Viral Stress-inducible Protein p56 Inhibits Translation by Blocking the Interaction of eIF3 with the Ternary Complex eIF2·GTP·Met-tRNAi*

Viral stress-inducible protein p56 is produced in response to viral stress-inducing agents such as double-stranded RNA and interferon, as well as other poorly understood mechanisms of viral infection. It has been shown previously that p56 is able to bind the eukaryotic initiation factor 3e(eIF3e) (p48/Int-6) subunit of the eukaryotic translation initiation factor eIF3 and function as an inhibitor of translation in vitro and in vivo. The exact mechanism by which p56 is able to interfere with protein synthesis is not understood. Based on the known roles of eIF3 in the initiation pathway, we employed assays designed to individually look at specific functions of eIF3 and the effect of p56 on these eIF3-mediated functions. These assays examined the effect of p56 on ribosome dissociation, the eIF3·eIF4F interaction, and enhancement of the ternary complex eIF2·GTP·Met-tRNAi formation. Here we report that p56 is able to inhibit translation initiation specifically at the level of eIF3·ternary complex formation. The effect of p56-mediated inhibition was also examined in two different contexts, cap-mediated and encephalomyocarditis virus internal ribosomal entry site-mediated translation. Whereas cap-dependent initiation was severely inhibited by p56, internal ribosomal entry site-mediated translation appeared to be insensitive to p56.

Viral stress-inducible protein p56 is a 56-kDa protein that is induced by various agents characteristic of viral infection, such as double-stranded RNA (dsRNA), 1 interferon, and by other less well understood mechanisms associated with viral infection (1). Previous work has characterized the transcriptional regulation of the p56 gene, also known as the 561 gene, ISG56, and IFIT1 (2), as well as the cellular function of the protein in vitro and in vivo (3). p56 belongs to a family of related viral stress-inducible proteins that includes p54, p56, p58, and p60 (4 -8), but the other members of the family are not as well characterized nor are their cellular functions known. Structurally, the only significant feature of p56 is the presence of eight tetratricopeptide (TPR) motifs, motifs shown in other proteins to mediate protein-protein interactions (9). One of the best characterized interactions with p56 identified thus far has been with Int-6/p48 (also known as eIF3e), a subunit of eukaryotic initiation factor 3 (eIF3) (10). Binding of p56 to the eIF3e subunit has been shown to have a functional effect by inhibiting overall cellular translation (3). Whereas the functional effect of p56 has been mapped to TPR motifs 6 -8, the mechanism by which p56 inhibits translation is not known (10).
eIF3 is one of the 11 or more initiation factors that are involved in the first stage of protein synthesis in eukaryotes. Mammalian eIF3 is the largest (650 kDa) of all the initiation factors and is composed of 11 or 12 individual subunits eIFa-eIFj (11). The exact interactions and stoichiometry of the eIF3 subunits are poorly understood. eIF3 is a multifunctional initiation factor that has been shown to operate at different levels of the initiation pathway. eIF3 serves as a dissociation factor by binding to the 40 S ribosomal subunit, thereby preventing reassociation with the 60 S subunit and thereby increasing the size of the 40 S subunit pool (12,13). eIF3 also functions by interacting with eIF2 and stabilizing the interaction between the ternary complex (composed of eIF2⅐GTP⅐Met-tRNA i ) and the 40 S ribosomal subunit, thereby forming the 43 S ribosomal complex (14,15). Finally, eIF3 has been shown to interact with initiation factor eIF4G of the heterotrimeric eIF4F complex, which also consists of eIF4A and eIF4E (16). The interaction between eIF3 and eIF4F⅐mRNA serves to bring the mRNA to the 43 S ribosomal complex, thereby forming the 48 S complex (17). eIF3 is also known to play a role in alternative, cap-independent mechanisms of translation that are still being elucidated. Many viral, as well as cellular, mRNAs have evolved internal ribosomal entry sites (IRES) to provide a site for ribosome attachment. Although the initiation factor requirements for translation appear to be different depending on the IRES, eIF3 has already been shown to be required for encephalomyocarditis virus (EMCV) and hepatitis C virus (HCV) IRES-mediated translation (18,19). Thus, eIF3 provides a possible site of regulation by p56 at both the cap-dependent and cap-independent levels.
Based on the known functions of eIF3, the role of p56 and the mechanism of inhibition were examined further by using inhibition of eIF3-function assays. It was found that p56 inhibits translation specifically at the level of eIF3⅐ternary complex formation. Furthermore, the inhibition appears to act preferentially on cap-mediated translation, whereas EMCV IRESmediated translation was apparently insensitive to inhibition.
The addition of exogenous eIF2 stimulated cap-mediated translation, but not IRES-mediated translation, suggesting that excess levels of eIF2 (and thereby ternary complex) may be able to compete with p56 for binding to either free eIF3 or 40 S bound eIF3.

MATERIALS AND METHODS
Antibodies-Rabbit polyclonal antibody against p56 was obtained by injection of purified p56 protein expressed in Escherichia coli as described previously (3). A polyclonal antibody raised in goat against purified rabbit eIF3 was also used in these studies. This antibody recognized primarily the p110 (eIF3c) subunit of eIF3 and other eIF3 subunits to a lesser extent.
Purification of Ribosomal Subunits-Free 40 S and 60 S ribosomal subunits were purified using high salt sucrose gradients as described previously (13,20).
Ribosome Dissociation Assay-The ribosome dissociation assay was performed as described in Ref. 13. Purified 40 S ribosomal subunits (0.7 A 260 units) were incubated with 1.4 A 260 units purified 60 S ribosomal subunits to form 80 S ribosomes in a 100-l reaction volume containing 100 mM KCl, 10 mM Tris-HCl, pH 7.5, 3 mM MgCl 2 , and 2 mM DTT. To dissociate ribosomes, 60 pmol of eIF3 (37.5 g, 600 nM) was added to the reaction, incubated for 10 min at 37°C, then layered on a 12-ml 10 -25% sucrose gradient (100 mM KCl, 20 mM Hepes-KOH, pH 7.5, 5 mM MgCl 2 , 2 mM DTT), and centrifuged for 16 h at 20,000 rpm at 4°C (Beckman SW40Ti rotor). To test the effect of p56, 60 pmol of purified p56 (3.2 g, 600 nM) was preincubated with 60 pmol of eIF3 for 10 min at 30°C prior to the addition of ribosomes. Gradients were unloaded via upward displacement using 60% sucrose, and UV absorbance was measured at 254 nm with an ISCO flow cell. For Western blot analysis of sucrose gradient fractions, 500 l of each fraction was precipitated with a final concentration of 10% trichloroacetic acid, subjected to 10% SDS-PAGE, electroblotted onto a nitrocellulose membrane (Millipore), and then probed with either eIF3 antibody (1:1000) or p56 antibody (1:2000) as described previously (3).
Ternary Complex Assay-Ternary complex formation was performed as described in Ref. 23. Purified eIF2 (8 pmol, 1.0 g) was incubated with 10 pmol of [ 14 C]Met-tRNA i and 100 M GTP in a 100-l reaction volume containing 20 mM Tris-HCl, pH 7.5, 100 mM KCl, 2 mM MgCl 2 1 mM DTT, 0.3 IU pyruvate kinase, and 4 mM phosphoenolpyruvate. To stimulate ternary complex formation, 30 pmol (20 g) of purified eIF3 was also added to the reaction. Reactions were incubated for 10 min at 37°C and then immediately quenched with 3 ml of ice-cold 20 mM Tris-HCl, pH 7.5, 100 mM KCl, and 5 mM MgCl 2 (quenching buffer). Ternary complex was bound to the nitrocellulose filter (25 mm, 0.45-mm pore size, Millipore) by adding the entire reaction mixture to the filter, followed by vacuum filtration. Filters were washed twice with 5 ml of quenching buffer, followed by vacuum filtration. Filters were dried under a heat lamp, and then radioactivity was determined by liquid scintillation spectrometry. To test the effect of p56 on ternary complex formation, 30 pmol (1.7 g) of purified p56 or 30 pmol (1.2 g) of mutant p56 was pre-incubated with eIF3 for 10 min at 30°C prior to their addition to the ternary complex reaction mixture.
Generation of a Bicistronic Reporter mRNA-A bicistronic plasmid, containing the EMCV IRES (420 nucleotides) intervening between the chloramphenicol acetyltransferase (CAT, ORF1) and firefly luciferase (LUC, ORF2) reporter genes and designated pGEM-CAT-EMCV-LUC (24), was a generous gift from Dr. Nahum Sonenberg, McGill University, Montreal, Canada. The bicistronic (CAT-EMCV-LUC) insert was introduced downstream of the viral T7 RNA polymerase promoter and contains a unique restriction site, XhoI, that was used to linearize the plasmid prior to transcription. In vitro transcription was carried out for 2 h at 37°C in the presence of m 7 G (5Ј)ppp(5Ј)G using Megascript transcription kit and procedures outlined by Ambion. Following the transcription reaction, samples were treated with RNase-free DNase for 10 min and phenol:chloroform:isoamyl alcohol extracted twice, ethanol-precipitated, and washed in 70% ethanol. RNA samples were recovered in water and quantitated using a Hewlett-Packard Chemstation 8453 spectrophotometer and monitored for integrity by ethidium bromide staining following electrophoresis on acrylamide/urea gels.
In Vitro Translation Inhibition Assay-In vitro translations in rabbit reticulocyte lysate were programmed with the capped CAT-EMCV-LUC transcript in the presence of [ 35 S]methionine. In vitro translations were performed with nuclease-treated rabbit reticulocyte lysate under conditions recommended by Promega. A typical 15-l reaction contained 10 l of lysate, 0.2 mM amino acid mixture (minus methionine), 6 Ci of [ 35 S]methionine (1200 Ci/mmol), 24 units of RNase inhibitor (Roche Applied Science), 0.4 g of transcript (constituting 25 g/ml). Purified eIF2 was added in increasing amounts as indicated in the figure legends (20,21,23). Translations were performed at 30°C for 1 h in the presence of increasing amounts of wild type or mutant p56. Following translation, an aliquot of the lysate was resolved by 12.5% SDS-PAGE, and gels were dried and exposed to a PhosphorImager screen, and incorporated radioactivity was quantitated using the ImageQuant software from Amersham Biosciences.

RESULTS
p56 Does Not Interfere with eIF3-mediated Ribosome Dissociation-Because eIF3 is an essential initiation factor whose function in the initiation of translation is implicated at the earliest step of the initiation pathway, studies with p56 began at the level of ribosome dissociation. Fig. 1 shows the results of a size-separation assay by using sucrose gradient centrifugation to analyze the dissociation of ribosomes by eIF3. In this assay, individual preparations of purified 40 S and 60 S ribosomal units were used rather than salt-washed ribosomes, because they produced clearer UV profiles, separating into distinct peaks (data not shown). The UV profile of the gradient remained unchanged in the presence of p56 protein buffer as p56 Inhibits eIF3⅐Ternary Complex shown in Fig. 1A. Because eIF3 serves as an anti-association factor, addition of eIF3 to the reaction would be expected to "dissociate" 80 S ribosomes by binding to the small pool of 40 S ribosomal subunits and prevent reassociation with 60 S subunits, effectively shifting the equilibrium to free 40 S subunits. In Fig. 1B, 60 pmol of eIF3 in an approximate 2:1 molar excess of ribosomes resulted in a dramatic decrease in the 80 S peak and a corresponding increase in the 60 S peak, which suggests an increase in the amount of 60 S formed. We did not observe an equivalent increase in the 40 S ribosomal subunit pool. In part, this appears to reflect a tendency of mammalian 40 S subunits to dimerize (see Fig. 4 of Ref. 25), and this peak did not resolve well from the peak of 60 S subunits.
To investigate the possibility of p56 interfering with the function of eIF3 as a dissociation factor, an equimolar amount of p56 protein was pre-incubated with eIF3 prior to the addition of the ribosomal subunits. After sucrose gradient centrifugation, the activity of eIF3 as a "dissociation factor" was unaffected, as the peak corresponding to 80 S subunits was still reduced, and the peaks corresponding to the 40 S and 60 S subunit pool increased (Fig. 1C). As expected, a mutant form of p56, lacking the interaction domain (TPR6 -8) (10), did not affect the dissociation of ribosomes, producing a profile similar to only eIF3 added (data not shown).
p56 Does Not Interfere with eIF3⅐40 S Ribosomal Subunit Interactions-To confirm that p56 was indeed binding to eIF3 and was subsequently able to form a complex with 40 S subunits, Western blot analysis of fractions from sucrose gradients containing 40 S subunits, eIF3, and p56 was performed. Presented in (Fig. 2A) is the general UV profile for 0.7 A 260 units of 40 S ribosomal subunits used in this assay. Under these conditions, the 40 S ribosomes migrated in fraction 8. When both 60 pmol of p56 and 60 pmol of eIF3 were added to the reaction containing ϳ20 pmol of 40 S ribosomal subunits, p56 was able to bind to eIF3, which formed a complex with the 40 S subunit, as shown by Western blot analysis of gradient fractions using an antibody to p56 (Fig. 2C). p56 migration with the 40 S ribosomal subunit was dependent on the interaction with eIF3, as p56 incubated alone with 40 S subunits did not bind and remained in the upper region of the gradient (Fig. 2B). Similarly, eIF3 alone remained at the top of the gradient in the absence of 40 S ribosomal subunits, as visualized by Coomassie staining of sucrose gradient fractions (data not shown).
To confirm that p56 was not interfering with the ability of eIF3 to bind to 40 S ribosomal subunits, a Western blot analysis was performed using a polyclonal antibody against eIF3. As seen in Fig. 3A, this antibody was able to recognize primarily the p110 subunit of eIF3 (eIF3c), but at longer exposure times other subunits could also be detected, although to a lesser degree. When fractions from a sucrose gradient containing 0.7 A 260 units of 40 S ribosomal subunits and 60 pmol of eIF3 (at an ϳ3-fold molar excess to ribosomal subunits) were subjected to Western blot analysis, eIF3 could be seen migrating in the fractions corresponding to 40 S subunits (Fig. 3B). This association remained unaffected, even in the presence of equimolar p56 (Fig. 3C), indicating that p56 had no effect on the ability of eIF3 to bind to 40 S subunits.
p56 Does Not Interfere with the eIF3⅐eIF4F Interaction-eIF3 is also known to play an important role in binding and stabilization of the eIF4F complex on the 40 S subunit. The eIF4F complex is responsible for binding and unwinding the mRNA prior to translation. Because p56 was shown not to interfere with the ability of eIF3 to dissociate ribosomes, inhibition could be mediated through the disruption of the eIF3⅐eIF4F interaction. To examine the effect of p56 on the eIF3⅐eIF4F interaction, a sucrose gradient separation assay was performed using radiolabeled eIF4F (200 kDa) and unlabeled eIF3 (650 kDa). As shown in Fig. 4A, 150 pmol of radiolabeled eIF4F remained in the upper region of a 5-18% sucrose gradient, as measured by liquid scintillation spectrometry of sucrose gradient fractions. When equimolar (150 pmol) eIF3 was incubated in the presence of eIF4F, a shift in radioactivity was observed from the upper region of the gradient, toward the middle of the gradient, corresponding to fractions containing eIF3 (Fig. 4B). The shift in radioactivity represents the formation of the complex of eIF3 and eIF4F. The efficiency of eIF4F binding to eIF3 was ϳ50%, which is consistent with previous findings (22) reporting the eIF3⅐eIF4F interaction.  3. p56 does not block eIF3⅐40 S ribosomal subunit interaction. Western blots of proteins using an antibody to eIF3 are shown. A, Western blot analysis of purified rabbit eIF3 (1 g) using a polyclonal antibody showed recognition of primarily one band corresponding to the p110 subunit of eIF3. B, the Western blot represents fractions from sucrose gradient analysis of a reaction mixture that contained 0.7 A 260 40 S subunits and 60 pmol of eIF3. The molar ratio of eIF3 to ribosomes was ϳ3:1. C, Western blot of fractions from sucrose gradient analysis of a reaction mixture that contained 0.7 A 260 40 S subunits, 60 pmol of eIF3, and 60 pmol of p56.

p56 Inhibits eIF3⅐Ternary Complex
To test the effect of p56 on the formation of the eIF3⅐eIF4F complex, an equimolar (150 pmol) amount of purified p56 was pre-incubated with eIF3 prior to incubation with eIF4F. After separation on the gradient, the radioactivity profile remained unchanged, showing that eIF4F was still able to form a complex with eIF3, even in the presence of p56 (Fig. 4C). To confirm that p56 was present in the eIF3⅐eIF4F complex, Western blot analysis of sucrose gradient fractions with the p56 antibody demonstrated the presence of p56 in the fractions containing eIF3 (data not shown). As expected, the mutant form of p56 did not change the radioactivity profile of the eIF4F (Fig. 4D). Therefore, the eIF3⅐eIF4F interaction is unaffected by p56 and is unlikely to be responsible for the p56-mediated inhibition of protein synthesis.
p56 Inhibits the Enhancement of Ternary Complex Formation by eIF3-Another step of the initiation pathway that eIF3 is known to operate at, and thus representing a possible target site for p56 inhibition, is the formation and stabilization of the ternary complex. eIF3 binds the ternary complex, which consists of eIF2, initiator Met-tRNA i , and GTP, and in doing so shifts the equilibrium toward ternary complex and ternary complex⅐eIF3 (15). To investigate the possible effect of p56 on ternary complex formation, a filter binding assay was employed. This assay is based on the principle that Met-tRNA i binding to a nitrocellulose filter is dependent on the binding to an initiation factor (26). Thus, in the presence of eIF2, which is retained on nitrocellulose, Met-tRNA i will be retained on the filter by binding to eIF2 in the presence of GTP. By using radiolabeled Met-tRNA i , formation of the ternary complex can be quantitatively measured.
Presented in Fig. 5 are the cumulative results of 5 separate filter-binding assays as described under "Materials and Methods." As expected, [ 14 C]Met-tRNA i was not retained on the nitrocellulose filter, and all other values were normalized to the background levels of radioactivity detected (Fig. 5, column A). In the presence of 8 pmol (1 g) of eIF2 and GTP, [ 14 C]Met-tRNA i was retained on the filter, representing ternary complex formation (Fig. 5, column B). The addition of 30 pmol of eIF3 to the reaction resulted in a 4-fold increase in radioactivity retained on the nitrocellulose filter (Fig. 5, column C), representing the effect of eIF3 on the stabilization of ternary complex formed. Of particular interest, when 30 pmol of p56 was pre-incubated with equimolar eIF3, the result was a complete abrogation of the eIF3 stimulatory effect on ternary complex formation (Fig. 5, column E), providing the first evidence of a possible mechanism of inhibition. Additional evidence supporting this observation was the relative inability of mutant p56 to inhibit the eIF3-mediated stimulation of ternary complex (Fig. 5, column F), presumably due to its inability to bind to eIF3 (10).
p56 Exhibits Differential Effects on m 7 G Cap and IRESmediated Translation Initiation-Once the specific stage of the initiation pathway at which p56 inhibits translation had been implicated, p56 was investigated for its ability to inhibit in vitro translation in two different contexts, cap-mediated and IRES-mediated translation initiation. The reporter used was a bicistronic mRNA, containing the EMCV IRES. The EMCV IRES requires all canonical initiation factors for translation to initiate internally (19). Because ternary complex is required for both cap-mediated and IRES-mediated translation, it would seem likely that both modes of translation would be equally inhibited by p56.
Surprisingly, p56 did not appear to inhibit IRES-mediated translation at the levels of p56 added. Translation of the capdependent open reading frame, CAT was inhibited in a dosedependent manner (Fig. 6A, lanes 1-6), which is consistent with our findings reported previously (3). However, translation of the IRES-dependent open reading frame, luciferase (LUC), was not inhibited upon addition of p56, suggesting that the EMCV IRES has an alternative mechanism of translation that is less sensitive to levels of ternary complex. Addition of exogenous eIF2 was able to partially restore cap-dependent translation (Fig. 6A, lane 7). Quantitation of the two products is shown in Fig. 6B, demonstrating the dose-dependent inhibitory effect on CAT translation by increasing amounts of p56, whereas LUC translation remained unaffected. Consequently, the ratio of CAT translation to LUC translation decreased (Fig.  6C). In a similar experiment, mutant p56 did not inhibit either cap-or IRES-mediated translation, even when tested at higher concentrations (560 nM; data not shown). DISCUSSION Previous efforts identified that p56 was capable of binding the p48 (Int-6) subunit of eIF3 and that this binding could lead to the inhibition of protein synthesis, although the mechanism of inhibition was not defined (3). eIF3 is a large, multisubunit complex and appears to coordinate a number of events in the 80 S initiation pathway (11). The best characterized of these interactions is . Column D same as column C plus p56 Buffer control. Column E same as column C plus equimolar amount of p56 (30 pmol, 1.7 g). Column F same as C plus equimolar mutant p56 (30 pmol, 1.2 g).
the ability of eIF3 to bind to 40 S subunits, eIF4F, and the ternary complex (11,15,21). We have systematically examined the ability of p56 to interfere with each of these processes and observed that only the third interaction was affected.
It has been shown that under conditions of limiting Met-tRNA i , eIF3 enhances retention of Met-tRNA i on nitrocellulose filters (15) presumably by forming an eIF3⅐ternary complex and thus pulling the following equilibrium (Reaction 1) to the right: Met-tRNA i ϩ eIF2⅐GTP 7 eIF2⅐GTP⅐Met-tRNA i ϩ eIF3 7 eIF3⅐eIF2⅐GTP⅐Met-tRNA i (ternary complex) REACTION 1 At present it is not clear whether this interaction reflects a sequence whereby eIF3 binds to 40 S subunits as an eIF3⅐ternary complex or whether this reflects the intimate interaction of these proteins on the surface of the 40 S subunit. As seen in Fig. 5, p56 completely inhibited this interaction. The consequence of limiting the eIF3⅐ternary complex interaction is depicted in Fig. 7. Arrows labeled A and B show the possible sites where p56 might have its effect, either by blocking the binding of the ternary complex to the 40 S⅐eIF3 complex (A) or by blocking the direct interaction between eIF3 and the ternary complex (B). Either of these steps would lead to reduced levels of 40 S⅐eIF3⅐ternary complexes and a general overall slowing of protein synthesis initiation. This study demonstrates the first example of differential regulation of the three functions of eIF3 in translation initiation. Although binding of p56 inhibited the interaction of eIF3 with the ternary complex, its other two interactions (either with 40 S subunits or eIF4F) remained unaffected. This observation suggests that different domains of eIF3 interface with different interacting partners, and these domains can function independently of one another. It remains to be seen if there are other cellular proteins that can bind to other subunits of eIF3 and regulate the functions unaffected by p56. Independent regulation of the different activities of eIF3 may be needed for selective, but not global, regulation of translation initiation as observed for the effect of p56 on translation of the bicistronic mRNA. eIF3-interacting proteins, such as p56, may be useful tools for mapping the functional topology of this multiprotein complex.
As a further test of p56 action, p56 was tested for its ability to inhibit the translation of the bicistronic mRNA, pGEM-CAT-EMCV IRES-LUC. Surprisingly, there appeared to be little or no inhibition of the IRES-mediated translation of LUC. This observation is consistent with preliminary studies using hemedeficient lysate or normal lysate supplemented with dsRNA. 2 Both of these translation systems contain reduced levels of ternary complex, relative to either normal lysate or lysate 2 R. Bhasker, unpublished observations. supplemented with exogenous eIF2, and only inhibition of the cap-dependent translation was observed.
Does this preferential inhibition of cap-dependent protein synthesis make sense? The answer would appear to be yes. First, the preferential inhibition of cap-dependent translation appears to be exactly the same as that observed with PKR, an interferon induced eIF2␣ kinase that is activated by dsRNA, the proverbial "calling card" of viral infection. Second, interferon action is not exclusively restricted to a protective effect against viral infections. Interferons have been shown to have a broad range of biologically protective functions. In this light, it should be noted that many, but not all, proteins associated with response to stressing events contain IRES elements (27). Although it is not clear that these cellular IRES all behave identically to the EMCV IRES, it could be imagined that many would. Thus, these IRES containing mRNAs would be resistant to the general reduction in ternary complexes that could have resulted from the action of any of the four eIF2␣ kinases (HRI, PKR, PERK, and GCN2).
In addition, many stressing agents also influence eIF4F activity, primarily through the eIF4E/4E-BP interaction (27). The reduction of eIF4F activity under stress directly down-regulates cap-dependent translation as expected and simultaneously allows for the up-regulation of IRES-mediated translation (28). In sum, the ability to maintain IRES-mediated translation for the expression of a variety of protective proteins in the presence of stressing agents or in response to interferon stimulation, which may influence ternary complex levels and/or eIF4F activity, ensures optimal protection for the cell.
The unanswered question is why is IRES-mediated translation less sensitive to ternary complex levels? The general 80 S initiation pathway would predict that both cap-dependent and IRES-mediated translation initiation should be reduced when the level of 40 S⅐eIF3⅐ternary complexes (43 S complexes) fall as the mRNA binding event requires the 43 S complex as a substrate (11). At present, there appear to be three possible alternatives. First, the IRES-containing mRNA may be more competitive and thus out-compete cap-dependent initiation for the limiting amounts of 43 S complexes. Second, due to the stickiness of the translation factors, a large factor aggregate, which is eIF3⅐ternary complex⅐eIF4F⅐mRNA, may form and bind directly to the 40 S ribosome. Finally, the IRES element may bind to either 40 S or 40 S⅐eIF3 (but not 40 S⅐eIF3⅐ternary complexes) and thereby form a more effective trap for the limiting amounts of ternary complex.
Currently, there is insufficient evidence to decide which of these is correct, either in general or for the pGEM-CAT-EMCV-LUC construct used in these studies. Preliminary competition studies indicate that for this construct, cap-dependent translation is more efficient 2 ; however, this will obviously vary depending on the 5Ј-untranslated region used for cap-dependent translation and the specific IRES used to direct internal initiation. Evidence for the binding of the mRNA to the 40 S subunit in the absence of the ternary complex is not well substantiated in general; however, for HCV and cricket paralysis virus mRNAs, the evidence is rather good (29 -31). Although there is ample evidence demonstrating a degree of "stickiness" between translation initiation factors (15,21,(32)(33)(34), there is no evidence that such larger aggregates are an integral part of any initiation pathway. Clearly, it is of extreme interest to examine in greater mechanistic detail the binding of IRES-containing mRNAs to the 40 S subunit. It will also be of interest to see if the effectiveness of interferon treatment in vivo is generally better against viruses that generate mRNAs that are translated in a cap-dependent manner than against viruses that rely on IRES-mediated translation of their mRNAs.
It is curious to note that the lack of inhibitory effects of p56 on EMCV IRES-mediated translation observed in the current study may not be universally applicable to all IRES elements in all cell types. Wang et al. (35) reported that in Huh7 cells, p56 could strongly inhibit translation directed by the HCV IRES. This inhibition was less pronounced for both cap-dependent and EMCV IRES-dependent translation. The higher sensitivity of HCV IRES to p56, as compared with cap-dependent translation, was also observed in vitro using rabbit reticulocyte lysates (35). Taken together with the conclusions of Wang et al. (35), our study suggests a hierarchy of p56 susceptibility; EMCV IRES is much less sensitive than cap-dependent initiation, and HCV IRES is more sensitive. The underlying biochemical mechanism responsible for this hierarchy is currently unclear, although the known difference in the ways EMCV IRES and HCV-IRES recruit the 40 S subunit is expected to play an important role (36).