A mini-hairpin shaped nascent peptide blocks translation termination by a distinct mechanism

Abstract

Protein synthesis by ribosomes produces functional proteins but also serves diverse regulatory functions, which depend on the coding amino acid sequences. Certain nascent peptides interact with the ribosome exit tunnel to arrest translation and modulate themselves or the expression of downstream genes. However, a comprehensive understanding of the mechanisms of such ribosome stalling and its regulation remains elusive. In this study, we systematically screen for unidentified ribosome arrest peptides through phenotypic evaluation, proteomics, and mass spectrometry analyses, leading to the discovery of the arrest peptides PepNL and NanCL in E. coli. Our cryo-EM study on PepNL reveals a distinct arrest mechanism, in which the N-terminus of PepNL folds back towards the tunnel entrance to prevent the catalytic GGQ motif of the release factor from accessing the peptidyl transferase center, causing translation arrest at the UGA stop codon. Furthermore, unlike sensory arrest peptides that require an arrest inducer, PepNL uses tryptophan as an arrest inhibitor, where Trp-tRNA^Trp reads through the stop codon. Our findings illuminate the mechanism and regulatory framework of nascent peptide-induced translation arrest, paving the way for exploring regulatory nascent peptides.

Fig. 1. Screening of uncharacterized ribosome arrest peptides by overexpression phenotype.

a Working hypothesis for screening uncharacterized ribosome arrest peptides. Serial dilution spot assay for evaluating the cytotoxicity when established (b) or recently identified (c) sORFs were overexpressed in E. coli cells. A representative of three independent experiments (n = 3) is shown. d Schematic illustration of a comparative quantitative proteomic analysis. Each sORF was overexpressed in E. coli cells, and all samples were subjected to the LC-MS/MS-based proteome analysis. The proteomic changes by the overexpression were quantified as the fold changes relative to the cells harboring an empty vector. e PCA score plot of PC1 and PC2. The numbers in the x- and y-axis labels represent the proportion of the variances of PC1 and PC2, respectively. Vector means the data from fold changes of another vector control against a vector control. Colors (green, blue, and red) correspond to the clustering by a hierarchical clustering analysis (Supplementary Fig. 1i).

Fig. 2. Identification of the ribosome-arresting peptidyl-tRNA by LC-MS/MS.

a Schematic illustration of the experimental procedure for the identification of peptidyl-tRNA-derived peptides. Each candidate sORF was overexpressed in E. coli cells, and the total RNA fraction in the cell lysate was concentrated with a silica column. The RNA fraction was hydrolyzed by alkaline treatment, and the obtained peptides were identified by protease digestion and subsequent LC-MS/MS. b–e MS/MS spectra of peptides derived from peptidyl-tRNAs. Amino acid sequences above the graphs represent the positions of the detected peptides (blue areas indicate the detected peptide). P and A below the amino acid sequence represent the plausible positions of the P- and A-sites in the ribosome during translation arrest, respectively. In the graphs, the peaks of the b- and y-fragment ions are shown in red and blue, respectively.

Fig. 3. PepNL nascent peptide turns back toward the entrance of the ribosome tunnel.

a Serial dilution spot assay to assess the cytotoxicity upon overexpression of wild-type PepNL (WT) or its frameshift mutant (FS) in E. coli cells. b The PepNL-tRNA accumulated in E. coli cells expressing the indicated pepNL variants was detected by northern blotting using an anti-tRNA^Asp probe. An asterisk indicates the unprocessed rrnC or rrnH transcripts that include the unprocessed tRNA^Asp sequence. (#) c The wild-type or frameshifted pepNL mRNA was translated by PUREfrex in the absence of tryptophan, and the PepNL-arrested ribosome was visualized by toeprint analysis. Thiostrepton, which inhibits translation elongation, was pre-included where indicated. The pepNL mRNA was translated in the absence of release factors (RF⁻) to prepare the ribosomes stalled at the stop codon for the position marker (lane 5). (#) d Overall cross section of the cryo-EM density map at the peptide exit tunnel of the 70S ribosome (gray), showing P-tRNA (green), RF2 (cyan), and PepNL peptide (orange). e, f PepNL peptide and interacting 23S rRNA nucleotides in the ribosome exit tunnel. The N-terminus and C-terminus of PepNL were labeled as “N” and “C”, respectively. Close-up views of the intramolecular interactions within the PepNL peptide (g) and intermolecular interactions between the PepNL peptide and 23S rRNA nucleotides (h). g, panel 1: A hydrophobic interaction between Ile3 and Tyr9. g, panel 2: β-sheet-like interactions involving Lys2 with Ala10 and Leu4 with Ile8. h, panel 1: A hydrophobic interaction between Ile3 and U2609. h, panel 2: A hydrophobic interaction between Leu4 and A2062. h, panel 3: Hydrophobic interactions between Ile8 and A2058-A2059. h, panel 4: A hydrophobic interaction between Tyr9 and U2610. i Structural comparison of the PepNL nascent peptide (orange) with nascent peptides that lack RAP activity {PDB: 8CVJ (nascent peptide sequence: fMSEAC, pink) and 8CVL (fMTHSMRC, purple)}. j Serial dilution spot assay to assess the cytotoxicity upon overexpression of wild-type PepNL (WT) or its variants carrying the indicated amino acid substitution in E. coli cells. (#) A representative of three independent experiments (n = 3) is shown.

Fig. 4. PepNL blocks RF2-mediated translation termination.

a Overview of the RF2 structure within the PepNL-arrested ribosome (cyan) compared with the active RF2 (purple, PDB: 6C5L). The P-tRNA^Asp (green) and pepNL mRNA (brown) are also shown. b Close-up view of the apical loop conformations of the rearranged and active RF2 states. Gln252s of the GGQ motif are highlighted as sticks with sphere models. The ribosome is shown as a surface model. Hydrogen bond interactions between the apical loop of the rearranged RF2 and 23S rRNA. c: Arg245 and G2508, Gln252 and G2583, Val254 and C2573, Arg256 and U2554, Gly251 and G2553. d: Glu258-Gln280 and U2492, His281 and U2460. e: Arg262 and G2557. f Superimposition of the active RF2 (purple) indicates a steric clash of its Gln252 with Ile13 of PepNL. The distorted Ile13 of the PepNL peptide occupies the pocket formed by A2451, C2452, and U2506 (not shown) of 23S rRNA, blocking the proper accommodation of the RF2 apical loop. The rRNA nucleotides and Ile13 are shown as sphere models. g Schematics representing the mechanism of RF2 inhibition by the PepNL nascent peptide.

Fig. 5. Read-through of stop codon serves as an arrest inhibition mechanism for PepNL.

a The wild-type (UGA) or mutated (UAG) pepNL mRNA was translated by PUREfrex in the presence or absence of tryptophan (Trp), and the PepNL-arrested ribosome was visualized by toeprint analysis as shown in Fig. 3c. (#) b Schematics of pepNL mRNAs analyzed. c The pepNL mRNAs indicated in Fig. 5b were translated by PUREfrex in the absence (lanes 3 and 4) or presence of tryptophan (lanes 1, 2, 5, 6, 7, and 8). The ³⁵S-methionine-labeled translation products were separated by neutral pH SDS-PAGE with optional RNase A (RN) pretreatment. The truncated pepNL NS−1 (14 aa) and NS−2 (27 aa) mRNA were also analyzed to serve as size markers (lanes 5 to 8). The PepNL (14 aa) or PepNL (read-through: RT) peptidyl-tRNA and PepNL (14 aa) or PepNL (RT) peptide are schematically indicated. The asterisk denotes the fMet-tRNA. (#) d The pepNL-stop (UGA/UAA/UAG)-lacZ mRNA was expressed in E. coli cells, and the frequency of stop codon read-through was calculated as described in the Methods section. The mean values ± SE estimated from three independent biological replicates (n = 3) are shown. e The pepNL mRNA was translated by PUREfrex without tryptophan and release factors　for 30 min at 37°C. Afterward, a final 25 µM of tryptophan was added and further incubated for the indicated duration. The ³⁵S-methionine-labeled translation products were analyzed as shown in Fig. 5c. The asterisk denotes the fMet-tRNA. (#) f Schematic illustration of the Trp-tRNA^Trp−dependent inhibition of the PepNL-induced translation arrest. RF2 inefficiently terminates the translation of pepNL due to the steric clash shown in Fig. 4. However, in the presence of sufficient tryptophan, the Trp-tRNA^Trp decodes the UGA of pepNL, leading to the stop codon read-through. Two potential scenarios could explain this event: 1) Trp-tRNA^Trp initiates read-through before the hairpin folds, which could otherwise inhibit Trp-tRNA^Trp accommodation (Supplementary Fig. 7f); or 2) Trp-tRNA^Trp releases the ribosome stalled by the hairpin-shaped PepNL, alleviating the translation arrest at a moderate rate. (#) A representative of three independent experiments (n = 3) is shown.