1

Bridging the gap between ribosome structure and

biochemistry by mechanistic computations

Johan Åqvist*, Christoffer Lind, Johan Sund and Göran Wallin

Department of Cell and Molecular Biology, Uppsala University, Biomedical Center, Box

596, SE-751 24 Uppsala, Sweden

*Corresponding author:

e-mail: aqvist@xray.bmc.uu.se

phone: +46 184714109

fax: +46 18530396

*ManuscriptClick here to view linked References

2

Abstract

The wealth of structural and biochemical data now available for protein synthesis on the

ribosome present major new challenges for computational biochemistry. Apart from technical

difficulties in modeling ribosome systems, the complexity of the overall translation cycle with a

multitude of different kinetic steps presents a formidable problem for computational efforts

where we have only seen the beginning. However, a range of methodologies including molecular

dynamics simulations, free energy calculations, molecular docking and quantum chemical

approaches have already been put to work with promising results. In particular, the combined

efforts of structural biology, biochemistry, kinetics and computational modeling can lead

towards a quantitative structure-based description of translation.

Introduction

Computational studies of the translation machinery based on experimental 3D structures actually

date back to the first crystallographic structure of a ribosomal component, namely the C-terminal

domain of the L12 protein [1]. Several molecular dynamics (MD) simulations of this molecule

were published [2-5] showing a harmonic-like collective motion of the -turn- motif that is

today believed to be involved in recruiting translation factors to the ribosome [6]. Otherwise, not

much could be learned from such simulations of isolated components, except that the L12 CTD

turned out to be an unusually stable protein in solution MD simulations with an -carbon RMSD

below 0.7 Å with respect to the crystal structure, commensurable with experimental B-factors

[4]. However, the situation of course changed drastically after the first crystal structures of entire

ribosomal subunits at medium-high resolution were published in 2000 [7-9], that were soon to be

followed by full 70S ribosome structures [10,11]. In principle, the mechanisms of the key events

taking place on the ribosome such as the peptidyl transfer reaction and the mRNA decoding by

tRNAs could now be addressed by quantitative computational methods based on molecular

mechanics/dynamics force field calculations or even quantum mechanical approaches. However,

the size and complexity of these molecular assemblies presented computational biochemists with

major challenges that, in many respects, still remain to be overcome.

3

One obvious problem is the sheer size of any realistic simulation system that includes entire

ribosomes. Another problem is the comparatively low resolution of the crystal structures

available for ribosome, which are of the order of 3 Å. Not only are there conformational

uncertainties about local conformation in structures derived from electron density maps having

resolutions that low, the location of bound water molecules and counterions cannot be reliably

determined. The higher resolution structures of the 50S subunit with transition state (TS) analogs

bound [12,13] are notable exceptions in this regard, reaching as far as 2.2 Å where the level of

detail is considerably higher.

Anyone who has tried to analyze detailed structural and energetic convergence in biomolecular

MD simulations, e.g., by free energy calculations, realizes that inaccurate solvent or ion positions

can completely invalidate the results. In this respect, longer simulations usually do not solve the

problem since waters and ions can be trapped on very long time scales and the likelihood that a

large number of them will find their stable “equilibrium” positions within a reasonable time is

small. This problem would seem to cast some doubt over the reliability of large scale ribosome

MD simulations, e.g., of entire 70S structures in aqueous solution, where the details are likely to

be wrong in many places. However, even though detailed mechanistic predictions at the atomic

level may be beyond the scope, such calculations could still yield relevant information on larger

scale movements and conformational dynamics. A parallel can perhaps be drawn to the

significant insight into many of the ribosome states during translation obtained by the relatively

lower resolution cryo-EM reconstructions [14,15]. An alternative is to break down the translation

cycle into sub-problems that can be computationally investigated with high precision. These

processes can then be analyzed with different types of methods, such as MD, docking, free

energy calculations and chemical reaction simulations, applied to limited structural regions of the

ribosome. The drawback is, of course, that larger scale motions cannot be treated with truncated

systems, so that the two above strategies are rather complementary to each other.

Peptide bond formation

The crystallographic determination of the 50S ribosomal subunit from Haloarcula marismortui

[7] opened up the possibility of computationally analyzing the peptidyl transfer reaction (Figure

4

1a) in detail. Although an early attempt of modeling a peptidyl transfer reaction with

semiempirical quantum chemistry had already been made [16], that study did not include any

elements of the ribosome and obtained energy barriers of over 100 kcal/mol. The first two

computational papers addressing peptidyl transfer in the ribosomal environment, including water

and counter ions, were published in 2005 [17,18]. Both of these studies utilized the empirical

valence bond (EVB) method and obtained reasonable activation free energies and similar

transition states. This demonstrated that the ribosome can catalyze a proton shuttle type of

mechanism where protons are relayed between the A-site amine nucleophile and O3’ leaving

group via the P-site A76 O2’ hydroxyl (it was also shown that A2451 cannot act as a general

base [17,19]). This type of mechanism had also been suggested based on experimental results

with 2’-deoxy or chemically modified substrates [20,21]. These computational studies as well as

kinetic experiments [22,23] further showed that a neighboring hydroxyl group (2’OH), per se,

does not confer any catalytic advantage. Hence, the term “substrate-assisted mechanism” is

probably preferable to “substrate-assisted catalysis”. The surprising entropy effect reported by

Sievers et al. for peptide bond formation with the puromycin substrate analog [22] indicated

that something unusual is going on in the ribosomal peptidyl transferase center (PTC). That is, a

huge difference in activation entropy was observed between the uncatalyzed and catalyzed

reactions [22]. This was also confirmed for the reaction with full length tRNAs where

for kcat (ribosome vs. solution) is about 18 kcal/mol at 25°C [24]. The corresponding prediction

from our computer simulations was 19 kcal/mol [17] but considerably smaller in the work by

Sharma et al. [18]. However, the two studies were in agreement with regard to the relatively

small contribution from alignment and proximity of the substrates (~4 kcal/mol) [17-19], as also

indicated by the small entropy effect on binding [22]. The similarity between the absolute

activation parameters for the small puromycin and large tRNA substrates [24] is also noteworthy

in this respect.

There has been some recent controversy regarding the magnitude of the 2’OH contribution to the

rate of peptide bond formation, or rather the rate reduction caused by its removal [25,26], where

its initially estimated large effect of about 106 [21] now seems to have been revised to about 100-

1000 [26]. However, the key issue is what the actual reaction mechanism is on the native

ribosome, rather than how big an effect removal of the 2’OH causes. If, for example, the 2’OH

5

can be effectively substituted by a water molecule, which is not unlikely, then the viability of

such a perturbed system in peptide bond formation is not inconsistent with the proposed role of

the 2’OH.

Computer simulations [17] also made detailed predictions regarding stereochemistry, product

structure and key water molecules in the PTC, that act as bridges in an H-bond network

involving the substrates and universally conserved bases. This H-bond network was further

identified as the main factor responsible for the favorable activation entropy in that it, unlike

bulk solvent molecules, does not need to reorganize significantly during the reaction. Subsequent

analysis of new high-resolution 50S structures with bound TS analogs [12,13] validated the

MD predictions [19]. Attempts to model the peptidyl transfer reaction with density functional

theory (DFT) methods without including any solvent molecules or parts of the ribosome have

also been made [27-29]. Typical of such minimal gas-phase models is the spurious H-bonding

often occuring between whatever donors and acceptors the model includes. Hence, both H-

bonding between the P-site 2’OH and the A-site aminoacyl-tRNA ester carbonyl [27] and

between the 2’OH and the P-site substrate carbonyl [28,29] were predicted to exist, in compact

four-membered transition states with unrealistically high energies. These types of transition

states, however, are not supported by the crystal structures with TS analogs [12,13] (Figure

1a).

The question of whether the ribosome is really a ribozyme surfaced again after publication of a

2.8 Å resolution Thermus thermophilus 70S complex [11], which showed ribosomal protein L27

protruding into the PTC. As that structure, however, had no visible electron density for the A-site

tRNA the possible role of L27 in peptidyl transfer remained unclear. This problem was

addressed by modeling and MD/EVB simulations of both the tRNA and puromycin reactions

with and without L27 present [30]. The results showed only a minor rate reduction upon deletion

of L27, in agreement with biochemical data [31], but that its N-terminus could affect tRNA

binding as it was predicted to interact with the phosphate groups of A76 and C2452. Very similar

interactions were seen in a subsequent 70S structure with both P-site and A-site tRNAs bound

[32]. The calculations further predicted that the ionization state of the L27 N-terminus has a

marked effect on the puromycin reaction that could reflect its particular pH-dependence with an

6

unknown group (in addition to the attacking amine) ionizing in the physiological pH range

[33,34].

The relevance of the puromycin reaction for peptide bond formation with tRNAs has been the

subject of some debate. That is, apart from the fact that the puromycin reaction has a different

pH-dependence with seemingly two ionizing groups affecting the rate [33,34], the response of

the reaction rate to active site mutations also differs significantly from the native tRNA reaction

[35]. Possible explanations for this behaviour come both from crystal structures with TS

analogs [12,13] and molecular simulations [19,30], where the non-native amide ribose linkage

of puromycin was seen to engage in H-bonding that would not be possible with a natural tRNA

substrate. Puromycin and derivatives thereof have nevertheless long been used to study the

mechanistic details of the peptidyl transfer reaction. A pioneering attempt to measure kinetic

isotope effects (KIEs) for peptide bond formation on the ribosome showed that 15

N substitution

of the attacking (CC-puromycin derivative) amine somewhat surprisingly gave rise to a normal

KIE (> 1), in contrast to nonenzymatic aminolysis reactions in solution [36]. This experiment,

carried out in a very slow 50S assay, indicated that the TS is early rather than late and associated

with NC bond making rather than CO bond breaking.

To shed further light on the transition state for ribosomal peptidyl transfer, quantum chemical ab

initio calculations were carried out on relatively large model systems containing, not only

substrates but also key elements of the PTC, including water molecules [37]. It was found that

unconstrained geometry optimizations converged to structures very close to those obtained

experimentally with TS analogs, and that no early transition states exist on the corresponding

potential energy surface. Both for six- and eight-membered mechanisms, the TS was predicted to

be late and mainly associated with CO bond breaking. The eight-membered mechanism

corresponds to a double proton shuttle, with an extra intervening water molecule that was earlier

predicted from MD simulations [17,19] and found in high-resolution 50S structures [12,13].

This mechanism also gave an activation enthalpy in excellent agreement with experiment [37].

As expected for this type of TS, however, the 15

N isotope effect was predicted to be inverse

rather than normal. A recent study also measured (H/D) kinetic solvent isotope effects (KSIEs)

on the 70S puromycin reaction [38], where the proton inventory suggests three protons in flight

7

in the TS and a corresponding large KSIE of about 8. These results are in very good agreement

with the eight-membered proton shuttle which involves three H-bonds with a predicted KSIE of

7.1 (Wallin & Åqvist, unpublished results).

Strobel and coworkers also recently reported additional heavy atom KIEs for the same 50S

reaction used earlier [39]. These results are still consistent with an early TS, showing a normal

15N KIE that is stronger than the corresponding

18O effect for the leaving 3’-oxygen, but are not

easy to reconcile with the proton inventory studies. They have also attempted to model the

observed KIEs with DFT calculations on minimal vacuum models by fitting observed and

calculated isotope effects in terms of geometry [36,39]. However, transition states cannot really

be determined in this way, but must be obtained by unconstrained search and optimization of

saddle points with vanishing gradients on the relevant potential energy surface. In this sense, no

true TS consistent with an 15

N KIE > 1 has yet been shown to exist. Further, the TS suggested in

[39] does not appear to be on the reaction path to products as it gives no clue of how the O3’

would become protonated. Clearly, these isotope effect experiments are extremely elegant and

carefully executed but it should be remembered that they do not pertain to the native ribosome

reaction. That is, apart from reported differences in pH-dependence and the effect of substituting

the -amino group for a hydroxyl [36, 40], the rate of reaction for the model substrate with only

large ribosomal subunits is over a 1000-fold slower than the native 70S reaction with tRNAs

[24,41,42]. This situation would seem to call for considerable caution in extrapolating detailed

conclusions regarding transition states to the native reaction.

Peptidyl transfer was further probed by measuring Bronsted coefficients for a series of

puromycin derivatives in both 50S and 70S assays [42]. Linear free energy relationships between

nucleophile pKa and the logarithm of the peptidyl transfer rate showed Bronsted coefficients

near zero, i.e. essentially no rate dependence on the nucleophilicity of the attacking amine. This

is usually qualitatively interpreted such that either the nucleophilic attack is not rate-limiting

(implying a late TS associated with leaving group departure), or that there is no charge build-up

on nucleophile in a rate-limiting transition state. While several types of TSs are compatible with

these results one can note that deprotonation of the attacking amine as the NC bond forms

provides a possible explanation, in agreement with the ab initio DFT calculations [37] that do

8

not show any significant charge build-up on the amine in the TS. In fact, these calculations also

predicted little charge development on the substrate “oxyanion” as the reaction proceeds towards

the TS (Fig. 5 of [37]). The interaction of the oxyanion with a (single) conserved water

molecule [12,17,37] is, however, clearly favourable and could therefore be said to

stabilize the oxyanion compared to the (gas-phase like) situation with no water present. Adding

more waters around the oxyanion, approaching the solution situation, was seen to polarize the

oxyanion in the TS and lower the activation enthalpy of the reaction towards the value observed

for the uncatalyzed reaction (presumably with a concomitant decrease in the activation entropy)

[37]. The fact that the activation enthalpy is larger for the ribosome reaction than in solution

[22] is thus consistent with the role of this key water molecule (that, however, by itself does not

provide as much solvation of the oxyanion as bulk solvent). Attempts to estimate the

corresponding interaction by binding of TS mimics [43] also appear to be consistent with this

interpretation. A recent DFT study using a different functional confirms the importance of

including waters and ribosomal hydroxyl groups for approaching realistic energetics [44].

Another proton shuttle variant was suggested there [44] that is high in energy and hard to

reconcile with crystal structures [12,13]. Interestingly, none of the quantum mechanical

calculations published so far supports any early rate-limiting TS.

tRNA accomodation

The issue of whether tRNA accommodation into the ribosomal A-site, after GTP hydrolysis and

release of EF-Tu, actually limits the rate of peptide bond formation has also been debated.

Rodnina and coworkers have reported rates of tRNA accommodation, based on fluorescence

measurements, that are essentially identical to their measured rates of dipeptide formation with

full-length tRNAs [45,46], suggesting that accommodation is rate-limiting for the overall

process. On the other hand, the reported rate constant for dipeptide formation at saturating

conditions (174 s-1

) is about four times faster than the maximal reported rate of accommodation

(40 s-1

), casting some doubt on the latter value [46]. Further evidence in support of a rate-limiting

accommodation step also seemed to be provided by the reported pH-independence of peptidyl

transfer with Phe-tRNAPhe

, suggesting that the expected ionization of the -amino group was

masked by a slower step [46]. However, pH-dependence of the reaction with six different tRNAs

9

was recently measured, instead suggesting that the chemical step is rate-limiting [47]. MD

simulations further showed that the observed pKa-shifts were clearly correlated with the

electrostatic interactions of the -amino groups of the aminoacyl-tRNAs, as would be expected

if their ionization is reflected in the pH-rate profiles. In this respect, the similarity of and

between the full-length tRNA [24] and puromycin reactions (for which there is no

accommodation) [22] also speaks against accommodation being rate-limiting. Sanbonmatsu and

coworkers have addressed the accommodation process by computer simulations in a series of

papers [48,49,50]. Initially they used targeted MD simulations (gradual restraining to target

structure) based on low-resolution structures to elucidate the overall path and proposed the

existence of an accommodation gate comprised of three universally conserved bases [48].

Mutation of C2573 to a purine was suggested to cause inhibition of accommodation but was,

interestingly, found experimentally to affect termination but not peptide bond formation [51].

Later, both so-called structure based simulations (i.e., utilizing Go-like potential energy

functions biased to the target structure with no solvent) and all-atom endpoint MD with explicit

solvent have been used to explore the time-dependence of accommodation [49,50]. These

simulations were also based on low-resolution models and the more recent atomic structures of

A/A and A/T states [11,52,53] would be interesting to use for validation.

Codon reading and GTPase activation

The initial tRNA selection process preceding peptide bond formation has also been addressed in

some computational studies. Sanbonmatsu and Joseph carried out MD simulations of the

decoding center based on the crystal structures of the 30S subunit with mRNA and tRNA

antistem loops, focusing on structural fluctuations of cognate and near-cognate codon-anticodon

pairs and their possible relation to tRNA discrimination [54]. Almlöf et al. used the same type of

system to evaluate binding free energies for six codon-anticodon pairs by MD simulations, in

combination with the linear interaction energy method [55]. They found that the ribosome

environment with the three monitoring bases A1492, A1493 and G530 amplifies the intrinsic

stability differences of codon-anticodon complexes in aqueous solution. With binding energies in

good agreement with experimental data for the first and third codon positions, they also

predicted the strongest discrimination at the second codon position which is consistent with its

10

generally more important role in determining amino acid properties [56,57]. These results were

recently confirmed by free energy perturbation (FEP) simulations based on more recent crystal

structures [58]. That work also used free energy calculations to address the effect of the

cmo5U34 modification in tRNA

Val and proposed alternate mismatched base-pair conformations

from those observed in the crystal structure [59]. The binding energetics obtained, however, does

not seem to explain the observed reading of all four synonymous codons by the tRNA. In

general, computational evaluation of base-paring energetics on the ribosome, including effects of

tRNA modifications, is of tremendous interest as it could allow an energetic interpretation of the

code table by quantifying the codon recognition contribution to the initial tRNA selection fidelity

[57]. We have probably only seen the beginning of such efforts and they are likely to be very

challenging, particularly if also possible keto-enol tautomerization of base-pairs have to be taken

into account [53].

A key step in initial selection is the irreversible hydrolysis of GTP on EF-Tu following ternary

complex binding to the ribosome. The presumably universal process by which the translational

GTPases become activated has been enigmatic, although components such as the universally

conserved ribosomal sarcin-ricin loop (SRL) and invariant His84 (EF-Tu numbering) of the

translational GTPases have been identified as being important for GTPase activation [60]. A

major breakthrough was the recent determination of the crystal structure of an EF-Tu ternary

complex bound to the 70S ribosome [52], showing the key interaction between the SRL and

His84 in the activated conformation. This led Vorhees et al. to suggest a GTP hydrolysis

mechanism where His84 acts a general base in the reaction. An analysis of this crystal structure

by Liljas et al., however, instead led to the suggestion that a substrate-assisted mechanism is

more likely [61]. Such a mechanism was also explored by Warshel and coworkers in a

simulation study that concluded that the histidine contributes to an allosteric effect [62]. While

they did not directly compare the different mechanistic possibilities, recent FEP simulation

results from our lab show the His84 general base mechanism to be strongly disfavoured [Åqvist

et al. unpublished].

Translocation

11

Following peptide bond formation, translocation of the ribosome relative to the mRNA and

bound tRNAs is the step that moves the next codon to be read into the ribosomal A-site. Like

accommodation, translocation involves large conformational changes and is thus difficult to

model in atomic detail. Translocation is rendered unidirectional by the hydrolysis of GTP on EF-

G, which induces a large conformational change in that factor that appears to force its fourth

domain into the A-site on the small subunit, pushing the message and its attached tRNAs across

the ribosome. A 3.6 Å crystal structure of the post-translocational state was obtained by

Ramakrishan and coworkers [63] but the pre-translocational configuration remains unknown.

However, several attempts have been made to model the translocation process computationally

using cryo-EM data [64,65]. MD simulations have also been done on free EF-GGDP in

solution [66]. The combined use of MD fitting and cryo-EM techniques hold considerable

promise for further elucidating the mechanics of translocation.

Termination and peptide release

Termination of translation is catalyzed by class-1 release factor (RF) proteins that decode stop

codons and trigger hydrolysis of the ester bond between the nascent polypeptide and P-site

tRNA. The first low resolution (~6Å) crystal structures of 70S complexes with RF1 and RF2

[67] showed no atomic detail for either of the two key loop motifs involved in stop codon

reading and peptide-tRNA hydrolysis. Trobro and Åqvist used molecular docking of a

heptapeptide, containing the universally conserved GGQ motif responsible for promoting

hydrolysis, to predict how this motif could interact with the PTC [68]. Docking solutions were

then subjected to MD/EVB/FEP simulations in order to gauge their catalytic potency. A single

solution emerged from these calculations, where the glutamine sidechain inserts deep into the A-

site and can coordinate a water molecule for attack on the P-site ester carbon, where again the 2’-

OH group plays a key role [68,69] (Figure 1b). This model explained the effects of several

mutants and the predicted position and conformation of the GGQ motif turned out to be in very

good agreement with subsequently determined medium-resolution (3-3.5Å) structures

[70,71,72]. The N-methylation of the catalytic Gln sidechain is important for peptide release

and its effect could also be rationalized [68,73].

12

The new crystal structures of 70S ribosomes with bound RFs [70,71] also made it possible to

decipher the energetics of stop-codon reading by computational analysis and to clarify the origin

of the high RF binding accuracy. Sund et al. used MD/FEP calculations to predict the relative

binding affinities of cognate and non-cognate termination complexes with RF1 and RF2 [74].

These simulations quantitatively explained the basic principles of decoding in all three codon

positions and revealed that stop codon reading is considerably more complex than the tripeptide

anticodon model [75] and has little to do with tRNA mimicry. Interaction networks were

identified that explain the high RF specificity and these were also found to be reflected by strong

conservation patterns among bacterial RFs (Figure 2a). Simulations of complexes with tRNATrp

further rationalized the observation of the “leaky stop codon” UGA (Figure 2b).

A remarkable feature of translation termination is the fact that prokaryotic and eukaryotic release

factors are completely different, except for the universally conserved GGQ motif involved in

peptidyl-tRNA hydrolysis. The bacterial RF1 and RF2 are specific for the UAA/UAG and

UAA/UGA stop codon pairs, respectively. In eukaryotes the single class-1 release factor eRF1

decodes all three stop codons and contains conserved NIKS and YxCxxxF motifs involved in the

recognition. Using the crystal structure of free eRF1 [76], Vorobjev and Kisselev have attempted

to model the eukaryotic termination complex based on the bacterial 70S complexes with tRNAs

and RFs, including some MD refinement [77,78]. Although these works were not based on the

higher resolution RF complexes and do not address the details of stop codon reading, the overall

structural models seem reasonable and consistent with experimental cross-linking data. The

emergence of the first eukaryotic ribosome structures [79,80] could now open the way for more

detailed modeling of the enigmatic stop codon reading by eRF1.

Outlook

Computational modeling and simulation of the key events in protein synthesis on the ribosome is

a relatively new field and, naturally, lagging behind the enormous progress in structural biology

of the translation system. Nevertheless, this area of computational biochemistry holds

considerable promise for the future, not the least because computations sometimes seem to be the

only way to really translate the multitude of 3D structures into coherent functional mechanisms.

13

In this respect, calculation of energetics is central since it provides the most important link

between structure and function that can be used to evaluate hypotheses based on structures alone.

Further, since the translational machinery occupies a dominant mass fraction of growing cells,

the ribosome and its auxiliary factors are under extraordinary selection pressure for optimal

performance. This means that many of the fundamental binding and rate constants of the system

will be closely connected to bacterial fitness. In this respect, computational techniques have the

potential to anchor kinetics in the structures of the macromolecular complexes and rationalize the

evolutionary design principles of the translation system.

14

References and recommended reading

1. Leijonmarck M, Eriksson S, Liljas A: Crystal structure of a ribosomal component at

2.6 Å resolution. Nature 1980, 286: 824-826.

2. Åqvist J, van Gunsteren WF, Leijonmarck M, Tapia O: A molecular dynamics study of

the C-terminal fragment of the L7/L12 ribsomal protein. Secondary structure

motion in a 150 picosecond trajectory, J Mol Biol 1985, 183: 461-477.

3. Åqvist, Leijonmarck M, Tapia O: A molecular dynamics study of the C-terminal

fragment of the L7/L12 ribsomal protein. II. Effects of intermolecular interactions

on structure and dynamics, Eur Biophys J 1988, 16: 327-339.

4. Åqvist J, Tapia O: Molecular dynamics simulation of the solution structure of the C-

terminal fragment of the L7/L12 ribsomal protein, Biopolymers 1990, 30, 205-209

5. Daggett V, Levitt M: A molecular dynamics simulation of the C-terminal fragment of

the L7/L12 ribosomal protein in solution, Chem Phys 1991, 158: 501-512.

6. Helgstrand M, Mandava, CS, Mulder FAA, Liljas A, Sanyal S, Akke M: The ribosomal

stalk binds to translation factors IF2, EF-Tu, EF-G and RF3 via a conserved region

of the L12 C-terminal domain, J Mol Biol 2007, 365: 468-479.

7. Ban N, Nissen P, Hansen J, Moore PB, Steitz TA: The complete atomic structure of the

large ribosomal subunit at 2.4 angstrom resolution, Science 2000, 289: 905-919.

8. Wimberly BT, Brodersen DE, Clemons WM Jr, Morgan-Warren RJ, Carter AP, Vonrhein

C, Hartsch T, Ramakrishnan V: Structure of the 30S ribosomal subunit, Nature 2000,

407: 327-339.

9. Harms J, Schluenzen F, Zarivach R, Bashan A, Gat S, Agmon I, Bartels H, Franceschi F,

Yonath A: High resolution structure of the large ribosomal subunit from a mesophilic

Eubacterium Cell 2001, 107: 679-688.

10. Yusupov MM, Yusupova GZ, Baucom A, Lieberman K, Earnest TN, Cate JHD, Noller

HF: Crystal structure of the ribosome at 5.5 angstrom resolution, Science 2001, 292:

883-896.

11. Selmer M, Dunham CM, Murphy FV, Weixlbaumer A, Petry S, Kelley AC, Weir JR,

Ramakrishnan V: Structure of the 70S ribosome complexed with mRNA and tRNA,

Science 2006, 313: 1936-1942.

15

12. Schmeing TM, Huang KS, Kitchen DE, Strobel SA, Steitz TA: Structural insights

into the roles of water and the 2’ hydroxyl of the P site tRNA in the peptidyl

transferase reaction, Mol Cell 2005, 20: 437-448.

A very elegant crystallographic work where a series of 50S structures with substrate and

TS analogs are reported at unusually high resolution. Key water molecules likely to be

important for the peptidyl transfer reaction are identified and different mechanistic

possibilities discussed.

13. Schmeing TM, Huang KS, Strobel SA, Steitz TA: An induced-fit mechanism to

promote peptide bond formation and exclude hydrolysis of peptidyl-tRNA, Nature

2005, 438: 520-524.

14. Mitra K, Frank J: Ribosome dynamics: Insights from atomic structure modeling into

cryo-electron microscopy maps, Ann Rev Biophys Biomolec Struct 2006, 35: 299-317.

15. Schuette JC, Murphy FV, Kelley AC, Weir JR, Giesebrecht J, Connell SR, Loerke J,

Mielke T, Zhang W, Penczek PA, Ramakrishnan V, Spahn CMT: GTPase activation of

elongation factor EF-Tu by the ribosome during decoding, EMBO J 2009, 28: 755-

765.

16. Das GK, Bhattacharyya D, Burma DP: A possible mechanism of peptide bond

formation on ribosome without mediation of peptidyl transferase, J Theor Biol 1999,

200: 193-205.

17. Trobro S, Åqvist J: Mechanism of peptide bond synthesis on the ribosome, Proc

Natl Acad Sci USA 2005, 102: 12395-12400.

The first computer simulation of the peptidyl transfer reaction making detailed structural

predictions that were later verified by experiments. The calculations showed the viability

of a 2’OH proton shuttle mechanism and that A2451 could not function as a base. A

preorganized H-bond network was identified as responsible for the unusual entropy effect

of the reaction.

18. Sharma PK, Xiang Y, Kato M, Warshel A: What are the roles of substrate-assisted

catalysis and proximity effects in peptide bond formation by the ribosome?

Biochemistry 2005, 44: 11307-11314.

16

An elegant computational study addressing the catalytic advantage of the A76 2’OH

group both in the uncatalyzed water reaction and in the PTC. It was shown that the

vicinal hydroxyl group confers no rate acceleration in solution but is effective only in the

ribosomal environment.

19. Trobro S, Åqvist J: Analysis of predictions for the catalytic mechanism of ribosomal

peptidyl transfer, Biochemistry 2006, 45: 7049-7056.

20. Dorner S, Panuschka C, Schmid W, Barta A: Mononucleotide derivatives as ribosomal

P-site substrates reveal an important contribution of the 2’-OH to activity, Nucleic

Acids Res 2003, 31: 6536-6442.

21. Weinger JS, Parnell KM, Dorner S, Green R, Strobel SA: Substrate-assisted catalysis of

peptide bond formation by the ribosome, Natur Struct Mol Biol 2004, 11: 1101-1106.

22. Sievers A, Beringer M, Rodnina MV, Wolfenden R: The ribosome as an entropy

trap, Proc Nat. Acad Sci USA 2004, 101: 7897-7901.

A groundbreaking study of the temperature dependence of ribosomal peptidyl transfer in

comparison to an uncatalyzed reference reaction. The results showed an unprecedented

reduction of the entropy contribution to the activation free energy. The similarity between

for kcat/KM and for kcat indicates that substrate alignment and proximity is not a

major part of the entropy effect.

23. Schroeder GK, Wolfenden R: The rate enhancement produced by the ribosome: An

improved model, Biochemistry 2007, 46:4037-4044.

24. Johansson M, Bouakaz E, Lovmar M, Ehrenberg M: The kinetics of ribosomal peptidyl

transfer revisited, Mol Cell 2008, 30:589-598.

25. Huang YW, Sprinzl M: Peptide bond formation on the ribosome: the role of the 2 '-

OH group on the terminal adenosine of peptidyl-tRNA and of the length of nascent

peptide chain, Angew Chem Int Ed 2011, 50: 7287-7289.

26. Zaher HS, Shaw JJ, Strobel SA, Green R: The 2 '-OH group of the peptidyl-tRNA

stabilizes an active conformation of the ribosomal PTC, EMBO J 2011, 30: 2445-

2453.

27. Gindulyte A, Bashan A, Agmon I. Massa L, Yonath A, Karle J: The transition state for

formation of the peptide bond in the ribosome, Proc Natl Acad Sci USA 2006,

103:13327-13332.

17

28. Wang QA, Gao J, Liu YJ, Liu CB: Validating a new proton shuttle reaction pathway

for formation of the peptide bond in ribosomes: A theoretical investigation, Chem

Phys Lett 2010, 501: 113-117.

29. Monajemi H, Zain SM, Abdullah WATW: Computational outlook on the ribosome as

an entropy trap, Comput Theor Chem 2011, 976: 148-152.

30. Trobro S, Åqvist J: Role of ribosomal protein L27 in peptidyl transfer, Biochemistry

2008, 47: 4898-4906.

31. Wower IK, Wower J, Zimmermann RA: Ribosomal protein L27 participates in both

50 S subunit assembly and the peptidyl transferase reaction, J Biol Chem 1998, 273:

19847-19852.

32. Voorhees RM, Weixlbaumer A, Loakes D, Kelley AC, Ramakrishnan V: Insights into

substrate stabilization from snapshots of the peptidyl transferase center of the intact

70S ribosome, Nature Struct Mol Biol 2009, 16: 528-533.

33. Katunin VI, Muth GW, Strobel SA., Wintermeyer W, Rodnina MV: Important

contribution to catalysis of peptide bond formation by a single ionizing group within

the ribosome, Mol. Cell 2002, 10: 339-346.

34. Brunelle JL, Youngman EM, Sharma D, Green R: The interaction between C75 of

tRNA and the A loop of the ribosome stimulates peptidyl transferase activity, RNA

2006, 12: 33-39.

35. Youngman EM, Brunelle JL, Kochaniak AB, Green, R: The active site of the

ribosome is composed of two layers of conserved nucleotides with distinct roles in

peptide bond formation and peptide release, Cell 2004, 117: 589-599.

The role of universally conserved bases in the PTC was investigated by mutagenesis and

kinetic measurements. Mutations in the innermost layer of the active site had large effects

on peptide bond formation with puromycin, but not with aminoacyl-tRNA, and show

defects in peptide release.

36. Seila AC, Okuda K, Núnes S, Seila AF, Strobel SA: Kinetic isotope effect analysis of

the ribosomal peptidyl transferase reaction, Biochemistry 2005, 44: 4018-4027.

37. Wallin G, Åqvist J: The transition state for peptide bond formation reveals the

ribosome as a water trap, Proc Natl Acad Sci USA 2010, 107: 1888-1893.

18

This high-level quantum chemical study could locate and characterize global transition

states for the reaction, described by a molecular model encompassing all the key elements

of the reaction center (~80 atoms), pointing to the feasibility of an eight-membered

“double proton shuttle” mechanism.

38. Kuhlenkoetter S, Wintermeyer W, Rodnina MV: Different substrate-dependent

transition states in the active site of the ribosome, Nature 2011, 476: 351–354.

Kinetic solvent isotope effects show that different numbers of protons are in motion in

the TSs of the peptidyl transfer and termination reactions. Three H-bonds are predicted to

be involved in the former TS but the proposed eight-membered structure seems difficult

to reconcile with the 1VQN and 1VQP analog structures [12,13]. Only one proton is

predicted to be in flight in the TS for hydrolysis which is consistent with the early TS

predicted in [68]. The observed pH-dependence indicates that a group with pKa > 9 is

active in its ionized form, suggesting that deprotonation of either a water molecule, the

A76 2’OH or another group in the PTC contributes to catalysis.

39. Hiller DA, Singh V, Zhong MH, Strobel SA: A two-step chemical mechanism for

ribosome-catalysed peptide bond formation, Nature 2011, 476: 236–239.

An elegant heavy atom KIE analysis of peptide bond formation with a puromycin

derivative on 50S subunits. The proposed early TS of a two-step mechanism has a doubly

protonated A76 2’-oxygen, which would seem highly unfavourable in terms of pKa.

40. Okuda K, Seila AC, Strobel SA: Uncovering the enzymatic pKa of the ribosomal

peptidyl transferase reaction utilizing a fluorinated puromycin derivative,

Biochemistry 2005, 44: 6675-6684.

41. Schmeing TM, Seila AC, Hansen JL, Freeborn B, Soukup JK, Scaringe, SA, Strobel,

SA, Moore, PB, Steitz, TA: A pre-translocational intermediate in protein synthesis

observed in crystals of enzymatically active 50S subunits, Nature Struct Biol 2002, 9:

225-230.

42. Kingery DA, Pfund E, Voorhees RM, Okuda K, Wohlgemuth I, Kitchen DE, Rodnina,

MV, Strobel SA: An uncharged amine in the transition state of the ribosomal

peptidyl transfer reaction, Chem Biol 2008, 15:493-500.

19

43. Carrasco N, Hiller DA, Strobel SA: Minimal transition state charge stabilization of

the oxyanion during peptide bond formation by the ribosome, Biochemistry 2011, 50:

10491-10498.

44. Acosta-Silva C, Bertran J, Branchadell V, Oliva A: Quantum-mechanical study on the

mechanism of peptide bond formation in the ribosome, J Am Chem Soc 2012, 134:

5817-5831.

45. Bieling P, Beringer M, Aido S, Rodnina M.: Peptide bond formation does not involve

acid-base catalysis by ribosomal residues, Nature Struct. Mol. Biol. 2006, 13: 423-428.

46. Wohlgemuth I, Pohl C, Rodnina MV: Optimization of speed and accuracy of decoding

in translation, EMBO J 2010, 29: 3701-3709.

47. Johansson M, Ieong KW, Trobro S, Strazewski P, Åqvist J, Pavlov MY, Ehrenberg M:

pH-sensitivity of the ribosomal peptidyl transfer reaction dependent on the identity

of the A-site aminoacyl-tRNA, Proc Natl Acad Sci USA 2011, 108: 79-84.

The first demonstration of the pH-dependence of the reaction with native substrates

showed downshifted pKa values of the aminoacyl-tRNAs that correlated well with those

predicted from MD simulations.

48. Sanbonmatsu KY, Joseph S, Tung CS: Simulating movement of tRNA into the ribosome

during decoding, Proc Natl Acad Sci USA 2005, 102: 15854-15859.

This large all-atom MD simulation of a solvated 70S ribosome used targeted MD to drive the

structure between models of the A/T (unaccomodated) and A/A (accommodated) states in

order to elucidate the conformational pathway.

49. Whitford PC, Geggier P, Altman RB, Blanchard SC, Onuchic JN, Sanbonmatsu KY:

Accommodation of aminoacyl-tRNA into the ribosome involves reversible

excursions along multiple pathways, RNA 2010, 16: 1196-1204.

50. Whitford PC, Onuchic JN, Sanbonmatsu KY: Connecting energy landscapes with

experimental rates for aminoacyl-tRNA accommodation in the ribosome, J Am Chem

Soc 2010, 132: 13170-13171.

51. Burakovsky DE, Sergiev PV, Steblyanko MA, Kubarenko AV, Konevega AL, Bogdanov

AA, Rodnina MV, Dontsova OA: Mutations at the accommodation gate of the

ribosome impair RF2-dependent translation termination, RNA 2010, 16: 1848-1853.

20

52. Voorhees RM, Schmeing TM, Kelley AC, Ramakrishnan V: The mechanism for

activation of GTP hydrolysis on the ribosome, Science 2010, 330: 835-838.

Groundbreaking crystal structure of an EF-Tu ternary complex, with a non-hydrolyzable

GTP analog, bound to the ribosome. The structure shows how the sarcin-ricin loop on the

ribosome positions the universally conserved histidine of the translational GTPases,

thereby activating EF-Tu for GTP hydrolysis.

53. Demeshkina N, Jenner L, Westhof E, Yusupov M, Yusupova G: A new understanding

of the decoding principle on the ribosome, Nature 2012, 484: 256-259.

54. Sanbonmatsu KY, Joseph S: Understanding discrimination by the ribosome: Stability

testing and groove measurement of codon-anticodon pairs, J Mol Biol 2003, 328: 33-

47.

55. Almlöf M, Andér M, Åqvist J: Energetics of codon-anticodon recognition on the small

ribosomal subunit, Biochemistry 2007, 46: 200-209.

56. Crick FH: The origin of the genetic code, J Mol Biol 1968, 38: 367-379.

57. Johansson M, Zhang JJ, Ehrenberg M: Genetic code translation displays a linear

trade-off between efficiency and accuracy of tRNA selection, Proc Natl Acad Sci USA

2012, 109: 131-136.

An extraordinary detailed kinetic analysis of the maximal accuracies that can be attained

in tRNA selection, showing differential accuracies at the different codon positions.

58. Allnér O, Nilsson L: Nucleotide modifications and tRNA anticodon-mRNA codon

interactions on the ribosome, RNA 2011, 17: 2177-2188.

Extensive MD free energy calculations were carried out to explore the effects of tRNA

modifications and the reading of the four synonymous codons by tRNAVal

. Many

interesting points are brought up in this study, which illustrates that the energetics of

codon reading is far from trivial.

59. Weixlbaumer A, Murphy FV, Dziergowska A, Malkiewicz A, Vendeix FAP, Agris PF,

Ramakrishnan V: Mechanism for expanding the decoding capacity of transfer RNAs

by modification of uridines, Nature Struct Mol Biol 2007, 14: 498-502.

60. Rodnina MV: Visualizing the protein synthesis machinery: New focus on the

translational GTPase elongation factor Tu, Proc Natl Acad Sci USA 2009, 106: 969-

970.

21

61. Liljas A, Ehrenberg M, Åqvist J: Comment on "the mechanism for activation of GTP

hydrolysis on the ribosome", Science 2011, 333: 37.

62. Adamczyk AJ, Warshel A: Converting structural information into an allosteric-

energy-based picture for elongation factor Tu activation by the ribosome, Proc Natl

Acad Sci USA 2011, 108: 9827-9832.

63. Gao YG, Selmer M. Dunham CM. Weixlbaumer A. Kelley AC, Ramakrishnan V: The

structure of the ribosome with elongation factor G trapped in the

posttranslocational state, Science 2009, 326: 694-699.

64. Ratje AH, Loerke J, Mikolajka A, Brunner M, Hildebrand PW, Starosta AL, Donhofer,

A, Connell, SR, Fucini P, Mielke T, Whitford PC, Onuchic JN, Yu YN, Sanbonmatsu

KY, Hartmann RK, Penczek PA, Wilson DN, Spahn CMT: Head swivel on the

ribosome facilitates translocation by means of intra-subunit tRNA hybrid sites,

Nature 2010, 468: 713-716.

An elegant analysis of structural substates from cryo-EM where pre- and post-

translocational intermediates are used to interpret the mechanics of the process.

65. Whitford PC, Ahmed A, Yu YA, Hennelly SP, Tama F, Spahn, CMT, Onuchic JN,

Sanbonmatsu KY: Excited states of ribosome translocation revealed through

integrative molecular modeling, Proc Natl Acad Sci USA 2011, 108: 18943-18948.

66. Li W, Trabuco LG, Schulten K, Frank J: Molecular dynamics of EF-G during

translocation, Proteins 2011, 79: 1478-1486.

67. Petry S, Brodersen DE, Murphy FV, Dunham CM, Selmer M, Tarry MJ, Kelley AC,

Ramakrishnan,V: Crystal structures of the ribosome in complex with release factors

RF1 and RF2 bound to a cognate stop codon, Cell 2005, 123: 1255-1266.

68. Trobro S, Åqvist J: A model for how ribosomal release factors induce peptidyl-

tRNA cleavage in termination of protein synthesis, Mol. Cell 2007, 27: 758-766.

A surprisingly accurate structural prediction from computer simulations of how the

catalytic GGQ motif inserts into the ribosomal A-site to promote the termination reaction.

69. Brunelle JL, Shaw JJ, Youngman EM, Green R: Peptide Release on the ribosome

depends critically on the 2’ OH of the peptidyl-tRNA substrate, RNA 2008, 14: 1526-

1531.

22

70. Laurberg M, Asahara H, Korostelev A, Zhu J, Trakhanov S, Noller HF: Structural

basis for translation termination on the 70S ribosome, Nature 2008, 454: 852-857.

The first medium resolution crystal structure of a release factor bound to the ribosome

showed the positioning of the GGQ loop and the unexpected mRNA stop-codon

conformation in complex with RF1.

71. Weixlbaumer A, Jin H, Neubauer C, Voorhees RM, Petry S, Kelley AC, Ramakrishnan,

V: Insights into translational termination from the structure of RF2 bound to the

ribosome, Science 2008, 322: 953-956.

This medium resolution structure of a 70S complex with RF2 gave valuable clues as to

how RF2 can read both A and G in the second stop-codon position.

72. Trobro S, Åqvist J: Mechanism of the translation termination reaction on the

ribosome, Biochemistry 2009, 48: 11296-11303.

73. Andér M, Åqvist J: Does glutamine methylation affect the intrinsic conformation of

the universally conserved GGQ motif in ribosomal release factors? Biochemistry

2009, 48: 3483-3489.

74. Sund J, Andér M, Åqvist, J: Principles of stop-codon reading on the ribosome,

Nature 2010, 465: 947-950.

These pioneering free energy perturbation calculations gave a quantitative picture of the

structural mechanisms underlying stop-codon reading and could also explain sequence

conservation patterns among bacterial RFs.

75. Ito K, Uno M, Nakamura Y: A tripeptide 'anticodon' deciphers stop codons in

messenger RNA, Nature 2000, 403: 680-684.

An impressive genetic engineering approach was used to identify characteristic RF

tripeptide motifs that later were verified to indeed be in the immediate vicinity of the

stop-codons.

76. Song H, Mugnier P, Das AK, Webb HM, Evans DR, Tuite MF, Hemmings BA, Barford

D: The crystal structure of human eukaryotic release factor eRF1-mechanism of

stop codon recognition and peptidyl-tRNA hydrolysis, Cell 2000, 100: 311-321.

77. Vorobjev YN, Kisselev LL: Model of the structure of the eukaryotic ribosomal

translation termination complex, Mol Biol 2007, 41: 93-101.

23

78. Vorobjev YN, Kisselev LL: Modeling of the positioning of eRF1 and the mRNA

stop codon explains the proximity of the eRF1 C domain to the stop codon in the

ribosomal complex, Mol Biol 2008, 42: 302-311.

A very challenging docking problem was attacked by computational modeling and MD

simulation guided by experimental cross-linking data and bacterial RF complex

structures. Although no detailed model for stop-codon reading was presented the overall

positioning of the eukaryotic RF appears largely consistent with experiments.

79. Ben-Shem A, de Loubresse NG, Melnikov S, Jenner L, Yusupova G, Yusupov M: The

structure of the eukaryotic ribosome at 3.0 angstrom resolution, Science 2011, 334:

1524-1529.

80. Rabl J, Leibundgut M, Ataide SF, Haag A, Ban N: Crystal structure of the eukaryotic

40S ribosomal subunit in complex with initiation factor 1, Science 2011, 331: 730-

736.

81. Schmeing TM, Voorhees RM, Kelley AC, Ramakrishnan V: How mutations in tRNA

distant from the anticodon affect the fidelity of decoding, Nature Struct Mol Biol

2011, 18: 432-436.

24

Figures

Illustration of the two reactions taking place in the ribosomal peptidyl transferase center: (a)

peptide bond formation and (b) peptidyl-tRNA hydrolysis. The depicted geometry of the peptidyl

transfer reaction (a) corresponds to that observed in the high-resolution 1VQP crystal structure

[12], while the model for the RF catalyzed hydrolysis reaction is based on computer

simulations [68,72] and medium-resolution crystallographic complexes of RFs bound to the

ribosome [70,71]. Water molecules are depicted as red spheres and possible H-bond

interactions are indicated by dashed lines.

25

(a) Prokaryotic RFs show a strong conservation of the residues involved in the stop-codon

reading mechanisms. Residues involved in the first, second and third position reading

mechanisms are coloured in green, cyan and magenta, respectively, while the characteristic PxT

and SPF tripeptide motifs [75] are in yellow. The inset shows the detailed interaction network

in RF2 that allows dual reading of A and G in the second codon position. (b) Comparison of the

predicted near-cognate interaction between the UGA stop-codon and tRNATrp

from MD

simulations [74] (green carbons) with the recent crystallographic structure of the same type of

complex [81]. The third position A–C mismatch was predicted not to involve a very strong

energetic penalty.

Highlights

Computational modeling and simulation are powerful tools for deriving functional

mechanisms from ribosome structures.

Evaluation of energies is most important for linking structure and function.

Methods ranging from quantum chemistry to MD simulation can be used to explore the

steps of translation.

Several predictions about ribosome function from computations have been confirmed

experimentally.

Quantitative structure-based descriptions of protein synthesis can be obtained by combining

computations with experimental kinetics.

*Highlights