Operon mRNAs are organized into ORF-centric structures that predict translation efficiency
- University of California, San Francisco, United States
- Howard Hughes Medical Institute, University of California, San Francisco, United States
Figures
Figure 1
DMS-seq effectively probes RNA structures in E. coli.
(A) Schematic for obtaining mRNA structure and translation efficiency using DMS-seq, mRNA-seq, and ribosome profiling from the same sample. (B) Plot showing the effect of DMS-seq read coverage on …
Figure 2 with 1 supplement
E. coli mRNAs have intrinsic ORF-wide secondary structures.
(A–C) Plots comparing the Gini indices of the first half of the ORF against those of the second half of the ORF for: A. in vivo modified mRNA from cells growing at 37°C; B. in vivo modified mRNA …
Figure 2—figure supplement 1
mRNA structure is organized around open reading frames.
(A) Relative 35S-methionine incorporation of WT and ΔgcvB cells after treatment of kasugamycin at 37°C, normalized against the total incorporated radioactivity measured immediately before treatment …
Figure 3 with 1 supplement
Translational efficiency (TE) is highly correlated with ORF mRNA structure.
(A) Plots comparing the Gini indices of ORFs in polycistronic operons calculated from in vivo DMS-seq to their TEs (N = 483). (B) Histograms of TE ratios between adjacent non-overlapping (N = 253) …
Figure 3—figure supplement 1
Correlation between the mRNA structural level and translation efficiency.
(A) An example showing the results of mRNA-seq, ribosome profiling and DMS-seq of the rpsF-priB-rpsR-rplI operon, with translation efficiency (TE) and Gini index of each ORF indicated. (B) Plot of …
Figure 4 with 2 supplements
Correlation of Other mRNA features with TE.
(A–C) Plots comparing tAI (tRNA adaptation index) of the entire ORF against: A. translation efficiency (TE, protein synthesis rate per mRNA); B. protein synthesis rate (average ribosome footprint …
-
Figure 4—source data 1
Linear regression model to predict TE based on different mRNA features.
A multiple linear regression model is applied to predict TE based on the following features: mRNA structure level of various portions of the ORF, codon usage predicted by tAI (dos Reis et al., 2004; Tuller et al., 2010), codon influence metric (Boël et al., 2016), and the strength of Shine-Dalgarno sequence (using the RBS Calculator established by Salis et al from https://github.com/hsalis/Ribosome-Binding-Site-Calculator-v1.0). (A) Comparison between the experimentally measured TE and the model-predicted TE. The red dashed line indicates the y = x diagonal line. (B) Relative contribution of the factors in predicting TE, calculated from stepwise regression. Y-axis: R2 of different models with stepwise addition of individual factors. Asterisks indicate significant improvement of model (based on ANOVA, with significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05).
- https://doi.org/10.7554/eLife.22037.009
Figure 4—figure supplement 1
Effect of SD strength, tAI, and codon influence on predicting TE of endogenous genes.
(A) Plot of predicted Shine-Dalgarno strength (see Materials and methods) against in vivo translation efficiency. Genes with Gini indices in a tight range (0.5–0.52) are indicated in cyan. (B–C) …
Figure 4—figure supplement 2
Comparison of the relative significance of different mRNA features in predicting TE.
Top: Bar plot comparing the absolute values of Spearman’s rank correlation coefficient (ρ) in various conditions with Gini indices of the dataset for each condition (gray bars), or for the 421 ORFs …
Figure 5 with 1 supplement
ORFs are isolated from each other by forming ORF-specific RNA structures.
(A) Cumulative distribution of spacing between adjacent ORFs within operons of E. coli. X-axis: distance from 3’ of the stop codon of upstream genes (gene A) to 5’ of the start codon of downstream …
Figure 5—figure supplement 1
Structural isolation between mRNA of adjacent ORFs on the same operons.
(A) The absolute value of correlation (Spearman's ρ) between computationally predicted local mRNA structure at the ORF boundary, quantified by predicted ∆G of minimum free-energy structure, and the …
Figure 6 with 1 supplement
Disruption of structural isolation between dusB and fis affects fis translation.
(A) mRNA structure at the 3’ end of dusB, with mutations M3 and M2 indicated. Translation efficiencies (TEs) of dusB and fis in WT cells are 0.02 and 2.06, respectively. (B) The dusB-M3 mutation …
Figure 6—figure supplement 1
mRNA secondary structure at the dusB-fis boundary region of WT cells and dusB mutants.
Gini indices calculating within 100 nt rolling windows (plotted at the center of windows) at the boundary region of dusB-fis operon. X-axis: distance from the 5’ of fis start codon. Different …
Figure 7
Model of operon mRNA structural organization.
Polycistronic mRNAs are organized into ORF-centric modules with characteristic but different extents of mRNA structure, punctuated by regions of low basepairing close to the translational start site …
Additional files
-
Supplementary file 1
In vivo data summary.
Data summary of ORFs that have ≥15 reads per nucleotide for in vivo DMS-seq (N = 1116).
- https://doi.org/10.7554/eLife.22037.017
-
Supplementary file 2
In vitro data summary.
Data summary of ORFs that have ≥15 reads per nucleotide for in vitro DMS-seq (N = 710).
- https://doi.org/10.7554/eLife.22037.018
-
Supplementary file 3
In vivo untranslated (ksg-treated) data summary.
Data summary of ORFs that have ≥15 reads per nucleotide for DMS-seq of in vivo untranslated (ksg-treated) mRNA (N = 480).
- https://doi.org/10.7554/eLife.22037.019
-
Supplementary file 4
Common ORFs data summary.
Data summary of ORFs that have ≥15 reads per nucleotide in all three conditions: DMS-seq of in vivo mRNA, in vivo untranslated mRNA and in vitro modified mRNA (N = 421).
- https://doi.org/10.7554/eLife.22037.020
