+G(d,p)//M06-2X/6-311+G(d,p) (SMD)	30.11	0.8 × 10^−10
	Wang et al.^32	Vacuum		B3LYP/6-31+G(d,p)	43.60	1.0 × 10^−19
	Gindulyte et al.^28	Vacuum		B3LYP/6-31+G(d,p)	35.50	0.9 × 10^−13
TS6	This study	Vacuum		M06-2X/6-31++G**//M06-2X/6-311++G**	45.41	0.4 × 10^−20
		Implicit water		M06-2X/6-31++G** (SMD)//M06-2X/6-311++G** (SMD)	38.92	2.7 × 10^−16
	Świderek et al.^35	Explicit water		M06-2X/6-31+G(d,p)	24.20	1.7 × 10^−5
	Acosta et al.^34	Implicit water		MP2/6-311+G(d,p)//M06-2X/6-311+G(d,p) (SMD)	33.30	3.6 × 10^−12
	Wang et al.^32	Vacuum		B3LYP/6-31+G(d,p)	44.30	3.0 × 10^−20
	Byun & Kang^33	Implicit water		M06-2X/6-31G*//B2PLYP-D/6-311++G(d,p) (SMD)	31.32	1.0 × 10^−10
	Xu et al.^38	Explicit water		B3LYP/6-31G*/MM	29.00	0.5 × 10^−8
	Wallin & Åqvist^37	Vacuum		B3LYP/6-311G**	34.8	2.9 × 10^−13
TS8	This study	Vacuum		M06-2X/6-31++G**//M06-2X/6-311++G**	35.01	0.2 × 10^−12
		Implicit water		M06-2X/6-31++G** (SMD)//M06-2X/6-311++G** (SMD)	30.41	0.5 × 10^−9
	Świderek et al.^35	Explicit water		M06-2X/6-31+G(d,p)	26.80	2.1 × 10^−7
	Acosta et al.^34	Implicit water		MP2/6-311+G(d,p)//M06-2X/6-311+G(d,p) (SMD)	24.57	0.9 × 10^−5
	Byun & Kang^33	Implicit water		M06-2X/6-31G*//B2PLYP-D/6-311++G(d,p) (SMD)	30.40	4.9 × 10^−10
	Xu et al.^38	Explicit water		B3LYP/6-31G*/MM	19.00	1.1 × 10^−1
	Wallin & Åqvist^37	Vacuum		B3LYP/6-311G**	22.3	4.3 × 10^−4
SW	This study	Vacuum	TS1-6	M06-2X/6-31++G**//M06-2X/6-311++G**	9.5	0.2 × 10^−6
			TS2-4		26.78
		Implicit water	TS1-6	M06-2X/6-31++G** (SMD)//M06-2X/6-311++G** (SMD)	13.8	1.3 × 10^−4
			TS2-4		23.01
	Świderek et al.^35	Explicit water	TS1-4	M06-2X/6-31+G(d,p)	33.40	0.9 × 10^−21
			TS2-4		46.40
	Wang et al.^32	Vacuum	TS1-6	B3LYP/6-31+G(d,p)	24.00	2.4 × 10^−5
			TS2-4		13.10
	Byun & Kang^33	Implicit water	TS1-6	M06-2X/6-31G*//B2PLYP-D/6-311++G(d,p) (SMD)	20.84	0.5 × 10^−2
			TS2-6		13.34
	Acosta et al.^34	Implicit water	TS1-6	MP2/6-311+G(d,p)//M06-2X/6-311+G(d,p) (SMD)	16.00	1.1 × 10^−4
			TS2-4		23.09
		Implicit water	TS1-6		16.00	0.6 × 10^−5
			TS2-6		24.77
		Implicit water	TS1-6		7.60	0.6 × 10^2
			TS2-8		13.86

---

Despite the mechanistic difference between the 4- and 6-membered transition structures, the reaction rates are almost similar between these two mechanisms in each study (Table 1). In the study of Acosta et al., the energy barrier for the 6-membered transition structure is even higher than that for the 4-membered mechanism. There are a few exceptions for the favorability of the 6-membered mechanism which can be regarded to the stabilizing effect of the water molecules in their active site model.^35,38 Even though the reaction rates of these studies are high relative to the other 6-membered mechanisms, they are not of particular significance among other mechanisms. For instance, in the study carried out by Xu and collaborators, presence of five water molecules is proposed in the active site, forming hydrogen bond network and stabilizing the transition structure. One of the water molecules acts as a proton shuttle to assist the 2′-OH in transferring its proton to the leaving 3′O in the 8-membered mechanism while no water molecule is involved in the chemical reaction of the 6-membered mechanism.³⁸ This would result in about 10⁷-fold enhancement in the rate of the former mechanism. Despite this big difference, the rate of 6-membered mechanism in their study is relatively high compared to the other 6-membered mechanisms which did not involve water molecules in their active site. Though, it is still rather low compared to the stepwise mechanisms. In fact, the relatively higher reaction rate of the 6-membered mechanisms in the two studies mentioned above is due to the stabilizing effect of the water molecules rather than effect of the P-site 2′-OH group as proton shuttle. This further supports the idea of Acosta et al., suggesting that the substrate assisted catalysis is not working at all and the process occurs through ribosomal catalysis and two water molecules.³⁴

Furthermore, despite the fact that the first step in all three of their stepwise mechanisms correspond to a 6-membered transition structure, the first step of the TS1-6 → TS2-8 occurs in 10⁶-fold higher rate than the first step of the TS1-6 → TS2-4 and TS1-6 → TS2-6 mechanisms which both occur with the same rate. Since the former goes through an 8-membered transition structure which involves a water molecule (Scheme 1B), even though the water does not directly participate in proton shuttle mechanism in the first step, it can be noted that the presence of a water molecule in the active site has a stabilizing effect on the transition structure.³⁴ However, the concerted 8-membered mechanism in their study has a similar rate as TS1-6 → TS2-6 and even lower rate than TS1-6 → TS2-4. Similar effect is also observed in our study where the rate of concerted 8-membered mechanism is 10⁶-fold (in gas phase) and 10⁵-fold (in water) lower than our stepwise mechanism, despite the absence of a water molecule in the latter. This concludes that the presence of a water molecule in the active site is not the sole reason for the rate enhancement and the main reason for the high favorability of the TS1-6 → TS2-8 is in fact the proper positioning of the active groups for facilitating the protonation of the P-site 2′-O group. For the concerted 8-membered mechanisms, this group is being protonated by the attacking nucleophile, and since the latter's activity depends on the peptidyl transfer from P-site 3′-O, there is a delay in the whole mechanism. On the other hand, the peptide bond is already been formed during the first step of the stepwise mechanism and the only thing left is the protonation of the P-site 3′-O during the second step which is followed by the ester bond dissociation and the latter is independent of the peptide bond formation. This indicates the importance of the P-site 3′-O protonation due to its occurrence in the rate determining step. Since the leaving 3′-O is in a closer proximity to the carbonyl OH rather than the adjacent 2′-OH, the TS2-4 takes place with a higher rate compared to the TS2-6.

Table 2 indicates the TS coordinates for the previously proposed mechanisms carried out in this study. In all concerted mechanisms, the ester bond dissociation and peptide bond formation occur rather late except for the stepwise mechanism, which explains the relatively higher rate of the latter, specifically the 8-membered mechanism. Similar structure is also observed in all the above mentioned computational studies, regardless of the type of their mechanism, level of theory or the medium.^{28,32–35,37,38} The only study which favors the concerted mechanism is the one carried out by Swiderek et al.³⁵ The difference in their study with the other ones is the formation of the zwitterionic species which is required before any of the transition structures take place. In that case, protonation of the P-site 3′-O for the 4-membered mechanism and P-site 2′-O for the 6- and 8-membered mechanisms would occur in a higher rate compared to protonation of the 3′-O in the stepwise mechanism since the latter occurs through the carbonyl oxygen, while the first three structures occur through the highly reactive positive nitrogen. It is however mentioned by Acosta et al., that the formation of zwitterionic species is not observed in the ribosome.³⁴

Table 2 The different TS coordinates for 4-member, 6-member, 8-member and stepwise proposed reaction mechanisms of peptide bond formation in this study

Mechanism	Medium	Bond name	Bond type	Distance (Å)
4-member	Vacuum	Namino–Ccarbonyl	Forming	1.53
		Ccarbonyl–3′O	Breaking	2.14
		3′O–Hamino	Forming	1.33
		Namino–Hamino	Breaking	1.18
	Implicit water SMD	Namino–Ccarbonyl	Forming	1.50
		Ccarbonyl–3′O	Breaking	2.17
		3′O–Hamino	Forming	1.57
		Namino–Hamino	Breaking	1.08
6-member	Vacuum	Namino–Ccarbonyl	Forming	1.51
		Ccarbonyl–3′O	Breaking	2.12
		3′O–2′H	Forming	1.50
		2′H–2′O	Breaking	1.04
		2′O–Hamino	Forming	1.58
		Namino–Hamino	Breaking	1.07
	Implicit water SMD	Namino–Ccarbonyl	Forming	1.50
		Ccarbonyl–3′O	Breaking	2.23
		3′O–2′H	Forming	1.63
		2′H–2′O	Breaking	1.00
		2′O–Hamino	Forming	2.38
		Namino–Hamino	Breaking	1.02
8-member	Vacuum	Namino–Ccarbonyl	Forming	1.53
		Ccarbonyl–3′O	Breaking	2.10
		3′O–Hwat	Forming	1.08
		Hwat–Owat	Breaking	1.38
		Owat–2′H	Forming	1.51
		2′H–2′O	Breaking	1.03
		2′O–Hamino	Forming	1.52
		Namino–Hamino	Breaking	1.10
	Implicit water SMD	Namino–Ccarbonyl	Forming	1.52
		Ccarbonyl–3′O	Breaking	2.12
		3′O–Hwat	Forming	1.13
		Hwat–Owat	Breaking	1.28
		Owat–2′H	Forming	1.48
		2′H–2′O	Breaking	1.03
		2′O–Hamino	Forming	1.50
		Namino–Hamino	Breaking	1.10
SW	Vacuum	Step 1
		Namino–Ccarbonyl	Forming	1.58
		Hamino–2′O	Forming	1.49
		Namino–Hamino	Breaking	1.11
		Ocarbonyl–2′H	Forming	1.17
		2′O–2′H	Breaking	1.26
		Step 2
		2′H–3′O	Forming	1.35
		3′O–Ccarbonyl	Breaking	1.99
		Ocarbonyl–2′H	Breaking	1.11
	Implicit water SMD	Step 1
		Namino–Ccarbonyl	Forming	1.56
		Hamino–2′O	Forming	1.59
		Namino–Hamino	Breaking	1.09
		Ocarbonyl–2′H	Forming	1.09
		2′O–2′H	Breaking	1.37
		Step 2
		2′H–3′O	Forming	1.44
		3′O–Ccarbonyl	Breaking	2.05
		Ocarbonyl–2′H	Breaking	1.06

---

The schematic form of the transition structures of the abovementioned mechanisms carried out in this study are indicated in Scheme 4. Aside from the differences mentioned above, these structures are similar with the previous studies in other aspects.

Scheme 4 The most well-known proposed mechanisms for the formation of peptide bond: (a) a 4-membered reaction mechanism where the ester bond dissociation and peptide bond formation take place in a concerted manner and the amino hydrogen (H₁) is directly being transferred to the P-site leaving group (O₁). (b) A stepwise reaction mechanism with a tetrahedral intermediate. The first step involves a proton transfer from the A76 2′-OH to the carbonyl oxygen (H₂), while simultaneously, the proton from attacking amino group is transferred to the 2′-O (H₁). The tetrahedral intermediate is formed with the carbonyl carbon as a central atom linking to both the attacking nucleophile (N₁) and the leaving 3′O (O₁). In the second step, the H₂ is transferred to the leaving 3′O while the ester bond is being broken. (c) A 6-member reaction mechanism where the A76 2′-OH acts as a proton shuttle between the attacking nucleophile (N₁) and the leaving 3′O (O₁). In this mechanism, the 2′O donates its proton to the leaving 3′O while it accepts a proton from attacking nucleophile simultaneously. (d) An 8-membered reaction mechanism where a water molecule is present at the active site. In this mechanism, three protons are in flight at the same time. The H₁ from attaching nucleophile is transferred to the 2′O which donates its proton (H₂) to the adjacent water, causing its proton (H₃) to transfer to the leaving 3′O.

Despite the observed role played by the 2′-OH in the experimental studies, the mechanism in which this group protonates the leaving 3′-O occurs in a very low rate, which is rather surprising. In order to further investigate this dilemma, we have proposed three new mechanisms based on three different structures of pre- and post-peptidyl transfer states. This requires the consideration of a bigger portion of the active site which further increases the complexity of the structure as well as the computational time.

3.2. Original proposed mechanisms in this study

The interactions in the initial geometries are taken from the structures of the previous studies.^7,10,17 Even though these studies highlight important roles played by water molecules as well as the A2451 of the 23S-rRNA in catalysis, they do not agree well in some key interactions of these functional groups with substrates in the pre- and post-peptidyl transfer structures. It however does not conclude the inaccuracy of any of the structures since they might have been observed in different time frames during the peptidyl transfer mechanism. As it was suggested by Sievers et al.,⁴¹ the active site of the ribosome goes through a significant structural change before the formation of transition structure in order to align the substrates in their proper position. This proper alignment results in a reactant complex which leads towards a stabilized transition structure. Our goal here is to figure out which one of these structures has more effect on the TS stabilization which would eventually give us a clearer picture of the actual peptidyl transfer mechanism.

3.2.1. Proposed mechanism based on the structure of Steitz¹⁰. Proposed by Steitz and collaborators,¹⁰ the α-amino group of the A-site is within hydrogen bond distance of the ribosome which is somehow activated to donate its proton to the A76 2′-OH group of the P-site tRNA. They have also suggested the presence of a water molecule to stabilize the oxyanion by acting as a hole during the carbon tetrahedral intermediate formation. Åqvist and collaborators have also suggested the presence of two water molecules in the active site, one participating in the reaction, while the other one is stabilizing the transition structure.³⁷ Based on these observations, we have optimized the pre-peptidyl transfer structure with presence of two water molecules in the active site. After the optimization, the water molecules are positioned adjacent to each other between the 2′ and 3′ OH of the P-site A76, while the 3′-OH group of A2451 is interacting with the α-amine. Even though the N3 of A2451 is been observed in a close proximity of the amino group, its interaction is not confirmed throughout the reaction. Furthermore, it was suggested by Acosta et al., that the elimination of the A2451 nucleobase and leaving only the sugar moiety does not have any significant effect on the rate of peptide bond formation since it is only the 2′-OH of A2451 which is of particular importance for participating in the mechanism.³⁴ Therefore we only used the sugar moieties for all three substrates (i.e. A-site P-site and A2451) where there is no interaction of the N3 with the amino group.

During the TS optimization using the reactant and product complexes, formation of a rather stable intermediate structure was observed, suggesting a two-step mechanism (Fig. 1). It is however not similar to the mechanism suggested by Steitz et al.¹⁰ The only similarity is the interaction of the A2451 3′-OH with the attacking nucleophile, stimulating it to donate its proton to the P-site A76 2′-O. This stimulation results in the transfer of protons in the following order: from A2451 3′-O → α-amine → P-site A76 2′-OH → wat1 → wat2 → P-site A76 3′-O. The growing polypeptide chain is then transferred from the latter to the A2451 3′-O which has just lost its proton to the α-amine. The main proton transfer occurs in this step, resulting in a 12-membered transition structure with five protons in flight (Fig. 1, TS1-12).

Fig. 1 The TS coordinates of a stepwise mechanism carried out in gas phase with two water molecules and A2451 ribose in the active site based on the structure of Steitz et al.¹⁰ (the schematic view of this mechanism can be observed in Scheme 3). The pre- and post-peptidyl transfer structures are based on the structure of Steitz and collaborators.¹⁰ However, some key interactions have changed after re-optimization to the reactant and product.

Since the polypeptide chain is now in a closer distance from the attacking nucleophile, the second step occurs quite fast through a 4-membered transition structure with the A2451 3′-O taking back its proton from the α-amine while donating the polypeptide chain to it. In other words, the 3′O of the A2451 acts as a peptidyl shuttle between the P-site A76 3′-O and the attacking nucleophile (Fig. 1, TS2-4). Since Steitz et al., Acosta et al., and Xu et al., have reported an improvement on the reaction rate due to the presence of two water molecules in the active site and participation of the A2451 in the mechanism, we expect to observe a higher reaction rate compared to the previously proposed mechanisms. Surprisingly, despite abundance of OH groups as both proton shuttle and stabilizing agents, there is not much of improvement in the reaction rate of this mechanism and the first step, the rate determining step, must go through 40.4 kcal mol⁻¹ energy barrier. The only fundamental difference between this mechanism and the previous ones is participation of both water molecules in the reaction as proton shuttle. For instance, Steitz et al., suggested a mechanism which goes through formation of a tetrahedral intermediate.¹⁰ The zwitterion is then neutralized through presence of a water molecule which is coordinated by the universally conserved rRNA bases A2602 and U2584, acting as a hole to stabilize the oxyanion. The second water molecule then interacts with the P-site 2′-O and 3′-O as a proton shuttle while the attacking nitrogen donates its proton to the P-site 2′-O. Similar to the step 1 of our proposed mechanism, this proton transfer occurs in a concerted manner rather than sequential. However, there is only one water molecule involved in the chemical reaction. Another example is the study of Acosta et al., in which the second water molecule avoids artificial stabilization of the reactant complex and stabilizes the transition structure through hydrogen bonding interaction with the active site and not through participating in the reaction, hence, resulting in 6.2 kcal mol⁻¹ decrease in the energy barrier.³⁴ Correspondingly in the study of Xu et al., despite having several water molecules in the active site, there is only one water molecule acting as proton shuttle and participating in the reaction mechanism.³⁸ All these observations indicate that additional to the proton shuttling role, the stabilizing role of the OH groups is also of importance.

The low reaction rate observed in this mechanism which is around 10⁻¹⁶ s⁻¹ (Table 3) suggests that there is more to ribosomal catalytic power than participation of several proton shuttling groups in the active site. In order to figure it out, we have obtained the optimized pre- and post-peptidyl transfer structures from another structure with a different orientation of the catalytic arrays in the active site.

Table 3 Indicating the activation energies, enthalpies and reaction rates for our proposed reaction mechanisms^a

Structure based on	Method	Mechanism	ΔG^‡	k/s
a The energy unit for ΔG^‡ is kcal mol^−1.
Structure of Steitz^10	M06-2X/6-31++G**//M06-2X/6-311++G**	TS1-12	40.40	0.2 × 10^−16
		TS2-4	32.52	0.1 × 10^−10
Structure of Ramakrishnan^7	M06-2X/6-31++G**//M06-2X/6-311++G**	TS-6	40.7	1.3 × 10^−17
Structure of Polacek^17	M06-2X/6-31++G**//M06-2X/6-311++G**	TS1-6	14.48	2.3 × 10^2
		TS2-4	19.48	0.5 × 10^−1

---

3.2.2. Proposed mechanism based on the structure of Ramakrishnan⁷. The structure of Ramakrishnan and collaborators⁷ include the snapshots of both pre- and post-peptidyl transfer states where the tRNAs in both A- and P-sites interact with the intact (70S) ribosome. No water molecule is observed in the active site of this structure which could be the indication of the ribosomal contribution to the TS stabilization by substrate catalysis and water exclusion.^12,41 Similar to the structure of Steitz et al., the N3 of the A2451 interacts with the A-site α-amine. However, instead of the 3′-OH, the 2′-OH of the A2451 residue interacts with the A-site α-amine. After optimization of the pre-peptidyl transfer structure in this study, the N3 of A2451 remains in hydrogen bonding distance of the α-amine, however, the 2′-OH of the A2451 interacts with the 2′-OH of the P-site tRNA instead (Fig. 2 reactant). This sequential change in the interaction of the catalytic arrays could be a clear representation of the series of conformational changes of the catalytic arrays in order to form a shape suitable for the transition structure of the substrates rather than the reactant itself.

Fig. 2 The TS coordinate for the 6-membered transition structure calculated in gas phase from our optimized pre- and post-peptidyl transfer structures based on the structure of Ramakrishnan and collaborators⁷ (the schematic view of this mechanism can be observed in Scheme 2, TS6). The N3 of A2451 is having a hydrogen bond interaction with the hydrogen of the attacking nucleophile in both pre- and post-peptidyl transfer structures.

Our optimized structure could clearly suggest a mechanism in which the 2′-OH of the A2451 transfers a proton from the P-site 2′-OH to its neighboring 3′-O, while the former is being protonated by the attacking nucleophile, resulting in an 8-membered transition structure. The difference between the A2451 residue and a water molecule in an 8-membered transition structure would be the latter's freedom of movement in the region, the movement of which is not observed in the former due to its interaction with the N3 of the P-site A76 through its 3′-H. This interaction is also observed in the post-peptidyl transfer structure (Fig. 2, product) suggesting its rigidness throughout the mechanism. It mainly limits the “proposed” proton shuttle group of A2451 (i.e. the 2′-OH group) from getting closer to the P-site 3′-O (not less than 2.8 Å), resulting in an absolutely not-satisfactory TS optimization. The most probable mechanism based on this optimized structure would be the 6-membered mechanism in which the A2451 does not directly participate in the reaction, though, it stabilizes the 6-membered transition structure by interacting with the hydrogen of both A-site attacking nucleophile and the P-site A76 2′-O through its N3 and 2′-O respectively. The stabilizing effect of this group on the transition structure is indeed observed by comparing the rate of this mechanism with that of previously proposed 6-membered mechanism (with a slightly similar transition structure in terms of the atoms involved in the mechanism) (Fig. 2). The presence of adenine bases of the A- and P-site ribose sugars and the A2451 ribosomal base results in 10³-fold enhancement in the reaction rate of our proposed 6-membered mechanism compared to that of previously proposed 6-membered mechanism carried out in this study in gas phase (Table 1). The interaction of the A2451 N3 with α-amine from one side and that of the A2451 3′-O with the P-site A76 N3 from the other side results in this rate enhancement by acting as an anchor to keep the reacting complexes in their proper position for peptidyl transfer mechanism.

Even though this mechanism advocates the importance of the ribosome in substrate positioning for catalysis, it does not highlight all catalytic power of the ribosome due to its relatively high reaction rate. It suggests that the position of the catalytic array might not be apt for formation of a stabilized transition structure. In order to get another perspective from the position of the catalytic arrays in the active site, we have investigated the structure of Polacek et al.

3.2.3. Proposed mechanism based on the structure of Polacek¹⁷. The presence of the A2451 between the A-site α-amine and P-site A76 2′-OH group in their structure suggests two types of interactions. One is the hydrogen bonding interaction between the 2′-O of A2451 and hydrogen of the α-amine, the other one is the hydrogen bonding interaction between the 2′-H of A2451 and the P-site A76 2′-O. However, due to the large distance between these active groups, the two interactions are less likely to be observed simultaneously and only the occurrence of one is probable during the TS. The occurrence of the former suggests the A2451 2′-OH group to act as proton acceptor from the A-site α-amine, though; it would result in a reduced nucleophilicity of the latter group. On the other hand, the occurrence of the latter indicates the significance of the A2451 2′-OH group on hampering the spontaneous transesterification between the P-site C-3′ and its adjacent C-2′ rather than acting as proton shuttle (proton donor to the P-site A76 2′-OH group). Thus, there is no proper participation of this group in the chemical reaction and the P-site A76 2′-OH group is the main proton shuttle group which deprotonates the attacking nucleophile.

In our optimized structure, the hydrogen bonding interaction of the A2451 2′-H with the N3 of the P-site adenine base positions the A2451 3′-OH group between the P-site 2′-OH and 3′-O. Therefore, instead of the A2451 2′-H⋯2′-O (P-site A76) interaction observed in the structure of Polacek et al., the A2451 3′-O⋯2′-H (P-site A76) interaction takes place. This would lead the A2451 3′-H in a proton shuttling position towards the P-site 3′-O (Fig. 3-reactant), leaving the lone pair of the P-site A76 2′-O free to interact with the hydrogen of the attacking nucleophile. A series of conformational changes in the active site groups is required for the formation of this reactant complex, which if correct, supports the idea of Sievers et al., on the ribosome as an entropy trap.⁴¹ Further investigation of the entropy cost of this active site reorganization requires a larger portion of the PTC and QM-MD method which is not within the scope of this study.

Fig. 3 The TS coordinate for the stepwise mechanism proposed in our study based on the structure of Polacek and collaborators¹⁷ (the schematic view of this mechanism can be observed in Scheme 2. TS1-6 → TS2-4). The calculations are carried out in gas phase. TS1-6 goes through a 6-membered mechanism where the A2451 acts as a proton shuttle by accepting a proton from P-site 2′-O and donating one to the P-site 3′-O, while the 3′-O releases the growing polypeptide chain to the 2′-O. This process occurs in a concerted way with a rather high rate. During the second transition, the growing polypeptide chain is transferred to the α-amine, while the latter group protonates the 2′-O. The A2451 is still interacting through hydrogen bond with the P-site 2′-OH, making it easy for the latter group to release the polypeptide chain to the α-amine.

Despite what have been suggested from the structure of Polacek et al. in terms of the role of A2451 adenosine base, our optimized structure indicates that not only the A2451 does not hamper the spontaneous transesterification between the 2′ and 3′ carbons of the P-site A76 base, it further induces it by shuttling a proton between the two, resulting in a 6-membered transition structure in the first step of the reaction. By protonation of the 3′-O and deprotonation the 2′-O with direct participation of the A2451 3′-O, the growing polypeptide chain is transferred to the 2′-O, forming an intermediate structure with the energy higher than that of reactant and product. According to our calculated reaction rate (Table 3) as well as previous studies,¹⁷ this isomerization reaction occurs so fast that does not allow the attacking nucleophile to protonate the 2′-O. This step is then followed by a 4-membered reaction mechanism between the attacking nucleophile and the P-site 2′-O. The reaction rate in the latter is 10³-fold lower than that in the first step, determining the overall rate of peptide bond formation in this mechanism. It is however only 10³-fold lower than the observed rate in the experiments (i.e. 20 s⁻¹).

As opposed to the previous studies suggesting the interaction of the A2451 2′-OH group with the P-site 2′-OH group, in our study, the A2451 2′-OH group is having a hydrogen bonding interaction with the N3 of the P-site adenine base throughout the whole mechanism. Therefore, we suggest that it would act as an anchor holding the base in a close proximity to the active site group, which in this case, is rather an important role. This anchoring role was played by the 3′-OH group of A2451 in our optimized structure in Section 3.2.2, while the A2451 2′-OH group was more of an inhibitor by holding the P-site 2′-OH back from participating in the reaction as a proton shuttle (Fig. 2).

After formation of the intermediate structure, due to the proton shuttle activity of A2451 3′-OH group, this hydrogen bonding interaction changes to the one where the P-site 3′-H interacts with the A2451 3′-O and the A2451 3′-H interacts with the P-site 2′-O. The former interaction is vital for the second TS stabilization and peptidyl transfer mechanism, resulting in a lower activation barrier.

4. Conclusion

Proton shuttle activity during peptide bond formation which is suggested to stabilize the transition structure is one of the most investigated and proposed activities in previous experimental and computational studies. Now whether this activity is being induced by ribosomal bases (ribosomal catalysis) or by the P-site A76 2′-OH group (substrate assisted catalysis) or simply just a few water molecules in the active site (ribosome as water trap) is still an open question. All these mechanisms however have one thing in common, that is, the presence of a proton shuttling group which puts distance between the attacking nucleophile and the 3′-O. Even though there are many debates on how the proton of the α-amine is being transferred to the leaving 3′-O, it is still not very clear that how the growing polypeptide chain should be transferred to the α-amine considering this distance.

The observed 100-fold reduction in the reaction rate by mutation of P-site A76 2′-OH group,²¹ however small, can be the indication of this group's activity during peptidyl transfer reaction. This activity is not due to the proton shuttling though. The observed high activation energy barrier for the proposed 6-membered mechanism in this study and previous studies is a valid reason for this argument. Even in our own proposed mechanism where the interaction of A2451 2′-H with the P-site 2′-O is expected to stabilize the transition structure, the reaction rate is very low. The abundance of OH groups in the active site also does not reduce the activation barrier as it can be observed in our stepwise mechanism in Section 3.2.1 where the reaction rate of the TS1-12 → TS2-4 is similar to the 6-membered mechanism in Section 3.2.2.

In our third proposed mechanism (Section 3.2.3) on the other hand, we can clearly observe a rather plausible mechanism in which the A2451 3′-O is actively present throughout the whole reaction. It was suggested by Polacek and collaborators that the A2451 2′-O hampers the spontaneous transesterification between the P-site 3′-O and 2′-O. In one point through the mechanism it might be true since we have also observed the very same interaction in Section 3.2.2. However, by moving forward through the mechanism, we observe that the A2451 2′-OH group is no longer interacting with the P-site 2′-OH group and gave its place to its neighboring 3′-OH group. As opposed to the interaction of the 2′-OH group (i.e. A2451 2′-H⋯O (2′) A76), the 3′-OH group is interacting with the P-site 2′-H through its oxygen atom (i.e. A2451 3′-O⋯H (2′) A76). This results in not only hampering the transesterification reaction; it further facilitates it by acting as a proton shuttle between the P-site 3′-O and 2′-O. The fast esterification of the 2′-O does not allow the attacking nucleophile to protonate the 2′-O, though, remains in hydrogen bonding distance with it. In the step two of the mechanism, the A2451 3′-OH group contributes one more time to facilitate the peptidyl transfer from 2′-O to the α-amine by hydrogen bonding with the 2′-O and further stabilizing the transition structure. In fact that is the purpose of receiving a proton from 2′-O during the first step, to stimulate it back with the very same proton to release the polypeptide chain without entropy cost for a reorganization of the active site groups. The term in Sievers' study was referred to as an entropy trap.⁴¹ Throughout all these steps, the A2451 2′-H keeps its interaction with the P-site N3 to hold the active site groups in close proximity of one another for a better TS stabilization.

The results of this study suggest the important role of a few active site groups which their presence was predicted by crystallographic data, though, have not been investigated in such detail in previous computational studies. These groups are the 2′-OH and 3′-OH groups of the 23S rRNA base, A2451, as well as the 2′-OH, 3′-O and the N3 of the P-site A76 adenosine base. Among them, the presence of the A2451 3′-OH throughout the reaction is extremely vital which would suggest further studies on the reaction rate of peptide bond formation where this group is mutated.

A very recent crystal structure of Steitz and collaborators which also highlights the importance of A2451 and three water molecules reveals a previously unseen network of proton wire from the attacking nucleophile all the way to the N terminus of protein L27 in the following order: NH(H)⋯(2′O) P-site A76 (2′H)⋯(2′O)' A2451 (2′H)⋯(O) water 1 (H)⋯(N)H₂ L27. Even though the protein L27 is around 8–10 Å away from the attacking nucleophile, its engagement in proton transfer mechanism could question whether the ribosome is in fact a ribozyme; a newborn riddle from Steitz's recent structure which requires further investigation.⁵⁴

Acknowledgements

This research is supported by High Impact Research MoE Grant UM.C/625/1/HIR/MoE/SC/04 from the Ministry of Education Malaysia and University Malaya Centre for Ionic Liquids (UMCiL). We would also like to thank our information technology officer in Information Technology Centre, University of Malaya, Mr Mohamad Safwan Bin Jusof for his hard work and effort in handling the computer cluster.

References

1. P. B. Moore and T. A. Steitz, RNA J., 2003, 9, 155–159.
2. V. Ramakrishnan, Biochem. Soc. Trans., 2008, 36, 567–574.
3. P. B. Moore, J. Biol., 2009, 8, 8.1–8.10.
4. T. M. Schmeing and V. Ramakrishnan, Nature, 2009, 461, 1234–1242.
5. M. Simonovic and T. A. Steitz, Biochim. Biophys. Acta, 2009, 1789, 612–623.
6. G. Blaha, R. E. Stanley and T. A. Steitz, Science, 2009, 325, 966–970.
7. R. M. Voorhees, A. Weixlbaumer, D. Loakes, A. C. Kelley and V. Ramakrishnan, Nat. Struct. Mol. Biol., 2009, 16, 528–533.
8. S. Uemura, C. E. Aitken, J. Korlach, B. A. Flusberg, S. W. Turner and J. D. Puglisi, Nature, 2010, 464, 1012–1017.
9. H. S. Zaher and R. Green, Nature, 2009, 457, 161–166.
10. T. M. Schmeing, K. S. Huang, D. E. Kitchen, S. A. Strobel and T. A. Steitz, Mol. Cell, 2005, 20, 437–448.
11. T. M. Schmeing, K. S. Huang, S. A. Strobel and T. A. Steitz, Nature, 2005, 438, 520–524.
12. R. M. Voorhees and V. Ramakrishnan, Annu. Rev. Biochem., 2013, 82, 203–236.
13. N. Ban, P. Nissen, J. Hansen, P. B. Moore and T. A. Steitz, Science, 2000, 289, 905–920.
14. P. Nissen, J. Hansen, N. Ban, P. B. Moore and T. A. Steitz, Science, 2000, 289, 920–930.
15. H. F. Noller, V. Hoffarth and L. Zimniak, Science, 1992, 256, 1416–1419.
16. G. W. Muth, L. Ortoleva-Donnelly and S. A. Strobel, Science, 2000, 289, 947–950.
17. K. Lang, M. Erlacher, D. N. Wilson, R. Micura and N. Polacek, Chem. Biol., 2008, 15, 485–492.
18. D. A. Hiller, V. Singh, M. Zhong and S. A. Strobel, Nature, 2011, 476, 236–239.
19. N. Polacek, M. Gaynor, A. Yassin and A. S. Mankin, Nature, 2001, 411, 498–501.
20. J. L. Hansen, T. M. Schmeing, P. B. Moore and T. A. Steitz, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 11670–11675.
21. H. S. Zaher, J. J. Shaw, S. A. Strobel and R. Green, Eur. Mol. Biol. Organ. J., 2011, 30, 2445–2453.
22. Y. Huang and M. Sprinzl, Angew. Chem., 2011, 50, 7287–7289.
23. M. Koch, Y. Huang and M. Sprinzl, Angew. Chem., 2008, 47, 7242–7245.
24. S. Dorner, C. Panuschka, W. Schmid and A. Barta, Nucleic Acids Res., 2003, 31, 6536–6542.
25. M. D. Erlacher, K. Lang, N. Shankaran, B. Wotzel, A. Huttenhofer, R. Micura, A. S. Mankin and N. Polacek, Nucleic Acids Res., 2005, 33, 1618–1627.
26. J. S. Weinger, K. M. Parnell, S. Dorner, R. Green and S. A. Strobel, Nat. Struct. Mol. Biol., 2004, 11, 1101–1106.
27. J. L. Brunelle, J. J. Shaw, E. M. Youngman and R. Green, RNA J., 2008, 14, 1526–1531.
28. A. Gindulyte, A. Bashan, I. Agmon, L. Massa, A. Yonath and J. Karle, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 13327–13332.
29. S. Trobro and J. Aqvist, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 12395–12400.
30. S. Trobro and J. Åqvist, Biochemistry, 2006, 45, 7049–7056.
31. M. A. Rangelov, G. N. Vayssilov, V. M. Yomtova and D. D. Petkov, J. Am. Chem. Soc., 2006, 128, 4964–4965.
32. Q. Wang, J. Gao, Y. Liu and C. Liu, Chem. Phys. Lett., 2010, 501, 113–117.
33. B. J. Byun and Y. K. Kang, Phys. Chem. Chem. Phys., 2013, 15, 14931–14935.
34. C. Acosta-Silva, J. Bertran, V. Branchadell and A. Oliva, J. Am. Chem. Soc., 2012, 134, 5817–5831.
35. K. Świderek, I. A. Tuñón, S. Martí, V. Moliner and J. Bertrán, J. Am. Chem. Soc., 2013, 135, 8708–8719.
36. I. Wohlgemuth, M. Beringer and M. V. Rodnina, Eur. Mol. Biol. Organ. J., 2006, 7, 699–703.
37. G. Wallin and J. Aqvist, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 1888–1893.
38. J. Xu, J. Z. H. Zhang and Y. Xiang, J. Am. Chem. Soc., 2012, 134, 16424–16429.
39. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J.-Y. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. John, A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian, Inc., Wallingford CT, 2010.
40. D. J. Tantillo, J. Chen and K. N. Houk, Curr. Opin. Chem. Biol., 1998, 2, 743–750.
41. A. Sievers, M. Beringer, M. V. Rodnina and R. Wolfenden, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 7897–7901.
42. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
43. Y. Zhao and D. G. Truhlar, Acc. Chem. Res., 2008, 41, 157–167.
44. R. Ditchfield, W. J. Hehre and J. A. Pople, J. Chem. Phys., 1971, 54, 724–728.
45. J. Klimes and A. Michaelides, J. Chem. Phys., 2012, 137, 120901–120912.
46. S. E. Wheeler and K. N. Houk, J. Chem. Theory Comput., 2010, 6, 395–404.
47. A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B, 2009, 113, 6378–6396.
48. A. V. Marenich, R. M. Olson, C. P. Kelly, C. J. Cramer and D. G. Truhlar, J. Chem. Theory Comput., 2007, 3, 2011–2033.
49. W. C. Still, A. Tempczyk, R. C. Hawley and T. Hendrickson, J. Am. Chem. Soc., 1990, 112, 6127–6129.
50. H. B. Schlegel, J. Comput. Chem., 1982, 3, 214–218.
51. C. Peng and H. B. Schlegel, Isr. J. Chem., 1993, 33, 449–454.
52. C. Peng, P. Y. Ayala and H. B. Schlegel, J. Comput. Chem., 1996, 17, 49–56.
53. S. Kuhlenkoetter, W. Wintermeyer and M. V. Rodnina, Nature, 2011, 476, 351–354.
54. Y. S. Polikanov, T. A. Steitz and C. A. Innis, Nat. Struct. Mol. Biol., 2014, 21, 181–793.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra02767e

---