Structure 14, 913–923, May 2006 ª2006 Elsevier Ltd All rights reserved
DOI 10.1016/j.str.2006.03.012
Two Structurally Atypical HEAT Domains in the C-Terminal Portion of Human eIF4G Support Binding to eIF4A and Mnk1 Lluı´s Bellsolell,1,4 Park F. Cho-Park,3,4 Francis Poulin,3,6 Nahum Sonenberg,3,5 and Stephen K. Burley1,2,5,7,* 1 Laboratories of Molecular Biophysics 2 Howard Hughes Medical Institute The Rockefeller University New York, New York 10021 3 Department of Biochemistry and McGill Cancer Center McGill University Montreal, Quebec H3G 1Y6 Canada
Summary The X-ray structure of the C-terminal region of human eukaryotic translation initiation factor 4G (eIF4G) has ˚ resolution, revealing two been determined at 2.2 A atypical HEAT-repeat domains. eIF4G recruits various translation factors and the 40S ribosomal subunit to the mRNA 50 end. In higher eukaryotes, the C terminus of eIF4G (4G/C) supports translational regulation by recruiting eIF4A, an RNA helicase, and Mnk1, the kinase responsible for phosphorylating eIF4E. Structure-guided surface mutagenesis and protein-protein interaction assays were used to identify binding sites for eIF4A and Mnk1 within the HEAT-repeats of 4G/C. p97/DAP5, a translational modulator homologous to eIF4G, lacks an eIF4A binding site in the corresponding region. The second atypical HEAT domain of the 4G/C binds Mnk1 using two conserved aromatic/ acidic-box (AA-box) motifs. Within the first AA-box, the aromatic residues contribute to the hydrophobic core of the domain, while the acidic residues form a negatively charged surface feature suitable for electrostatic interactions with basic residues in Mnk1. Introduction Initiation of protein synthesis in eukaryotes is mediated by a large group of eukaryotic translation initiation factors (eIFs) that recognize and prepare the mRNA, support recruitment of the small ribosomal subunit to the initiation codon, and promote joining of the large ribosomal subunit (Hershey and Merrick, 2000). In cap-dependent translation initiation, eIF4F, a trimolecular complex of eIF4E, eIF4G, and eIF4A, serves to recruit other components of the translational machinery. eIF4E recognizes the mRNA 50 cap structure (Sonenberg et al., 1979). eIF4A, an ATP-dependent RNA helicase (Rozen et al., 1990), is thought to unwind secondary structures in the 50 -untranslated region (UTR) of the mRNA to facil-
*Correspondence:
[email protected] 4 These authors contributed equally to this work. 5 These authors contributed equally to this work. 6 Present address: Department of Integrative Biology, University of California, Berkeley, Berkeley, California 94720. 7 Present address: SGX Pharmaceuticals, Inc., 10505 Roselle Street, San Diego, California 92121.
itate small ribosomal subunit binding. eIF4G, the subject of this paper, acts as a molecular bridge between eIF4E and eIF4A. In addition, eIF4G acts as a scaffold for recruitment of eIF3, which in turn binds the small ribosomal subunit. eIF4G also binds the poly(A) binding protein (PABP), allowing circularization of the mRNA, and the Mnk1 protein kinase, which phosphorylates eIF4E (reviewed by Gingras et al., 1999). Human eIF4G is a modular protein consisting of three regions, as determined by viral protease cleavage patterns, of approximately 500 amino acids each (Figure 1A) (Lamphear et al., 1995). The Tyr-X4-Leu-f binding site for eIF4E and the middle region (4G/M) are essential for cap-dependent translation. The C-terminal region (4G/C) is restricted to higher eukaryotes and is known to modulate protein synthesis, making it a target for a variety of proteins (Morino et al., 2000). Both 4G/M and 4G/C bind eIF4A (Imataka and Sonenberg, 1997). Although the biological significance of 4G/C binding to eIF4A is not well understood, it was previously shown that the 4G/C-dependent inhibition of 4G/M activity is relieved when 4G/C interacts with eIF4A (Morino et al., 2000). In higher eukaryotes, 4G/M and the eIF4E binding domain suffices for recruitment of eIF3 and the small ribosomal subunit to mRNAs (Morino et al., 2000). Simpler eukaryotes, such as yeast, lack the C-terminal region of eIF4G altogether. 4G/C is thought to play a modulatory role in translation initiation (De Gregorio et al., 1999; Morino et al., 2000). A mutation in the C-terminal region of human eIF4GI that abolishes eIF4A binding decreases ribosome binding by approximately 3- to 4-fold (Morino et al., 2000). 4G/C is also dispensable for capindependent or internal ribosome entry site (IRES)-mediated translation initiation, but it does potentiate UV cross-linking of eIF4A to the IRES (Pestova et al., 1996). Mnk1 binding to eIF4G is required for efficient phosphorylation of eIF4E (Ser209 in human eIF4E), because it recruits the kinase to the eIF4F complex (Pyronnet et al., 1999; Waskiewicz et al., 1997, 1999). Stimulation of cap-dependent mRNA translation by mitogens, growth factors, serum, and other nutrients is correlated with phosphorylation of eIF4E (Gingras et al., 1999), which decreases the affinity of eIF4E for the cap structure (Scheper et al., 2002; Zuberek et al., 2003). Mnk1 activation depends on the ERK and p38 MAP kinase signaling pathways (Fukunaga and Hunter, 1997; Waskiewicz et al., 1997). Phosphorylation of a mutant form of eIF4E that does not bind to eIF4G is severely impaired in cells (Pyronnet et al., 1999), and overexpression of a kinase-inactive Mnk1 mutant decreases eIF4E phosphorylation w8-fold, and cap-dependent translation by w20-fold (Cuesta et al., 2000). p97 (Imataka et al., 1997) (also known as DAP5; LevyStrumpf et al., 1997) is a modulator of translation initiation with homology to the middle and the C-terminal regions of eIF4G (Figure 1). This eIF4G paralog lacks the Tyr-X4-Leu-f binding site for eIF4E. It can, however, bind eIF4A, Mnk1, and eIF3, and may function like a dominant negative inhibitor of translation initiation by acting as a molecular mimic of eIF4G (Imataka et al.,
Structure 914
Figure 1. Domain/Motif Organization of eIF4G (A) Comparison between eIF4G and p97/DAP5 domain/motif organization in different species. Colored blocks identify homologous regions with % sequence identity to human eIF4GI (4G/M, blue; phosphoregion, orange; 4G/C1, green; 4G/C2, red). The small purple box denotes the eIF4E binding site and the yellow boxes refer to AA-boxes. Cleavage sites and interaction regions identified for human eIF4GI and p97/DAP5 are also shown. Fly eIF4G and p97/DAP5 lack canonical AAboxes. (B) C-terminal region of human eIF4GI with three phosphorylation sites indicated. The green and red blocks correspond to the two HR domains (4G/C1 and 4G/C2). The bars above them indicate the segments required for interaction with eIF4A and Mnk1, respectively.
1997; Yamanaka et al., 1997), but it has also been shown to stimulate IRES-dependent translation (Henis-Korenblit et al., 2002; Nevins et al., 2003). Unlike eIF4G, the C-terminal (4G/C-like) domain of p97/DAP5 cannot form a stable complex with eIF4A (Imataka and Sonenberg, 1997). The clinical importance of p97/DAP5 is not well documented, but reduced protein levels, secondary to aberrant RNA editing when apolipoprotein B mRNAediting enzyme-1 is overexpressed, does appear to contribute to oncogenesis (Yamanaka et al., 1997). Large portions of eIF4G appear to be unstructured, as judged by primary sequence analyses and limited proteolysis studies (data not shown), effectively preventing an X-ray crystallographic approach to the full-length protein. The Tyr-X4-Leu-f region, within the N terminus of eIF4G, undergoes a disorder-to-order transition when it binds to eIF4E (Gross et al., 2003; Marcotrigiano et al., 1999). The proteolytically resistant core of 4G/M forms a well-defined globular a-helical domain consisting of five HEAT repeats (HRs) (Marcotrigiano et al., 2001). In this paper, we report the X-ray structure of the C-terminal region of eIF4G, which consists of two atypical HR domains. Complementary biophysical stud-
ies were used to identify the binding site for eIF4A, and the functions of 4G/C and p97/DAP5 in higher eukaryotes are discussed in some detail. Results and Discussion Crystallization and Structure Determination Sequence comparisons (Figure 1A), secondary structure predictions (data not shown), and limited proteolysis combined with mass spectrometry (data not shown) permitted identification of a protease-resistant C-terminal portion of human eIF4GI that supports binding to eIF4A and Mnk1 (data not shown). Further truncation (eIF4GI[1235–1572]) yielded a two-domain protein that gives high-quality crystals with two protomers per asymmetric unit (Experimental Procedures). The X-ray structure of 4G/C was determined via multiwavelength anomalous dispersion (MAD) with Se-Met protein (Hendrickson, 1991). Experimental phases obtained at 2.6 A˚ resolution (Table 1) gave a poor-quality electron density map, which was improved by multiple cycles of density modification and phase combination. The current refinement model has an R factor of 24.4% and a free R value
C-Terminal Region of eIF4G 915
Table 1. Statistics of the Crystallographic Analysis
Data Set MAD analysis (14 sites) l1 (0.97973 A˚) l2 (0.97965 A˚) l3 (0.96410 A˚) Overall MAD figure of merit = 0.54 Native
Resolution (A˚) (Overall/Outer Shell)
No. of Reflections (Measured/Unique)
CI/sD
Completeness (%) (Overall/Outer Shell)
Rsym (%)a (Overall/Outer Shell)
Phasing Powerb
28.0–2.60/2.69–2.60 28.0–2.60/2.69–2.60 28.0–2.60/2.69–2.60
119,486/21,256 120,947/21,469 119,449/21,408
5.9 6.2 6.0
88.4/57.8 89.3/67.2 89.1/66.3
8.8/18.4 9.7/16.7 8.0/15.2
3.5 3.7 —
30.0–2.24/2.31–2.24
259,674/36,093
9.0
96.4/95.0
7.7/20.0
jFj/s(jFj) cutoff = 0
Rf = 24.4%
Rfreec = 29.2%
Protein: 5150 Bond lengths: 0.006 A˚
Solvent: 224 Bond angles: 1.18 º
Refinement Statistics Resolution: 30–2.24 A˚ No. of atoms Rmsd
Thermal parameters: 4.3 A˚
Rsym = SjI 2 CIDj/SI, where I = observed intensity, CID = average intensity obtained from multiple observations of symmetry related reflections. Phasing power = rmsd (jFHj/E), where jFHj = heavy atom structure factor amplitude and E = residual lack of closure. c Rfree was calculated with 5% of data, in thin resolution bins, omitted from the structure refinement. a
b
of 29.2% (Brunger, 1992) at 2.24 A˚ resolution with excellent stereochemistry (Table 1).
ing proteins (4E-BPs) and eIF4GI (Gingras et al., 1999; Raught et al., 2000).
Structural Overview The C-terminal region of human eIF4GI consists of two a-helical domains, 4G/C1 and 4G/C2, joined by a 13 residue linker (Figures 1B and 2). Both are HR domains, an all-helical fold observed in a number of scaffolding proteins, such as the PR65/A subunit of protein phosphatase 2A (PP2A) (Groves et al., 1999), importin-b (Chook and Blobel, 1999; Cingolani et al., 1999), and 4G/M (Marcotrigiano et al., 2001). HRs derive their name from the first four proteins in which this characteristic sequence pattern was identified: Huntingtin, EF3, PR65/A subunit of PP2A, and Tor1 (Andrade and Bork, 1995). It is remarkable that HRs have been detected in a substantial number of proteins involved in translation, including the eIF4Gs, p97/DAP5, poly(A) binding protein interacting protein-1 (Paip-1), prokaryotic EF3, GCN1, and FRAP/ mTOR, which regulates phosphorylation of eIF4E bind-
4G/C1 and 4G/C2 Are Atypical HEAT Domains Canonical HEAT domains contain a large number of consecutively packed HRs, resulting in right-handed superhelical or solenoid structures with extended quasicylindrical shapes. The HEAT domains of 4G/C are much more compact (Figure 2A): 4G/C1 contains five HRs (overall dimensions 30 A˚ 3 30 A˚ 3 50 A˚) and 4G/C2 contains three and one half HRs (30 A˚ 3 30 A˚ 3 30 A˚). In all, there are 17 lengthy a helices (H1–H17), comprising eight and one half HRs (Figure 3), plus several short a and 310 helices in loops within and between HRs, and a short b-ribbon connecting a helices H13 and H14. In a typical HEAT domain, most interrepeat twist angles are close to +20º, with occasional values of w 245º, as observed in the PR65/A regulatory subunit of PP2A (Groves et al., 1999). The HEAT domain of 4G/M (Marcotrigiano et al., 2001) contains five repeats, with Figure 2. Structures of 4G/C and 4G/M (A) Two orthogonal views are shown. The two HR domains are colored as in Figure 1. Individual HRs are colored alternatively in dark and light tones. Ten residues in the interdomain linker of 4G/C (left) were not observed in the electron density maps and are shown as a polyalanine trace (violet). (B) Structure of 4G/M (right). Two unobserved loops in 4G/M are not shown.
Structure 916
Figure 3. Sequence Alignment of the eIF4G and p97/DAP5 Superfamily in the C-Terminal Region Positions identical to human eIF4GI in at least two other sequences are colored dark blue, and other conserved positions are colored light blue. Information about human eIF4GI is indicated above the sequences: numbering, secondary structure elements (colored green and red, as in Figure 2; cylinders represent a helices and arrows represent b strands), and PSIpred secondary structure prediction (helices are shown as boxes). Below the sequences, the numbering and PSIpred prediction for human p97/DAP5 are shown. Figure prepared with ALSCRIPT (Barton, 1993).
two interrepeat twist angles close to +20º and two close to 245º (Figure 2B). These values are characteristic of preferred helix-helix packing arrangements that permit the so-called knobs-in-holes type interactions (Bowie, 1997; Chothia et al., 1981). Both HEAT domains of 4G/C utilize atypical interrepeat twist angles that are exclusively approximately 245º, which explains why the X-ray structures of 4G/C and other HEAT domains cannot be superimposed on one another. A search using the DALI Server (Dietmann and Holm, 2001; Holm and Sander, 1993) revealed only one structurally similar entry in the Protein Data Bank (PDB; http://www.rcsb.org/pdb). The first three HRs of 4G/C1 can be superimposed on the a-helical core of GTPase activating protein (GAP) 120 (PDB code: 1WER; rmsd = 3.6 A˚; 108 a-carbon pairs). This GAP is not, however, a HEAT protein, because the six superimposed a helices are not consecutive within the GAP120 polypeptide chain. The first three HRs of 4G/C1 and 4G/C2 are true structural paralogs of one another (rmsd = 3.0 A˚ ; 111 a-carbon pairs). Therefore, we propose that the structures of 4G/C1 and 4G/C2 represent a new subclass of HEAT proteins, which exclusively uti-
lize interrepeat twist angles of approximately 245º, giving rise to more compact domain shapes that are distinct from the +20º quasicylindrical HEAT domains. The paired a helices found within 4G/C1 and 4G/C2 occur in similar local environments; each a helix within the body of an atypical HEAT domain is packed among three antiparallel a helices: one from the same repeat and one each from the flanking HRs. Most of the stabilizing interhelical contacts involve hydrophobic residues that lie within the two and one half a-helical turns that comprise the hydrophobic core of the atypical HEAT domain. Unlike typical HRs, the long a helices of 4G/C1 and 4G/C2 are straight, making the first and second halves of each HR quasiequivalent. From structural superpositions of individual a helices, we were able to deduce a consensus sequence for an atypical HR (Figure 4). The proposed atypical HR consensus sequence contains three clusters of hydrophobic residues that map to the inner-facing half of the helix surface, over two and one half helical turns. The three clusters correspond to three layers of interhelix contacts within an HR. Together with additional contacts to the flanking HRs, they build up the hydrophobic core of the domain. The opposite
C-Terminal Region of eIF4G 917
Figure 4. Structure-Based Alignment of Atypical HEAT Repeats The sequence of the 17 lengthy a helices (black boxes) comprising the atypical HRs in 4G/C1 and 4G/C2 are structurally aligned. Other small helices and the b ribbon are indicated by red boxes. H18, the putative C-terminal a helix that was omitted from the expression construct, is colored blue. Where appropriate, a consensus character of each position is indicated by the background color (yellow, large hydrophobic; green, small hydrophobic; blue, polar). Conserved aromatic and acidic residues in the AA-boxes are colored in bold magenta and red, respectively.
face of the helix faces the solvent, and frequently includes a small side chain (Ala or Gly) that interacts with a side chain from the previous HR. This semiconserved position occurs between the first and second hydrophobic clusters (Figure 4). In the same face, but on the following helical turn, there is a conserved polar residue that lies between the second and third hydrophobic clusters of the consensus sequence. The C-terminal HR of 4G/C1 has a unique structure. While a helix H9 follows the geometric pattern of previous repeats, H10 is shorter, packs loosely against the remainder of the domain, and is flanked by two small 310 helices, H9b and H10b. DALI Server searches of the PDB revealed no structurally similar motifs. The second atypical HEAT domain (4G/C2) contains four complete HRs followed by a single hydrophobic a helix, H17. Our structure was obtained from crystals of a truncated form of the C terminus of eIF4GI lacking the last 28 residues. We think it likely that the missing polypeptide chain segment forms a terminal 18th a helix, which would pack against 2 conserved aromatic residues on the surface of H17 (Tyr1551 and Trp1564; Figure 5). The sequence of the putative a helix, H18, conforms to the consensus for the atypical HRs defined in Figure 4, and we believe that a full-length 4G/C2 contains four HRs. The linker connecting domains 4G/C1 and 4G/C2 is both highly polar and apparently flexible (10/13 residues were not visible in our electron density maps). The interdomain interface observed in our crystals is very small (275 A˚2). Moreover, the apposed surfaces are not particularly hydrophobic, and the interfacial residues are not conserved among eIF4Gs from different organisms. We suggest that the interdomain interface and the relative spatial arrangement of the two domains observed in
our 4G/C crystals result from lattice packing interactions and do not reflect the solution behavior of the protein. Although the structural similarity of 4G/C1 and 4G/C2 suggests that both domains might coalesce to form a single atypical HEAT domain in solution, we think this possibility unlikely. The exposed face of the first HR of 4G/C2 does not fit the consensus depicted in Figure 4. A similar suggestion could be made concerning potential interactions between the last HR of 4G/M (Marcotrigiano et al., 2001) and the first HR of 4G/C1. However, the N-terminal face of 4G/C1 is hydrophilic, so it appears unlikely that it could form a stable interaction with the C-terminal face of 4G/M and generate a continuous linear array of HRs extending throughout the C terminus of eIF4G. It is, of course, possible that interactions between eIF4G and one or more of its binding partners (i.e., eIF4A, Mnk1, etc.) could stabilize a structurally unique supramolecular assembly of HEAT domains and other proteins. Comparative Protein Structure Modeling with 4G/C1 and 4G/C2 In an attempt to identify other atypical HR proteins within the 4G/C subfamily, we used ModWeb (Sanchez et al., 2000), an automated modeling software pipeline based on MODELLER (Sali and Blundell, 1993). Reliable models were produced for 48 protein sequences (model score > 0.7), for which no three-dimensional structural information was previously available. Not surprisingly, 32 of these sequences were previously annotated as eIF4Gs, putative eIF4Gs, or eIF4G paralogs (p97/DAP5 and PDCD4). Models were also obtained for 5 putative topoisomerases and for 11 other hypothetical protein sequences of unknown function from eukaryotes. The typical HEAT domains in Paip-1, the middle region of eIF4G, the N-terminal region of p97/DAP5, and other HEAT proteins, were not selected. All results from our comparative protein structure modeling exercise with the C-terminal portion of eIF4GI are available from ModBase (http:// alto.rockefeller.edu/modbase/; Sanchez et al., 2000). A phylogenetic tree for the 4G/C family was constructed from our structure-based sequence alignment (data not shown). Not surprisingly, proteins involved in translation initiation (i.e., eIF4GI and II, and p97/DAP5) cluster together. Homology Modeling of p97/DAP5 Our homology model of the C-terminal portion of p97/ DAP5 merits further discussion, because 4G/C is closely related to it (27% identical in amino acid sequence). There are significant differences between 4G/C and p97/DAP5 within the linker connecting the two atypical HEAT domains (Figure 3). Not only is the H10-H11 loop in p97/DAP5 (5 versus 13 residues) too short to permit the arrangement of the two HEAT domains observed in 4G/C, but it is also markedly hydrophobic and is probably not solvent exposed. We suggest that the entire Cterminal region of p97/DAP5 could form a continuous HEAT domain. Aromatic and Acidic Boxes in eIF4G and p97/DAP5 Four proteins involved in eukaryotic translation initiation (eIF4G, p97/DAP5, eIF2B3, and eIF5) contain tandem conserved motifs known as AA (aromatic and acidic
Structure 918
Figure 5. Structure of the First AA-Box of eIF4GI (A) Ribbon representation of the C terminus of 4G/C2. Side chains conserved in AA-boxes are shown: universally conserved aromatic (magenta) and acidic (red) residues, and other conserved hydrophobic residues (yellow). (B) (2jFobservedj 2 jFcalculatedj) electron density map and refined atomic model for the region framed in (A). Atoms are color coded in yellow, red, blue and green. (C) Sequence alignment for the AA-box region. Selected representatives of the four proteins bearing AA-boxes are included: eIF4G, p97/DAP5, eIF2B3, and eIF5. Numbering and secondary structure elements of human eIF4GI are indicated above the sequences, with the same color scheme as in Figure 2, and the predicted a helix H18 is indicated as a black box. Conserved aromatic and acidic residues are boxed in magenta and red, respectively. Other conserved positions are boxed orange (acidic) and yellow (hydrophobic).
residues)-boxes at their C termini (Asano et al., 1999; Koonin, 1995). These motifs mediate specific binding to proteins containing one or more segments of positively charged residues: eIF4G and p97/DAP5 bind to Mnk1/Mnk2 (Pyronnet, 2000; Waskiewicz et al., 1999), and eIF5 and eIF2B3 bind to eIF2 (Asano et al., 1999). eIF2 contains three segments of about seven lysine residues each (K-boxes), whereas Mnk1/Mnk2 contain a single stretch of eight basic amino acids, composed of Arg and Lys (RK-box). Both AA-boxes in eIF4G are necessary for binding to human Mnk1 (Morino et al., 2000). To our knowledge, our work provides the first structure of an AA-box, which extends from the middle of a helix H16 to the end of a helix H17 (Figure 5). With the exception of Tyr1551 and Tyr1564, the eponymous aromatic residues contribute to the hydrophobic core. As expected, the acidic residues occur in the loop connecting a helices H16-H17 and surrounding positions, where they cluster to form an elongated, solvent-
exposed acidic patch suitable for interactions with either K- or RK-boxes. The aromatic residues in the second AAbox map to the end of putative a helix H18, which is followed by five acidic residues (Figure 5). If residues 1577–1591 do indeed form an a helix (H18) that pairs with H17 to generate a complete C-terminal HR, acidic residues 1592–1598 would lie close to their negatively charged counterparts in the H16-H17 loop. Each AA-box found in eIF4G and p97/DAP5 encompasses 14 conserved hydrophobic residues, 4 of which are aromatic (Figure 5C). The number of acidic residues varies: 3–6 in the H16-H17 loop (positions 1552–1554 and 1557–1559 in human eIF4GI), and 5–14 following the putative a helix H18. Although our structure of 4G/C does not include both AA-boxes, it does provide some insight into the function(s) of these conserved hydrophobic/acidic residues. For example, our homology model of Drosophila melanogaster eIF4G does not display the negatively charged surface feature overlying the first AA-box of human 4G/C.
C-Terminal Region of eIF4G 919
Unstructured Regions of the eIF4Gs and p97/DAP5 In mammalian eIF4Gs, the divergent linker connecting domains 4G/M and 4G/C1 appears to be unstructured (data not shown). This 155 residue segment contains a cleavage site for Caspase-3 (Asp-X-X-Asp1176 in human eIF4GI) that mediates inactivation of eIF4G during apoptosis (Bushell et al., 2000). Within this same region in human eIF4GI, but not eIF4GII and p97/DAP5, three serum-stimulated phosphorylation sites have been mapped to Ser1148, Ser1188, and Ser1232 (Raught et al., 2000). The relationship, if any, between the phosphorylation state of eIF4GI and apoptosis remains to be established. We suggest that eIF4GI and eIF4GII carry out similar biochemical roles, which are largely mediated by the conserved HEAT domains (4G/M, 4G/C1, and 4G/C2). Distinct biological functions appear to be mediated by divergent unstructured regions that are targets for both phosphorylation and proteolysis. Surface Mutagenesis of 4G/C Previously published scanning deletion mutagenesis studies of human eIF4GI mapped the minimal eIF4A binding region of 4G/C to residues 1241–1451 (Morino et al., 2000). In our structure, this segment corresponds to the first of the two HEAT domains (4G/C1). Surface accessible residues of 4G/C1 that are conserved amongst mammalian eIF4Gs map to one face of this atypical HEAT domain (front face in the left panel of Figure 6B). Triple-alanine mutagenesis studies with human eIF4GI suggest that this conserved face is responsible for binding eIF4A. Phe-Val-Arg1281/Ala-Ala-Ala, which disrupts eIF4A binding, maps to the conserved face, whereas Lys-Lys-Val1393/Ala-Ala-Ala on the opposite face has no effect (Morino et al., 2000). Phe1279 and Val1280 are, however, completely buried in the hydrophobic core, and loss of eIF4A binding could be explained by incorrect folding of 4G/C1. A structure-based mutational analysis of 4G/C1 was performed to identify solvent exposed amino acids that support recognition of eIF4A (Experimental Procedures; Figure 6C). The entire C-terminal region of human eIF4GI (residues 1080–1600) was used as a positive control for eIF4A binding (Figure 6D, lane 2). Initially, we focused our attention on Arg1281, because it had been previously implicated in eIF4A binding by Morino et al. (2000). Arg1281 is also an attractive candidate for mutagenesis, because it is surrounded by four conserved acidic residues (Glu1240, Glu1244, Glu1285, and Glu1322). We reasoned that Arg1281 acts as an organizer of the glutamate cluster, which would otherwise be unstable due to electrostatic repulsion. The resulting negative electrostatic potential patch (Figure 6A) may represent a surface feature for binding of eIF4A, which is positively charged. The Arg1281/Met substitution was designed to preserve packing interactions with the aliphatic portion of Arg1281, while eliminating the positive charge and its predicted role in stabilizing the observed glutamate tetrad. This mutant protein showed reduced affinity for eIF4A (Figure 6D; compare lane 3 to lane 2). Removal of a single glutamic acid residue from the cluster (Glu1285/Lys) had little, if any, effect on eIF4A binding (Figure 6D, lane 4), suggesting that removal of more than one negatively charged side chain is required to perturb favorable electrostatic interactions with eIF4A.
Other candidates for site-directed mutagenesis were selected from the results of a surface dissimilarity analysis that compared the structure of human 4G/C1 with our homology models of the corresponding domains in other eIF4Gs and p97/DAP5 (Experimental Procedures). We reasoned that surface features common to the eIF4Gs that are not conserved in p97/DAP5 represent putative eIF4A binding sites. While a mutant of a solvent-accessible residue in a helix H5 (Leu1326/Trp) had minimal effect on eIF4A binding (Figure 6D, lane 5), the Glu1379/Lys mutant in a helix H8 resulted in a clear increase in eIF4A binding (lane 6). It is possible that the Glu1379 residue interacts with a negative potential surface area of eIF4A. The loop connecting a helices H7 and H8 includes a helix H7b, which corresponds to a five residue gap in the p97/DAP5 family. To test the relevance of this loop in the 4G/C::eIF4A interaction, we designed a mutant that substitutes the Thr1364–Ala1373 residues by tripeptide Gly-Ala-Gly with no strain imposed in the surrounding a helices. The corresponding mutant showed a dramatic reduction in eIF4A binding (Figure 6D; compare lane 7 to lane 2). Finally, we examined the last HR of 4G/C1 (a helices H9 and H10). The Cterminal boundary of the eIF4A binding region of eIF4G was previously mapped to a polypeptide chain segment between residues 1410 (located N-terminal to a helix H10) and 1451 (located in a helix H11 within 4G/C2) (Morino et al., 2000). Given that 4G/C2 is not thought to be involved in eIF4A binding, and candidate residues in a helix H10 were previously evaluated for eIF4A binding and were ruled out (mutant Phe-Val-Ala-Glu-Gln1421/AlaVal-Ala-Ala-Ala [Morino et al., 2000]), we focused our attention on the loop between a helices H9 and H10. Substitution of this loop with a tripeptide (mutant Leu1408–Gln1412/Gly-Ala-Gly) yielded an internal deletion that does not bind to eIF4A (Figure 6D; compare lane 8 to lane 2). In summary, previous mutational analyses and our dissimilarity index (DI) mapping exercise identified a number of putative eIF4A binding regions. Our work ruled out some of the predicted eIF4A binding sites (Figure 6C, yellow), and identified three previously unknown surface patches (Figure 6C, magenta) that support eIF4A binding. Two of these confirmed eIF4A binding sites (the Arg1281/glutamate tetrad and helix H7b, within the H7H8 loop), and are highly dissimilar in the p97/DAP5 family, which may explain why they do not bind eIF4A. Our third validated eIF4A binding site (the H9b-H10 loop) is relatively conserved among both eIF4G and p97/DAP5 family members. Conservation between the eIF4G and p97/DAP5 families also provides insights into binding of other parts of the translation machinery. For example, the H5-H6 loop (encompassing helix H5b) represents a conserved convex feature on the surface of 4G/C1 (Figure 6B, left panel) flanked by two residues (Leu1326 and Glu1379) that are highly dissimilar between eIF4G and p97/DAP5. Neither of these residues is required for eIF4A binding, and we propose that this part of the conserved surface of 4G/C1 participates in interactions with other components of the translation initiation machinery. Conclusions The C-terminal region of eIF4G contains two atypical HR structural domains. Its modulatory effect in the function
Structure 920
Figure 6. Surface Properties of 4G/C GRASP (Nicholls et al., 1993) surface representation of the two domains of 4G/C. The orientation of images on the left side of the panels is similar to Figure 2A, top. The a-carbon backbone is also shown. (A) Surface map of the electrostatic potential calculated at 0.15 M ionic strength. Negatively charged regions are colored in red and positively charged regions are colored in blue. (B) Surface map of the calculated DI indicating residues that are conserved in eIF4Gs but not conserved in other members of the superfamily. Solvent-accessible residues conserved among mammalian eIF4Gs are colored for DI (DI4G, see Experimental Procedures): green (low)/brown/red (high). (C) Sites of mutations that abrogate binding to eIF4A are shown in magenta, and those that do not affect binding are shown in yellow. 4G/C1 and 4G/C2 are colored blue and gray, respectively. (D) The pcDNA3-HA expression vector containing 4G/C (1080–1600) wild-type (lane 2) or mutants (Phe-Val-Arg1281/Ala-Ala-Ala, lane 1; Arg1281/Met, lane 3; Asp1285/ Lys, lane 4; Leu1326/Trp, lane 5; Glu1379/Lys, lane 6; Thr1364–Ala1373/ Gly-Ala-Gly, lane 7; Leu1408–Gln1412/GlyAla-Gly, lane 8) was transfected in 293T cells with pcDNA3-FLAG-eIF4A. Immunoprecipitation was performed with an anti-FLAG affinity resin, and retained proteins were resolved by SDS-PAGE and analyzed by Western blotting for the presence of HA (first panel) or FLAG-tagged protein (second panel). A total of 5% of the immunoprecipitated total cell extract was analyzed by SDS-PAGE, followed by Western blotting with anti-HA (third panel) and anti-FLAG (fourth panel) antibodies. Chemiluminescence was quantitated using a LAS-1000 Plus (Fuji) luminescent image analyzer. (E) eIF4A binding is reported as the relative amount of immunoprecipitated 4G/C compared to the level of 4G/C present in the total cell extract.
of the eIF4F complex is related to molecular interactions with eIF4A and Mnk1. Three adjacent eIF4A binding sites on the surface of 4G/C1 were identified: Arg1281 and the surrounding glutamate tetrad, the loop containing a helix H7b (Thr1364–Ala1373) and loop H9-H10 (Leu1408–Gln1412). Domain 4G/C2, responsible for re-
cruiting Mnk1 to the eIF4F complex, contains two AAboxes at the C terminus. The conserved aromatic residues contribute to the hydrophobic core, and the conserved acidic residues are solvent exposed, making up an acidic cluster that most likely interacts with positively charged motifs in Mnk1. Our results allow a better
C-Terminal Region of eIF4G 921
understanding of the intricate net of protein-protein interactions that occur between the three structural domains of eIF4G, 4G/M, 4G/C1, and 4G/C2, and three proteins, eIF4A, eIF4E, and Mnk1, which together play a role in the activation of the translation initiation process. Experimental Procedures Protein Characterization, Expression, and Purification Limited proteolysis of the C-terminal region of human eIF4GI (residues 1080–1600) with a panel of specific endoproteases yielded a resistant core of approximately 40 kDa, as judged by gel electrophoresis (data not shown). Matrix-assisted laser desorption ionization mass spectrometry and N-terminal sequencing of proteolytic fragments mapped a large number of proteolytic cleavage sites to residues 1080–1234. Fragments of human eIF4GI and eIF4GII encompassing the proteolytically resistant C-terminal core, each with slightly different N and C termini, were overexpressed in Escherichia coli BL21(DE3), as determined with a modified pET-28 (Novagen) vector that furnished a hexahistidine tag at the N terminus of the protein, followed by a rhinoviral protease cleavage site. Following Q-Sepharose and Ni2+-chelating chromatographies, the affinity tag was removed with GST-PreScission Protease (Pharmacia), and the protein was purified to homogeneity by Ni2+-chelating and Superdex200 gel filtration chromatographies. The desired protein concentrated to 60 mg/ml in 0.1 M KCl, 20 mM TrisHCl (pH 7.5), 0.5 mM EDTA, and 5 mM DTT for crystallization. Se-Met protein was expressed in E. coli as described by Marcotrigiano et al., 2001. Crystallization Hanging-drop vapor diffusion of protein (60 mg/ml in 0.1 M KCl, 20 mM TrisHCl [pH 7.5], 0.5 mM EDTA, and 5 mM DTT) against a reservoir solution containing 30% PEG MME 5000, 0.1 M MES (pH 6.5), 0.2 M ammonium sulfate at 4ºC, yielded plate-like crystals of human eIF4GI (residues 1235–1572, typical dimensions of 0.040 mm 3 0.150 mm 3 0.500 mm) in the monoclinic space group P21 (unit cell: a = 78.6 A˚, b = 50.6 A˚, c = 99.9 A˚, b = 101.6º; 2 protomers/asymmetric unit). B-centered translational pseudosymmetry was evident from systematic weakness of the odd (h + l) parity class of reflections and the presence of an origin-like peak in the native Patterson synthesis at u = 1/2, v = 0, w = 1/2. Structure Determination and Refinement Se-Met protein was prepared and crystallized by the same methods described for the S-Met protein, and complete substitution of all 8 Met residues was confirmed by electrospray mass spectrometry (data not shown). A MAD experiment was conducted by measuring diffraction data in the vicinity of the Se absorption edge by beamline ID-32 at the Advanced Photon Source (Table 1). Data obtained at each X-ray wavelength were integrated and scaled with Scalepack/HKL (Otwinowski and Minor, 1997), and 14 of 16 possible Se sites were located with SnB (Weeks and Miller, 1999). SHARP (de La Fortelle and Bricogne, 1997) phase determination followed by density modification with DM (CCP4, 1994) yielded a partially interpretable experimental electron density map. Visual inspection of the map showed two a-helical molecules within the asymmetric unit, related by B-centered pseudosymmetry. Polyalanine a helices were built with O (Jones et al., 1991) to give a partial model for phase combination that yielded an interpretable electron density map. Only one molecule per asymmetric unit was built, and ideal noncrystallographic symmetry (NCS) operators were used in the initial stages of refinement employing CNS (Brunger et al., 1998). After several iterative cycles of model building and refinement, the model gave an R factor of 28.6% with R free = 30.9%. A set of reflections for cross-validation testing (5% of total) was selected with DATAMAN (Kleywegt and Jones, 1996) to prevent ‘‘NCS-related’’ reflections from appearing in both the test and work sets. At the final stages of refinement, NCS operators were re-evaluated, and restrained NCS refinement yielded a final model (residues 1234–1566, excluding 1428–1437 from the interdomain linker, rmsd = 0.2 A˚ for 323 a-carbon pairs) with an R factor of 24.4% and an R free of 29.2% (Table 1). PROCHECK (Laskowski et al., 1993) gave an overall G factor of 0.45 with excellent stereochemistry.
4G/C Superfamily Sequence Alignments and Dissimilarity Index Calculations Alignment of the sequences of the C-terminal regions of eIF4G and p97/DAP5 from various organisms were complicated by ambiguities between residues 1347 and 1501 in human eIF4GI (i.e., between a helices H6 and H14 of 4G/C). A convincing alignment was obtained following recourse to secondary structure predictions for p97/DAP5 using PSIpred (McGuffin et al., 2000). The resulting superfamily alignment was used to calculate the DI4G of the eIF4G family to the eIF4G + p97/DAP5 superfamily. For that purpose, the eIF4G family included human eIF4GI, human eIF4GII, and rabbit eIF4GI, and the superfamily included these three sequences plus human p97/DAP5. Fly sequences were omitted, because they have low similarity to other members of their families and added noise to the calculation. Conservation values were calculated at each position for the eIF4G family (C4G) and the superfamily (CS), independently, with the BLOSUM62 substitution matrix. The DI was defined in the simplest way: DI4G = C4G 2 CS. Positions with gaps in p97/DAP5 were assigned a CS = 0. Tests showed that the results are robust with respect to the choice of substitution matrix or the definition of DI4G, but they are obviously very sensitive to the superfamily alignment employed. Comparative Protein Structure Modeling We aimed to attribute three-dimensional structural information to as many amino acid sequences as possible, by using our 4G/C structure as a template for protein fold assignment and comparative protein structure modeling. The nonredundant TrEMBL database (Bairoch and Apweiler, 1999) was searched with the amino acid sequence of the C-terminal region of human eIF4GI by c-BLAST (Altschul et al., 1997). A total of 210 protein sequences matched the C-terminal region over at least 57 residues with an E value better than 100. By using the c-BLAST alignments, comparative protein structure models for these sequences were built by satisfaction of spatial restraints, as implemented in the program MODELLER (Sali and Blundell, 1993). The resulting models were evaluated by a model assessment procedure that relies on a Z score calculated from a statistical energy function, sequence similarity on which the model is based, and estimates of model compactness. Control calculations demonstrate that this approach has a false-positive rate of about 4% (Sanchez and Sali, 1998). Preparation of eIF4G Mutants and Assays of eIF4A Binding eIF4GI point mutants were generated using polymerase chain reaction cloning with oligonucleotides containing EcoRI (50 oligo) and XhoI (30 oligo) restriction enzyme sequences, which were used to subclone them into the pcDNA3-HA vector (Imataka et al., 1998). All the constructs were made in the context of residues 1080–1600 of eIF4GI. Wild-type eIF4A was inserted into the pcDNA3-FLAG vector (Imataka et al., 1998). All inserts were sequenced to confirm the presence of the correct mutation. Cationic lipid reagent (20 ml of Lipofectamine; Life Technologies) was diluted in serum-free media (Opti-MEM; Life Technologies) to cotransfect eIF4GI mutants and wild-type eIF4A into human embryonic kidney 293T cells (100 mm dish). Following 5 hr incubation, the medium was replaced with Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Transfected cells were harvested in phosphate-buffered saline (PBS) 24–36 hr after the addition of serum-containing media. Following centrifugation, harvested cells were resuspended in 600 ml of lysis buffer (20 mM HEPESKOH [pH 7.6], 200 mM KCl, 0.5 mM EDTA, 10% glycerol, 1% Triton X-100, 50 mg/ml RNase A, and protease inhibitor cocktail [Complete; Roche]) and lysed by repeated freeze/thaw cycles. Cell debris was pelleted by centrifugation, and the protein concentration in the supernatant was determined using the Bio-Rad assay. For coimmunoprecipitation experiments, 200 ml of total cell extract was diluted to 1 ml and precleared for 1 hr at 4ºC with 25 ml of protein A Sepharose. The reaction mixture was centrifuged, and the resulting supernatant transferred into a fresh tube with 25 ml of Anti-FLAG M2-Agarose affinity gel (Sigma), which had been precoated for 1 hr with bovine serum albumin (100 mg/ml). The mixture was incubated for 1 hr at 4ºC. Following incubation, beads were washed twice with lysis buffer and a third time with lysis buffer containing 300 mM KCl. Immunoprecipitates were eluted in 23 Laemmli sample buffer, and proteins were subjected to SDS-PAGE.
Structure 922
Polypeptides were resolved on SDS-10% polyacrylamide gels and transferred onto a 0.45 mm nitrocellulose membrane. The membrane was blocked for 16 hr at 4ºC with 5% milk in PBS containing 0.5% Tween 20. The membrane was incubated for 90 min with monoclonal anti-HA (1:5000; Babco) or anti-FLAG (1:5000; Sigma). Incubation with secondary antibody was performed with horseradish peroxidase-coupled sheep anti-mouse Ig (1:5000; Amersham). Detection was performed with ECL (NEN), and the signal was quantitated with a luminescent image analyzer LAS-1000 Plus (Fuji). Acknowledgments We thank Dr. J.B. Bonanno for assistance with X-ray data measurement and phasing, Dr. G. He for preparation of eIF4G expression clones, and Ms. T.B. Niven for editorial assistance. We are also grateful to Drs. G. Bricogne, J. Cabral, R. Deo, C. Edo, A.-C. Gingras, D. Jeruzalmi, J. Kuriyan, D. Lubell, J. Marcotrigiano, M. Miron, S.K. Nair, G.A. Petsko, and B. Raught for useful discussions, and Drs. S. Wasserman and K. D’Amico for access to the COM-CAT beamline. N.S. is a CIHR Distinguished Investigator and a Howard Hughes Medical Institute International Scholar. This work was supported by National Institutes of Health grant GM61262 to S.K.B., and a grant from the CIHR to N.S. L.B. was supported by a Ministerio de Educacion y Ciencia (Spain) and a Human Frontiers Science Program fellowship. F.P. was supported by a Doctoral Research Award from the CIHR. Received: August 17, 2005 Revised: March 20, 2006 Accepted: March 21, 2006 Published: May 16, 2006 References
Cingolani, G., Petosa, C., Weis, K., and Muller, C.W. (1999). Structure of importin-beta bound to the IBB domain of importin-alpha. Nature 399, 221–229. Cuesta, R., Xi, Q., and Schneider, R.J. (2000). Adenovirus-specific translation by displacement of kinase Mnk1 from cap-initiation complex eIF4F. EMBO J. 19, 3465–3474. De Gregorio, E., Preiss, T., and Hentze, M.W. (1999). Translation driven by an eIF4G core domain in vivo. EMBO J. 18, 4865–4874. de La Fortelle, E., and Bricogne, G. (1997). Maximum-likelihood heavy atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Methods Enzymol. 276, 472–494. Dietmann, S., and Holm, L. (2001). Identification of homology in protein structure classification. Nat. Struct. Biol. 8, 953–957. Fukunaga, R., and Hunter, T. (1997). MNK1, a new MAP kinase-activated protein kinase, isolated by a novel expression screening method for identifying protein kinase substrates. EMBO J. 16, 1921–1933. Gingras, A.C., Raught, B., and Sonenberg, N. (1999). eIF4 initiation factors: effectors of mRNA recruitment to ribosomes and regulators of translation. Annu. Rev. Biochem. 68, 913–963. Gross, J.D., Moerke, N.J., von der Haar, T., Lugovskoy, A.A., Sachs, A.B., McCarthy, J.E., and Wagner, G. (2003). Ribosome loading onto the mRNA cap is driven by conformational coupling between eIF4G and eIF4E. Cell 115, 739–750. Groves, M.R., Hanlon, N., Turowski, P., Hemmings, B.A., and Barford, D. (1999). The structure of the protein phosphatase 2A PR65/A subunit reveals the conformation of its 15 tandemly repeated HEAT motifs. Cell 96, 99–110. Hendrickson, W.A. (1991). Determination of macromolecular structures from anomalous diffraction of synchrotron radiation. Science 254, 51–58.
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.J. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402.
Henis-Korenblit, S., Shani, G., Sines, T., Marash, L., Shohat, G., and Kimchi, A. (2002). The caspase-cleaved DAP5 protein supports internal ribosome entry site-mediated translation of death proteins. Proc. Natl. Acad. Sci. USA 99, 5400–5405.
Andrade, M.A., and Bork, P. (1995). HEAT repeats in the Huntington’s disease protein. Nat. Genet. 11, 115–116.
Hershey, J.W.B., and Merrick, W.C. (2000). Pathway and mechanism of initiation of protein synthesis. In Translational Control of Gene Expression, N. Sonenberg, J.W.B. Hershey, and M.B. Mathews, eds. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), pp. 33–88.
Asano, K., Krishnamoorthy, T., Phan, L., Pavitt, G.D., and Hinnebusch, A.G. (1999). Conserved bipartite motifs in yeast eIF5 and eIF2Bepsilon, GTPase-activating and GDP-GTP exchange factors in translation initiation, mediate binding to their common substrate eIF2. EMBO J. 18, 1673–1688.
Holm, L., and Sander, C. (1993). Protein structure comparison by alignment of distance matrices. J. Mol. Biol. 233, 123–138.
Bairoch, A., and Apweiler, R. (1999). The SWISS-PROT protein sequence data bank and its supplement TrEMBL in 1999. Nucleic Acids Res. 27, 49–54.
Imataka, H., and Sonenberg, N. (1997). Human eukaryotic translation initiation factor 4G (eIF4G) possesses two separate and independent binding sites for eIF4A. Mol. Cell. Biol. 17, 6940–6947.
Barton, G.J. (1993). ALSCRIPT: a tool to format multiple sequence alignments. Protein Eng. 6, 37–40.
Imataka, H., Olsen, H.S., and Sonenberg, N. (1997). A new translational regulator with homology to eukaryotic translation initiation factor 4G. EMBO J. 16, 817–825.
Bowie, J.U. (1997). Helix packing in membrane proteins. J. Mol. Biol. 272, 780–789. Brunger, A.T. (1992). The free R value: a novel statistical quantity for assessing the accuracy of crystal structures. Nature 355, 472–474. Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al. (1998). Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921. Bushell, M., Poncet, D., Marissen, W.E., Flotow, H., Lloyd, R.E., Clemens, M.J., and Morley, S.J. (2000). Cleavage of polypeptide chain initiation factor eIF4GI during apoptosis in lymphoma cells: characterisation of an internal fragment generated by caspase-3mediated cleavage. Cell Death Differ. 7, 628–636. CCP4 (Collaborative Computational Project, Number 4) (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50, 760–763. Chook, Y.M., and Blobel, G. (1999). Structure of the nuclear transport complex karyopherin-beta2-Ran x GppNHp. Nature 399, 208–210. Chothia, C., Levitt, M., and Richardson, D. (1981). Helix to helix packing in proteins. J. Mol. Biol. 145, 215–250.
Imataka, H., Gradi, A., and Sonenberg, N. (1998). A newly identified N-terminal amino acid sequence of human eIF4G binds poly(A)binding protein and functions in poly(A)-dependent translation. EMBO J. 17, 7480–7489. Jones, T.A., Zou, J.Y., Cowan, S.W., and Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47, 110–119. Kleywegt, G.J., and Jones, T.A. (1996). xdlMAPMAN and xdlDATAMAN—programs for reformatting, analysis and manipulation of biomacromolecular electron-density maps and reflection data sets. Acta Crystallogr. D Biol. Crystallogr. 52, 826–828. Koonin, E.V. (1995). Multidomain organization of eukaryotic guanine nucleotide exchange translation initiation factor eIF-2B subunits revealed by analysis of conserved sequence motifs. Protein Sci. 4, 1608–1617. Lamphear, B.J., Kirchweger, R., Skern, T., and Rhoads, R.E. (1995). Mapping of functional domains in eukaryotic protein synthesis initiation factor 4G (eIF4G) with picornaviral proteases: implications for cap-dependent and cap-independent translational initiation. J. Biol. Chem. 270, 21975–21983.
C-Terminal Region of eIF4G 923
Laskowski, R.J., MacArthur, M.W., Moss, D.S., and Thornton, J.M. (1993). PROCHECK: a program to check stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 283–290. Levy-Strumpf, N., Deiss, L.P., Berissi, H., and Kimchi, A. (1997). DAP-5, a novel homolog of eukaryotic translation initiation factor 4G isolated as a putative modulator of gamma interferon-induced programmed cell death. Mol. Cell. Biol. 17, 1615–1625. Marcotrigiano, J., Gingras, A.C., Sonenberg, N., and Burley, S.K. (1999). Cap-dependent translation initiation in eukaryotes is regulated by a molecular mimic of eIF4G. Mol. Cell 3, 707–716. Marcotrigiano, J., Lomakin, I.B., Sonenberg, N., Pestova, T., Hellen, C.U.T., and Burley, S.K. (2001). A conserved HEAT domain within eIF4G directs assembly of the translation initiation machinery. Mol. Cell 7, 193–203. McGuffin, L.J., Bryson, K., and Jones, D.T. (2000). The PSIPRED protein structure prediction server. Bioinformatics 16, 404–405. Morino, S., Imataka, H., Svitkin, Y.V., Pestova, T.V., and Sonenberg, N. (2000). Eukaryotic translation initiation factor 4E (eIF4E) binding site and the middle one-third of eIF4GI constitute the core domain for cap-dependent translation, and the C-terminal one-third functions as a modulatory region. Mol. Cell. Biol. 20, 468–477. Nevins, T.A., Harder, Z.M., Korneluk, R.G., and Holcik, M. (2003). Distinct regulation of internal ribosome entry site-mediated translation following cellular stress is mediated by apoptotic fragments of eIF4G translation initiation factor family members eIF4GI and p97/ DAP5/NAT1. J. Biol. Chem. 278, 3572–3579. Nicholls, A., Bharadwaj, R., and Honig, B. (1993). GRASP: graphical representation and analysis of surface properties. Biophys. J. 64, A166. Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326. Pestova, T.V., Shatsky, I.N., and Hellen, C.U. (1996). Functional dissection of eukaryotic initiation factor 4F: the 4A subunit and the central domain of the 4G subunit are sufficient to mediate internal entry of 43S preinitiation complexes. Mol. Cell. Biol. 16, 6870–6878. Pyronnet, S. (2000). Phosphorylation of the cap-binding protein eIF4E by the MAPK-activated protein kinase mnk1. Biochem. Pharmacol. 60, 1237–1243. Pyronnet, S., Imataka, H., Gingras, A.C., Fukunaga, R., Hunter, T., and Sonenberg, N. (1999). Human eukaryotic translation initiation factor 4G (eIF4G) recruits mnk1 to phosphorylate eIF4E. EMBO J. 18, 270–279. Raught, B., Gingras, A.C., Gygi, S.P., Imataka, H., Morino, S., Gradi, A., Aebersold, R., and Sonenberg, N. (2000). Serum-stimulated, rapamycin-sensitive phosphorylation sites in the eukaryotic translation initiation factor 4GI. EMBO J. 19, 434–444. Rozen, F., Edery, I., Meerovitch, K., Dever, T.E., Merrick, W.C., and Sonenberg, N. (1990). Bidirectional RNA helicase activity of eucaryotic translation initiation factors 4A and 4F. Mol. Cell. Biol. 10, 1134– 1144. Sali, A., and Blundell, T.L. (1993). Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779–815. Sanchez, R., and Sali, A. (1998). Large-scale protein structure modeling of the Saccharomyces cerevisiae genome. Proc. Natl. Acad. Sci. USA 95, 13597–13602. Sanchez, R., Pieper, U., Mirkovic, N., de Bakker, P.I., Wittenstein, E., and Sali, A. (2000). MODBASE, a database of annotated comparative protein structure models. Nucleic Acids Res. 28, 250–253. Scheper, G.C., van Kollenburg, B., Hu, J., Luo, Y., Goss, D.J., and Proud, C.G. (2002). Phosphorylation of eukaryotic initiation factor 4E markedly reduces its affinity for capped mRNA. J. Biol. Chem. 277, 3303–3309. Sonenberg, N., Rupprecht, K.M., Hecht, S.M., and Shatkin, A.J. (1979). Eukaryotic mRNA cap binding protein: purification by affinity chromatography on sepharose-coupled m7GDP. Proc. Natl. Acad. Sci. USA 76, 4345–4349. Waskiewicz, A.J., Flynn, A., Proud, C.G., and Cooper, J.A. (1997). Mitogen-activated protein kinases activate the serine/threonine kinases Mnk1 and Mnk2. EMBO J. 16, 1909–1920.
Waskiewicz, A.J., Johnson, J.C., Penn, B., Mahalingam, M., Kimball, S.R., and Cooper, J.A. (1999). Phosphorylation of the cap-binding protein eukaryotic translation initiation factor 4E by protein kinase Mnk1 in vivo. Mol. Cell. Biol. 19, 1871–1880. Weeks, C.M., and Miller, R. (1999). The design and implementation of SnB v2.0. J. Appl. Crystallogr. 32, 120–124. Yamanaka, S., Poksay, K.S., Arnold, K.S., and Innerarity, T.L. (1997). A novel translational repressor mRNA is edited extensively in livers containing tumors caused by the transgene expression of the apoB mRNA-editing enzyme. Genes Dev. 11, 321–333. Zuberek, J., Wyslouch-Cieszynska, A., Niedzwiecka, A., Dadlez, M., Stepinski, J., Augustyniak, W., Gingras, A.C., Zhang, Z., Burley, S.K., Sonenberg, N., et al. (2003). Phosphorylation of eIF4E attenuates its interaction with mRNA 50 cap analogs by electrostatic repulsion: intein-mediated protein ligation strategy to obtain phosphorylated protein. RNA 9, 52–61.
Accession Numbers The atomic coordinates for the C-terminal region of eIF4GI have been deposited in the Protein Data Bank under ID code 1UG3.
