PMC

43S complex model.
(a) Docking aIF2β on the 40S–aIF2γ–Met-tRNA_i^Met complex model. aIF2β from the aIF2βγ heterodimer structure (pdb code: 2QMU) was docked on the 40S–aIF2γ–Met-tRNA_i^Met complex in . Helix H1 of aIF2β, which forms the only rigid body interaction with aIF2γ, is boxed. The aIF2β location corresponding to eIF2β-S264 is shown as purple spheres, and the Met-tRNA_i^Met residues cleaved by Fe(II)-BABE linked to eIF2ΔC-βS264C are colored purple.
(b) Docking of aIF2α, eIF1 and eIF1A on the 40S–aIF2βγ–Met-tRNA_i^Met complex. aIF2α is from the aIF2αγ heterodimer structure (pdb code: 2AHO); eIF1 is from the 40S–eIF1 co-crystal structure (pdb code: 2XZM); and eIF1A (pdb code: 1D7Q) was positioned based on the bacterial 30S–IF1 structure (pdb code: 1HR0) and hydroxyl radical mapping data. Only the eIF1A core structure is shown.
(c) Schematic depicting the alternate tRNA T-stem (orange) and rRNA helix h44 (black) interactions of domain III (green) from EF-Tu (left) and aIF2γ(eIF2γ) (right).

Structures of EF-Tu TC and aIF2γ TC model.
Ribbons representation of EF-Tu–GDPNP–Phe-tRNA^Phe TC (pdb code: 1TTT, left). The aIF2γ structure (pdb code: 2AHO) was aligned to the EF-Tu structure using PyMOL software (DeLano Scientific) to make an aIF2γ–GDPNP–Phe-tRNA^Phe TC model (right). Three interaction points between EF-Tu and Phe-tRNA^Phe (T stem, 5′ end, and 3′CCA) are boxed and labeled. The aIF2γ residues corresponding to residues K507 and R118 in S cerevisiae eIF2γ, which were used to tether Fe(II)-BABE for hydroxyl radical cleavage of tRNA_i^Met (see ), are shown as blue and magenta spheres, respectively. GTP binding domain (G), domain II (II), and domain III (III) of both EF-Tu and aIF2γ are depicted using different shades of blue and green, respectively, as indicated in the inset beside each figure, and Phe-tRNA^Phe is shown in orange. The same color schemes for EF-Tu, aIF2γ, and Phe-tRNA^Phe are used in all figures.

eIF2γ binding site for 3′ end of Met-tRNA_i^Met.
(a) aIF2γ–Met-tRNA_i^Met complex model, generated as described in . The RIT1 catalyzed 2′-O-ribosyl phosphate modification at residue 64 of tRNA_i^Met is shown in orange.
(b) Ribbons representation of EF-Tu–GDPNP–Phe-tRNA^Phe complex (pdb code: 1TTT). The site of the T394C mutation, which is analogous to the eIF2γ-K507C mutation, is shown as red spheres.
(c) Magnified view of proposed eIF2γ binding site for the 3′ end of Met-tRNA_i^Met (box in Fig. 5a). Locations of yeast eIF2γ mutations that impair eIF2 function in vivo are labeled and shown as blue sticks: Y142H (corresponds to S. sol. aIF2γ-Tyr51), E383K (Gly282), and G397S (Ala296).
(d,e) Purified WT, γY142H (D), or γK507C (E) mutant forms of eIF2 were assayed for eIF2 TC formation by filter binding assay. Fractions of [³⁵S]Met-tRNA_i^Met bound to eIF2 were plotted as a function of eIF2 concentration; points and s.d. are averages of at least three independent experiments.
(f) Purified WT or T439C mutant forms of EF-Tu (T. therm) were assayed for TC formation by filter binding assay. Fractions of [¹⁴C]Phe-tRNA^Phe bound to EF-Tu were plotted as a function of EF-Tu concentration; points and s.d. are averages of at least three independent experiments.

Directed hydroxyl radical cleavage of Met-[³²P]tRNA_i^Met by Fe(II)-BABE-derivatized eIF2 in 48S complexes.
(a) eIF2ΔC-γR118C-Fe(II)-BABE cleavages. Hydroxyl radical cleavage products from 48S complexes were resolved on 10% (w/v) denatured polyacrylamide gels, and cleavage sites on [³²P]tRNA_i^Met were determined by comparison with samples containing eIF2ΔC (WT) (lane 2). Sites of enhanced cleavage with eIF2ΔC-R118C are boxed. The tRNA ladders were prepared by digesting Met-[³²P]tRNA_i^Met with RNase T1 (cleaves 3′ of G residue, lane 3) or by base cleavage (lane 4). The tRNA residue numbers are shown at the left.
(b) Sites (colored magenta) of eIF2ΔC-γR118C-Fe(II)-BABE directed hydroxyl radical cleavage on Met-[³²P]tRNA_i^Met are shown on the secondary (left) and three-dimensional (pdb code: 1YFG, middle and right) structures of tRNA_i^Met.
(c) Met-[³²P]tRNA_i^Met cleavages in 48S complexes assembled with eIF2ΔC-γK507C-Fe(II)-BABE and eIF2ΔC-βS264C-Fe(II)-BABE, as in panel a. Cleavage products were resolved on 10% (left) or 20% (right) (w/v) denatured polyacrylamide gels.
(d,e) Summary of Met-[³²P]tRNA_i^Met cleavages in panel c, depicted as in panel b.
(f) aIF2γ-GDPNP-tRNA_i^Met TC model showing the sites of tRNA_i^Met (cyan) cleavages by hydroxyl radicals generated at residues K507 (blue) and R118 (magenta). The model was generated by docking the structure of yeast tRNA_i^Met (pdb code: 1YFG) on the aIF2γ TC structure from . The RIT1 catalyzed 2′-O-ribosyl phosphate modification at residue 64 of tRNA_i^Met is shown as orange spheres.

Directed hydroxyl radical cleavage of 18S rRNA by Fe(II)-BABE-derivatized eIF2 in 48S complexes.
(a,b) Primer extension analysis of 18S rRNA helix h44 cleavages by Fe(II)-BABE linked to the indicated positions in eIF2α, eIF2β or eIF2γ. Lanes marked “U” and “C” present 18S rRNA sequencing reactions using reverse transcriptase and the indicated dideoxynucleotide. Positions of cleaved nucleotides are boxed and the numbering of helix h44 residues is shown on the left.
(c) Sites of 18S rRNA helix h44 cleavages by eIF2ΔC-γK507C-Fe(II)-BABE (left, blue), eIF2ΔC-γD446C-Fe(II)-BABE (middle, black), and eIF2ΔC-γA480C-Fe(II)-BABE (right, red) are displayed on secondary and tertiary, taken from the crystal structure of the yeast ribosome (pdb code: 3O30), structure models of helix h44. aIF2γ and Met-tRNA_i^Met are docked on helix h44 as described in the text.
(d–f) 40S–aIF2γ–Met-tRNA_i^Met complex model viewed from the A-site, d; from above, e; and from below, f. Met-tRNA_i^Met was docked in the P-site of the yeast 40S ribosome (pdb code: 3O30) as observed in the structure of the bacterial 70S ribosome (pdb code: 2J00). aIF2γ was docked on the acceptor stem of Met-tRNA_i^Met as in the EF-Tu TC (); however, domain III of aIF2γ was positioned toward helix h44 and aligned consistent with the cleavage data. aIF2γ residues corresponding to K507 (blue), D446 (black), A480 (red) and R118 (magenta) are shown as spheres and the helix h44 and Met-tRNA_i^Met cleavage sites are shown in matching colors for the three sites of Fe(II)-BABE modification.

Construction and analysis of eIF2 Cys mutants.
(a) Removal of native Cys residues to generate Cys-lite eIF2ΔC. The designated Cys residues were mutated as indicated; note that Cys residues in the zinc binding domains (ZBD, pink box) of eIF2β and eIF2γ were left intact.
(b) Ribbon diagram of aIF2 showing sites of single Cys mutations (red spheres) in yeast eIF2ΔC. The aIF2 structure was assembled by aligning the structures of aIF2αγ (pdb code: 2AHO) and aIF2γβ (pdb code: 2QMU). The indicated mutations were introduced individually into S. cerevisiae eIF2 subunits; the S80C mutation in the eukaryote-specific N-terminal domain of eIF2β is not depicted because this element is not found in aIF2β. aIF2 subunit domains are colored as in panel a.
(c) Preformed TCs containing Fe(II)-BABE-derivatized (lanes 1–7, 9–16) or unmodified (lane 8) WT or the indicated mutant forms of eIF2 were assayed for 48S complex formation. The staggering of the bands is because the reactions were quenched by loading at 5 or 15 min onto a running native gel. The positions of 48S complexes and free Met-tRNA_i^Met are indicated. See also for rate constants obtained from the assay.
(d) Summary of K_d values for eIF2 TC binding to 40S subunits. Errors are s.d. from three independent measurements.

eIF2γ domain III is involved in 40S binding.
(a) Locations of conserved Arg residues (blue spheres, labeled as in yeast eIF2γ) in the proposed 40S binding surface of eIF2γ domain III are shown in the 40S–aIF2γ–Met-tRNA_i^Met complex model.
(b) Altered GCN4 translational control in yeast cells expressing eIF2γ domain III mutants. Derivatives of the gcn2Δ yeast strain J551 expressing wild type or the indicated mutant forms of eIF2γ were spotted on minimal medium with essential nutrients (SD, rows 1–3) or SD medium containing 0.3 μg per ml sulfometuron methyl (SM, rows 4–6).
(c) Western blot analysis of eIF2γ expression. Whole-cell extracts of strains described in (b) were subjected to immunoblot analysis using anti-yeast eIF2γ (upper panel) or anti-yeast eIF2α (lower panel) antiserum. Immune complexes were visualized using enhanced chemiluminescence.
(d) Analysis of GCN4-lacZ expression. The GCN4-lacZ plasmid p180 was introduced into derivatives of strain J551 expressing WT eIF2, eIF2γ-R439A, or eIF2γ-R510H. Cells were grown and β-galactosidase activities were determined as described previously. The β-galactosidase activities are the averages of three independent transformants and the errors are s.d.
(e) K_d values and standard deviations for WT or γR439A mutant forms of eIF2 TC binding to WT or 18S-A1152U mutant forms of yeast 40S–eIF1–eIF1A complexes were measured by 43S gel shift assays. Fitting curves are shown in .