Initiation of mRNA translation in bacteria: structural and dynamic aspects

Abstract

Initiation of mRNA translation is a major checkpoint for regulating level and fidelity of protein synthesis. Being rate limiting in protein synthesis, translation initiation also represents the target of many post-transcriptional mechanisms regulating gene expression. The process begins with the formation of an unstable 30S pre-initiation complex (30S pre-IC) containing initiation factors (IFs) IF1, IF2 and IF3, the translation initiation region of an mRNA and initiator fMet-tRNA whose codon and anticodon pair in the P-site following a first-order rearrangement of the 30S pre-IC produces a locked 30S initiation complex (30SIC); this is docked by the 50S subunit to form a 70S complex that, following several conformational changes, positional readjustments of its ligands and ejection of the IFs, becomes a 70S initiation complex productive in initiation dipeptide formation. The first EF-G-dependent translocation marks the beginning of the elongation phase of translation. Here, we review structural, mechanistic and dynamical aspects of this process.

Keywords: GTP; Protein synthesis; Translation initiation factors; fMet-tRNA; mRNA initiation region.

Fig. 1

Distribution of SD-containing and SD-lacking mRNAs in the bacterial kingdom and deficit of RNA secondary structure near the start codon. a Normalized distributions of energies assessed for hybridizations between the anti-SD of 16S rRNAs and the −22 to −2 sequences of 5′ UTRs of α-proteobacterial, γ-proteo-bacterical, cyanobacterial and plastid genes. Four major peaks at −5.9, −3.6, −1.4 and +1.5 kcal mol⁻¹ are visible in all taxonomic groups. They correspond, from left to right, to: (1) mRNAs with SD sequence AGGAG, (2) mRNAs with SD sequence GAGG, AGGA or GGAG, (3) mRNAs with short SD-like sequences (AGG, GAG or GGA), which may engage in SD-type interactions with the 3′ end of 16S rRNA and (4) mRNAs without SD sequences. b Predicted amount of RNA secondary structure around the start codon in α-proteobacteria, γ-proteobacteria, cyanobacteria, plant, metazoan and fungal mitochondria, and plastids. The line shows the running mean Minimum Free Energy (standard error of the mean is indicated by the shaded area) of 5000 genes with (blue) and without (green) an SD sequence, the difference in their MFE upstream and downstream of the initiation region (0 = first start codon nucleotide) in the three bacterial groups being largely due to differences in AT-content between genomes. In metazoan mitochondria, most transcripts are leaderless and lack a 5′ UTR so that the minimum free energy peak is shifted into the coding region. Reproduced with permission from [18]

Fig. 2

SD helix and mRNA movements on the 30S ribosomal subunit during translation initiation. a Location of the SD–aSD duplex (yellow) with respect to the 16S rRNA (light blue) within a 70SIC. The SD helix contacts h23a, h26 and h28 (dark blue); b close-up of the interaction between the SD helix (yellow) and h23a, h26 and h28 (cyan) and ribosomal protein S18 (dark blue). The bulged U723 that interacts with minor groove of the C1539·G10 bp and A1534 that binds to h28 in the 30S neck are indicated. The position of P-site-bound tRNA (orange) is also shown (reproduced with permission from [52]). Initiation factors-dependent and fMet-tRNA-dependent mRNA shift from “stand-by” to “P-decoding” site on the 30S subunit as evidenced by c site-directed cross-linking (redrawn from [60]) and d X-ray crystallography (reproduced from [55] with permission from Elsevier)

Fig. 3

Unique characteristics of the initiator tRNA_fMet anticodon stem loop (ASL). Comparison between the ASL of a E. coli initiator tRNA_fMet and b elongator tRNA_Phe. The ASL of initiator tRNA_fMet contains c a peculiar Cm32·A38 wobble base pair and d the A37·G29·C41 base triple. The anticodon bases undergo different stacking interactions when the tRNA is e free, f transformylase-bound or g P-site-bound (reproduced from [82] with permission from Oxford University Press)

Fig. 4

Structures of the initiation factors IF1, IF2 IF3. Structures of: a E. coli IF1 as determined by NMR spectroscopy (PDB 1AH9) [83]; b the N-terminal 2–50 residues of E. coli IF2 as determined by NMR spectroscopy (PDB 1ND9) [103]; crystallographic structures [107] of T. thermophilus c IF2-G2·GTP and d IF2-G2·GDP. The guanine nucleotides binding elements P-loop/G1, G2, G3 and G4 (cyan), switch I and switch II (green) are indicated; residue His130, α-helices H1, H4 and H6 as well as the position of domain G3 are also indicated (reproduced from [107]; e structure of the apo form of G. stearothermophilus IF2-G2 as determined by NMR spectroscopy (PDB 2LKC) [93]); f structure of G. stearothermophilus IF2-C1 as determined by NMR spectroscopy (PDB 1Z9B) [105]; g structure of G. stearothermophilus IF2-C2 as determined by NMR spectroscopy (PDB 2LKC) [106]; crystallographic structure of E. coli IF3; h N-terminal domain (PDB 1TIF) [113] and i C-terminal domain (PDB 1TIG) [113]. The N-terminus and C-terminus of the structures are indicated with N and C, respectively, the α-helices and the β-strands are shown in green and blue, respectively, and indicated with H and B letters followed by numbers, as appropriate. With the exception of c and d, molecular images were generated from PDB data using the UCSF Chimera package (http://www.cgl.ucsf.edu/chimera developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco (supported by NIGMS P41-GM103311)

Fig. 5

Domain composition, structure and ribosomal localization in 30SIC of translation initiation factor IF2. a Scheme illustrating the structural/functional domains constituting G. stearothermophilus IF2, and E. coli IF2α and E. coli IF2β. Domains G1 (light gray), G2 (green), G3 (yellow), C1 (orange) and C2 (red) are fairly conserved, whereas size and sequence of the N-terminal part of the molecule are not conserved although the N-terminal domain of both G. stearothermophilus and E. coli shares the property of anchoring IF2 to the ribosome [91, 92]. The number of residues constituting the IF2 molecules can be deduced from the bar above the scheme. b Overall architecture of IF2 as derived from the available crystal structure [96, 107, 108] of T. thermophilus IF2 (N through C1) and NMR structure [106] of G. stearothermophilus C2. The color code for G2, G3, C1 and C2 is the same as in a. The N-domain (blue) of T. thermophilus IF2 does not correspond to that of either G. stearothermophilus or E. coli IF2, but corresponds in part to N-terminal and G1 domains as described in the text. Localization of G2-bound GTP and of important structural elements such as helices H6 and H8 and switch II are indicated. Insert: localization of IF2 (green) in the 30SIC (lacking IF3). The 30S subunit, fMet-tRNA and IF1 are indicated in ochre, red, and blue, respectively. Reproduced with permission from [96]

Fig. 6

Conformational changes involving select regions of IF2 and positional adjustments of IF2 and fMet-tRNA during the assembly of a 70S initiation complex. Positions occupied by T. thermophilus His130 (corresponding to His448 in E. coli and His301 in G. stearothermophilus) in the IF2-G2 domain carrying: a GDP or b GTP [107]. This conserved residue, located immediately after the G2-box, is the first N-terminal residue of switch II and is implicated in GTP hydrolysis like its equivalent His80 of EF-Tu [123]. c A 180° rotation (dotted arrow) around helix H8, occurring upon binding of IF2 to the 30S subunit, changes the orientation of IF2-C1 domain on the ribosome and brings this domain close to switch II [96]; d backbone representation of G. stearothermophilus IF2-C1 (cyan and blue) and IF2-C2 (red) joined by the flexible connector (H12) (green) showing the positional mobility of the two domains with the arrows indicating the motional freedom of IF2-C2 with respect to IF2-C1 [93]; e P/I position occupied by fMet-tRNA (red, but for the acceptor end shown in blue) in the 30SIC with respect to the positions of A-site (light gray), E-site (black) tRNA and to the final P-site position attained in initiation dipeptide-productive 70SIC (dark gray) as deduced from cryoEM reconstitutions [95]. f Positions occupied by IF2 on the ribosome at different stages of the translation initiation pathway; IF2·GTP in the 30SIC (green), IF2·GDPNP in the 70SIC (yellow) and “ready to leave” IF2·GDP in the 70SIC (red) as deduced from cryoEM reconstitutions [95]. Reproduced with permission from [107] (a, b, e, f); [96] (c); [93] (d)

Fig. 7

Scheme of the pathway leading to 30SIC and 70SIC formation. a 30SIC formation. Step 1: a vacant 30S ribosomal subunit binds IF3 and IF2. In both cases the binding is biphasic. In the case of IF3, a very rapid step (1000 μM⁻¹ s⁻¹) is followed by a fast first-order rearrangement (34–55 s⁻¹). The off rates of the first and second step, regardless of the presence of IF2, are approximately 35 and 0.8 s⁻¹ [147]. This biphasic IF3 binding mechanism may reflect the fact that the two domains of IF3 bind sequentially to the 30S subunit, the binding IF3CTD occurring before binding of IF3NTD [159, 160]. IF2 binding to a 30S subunit already carrying IF3 occurs with k _on = 220–320 μM⁻¹ s⁻¹ followed by a rearrangement (2–6 s⁻¹). The off rates in the presence of IF3 alone are ~12  and ~1 s⁻¹, respectively [147]. Step 2: IF1 binds in a single event with on and off rates in the presence of both IF3 and IF2 of 10–12 μM⁻¹ s⁻¹ and 0.02 s⁻¹, respectively [141]. Steps 3 and 3′: in the presence of all three factors fMet-tRNA is recruited with k _on = 5 μM⁻¹ s⁻¹ and k _off = 1.5 s⁻¹ [139]. Steps 4 and 4′: the mRNA is bound with different on and off rates depending on its TIR structure; mRNAs with strong secondary structures are bound more slowly than those having little or no secondary structure. On the other hand, the presence of an SD sequence and IFs does not influence either on or off rates that typically range from k _on = 6–158 μM⁻¹ s⁻¹ and k _off = 0.003–4 s⁻¹ [62]. Step 5: mRNAs containing secondary structures must be unfolded in a process that is facilitated by IF2 bound to GTP or GDPNP and antagonized by IF3 [62]. Step 6: the isomerization of the 30S pre-IC allows P-site codon–anticodon interaction to yield a more stable 30SIC from which mRNA and fMet-tRNA are more stably bound. The locking step is under kinetic control of the IFs among which IF2 is mainly responsible for increasing the k _on whereas IF3 strongly increases the k _off (k _off = 0.004 s⁻¹ with canonical ligands) [147] when the 30S ligands are non-canonical. b 70SIC formation. Step 7: a 30SIC, containing IF1, IF2·GTP, IF3 and mRNA whose initiation triplet is P-site decoded by fMet-tRNA, is docked by a 50S subunit with k _on = 34 μM⁻¹ s⁻¹ and k _off = 35 s⁻¹ [132]; a very similar k _on = 12.2 μM⁻¹ s⁻¹ was reported in a previous study [149]. In this process, IF2 changes its conformation [94, 95, 132] and the stepwise dissociation of IF3 [160] begins. Step 8: upon contact with the GAC and SRL of the 50S subunit, the GTPase function of IF2 is activated and GTP is rapidly (k = 35–44 s⁻¹) hydrolyzed leaving GDP+Pi bound to IF2 [124, 132]; as the inter-subunit bridges are progressively formed [161], the IF3NTD loses its contacts with the ribosome [160] reducing the overall ribosomal affinity of the factor by ~2 orders of magnitude [166]. Step 9: this reversible conformational transition (k _on = 24 s⁻¹, k _off = 2.1 s⁻¹) [132] represents the last kinetic checkpoint of translation initiation fidelity by IF3 and IF1 [23, 144] and likely coincides (at least time-wise) with the formation of the final inter-subunit bridges [132, 161]; if the ribosomal ligands are canonical IF3 and IF1 readily dissociate from the ribosome [23], IF2 undergoes a conformational change and the resulting complex is stabilized. Step 10: during this first-order isomerization (k _on = 1.5–2.3 s⁻¹) that represents the rate-limiting step in 70SIC formation [124, 132], fMet-tRNA is adjusted on the ribosome occupying a P/I position intermediate between P/P and P/E, [94, 101, 132]. In the presence of non-hydrolyzable GTP analogs, switch II of IF2-G2 remains “frozen” in a rigid α-helical structure and the complex remains stuck in this non-productive conformation [94, 132]. Step 11: Pi is dissociated from IF2·GDP (k _on = 12 s⁻¹) [132] promoting helix-coil transition in switch II and allowing IF2 to change its conformation, thereby losing its contact with the acceptor end of fMet-tRNA that is adjusted in a productive P-site position [94]. Step 12: IF2 leaves the ribosome (or moves away from the A-site) clearing the way for EF-Tu binding. Step 13: the EF-Tu·GTP·aminoacyl-tRNA complex binds to the 70SIC (k _on ~85 μM⁻¹ s⁻¹) and through a number of steps [158] (not represented here) delivers to the ribosomal A-site the aminoacyl-tRNA encoded by the second mRNA codon. Step 14: the fMet-tRNA bound in the P-site of the peptidyl transferase center of the 50S subunit donates its formyl-methionine to the A-site-bound aminoacyl-tRNA to yield the initiation dipeptide fMet-aa (k = 0.2–2 s⁻¹) [124]