Structural Insights into tRNA Dynamics on the Ribosome

Abstract

High-resolution structures at different stages, as well as biochemical, single molecule and computational approaches have highlighted the elasticity of tRNA molecules when bound to the ribosome. It is well acknowledged that the inherent structural flexibility of the tRNA lies at the heart of the protein synthesis process. Here, we review the recent advances and describe considerations that the conformational changes of the tRNA molecules offer about the mechanisms grounded in translation.

Figure 1

Tertiary structure and computational analysis of tRNA’s structural plasticity. (a) Color-coded B factor information, which can be described as flexibility estimation, is displayed for the phosphorus atoms of tRNA^Asp X-ray structure (PDB 2TRA) [8]; (b) Color-coded flexibility prediction by constraint counting for the same tRNA species, in which overconstrained regions are indicated in blue, rigid regions in green, and flexible regions in red. (Figure adapted from [9], Copyright 2008 Elsevier).

Figure 2

A gallery of tRNA atomic structures, isolated or in complex with EF-Tu, aminoacyl tRNA synthetase (RS) or the ribosome. (a) PDB 4TRA [8]; (b) PDB 2TRA [8]. (c) PDB 1TTT [11]; (d) PDB 1ASY [12]; (e,f) PDB 4V51 [13]. The different regions of the tertiary structure of tRNA are color-coded. Variability is mainly manifested on the distal parts (i.e., ASL and CCA end), as well as in the relative angle between the anticodon and acceptor stems. Plasticity is also readily observed in the D- and T-loop regions. Measured distances between some phosphorus atoms (in Å) are shown on the left side of the structures. (Figure adapted from [14], Copyright 2008 Nature Publishing Group).

Figure 3

Comparison of tRNA structures from different eukaryotic initiation complexes. The shown tRNA structures are the following: eP/I’ (PDB 3J81 [97]), eP/I (archaeal Met-tRNAiMet [102] fitted as rigid body in EMD 5658 [95]) and P/I (PDB 4KZZ [94]), which are compared to classical P/P (PDB 4V51 [13]) and hybrid P/E (PDB 4V9H [103]) tRNAs. In the view at the right side, eIF2a is also shown. (Figure adapted from [97], Copyright 2014 Elsevier).

Figure 4

The A/T state of tRNA binding. (a) Comparison of the A/T tRNA (PDB 4V5G [129]) with the A/A tRNA that mimics the accommodated state (PDB 4V5D [133]); (b) Major changes are observed in the ASL region, as the helical twist is reduced after base pair 30:40. Also, the helical strands split at nucleotides 25 to 45 and 26 to 44. The well-known 27:43 base pair mutation (highlighted in pink) results in an error-prone phenotype [134], as it probably weakens the ASL stem and thus, facilitates the strand separation that leads to the final distorted form of the incoming tRNA; (c) Comparison of the A/T tRNA (PDB 4V5G [129]) with the EF-Tu bound form (TC) in the absence of the ribosome (PDB 1TTT [11]); (d) A gallery of mutations/manipulations in the tRNA molecule that affect the accuracy of decoding are highlighted: Cross-linking of nucleotides 8 and 13 [135], and mutations at 9:12:23 [136] and 24:11 [136,137,138]. Note that contrary to previous assumptions, subsequent structural analysis by Schmeing and coworkers [130] discarded the idea that the so-called Hirsh mutation influences the flexibility/deformability of the tRNA. This framework also puts into context previous kinetic experiments: see proflavin insertions at positions 16 and 17, e.g., see [20]. (Figure adapted from [129], Copyright 2009 Am. Assoc. Adv. Sci.).

Figure 5

Graphic representation of the sequential stages during tRNA incorporation. (a) Distortion of the aa-tRNA during the initial, reversible codon sampling stage; (b) After codon recognition, the tRNA is stabilized in an A/T form. The distorted A/T state is known to be critical for the subsequent activation of GTP hydrolysis on the ribosome [138,139]. Accompanying global changes of the 30S subunit (mostly comprising the shoulder region in 16S rRNA) lead to the conformational change of domain II’s β loop; (c) which causes the distortion of the acceptor end; (d) Concomitant disruption of the contacts with switch I region of EF-Tu relocate the catalytic residue His84 [140] into the proper orientation. A2662 of the sarcin-ricin loop (SRL) of the 23S rRNA is also involved. Note that some details about the GTPase activation mechanism are still being discussed (see [131,141,142] for more details); Release of Pi is coupled to the (e) disordering of the switch I loop and (f) subsequent conformational change of EF-Tu [143]; (g) Finally, EF-Tu is released and the tRNA molecule relaxes its conformation to accommodate in the vacant A/A site. All in all, the different decoding stages are governed by the dynamic nature of the tRNA molecule within the A site [27]. Note that as the sequence and structure of each aa-tRNA is tuned according to the nature of the amino acid they carry, as well as to the codon-anticodon strength, the different cognate tRNA species display unvarying decoding properties (i.e., uniform rates of acceptance by the ribosome) [144,145]. (Figure adapted from [115], Copyright 2013 Annual Reviews).

Figure 6

Comparison of mammalian and bacterial ternary complexes. CryoEM maps of eukaryotic (a) codon sampling and (b) GTPase activation steps. Fitted archaeal aEF1α (PDB 3VMF [148]) and A/T tRNAs (ASL part from PDB 4V5L [131], body from PDB 1TTT [11]) are shown in red and pink ribbons, respectively; (c) Superposition of codon sampling (grey) and GTPase activation state ternary complexes after alignment of 40S subunits. Superposition of bacterial GDPCP stalled ternary complex, in grey (PDB 4V5L [131]) with eukaryotic (d) codon sampling and (e) GTPase activation state ternary complexes after alignment of 18S/16S rRNA’s conserved parts (the electron density corresponding to the 40S subunit surface is shown in white); (e) The relative orientations of factor and tRNA diverge in the two states of the eukaryotic decoding complex, as well as in their bacterial counterpart, which leads to a different interaction mode with the ribosome. The largest differences are observed in the tRNA elbow region due to interactions with the SRL and H89 from the large subunit, interactions not observed in bacteria. Measured distances (in Å) between the ternary complexes shown in panels (d,e) are color coded and shown on the right side of panel. (Figure adapted from [146], Copyright 2014 Elsevier).

Figure 7

Comparison of translocation intermediates obtained by cryoEM and particle sorting. (a–f) Maximum likelihood methods were applied to a dataset of pre-translocational complexes in the absence of EF-G, obtaining an ensemble of substates that encompass a large conformational space. Out of the initial six classes, one subset (class 4) was further subdivided due to residual heterogeneity (class 1 is not shown as it was deemed to be artifactual due to bias in particle orientations). Class 2 (a) and class 4A (c) correspond to the classic A/A and P/P tRNA configuration, while class 3 (b) corresponds to ribosomes bearing a single tRNA configuration (P/P). Classes 5 (e) and 6 (f) are similar to each other, and to the previously characterized A/P and P/E hybrid configurations [160,161]. Class 4B (d) represents a new intermediate in which the P-site tRNA’s elbow is halfway through its transition from classic to hybrid state, a configuration that is coupled to an intermediate intersubunit rotation and L1 stalk movement; (g) Stereo-view of the superposition of A- and P-site tRNAs from class 2 (classic state tRNAs, in magenta and green), class 4B (new intermediates, in olive) and class 6 (hybrid state tRNAs, in gray), aligned with respect to the 70S ribosomes. (Figure adapted from [168], Copyright 2012 Proc. Natl. Acad. Sci. USA).

Figure 8

EF-G mediated mRNA-tRNA translocation intermediate. (a) Classic state pre-translocational complex (PDB 4V6F [188]); (b) Translocation intermediate in the presence of EF-G, bearing chimeric hybrid ap/ap and pe/E state tRNAs (lower-case letters imply that the corresponding tRNA is bound in a chimeric configuration, i.e., halfway through the path from one canonical site to the next). Comparison of tRNA positions in classic and intermediate states after alignment of (c) 50S subunits; (d) 30S subunit bodies or (e) 30S subunit heads. (Figure adapted from [189], Copyright 2014 Am. Assoc. Adv. Sci.).

Figure 9

Sensing amino acid starvation. (a–d) 3D cryo-EM structure of the 70S–RelA complex. The molecular structures fitted to the density are shown in ribbons; (e) Stereo-view of the superposition of the A/T state tRNA in the presence of EF-Tu (PDB 4V5G [129]), the classic state A/A tRNA (PDB 4V5D [133]) and the A/T-like structure in the presence of RelA. (Figure adapted from [207], Copyright 2013 John Wiley & Sons).