Remodeling of mRNA by eIF4F in human translation initiation

Figures

• 

  Figure 1.. Dynamics of eIF4F binding proximal to the cap a. Schematic of single-molecule 5′ end-to-eIF4A FRET assay where Cy5-labeled eIF4A (red) alongside eIF4E, eIF4G, eIF4B, and ATP was delivered to Cy3-5′-cap-labeled β-globin mRNA (green) tethered on a ZMW imaging surface. b. Example single-molecule data for the assay, FRET corresponds to the anti-correlated decrease in green signal and increase in red signal. The time taken for FRET to appear is the eIF4A association time, the lifetime of FRET is designated the eIF4A bound state lifetime. c. Plot of observed eIF4A association rate across different concentrations. Linear regression analysis (solid line) was used to derive the association rate constant. The error bars represent the 95% confidence interval (CI) of the observed association rate constants. From lowest to highest concentration: n = 146, 126, 131, and 121. d. Plot of mean eIF4A association time across various conditions. Error bars represent the derived 95% CI of the mean association times. From top to bottom: n = 115, 99, 202, 105, 196, 142, and 77. e. Plot of mean eIF4A bound state lifetimes at the indicated concentrations of eIF4B. Error bars represent the derived 95% CI…

• 

  Figure 2.. Features of mRNA modulate eIF4F binding kinetics a. Schematics of single-molecule FRET assay for measuring Cy5-eIF4A binding to uncapped β-globin mRNA (left) or measuring binding in competition with m7GTP (right). b, c. Cumulative probability plots of the indicated parameters, comparing capped β-globin to the conditions denoted in (a). The lines represent fits to exponential functions, and the denoted values are derived mean times and 95% CI from the fit. For capped, uncapped, and m7GTP competition: n = 121, 141, and 142 respectively. d. Schematic of predicted RNA secondary structures in the 5′ UTR of the indicated mRNAs. e. Cumulative probability plots of the indicated parameters, comparing the different mRNA denoted in (d). The lines represent fits to exponential functions, and the denoted values are approximate mean times, which were comparable between all conditions. For SL-50 and CAA: n = 101 and 110 respectively.

• 

  Figure 3.. Remodeling of mRNA by eIF4F/eIF4B is fast on the timescale of translation a. Schematic of single-molecule cap-to-probe FRET assay where eIF4A, eIF4E, eIF4G, eIF4B, and ATP are delivered to a Cy3-5′-cap-labeled β-globin mRNA with Cy5-probe 124 nt from the cap. Schematic denotes the two observed FRET states. b. Example single-molecule data for the assay. The top plot demonstrates the characteristic FRET in the absence of proteins. The bottom demonstrates the transition between FRET (compact) to loss of FRET (extended) in the presence of eIF4F/eIF4B. The lifetime of the first FRET state was denoted as extension time. Time to subsequent FRET was denoted as the extended state lifetime. c. Cumulative probability plot of the extension time, comparing the conditions denoted in (b). The lines represent fits to exponential functions, and the denoted values are derived mean times and 95% CI from the fit. d, e. Plots of mean extension times and mean extended state lifetimes across various conditions. Error bars represent the derived 95% CI of the mean times. From top to bottom: n = 131, 94, 106, 103, 117, 124, 111, and 121. f. Example single-molecule data for the eIF4B dropout condition…

• 

  Figure 4.. Features of an mRNA can modulate its conformational remodeling a. Schematics of single-molecule FRET assay for measuring mRNA extension to Cy3-labeled uncapped β-globin mRNA (left) or measuring extension by eIF4F in competition with m7GTP (right). Cy5-probe is at position +124. b, c. Cumulative probability plots of the indicated parameters, comparing capped β-globin to the conditions denoted in (a). The lines represent fits to exponential functions, and the denoted values are derived mean times and 95% CI from the fit. For uncapped β-globin and m7GTP competition n = 98 and 116 respectively. d. Schematic of predicted RNA secondary structures in the 5′ UTR of SL-50 mRNA and the position of the different fluorescent probes utilized. The green star reflects the probe on the 5′ cap, the red stars denote the position of the Cy5 dyes on the different probes used. e, f. Cumulative probability plots of extension times comparing the effects of eIF4 proteins on extension of SL-50 mRNA, using the two probe positions shown in (d). The lines represent fits to exponential functions, and the denoted values are derived mean times and 95% CI from the fit. For mRNA only and eIF4F + eIF4B: in (e) n…

• 

  Figure 5.. The extended mRNA state is necessary for 43S PIC loading a. Schematic of single-molecule 43S PIC loading assay. All shown initiation factors and ribosomal subunits are delivered to a tethered β-globin mRNA in real-time. b. Example single-molecule data for the assay. The burst in green intensity represents 43S PIC loading and the time taken for this burst is the 43S PIC loading time. c. On the left, a cumulative probability plot of the experimental setup shown in (a). The line represents the fit to a hypoexponential function. On the right, a bar plot of the mean times for both kinetic steps derived from the fit. Error bars denote 95% CI from the fit, n = 156. d. Schematic of single-molecule mRNA extension and 43S PIC loading assay. Cy5-label on β-globin mRNA is at position +124 from the cap. e. Example single-molecule data for the assay in (d). The end of FRET between green and red denotes mRNA extension. The burst in yellow intensity represents Cy3.5-43S PIC loading. f. Heat maps of normalized intensities of 40S-Cy3.5 (top) and mRNA-Cy5 (bottom) for all analyzed events, n = 109. mRNA extension is labeled. g. On the left, cumulative probability plots of PIC loading…

• 

  Figure 6.. Regulation of 43S PIC loading by eIF4F composition and mRNA features a. Bar plot of derived mean times for both kinetic steps (fast and slow) of 43S PIC loading at the indicated concentrations of 43S PIC. Error bars denote 95% CI from the fit, for 5, 10, 20 nM: n = 156, 119, and 98 respectively. b. Stacked bar plot of derived mean times of both kinetic steps (fast and slow) of 43S PIC loading for various conditions. Error bars denote 95% CI from the fit. From top to bottom n = 156, 118, 118, and 115. c. Bar plot of derived mean times for both kinetic steps (fast and slow) of 43S PIC loading with the indicated mRNAs. Error bars denote 95% CI from the fit, for CAA and SL-50: n = 160 and 120 respectively. d. Bar plot of derived mean times for both kinetic steps (fast and slow) of 43S PIC loading with the capped state of β-globin mRNAs. Error bars denote 95% CI from the fit, for uncapped mRNA: n = 105.

• 

  Figure 7.. Kinetic model for the early steps of translation initiation On the left, compact mRNA is represented as a cartoon with a 5′ cap, secondary structure, and start codon. Blue backgrounds denote on-pathway steps. In the first step, eIF4F binds to the 5′ end of an mRNA, guided by the 5′ cap and facilitated by eIF4G/E. eIF4A dissociates independently of eIF4F, with eIF4G/E possibly remaining bound. Upon initial binding of eIF4F, mRNA conformation is very rapidly remodeled into an extended conformation in an ATP-dependent manner, with eIF4B likely binding prior to this rapid extension step. This extended state is stabilized by the proteins, with recompaction occurring slowly. Critically, the extended state is the amenable conformation for 43S PIC loading at the 5′ end. We speculate that the extended state is maintained by the continuous presence of at least one eIF4A molecule at the 5′ end. This continuity is likely sustained by the overlapping binding cycles of multiple eIF4A molecules.