Functional and biochemical characterization of human eukaryotic translation initiation factor 3 in living cells

Abstract

The main role of the translation initiation factor 3 (eIF3) is to orchestrate formation of 43S-48S preinitiation complexes (PICs). Until now, most of our knowledge on eIF3 functional contribution to regulation of gene expression comes from yeast studies. Hence, here we developed several novel in vivo assays to monitor the integrity of the 13-subunit human eIF3 complex, defects in assembly of 43S PICs, efficiency of mRNA recruitment, and postassembly events such as AUG recognition. We knocked down expression of the PCI domain-containing eIF3c and eIF3a subunits and of eIF3j in human HeLa and HEK293 cells and analyzed the functional consequences. Whereas eIF3j downregulation had barely any effect and eIF3a knockdown disintegrated the entire eIF3 complex, eIF3c knockdown produced a separate assembly of the a, b, g, and i subunits (closely resembling the yeast evolutionary conserved eIF3 core), which preserved relatively high 40S binding affinity and an ability to promote mRNA recruitment to 40S subunits and displayed defects in AUG recognition. Both eIF3c and eIF3a knockdowns also severely reduced protein but not mRNA levels of many other eIF3 subunits and indeed shut off translation. We propose that eIF3a and eIF3c control abundance and assembly of the entire eIF3 and thus represent its crucial scaffolding elements critically required for formation of PICs.

FIG 1

eIF3c and eIF3a control expression of module ii and iii subunits of eIF3 and link all three modules tightly together in HeLa cells. (A) A schematic model of human eIF3 and its binding partners combining findings presented here and elsewhere (7, 10, 13, 14, 26, 46). Individual eIF3 modules as well as eIF3 subunits forming the PCI/MPN octamer are indicated. An arrow indicates a potential interaction between eIF3i and -g as proposed in reference . (B) Effect of siRNA targeting eIF3j (2 nM), eIF3c (5 nM), and eIF3a (5 nM) mRNAs on protein levels of eIF3 subunits and other eIFs determined by Western blotting 72 h posttransfection (this experiment was repeated numerous times with similar results). NT, control nontargeted cells; Tub, loading control, anti-TUBA4A from Sigma, catalog no. T6074. WCEs were prepared as described in Materials and Methods. Western blots were quantified using Quantity One and/or NIH ImageJ, and the signals were normalized to α-tubulin (Tub). The resulting values obtained with the control NT cells were set to 1.00, and those obtained with eIF3j^KD, eIF3c^KD, and eIF3a^KD cells were expressed relative to the NT. Standard deviations (SD) from at least three individual experiments are given. (C) Effect of siRNA targeting eIF3j (2 nM), eIF3c (5 nM), and eIF3a (5 nM) mRNAs on integrity of eIF3 determined by the coimmunoprecipitation assay using eIF3b as a co-IP bait (anti-eIF3b antibody from Santa Cruz, catalog number sc-16377) followed by Western blotting 72 h posttransfection (see also Fig. S3 in the supplemental material). nc, negative control (beads only, no antibody). Western blots were quantified using Quantity One and/or NIH ImageJ, and the signals were normalized to eIF3b. The resulting values obtained with the control NT cells were set to 1.00, and those obtained with eIF3j^KD, eIF3c^KD, and eIF3a^KD cells were expressed relative to the NT (SD are given).

FIG 2

Downregulation of eIF3c and eIF3a but not eIF3j diminishes initiation rates and ceases cell growth. (A) The effect of knockdowns of eIF3j, eIF3c, and eIF3a using siRNAs at indicated concentrations on cell proliferation was assessed by the MTT assay 24, 48, or 72 h posttransfection in HeLa cells. Figures represent the results from three independent experiments ± SD. (B and C) The effect of knockdowns of eIF3j, eIF3c, and eIF3a using siRNAs at indicated concentrations on translation rates was assessed by the [³⁵S]methionine incorporation (B) and polysome profile (C) analyses 72 h posttransfection. Positions of 40S, 60S, and 80S species are indicated by arrows. Graphs represent the results from three independent experiments ± SD. P/M, polysome-to-monosome ratio.

FIG 3

eIF3 module i is capable of binding to the 40S ribosome on its own. HeLa cells transfected with control NT siRNA (A) and siRNAs against eIF3c (B), eIF3a (C), and eIF3j (D) were treated with 0.2% formaldehyde (72 h posttransfection), and the WCEs were separated on 7 to 30% sucrose gradients. Collected fractions were subsequently subjected to Western blot analysis. These experiments were performed three times; for quantifications and SD, see Fig. S7 in the supplemental material. For description of boxes, see the text. Arrows indicate the eIF3l-specific band.

FIG 4

eIF3 module i significantly contributes to the eIF3 role in mRNA recruitment to 43S PICs in vivo. HeLa cells transfected with control NT siRNA and siRNAs against eIF3c and eIF3a were harvested without formaldehyde treatment 72 h posttransfection, and the WCEs were separated on 7 to 30% sucrose gradients. Total RNA was isolated from all collected fractions, and amounts of RPL41 mRNA associated with native 48S PICs (visualized by 18S rRNA) were measured using real-time PCR. The amounts of mRNA were calculated as 2^−Cq × 10⁶ for RPL41 mRNA and 2^−Cq × 10⁴ for 18S rRNA, where “2” represents the PCR efficiency and Cq the quantitation cycle.

FIG 5

The eIF3c^KD relaxes the stringency of the AUG start codon selection in vivo. (A) Schematic of the firefly luciferase reporter with alternative start sites AUG or GUG. (B) Schematic of the ATF4-firefly luciferase reporter showing positions of its two short uORFs, the first of which is mutated at the AUG start site and the other naturally extends into the coding region. (C) Quantification of the Sui⁻ phenotype. HeLa cells were transfected with the indicated siRNAs, and 48 h later, the second round of transfection with individual reporters (AUG or GUG) together with a plasmid expressing the Renilla luciferase (used for normalization) was carried out. Luciferase activities were measured in WCEs 24 h after the second transfection. The mean percentages ± SD of the GUG initiation rates relative to those of AUG were calculated from three experiments. (D) Quantification of the leaky scanning phenotype. Same as described for panel C, except that the ATF4 reporter was transfected during the second round of transfection. The mean percentages ± SD of the increased or decreased skipping of the AUG start site of uORF2 in cells with knocked-down expression relative to control NT cells were calculated from three experiments.

FIG 6

Schematics illustrating the effects of eIF3c and eIF3a knockdowns on integrity of the human eIF3 holocomplex and its ability to stimulate formation of the 48S PICs. The separate, intact eIF3 module i, resembling the budding yeast 5-subunit eIF3, is to a certain degree capable of stimulating formation of the 48S PICs. The 40S binding ability of the separate eIF3 module iii is limited. eIF3j associates with 40S ribosomes independently of the eIF3 holocomplex. The eIF3b subunit, perhaps still in complex with eIF3g, shows a weak 40S binding affinity on its own. For further details, see the text.