eIF4G--A Multipurpose Ribosome Adapter?

See allHide authors and affiliations

Science  24 Jan 1997:
Vol. 275, Issue 5299, pp. 500-501
DOI: 10.1126/science.275.5299.500

To be joined as partners in protein synthesis, mRNAs and ribosomes must be presented to each other in the ritual of translation initiation. Protein-mRNA interactions precede the engagement of the small ribosomal subunit (40S) with the mRNA. This rate-limiting step is followed by the joining of the large (60S) subunit at the translation initiation codon of the mRNA and the beginning of protein synthesis (1). Recent data from several laboratories have implicated a long-known translation initiation factor—eIF4G (formerly p220 or eIF-4γ)—as a critical link between mRNAs and 40S subunits during the initial engagement process.

Cellular mRNAs receive a 5' cap structure (7mGpppN) and a 3' polyadenylated [poly(A)+] tail as a nuclear dowry before their export into the cytoplasm as mature mRNAs. Both these moieties activate translation initiation in concert with cytoplasmic binding proteins, the cap-binding protein eIF4E, and the poly(A)+ binding protein (Pab1p in yeast). However, some picornaviral RNAs and a few cellular mRNAs have specialized sequences within their 5' untranslated regions that directly promote ribosome binding independent of a cap structure. These are the internal ribosome entry sequences (IRES) (2). How can all these different modes of connecting mRNAs and 40S ribosomal subunits be used for translation initiation? The cell's answer may resemble the solution found by avid international travelers for the difficulties of using electrical appliances in foreign countries: the use of an adapter.

Picornaviruses such as poliovirus or foot-and-mouth-disease virus encode proteases that specifically clip the NH2-terminal third off eIF4G to inhibit the cap-dependent translation of cellular mRNAs (3). The viral RNAs escape this inhibition by virtue of their IRES (2), indicating that the integrity of eIF4G is important for cap-dependent, but not for IRES-mediated, translation initiation. The cap-binding protein, eIF4E, can directly plug into the NH2-terminal third of eIF4G (4, 5), while the central third associates with eIF3, a multimeric complex that is bound to the 40S subunit (5) (see figure). Therefore, eIF4G forms a bridge between the cap structure (via eIF4E) and the 40S subunit (involving eIF3) that is broken by the picornaviral proteases. The proteolytic clip kills two birds with one stone, because it not only uncouples the cap from eIF4G, but the liberated COOH-terminal cleavage product actively supports IRES-mediated translation (6). Similar to the cap in conjunction with eIF4E, the IRES might also be connected with eIF4G, perhaps by means of IRES-binding proteins (X in the figure). Recent data suggest that, at least in some instances, X could be part of eIF4G itself (7). The central domain of eIF4G binds strongly and specifically to a structural element within the encephalomyocarditis virus IRES upstream of the initiation codon. Thus, eIF4G can also build a molecular bridge between this IRES and the ribosome.

More than gadgetry.

Several ways of engaging the small ribosomal subunit (40S) with mRNAs (red line) are mediated by the adapter protein eIF4G. For clarity, only translation initiation factors mentioned in the text are shown.

Once bound, the 40S subunit scans the mRNA in a 3' direction to identify the initiation codon (1). The directionality of the scanning process suggests that 40S subunits should contact the mRNAs 5' of the initiation codon, a requirement that the locations of both the cap structure and IRES help to meet.

Studies in cell-free translation extracts from yeast indicated that the poly(A)+ tail stimulates 40S subunit recruitment to the mRNA (8). This recruitment is mediated by Pab1p. Pab1p acts synergistically with the cap structure, but can function independently as well. Once more, eIF4G is the responsible molecular adapter (9): Pab1p can bind to a region of <120 amino acids that is conserved between the two yeast homologs of eIF4G (see figure). Mutations that disrupt Pab1p binding affect poly(A)+ tail-stimulated translation in vitro, with little effect on cap-mediated protein synthesis. Interestingly, Pab1p and eIF4G interact only in the presence of RNA, perhaps to selectively favor Pab1p that is bound to the poly(A)+ tail. eIF4G thus integrates the functions of the cap and the poly(A) + tail in translation. Somehow the 3' end of the mRNA must bend over backwards to deliver 40S subunits 5' of the translation initiation codon. A simultaneous interaction of eIF4G with the cap-binding protein eIF4E and the poly(A)+ binding protein Pab1p could be one mechanism to accomplish this and would effectively circularize the mRNA (see figure). It will be important to determine whether the mammalian poly(A)+ binding protein and eIF4G also interact to activate translation.

A role as a multipurpose adapter between mRNAs and ribosomes would make eIF4G an ideal target for translational control. The cellular protein DAP-5/p97 is a candidate for such a function (10, 11). To a first approximation, DAP-5/p97 resembles the larger of the cleavage products generated by picornaviral proteases and is homologous to the COOH-terminal two-thirds of eIF4G. It binds to eIF3 and eIF4A (a translation initiation factor with RNA-dependent helicase activity that is required for both cap- and IRES-dependent translation), but lacks the domain that binds the cap-binding protein eIF4E (10). Whether IRESs exist that can bind DAP-5/p97 remains to be determined, but transfection experiments have suggested that DAP-5/p97 suppresses both cap-dependent and encephalomyocarditis virus IRES-mediated translation (10), probably by forming a translationally inactive protein complex that includes eIF3 and eIF4A but excludes eIF4E. If DAP-5/p97 could bind to (currently unrecognized) specific mRNA sequences, it may act as a positive translation factor for these mRNAs (11). Although DAP-5/p97 appears to be involved in modulating interferon-γ-induced programmed cell death (11), the physiological targets and mode of action are still enigmatic.

Within a short while, these new, primarily biochemical data have changed eIF4G from an unknown entity to a center-stage player during translation initiation. eIF4G may also unlock some of the mysteries of how regulatory elements in the 3' untranslated regions of mRNAs and their binding proteins can control ribosome binding at the 5' end of mRNA. Such regulatory mechanisms underlie numerous genetic switches during early embryonic development and in cell differentiation (12). Much work still needs to be done to rigorously test whether, when, and how eIF4G serves as a multipurpose adapter for ribosomes. Nonetheless, eIF4G function already offers a conceptual framework for a better understanding of the initiation rites between mRNAs and ribosomes.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
View Abstract

Navigate This Article