Translational Control by Messenger RNA Competition for Eukaryotic Initiation Factor 2 *

Translation of globin mRNA in a micrococcal nu- clease-treated reticulocyte lysate was studied in the presence of increasing amounts of Mengovirus RNA, under conditions in which the number of translation initiation events remains constant as judged by the transfer of label from N-f~rmyl[~~S]methionyl-tRNA~ into protein. The translation of globin mRNA is pro- gressively inhibited by low concentrations of Mengovirus RNA, free of detectable traces of double-stranded RNA, concomitant with the increasing synthesis of Mengovirus RNA-directed products. On a molar basis, Mengovirus RNA apparently competes about 35 times more effectively than globin mRNA for a critical component in translation. The competition is relieved by the addition of highly purified eukaryotic initiation factor 2 (eIF-2). Addition of eIF-2 does not stimulate overall protein synthesis, but shifts it in favor of globin synthesis. No stimulation of globin mRNA translation by eIF-2 is seen when Mengovirus RNA is absent. These experiments show that Mengovirus RNA competes, di- rectly or indirectly, with globin mRNA for eIF-2. In direct binding experiments using isolated mRNA and eIF-2, Mengovirus RNA is shown to compete with globin mRNA for eIF-2 and to exhibit a 30-fold higher affinity for this factor. The binding of Mengovirus RNA to eIF-2 is much more resistant to increasing salt con- centrations than is the binding of globin mRNA, again reflecting its high affinity. These results reveal a direct correlation between the ability of these mRNA species to compete in translation and their ability to bind to initiation factor eIF-2.

Mengovirus RNA apparently competes about 35 times more effectively than globin mRNA for a critical component in translation. The competition is relieved by the addition of highly purified eukaryotic initiation factor 2 (eIF-2). Addition of eIF-2 does not stimulate overall protein synthesis, but shifts it in favor of globin synthesis. No stimulation of globin mRNA translation by eIF-2 is seen when Mengovirus RNA is absent. These experiments show that Mengovirus RNA competes, directly or indirectly, with globin mRNA for eIF-2.
In direct binding experiments using isolated mRNA and eIF-2, Mengovirus RNA is shown to compete with globin mRNA for eIF-2 and to exhibit a 30-fold higher affinity for this factor. The binding of Mengovirus RNA to eIF-2 is much more resistant to increasing salt concentrations than is the binding of globin mRNA, again reflecting its high affinity.
These results reveal a direct correlation between the ability of these mRNA species to compete in translation and their ability to bind to initiation factor eIF-2. They suggest that the affinity of a given mRNA species for eIF-2 is essential in determining its translation, relative to that of other mRNA species. Messenger RNA competition for eIF-2 may contribute significantly to the selective translation of viral RNA in infected cells.
Gene expression in eukaryotic organisms often involves the selective translation of certain mRNA templates over other ones. This type of regulation is used frequently in virus infection and cellular differentiation. Messenger RNA discrimination is thought to occur mainly at the initiation step, which involves the recognition of mRNA and its binding of ribosomes. Thus, translational competition between host mRNA * This work was supported by grants from the National Council for Research and Development (Israel), and the Gesellschaft fuer Strahlenforschung (Muenchen), and from the Israel Commission for Basic Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. + Present address, Institute of Cellular Biology, Consiglio Nazionale delle Ricerche, Via Romagnosi MA, Rome, Italy. and picornavirus RNA in a cell-free system from mouse ascites cells occurs at the level of initiation (Lawrence and Thach, 1974;Golini et al., 1976). During erythroid development, protein synthesis on P-globin mRNA is initiated more frequently than on a-globin mRNA, because /$globin mRNA competes more effectively for a critical component in translation (Lodish, 1971(Lodish, , 1974Lodish and Jacobson, 1972). The translational competition between the a-and P-globin mRNA species occurring in a micrococcal nuclease-treated reticulocyte lysate is relieved by the addition of eukaryotic initiation factor 2 (Di Segni et al., 1979). In addition to its indispensable role in the binding of Met-tRNAf to 40 S ribosomal subunits during initiation (Trachsel et al., 1977), this initiation factor possesses mRNA-binding properties (Kaempfer, 1974;Rosenfeld, 1977, 1978;Kaempfer et al., 1978Kaempfer et al., , 1979. Indeed, /3-globin mRNA binds to eIF-2 with higher affinity than does a-globin mRNA (Di Segni et al., 1979); the results of that study suggest strongly that mRNA interacts directly with eIF-2l during translation.
Here we show that a molecule of Mengovirus RNA apparently competes 35 times more effectively than a molecule of globin mRNA for a component at the initiation step of translation. The competition is relieved by eIF-2. We demonstrate, in binding experiments, that Mengovirus RNA and globin mRNA compete directly for eIF-2, and that Mengovirus RNA binds to this protein with a 30-fold higher affinity. These experiments reveal a quantitative correlation between the ability of a given mRNA species to compete in translation, and its affinity for initiation factor eIF-2.
Direct evidence for a specific interaction between eIF-2 and mRNA comes from studies with STNV RNA, as well as Mengovirus RNA. In the 1,239-nucleotide-long STNV RNA, eIF-2 recognizes specifically the 5'-terminal44-nucIeotide sequence that contains the ribosome binding site . In Mengovirus RNA, eIF-2 by itself protects specific sequences against nuclease attack that are identical with those protected in 40 S and 80 S initiation complexes.' These sequence studies point to an essential role for eIF-2 in directing the binding of ribosomes to mRNA during initiation of translation. The present results support the concept that this direct and specific interaction between eIF-2 and mRNA occurs during protein synthesis and reveal its importance for translational control.

EXPERIMENTAL PROCEDURES
Reticulocyte Lysate-Preparation of micrococcal nuclease-treated rabbit reticulocyte lysate was as described by Di Segni et al. (1979). 94 7 rnRNA-Globin mRNA was prepared as described (Kaempfer, 1979 (Kaempfer. 1979). 90 nlM of added KCI, 4 m v creatine phosphate. 20 mM Hepes buffer. pH 7.5, and [%]methionine (700 Ci/mmol; Amersham). Mixtures were incubated for 45 min a t :10 "C. and frozen until analysis of the products. Maximal translation o f globin mHNA was seen at about 90 mM ofadded KCI. l'he salt concentration contributed by the lysate is about 55 mM. as estimated from translation requirements of a gel-filtered lysate (not shown). Maximal translation of Mengovirus RNA was obtained at I25 mM o f added KCI. but the optimum is broad and translation at 90 mM of added KC1 is nearly as extensive. Hot CCl.lCOOH-precipitable radioactivity was determined in aliquots spotted onto I-inch square filters (Whatman No. 1 paper). In the conditions used (Kaempfer.  Preparation of [ "S]Met-/RNA, a n d Assqv of Ternary Comp1c.r Forrnation-[""S]Met-t~~NA, was prepared and ternary complex formation with eIF-2 and GTP was assayed as described (Kaempfer et al.. 1978a). N-Formyl-[ "'S]Met-tHNA1 was prepared as described by Rosen e/ af. (1981 ).
Purification of elF-2-The procedure is described by Kaempfer (1979). 1 3 Segni et al. (1979), . eIF-2 purified by this procedure is at least 98"; pure as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and is free of detectable protein at 80.OOO. the molecular weight assigned to factor eIF-4B; its mRNA-binding activity is completely sensitive t o competitive inhibition by Met-tRNA, and GTP but not by uncharged tRNA.
showing that the only mRNA-binding component in the p~c*paration is eIF-2 itself (Hosen and Kaempfer, 1979). The purified elF-2 bound 0.3 to 1.0 pmol of Met-tRNA,/pg of protein.
Radioiodination of rnRNA-The method of Commerford (1971) was followed, with the modifications described by 1 3 Segni e/ af. (1979). The "..'I-labeled mRNA obtained in this procedure is labeled to a relatively low specific activity, up to 2 X 10" cpm/pg. This low specific activity is important in determining the properties o f the labeled RNA. The RNA is fully intact as judged. for globin mRNA. by polyacrylamide gel electrophoresis and autoradiography. Both globin mRNA and Mengovirus RNA labeled in our procedure bind to eIF-2 with an affinity equal to that of the unlabeled, native RNA species, as determined by self-competition ( Fig. 9; Kaempfer, 1979: Kaempfer et af.. 1979).
Binding of rnHNA to elF-2-Reaction mixtures of 50 pl. containing 150 mM KCI, 20 mM Tris-HCI. pH 7.8, 2 mM Mg-acetate, 6 mM 2mercaptoethanol, eIF-2. and RNA, were incubated for 10 min a t 25 "C and then cooled for 10 min at 0 "C before the addition of 1 ml of icecold buffer A (20 mM Tris-HCI, pH 7.8, 6 mM 2-mercaptoethanol, 50 mM KC1. 2 mM Mg-acetate). The samples were passed through 25mm nitrocellulose filters (0.45 pm pore diameter) at a flow rate of 1 ml/min, and washed three times with 1 ml of buffer A. Dried filters were counted by 7 radiation spectrometry. For a more detailed description, see Kaernpfer (197%. Phosphorylation of e1F-2 a n d M , = 67,000 Polypeptide-Doublestranded RNA-dependent phosphorylation was studied in the ribosomal system. as described by Rosen et af. (1981).

Translational Competition between Globin mRNA and
Mengovirus RNA-The micrococcal nuclease-treated rabbit reticulocyte lysate (Pelham a n d Jackson, 1976) w a s used to study translational competition between globin mRNA and Mengovirus RNA. The products of translation were analyzed by cellulose acetate electrophoresis in the presence of urea and 2-mercaptoethanol. Fig.   1 illustrates the autoradiogram of the products synthesized in a series of lysates containing globin mRNA and increasing amounts of Mengovirus RNA. It is seen that a-and /3-globin are well resolved from each other and from the Mengovirus RNA-directed products of translation which move t o the opposite electrode. The protein products synthesized under the direction of Mengovirus RNA move as a single, broad hand. By varying the conditions of electrophoresis. it can he shown that this broad hand includes all the products of Mengovirus RNA translation. Mengovirus RNA, when translated alone, directed the synthesis of products that all moved with the latter band; these products are M , = 60,000 t o 100,000, as judged by electrophoresis in sodium dodecyl sulfate-polyacrylamide gels (not shown).   (0). Aliquots of 25 p1 were diluted with an equal volume of 1 M NaOH and incubated for 30 min at 37 "C before determination of CCLCOOHprecipitable radioactivity. Mengovirus RNA alone (1.5 pg) directed the incorporation of 3,500 cpm. strongly suggest that Mengovirus competes with globin mRNA in translation.
From the results of Fig. 3, it can be calculated that halfmaximal inhibition of globin synthesis occurs when the lysate contains 35 molecules of globin mRNA for every molecule of Mengovirus RNA. If we assume that both mRNA preparations contain an equal proportion of active mRNA molecules, then a molecule of Mengovirus RNA competes apparently 35 times more effectively than a molecule of globin mRNA for a rritical component in translation.
Relief of Translational Competition by eZF-2.-The effect of addition of highly purified initiation factor eIF-2, free of detectable traces of other initiation factors , to a reaction mixture for translation containing both Mengovirus RNA and globin mRNA is illustrated in Fig. 5. The addition of eIF-2 does not lead to any increase in total protein synthesis, but does lead to an increase in the amount of globin synthesis, concomitant with a decrease in the synthesis of Mengovirus protein. The total globin/Mengovirus protein synthetic ratio increases from 0.74 to 1.55. Comparison of Fig. 5 with Fig. 2 shows that the addition of eIF-2 to a lysate containing 1 pg of Mengovirus RNA caused a shift in the pattern of proteins synthesized to one resembling that seen in the presence of 0.5 pg of Mengovirus RNA. Therefore, the amount of eIF-2 added in the experiment of Fig. 5 reduced the inhibitory effect of 1 pg of Mengovirus RNA to that caused by about one-half this amount. The finding that the eIF-2-dependent increase in globin synthesis is coupled with a concomitant decrease in Mengovirus RNA translation, while overall translation remains constant, supports the explanation that eIF-2 acts to relieve mRNA competition.
Low concentrations of dsRNA lead to the inactivation of eIF-2 and the resulting inhibition of translation can be relieved by the addition of eIF-2 (Kaempfer, 1974;Clemens et ul., 1975). Since translation of Mengovirus RNA, in contrast to translation of globin mRNA, is resistant to the inhibitory effect of dsRNA (Rosen et al., 1981), and since RNA from vaccinia, reo-or vesicular stomatitis virus was shown to contain contaminating traces of dsRNA (Lenz and Baglioni, 1978), it was necessary to eliminate the possibility that both the inhibitory effect of Mengovirus RNA on translation of globin mRNA, and its relief by eIF-2 are accounted for by the presence of contaminating traces of dsRNA in the Mengovirus RNA preparation.
To examine this point, the dsRNA-dependent phosphorylation of the M , = 38,000 subunit of eIF-2 and of a M,. = 67,000 polypeptide was studied by incubating ribosomes isolated from a micrococcal nuclease-treated reticulocyte lysate with [y-'"P]ATP (Rosen et ul., 1981). The reaction mixtures were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. As seen in the autoradiogram of Co-purification of the activity that relieves translational competition with eIF-2. For purification of eIF-2, see "Experimental Procedures." The gradient portion of the phosphocellulose column is shown. Aliquots of 5 p1 were used to assay GTP-dependent binding of '"S-labeled Met-tRNA( (input 5,500 cpm). Aliquots of 0.5 p1 were used to assay binding of ""I-labeled Mengovirus RNA (8 X 1 0 cpm/pg; input, 7,100 cpm). No background was subtracted in either assay. Aliquots of 3 pI were added to reaction mixtures for protein synthesis containing 1 pg of globin mRNA and 0.5 pg of Mengovirus RNA; the KC1 concentration was adjusted to 95 mM in each sample. Incorporation of ["S]methionine into total protein was determined in 2 pI of reaction mixture. The products of translation were analyzed a in Figs. 1 and 2. The total amount of globin formed is depicted. Triangles on right indicate globin synthesis in reaction. Arrows indicate amount of globin synthesized in reaction mixtures lacking Mengovirus RNA, incubated with (A) and without (A) 3 pl of material from tube 11; total incorporated into protein in these controls was 3,900 cpm and 3,700 cpm, respectively. competition between Mengovirus RNA and globin mRNA copurifies with eIF-2. The experiment illustrates the phosphocellulose chromatography step in the purification, which involves elution of eIF-2 with a linear gradient from 0.4 to 0.8 M KC]. Two characteristic activities of eIF-2, its ability to form a ternary complex with [:%]Met-tRNA( and GTP, and its ability to bind mRNA, assayed by the retention of '""Ilabeled Mengovirus RNA on nitrocellulose filters, are seen to co-elute in a single peak near 0.6 M KCI. The fractions of this gradient were tested for their ability to stimulate globin synthesis in reaction mixtures containing globin mRNA and 0.5 pg of Mengovirus RNA. As seen in Fig. 7, this amount of Mengovirus RNA caused a 2-to 2.5-fold reduction in globin synthesis when compared to a control sample containing only globin mRNA. Although there is no stimulation of total protein synthesis by any fraction across the gradient, the synthesis of globin is stimulated to the level seen in the control lacking Mengovirus RNA by an activity that elutes precisely in t h e position of eIF-2. It should be noted that other basic proteins do elute ahead of eIF-2 or behind it, yet these fail to stimulate globin synthesis in conditions of competition. T h e stimulation of globin mRNA translation by eIF-2 is seen only in the presence of competing amounts of Mengovirus RNA, but not in the control containing only globin mRNA. Thus, the effect of eIF-2 is to relieve competition, and not to stimulate globin mRNA translation specifically. Fig. 8 illustrates the effect of increasing amounts of eIF-2 on lysates containing, besides globin mRNA, different amounts of Mengovirus RNA. At each Mengovirus RNA level tested, total protein synthesis is not stimulated, but the ratio of globin to Mengovirus protein synthesis increases as a func-  (filled symbols) or 1.5 pg (open symhols) of Mengovirus RNA, were incubated in the presence of the indicated amounts of eIF-2 (0.08 pg/ pl). Incorporation of ["%]methionine into total protein was determined in 5-pl aliquots. The products were analyzed as in Figs. 1 and tion of the amount of eIF-2 added. This increase is more pronounced, the lower the concentration of Mengovirus RNA, a finding also evident from the results of Figs. 5 and 7. Indeed, the slopes of the eIF-2 dose-response curves shown in Fig. 8 are approximately inversely proportional to the amount of Mengovirus RNA present. This is the result expected if Mengovirus RNA and globin mRNA compete for eIF-2 in translation.
Preferential Binding of Mengovirus RNA to eIF-2"If the translational competition observed between Mengovirus RNA and globin mRNA involves direct competition for initiation factor eIF-2, then one would predict Mengovirus RNA to possess a greater affinity than globin mRNA for eIF-2. To examine this point, we have labeled Mengovirus RNA and globin mRNA with lZ5I, and have studied the binding of these labeled RNA species to eIF-2 in the presence of competing amounts of the unlabeled ones. In the experiment of Fig. 9A, lZ5I-labeled Mengovirus RNA was incubated with a limiting amount of eIF-2 and increasing amounts of unlabeled Mengovirus RNA or globin mRNA. It is seen that, on a molar basis, Mengovirus RNA competes 30 times more effectively than globin mRNA. Unlabeled Mengovirus RNA competes GLOBIN mRNA with labeled Mengovirus RNA on an equimolar basis, showing that radioiodination did not affect the affinity of this RNA for eIF-2 (Kaempfer, 1979). The reciprocal experiment, using lZ5Ilabeled globin mRNA, is presented in Fig. 9B. Again, Mengovirus RNA is seen to compete 30 times more effectively than globin mRNA for eIF-2.
In Fig. 10, the extent of complex formation between eIF-2 and labeled globin mRNA or Mengovirus RNA is analyzed as a function of increasing KC1 concentration. Binding of globin mRNA to eIF-2 decreases sharply above 100 mM KC1, and is inhibited 50% at 160 mM KC1. By contrast, binding of Mengovirus RNA to eIF-2 is much more resistant to increasing salt concentrations. This result provides independent evidence that Mengovirus RNA binds more strongly to eIF-2.

DISCUSSION
The major result of these experiments is that Mengovirus RNA and globin mRNA compete during translation in a manner that can be relieved by eIF-2, and that each of these mRNA species binds to this initiation factor with an affinity that closely reflects its ability to compete in translation. Direct binding experiments show that Mengovirus RNA possesses a 30-fold greater affinity for eIF-2 than does globin mRNA. This affinity ratio is in striking agreement with translation competition experiments that reveal about a 35-fold greater apparent affinity of Mengovirus RNA for a critical component in protein synthesis. Together with our finding that eIF-2 relieves the translational competition between Mengovirus RNA and globin mRNA, these experiments strongly support the concept that Mengovirus RNA and globin mRNA compete directly for eIF-2 in protein synthesis. More generally, these results reveal a direct correlation between the ability of a given mRNA molecule to compete in translation and its ability to bind to initiation factor eIF-2. Such messenger RNA competition for eIF-2 may contribute significantly to the selective translation of viral RNA in infected cells.
Messenger RNA Competition for eIF-2-The concept that mRNA species differ in their efficiency of translation, apparently because of a different affinity for one or more critical components in the initiation step, was fist suggested by Lodish (1974Lodish ( , 1976 who showed that initiation of protein synthesis on a molecule of a-globin mRNA occurs with lower frequency than that on a molecule of P-globin mRNA (Lodish, 1971). Indeed, the addition of eIF-2 leads to relief of the 95 1 translational competition between a-and ,&globin mRNA occurring in a micrococcal nuclease-treated lysate identical with that used in these experiments (Di Segni et al., 1979). Such relief is observed both when competition is caused by a high concentration of globin mRNA and when it is sharpened by a high concentration of KC1. Results of that study suggest strongly that mRNA interacts directly with eIF-2 during protein synthesis, and that a-globin mRNA possesses a lower affinity for eIF-2 than does ,&globin mRNA.
In the present experiments, the result of translational competition analysis between globin mRNA and Mengovirus RNA are shown to correlate quantitatively with the results of direct binding competition analysis between these mRNA species and isolated eIF-2. Mengovirus RNA competes directly with globin mRNA for this factor, and binds it with a 30-fold higher affinity (Fig. 9). Independent verification that eIF-2 itself binds 30-fold more tightly to Mengovirus RNA than to globin mRNA is provided by the finding that Mengovirus RNA is 30 to 40 times more effective than globin mRNA as a competitive inhibitor of ternary complex formation between Met-tRNAf, GTP, and eIF-2 (Rosen et al., 1981). Mengovirus RNA acts as a powerful inhibitor of globin mRNA translation (Figs. 1 to 3), in conditions where the overall number of initiation events remains constant (Fig. 4). This inhibition of globin mRNA translation is relieved by the addition of highly purified eIF-2 ( Figs. 6 to 8). Yet, addition of this initiation factor does not stimulate total translation, and it stimulates globin mRNA translation only when Mengovirus RNA is present, but not in its absence (Fig. 7). This explains why, in the presence of competing amounts of Mengovirus RNA, the increase in globin mRNA translation caused by addition of eIF-2 is matched by a decrease in Mengovirus RNA translation (Fig. 5). This observation suggests strongly that the addition of eIF-2 does not lead to an increase in the total number of initiations of translation, a point shown already for the case of a-and pglobin mRNA competition studied in the same circumstances (Di Segni et al., 1979). The fact that eIF-2 acts to shift translation in favor of globin synthesis shows clearly that globin mRNA and Mengovirus RNA compete for eIF-2. The addition of eIF-2 leads to relief of competition because it increases the number of eIF-2 molecules available, allowing thereby a relatively greater number of the more weakly binding, but more numerous, globin mRNA molecules to interact with eIF-2. Thus, although the total number of initiation events does not increase, the proportion of globin mRNA molecules entering initiation complexes will increase, up to the point where binding of globin mRNA and Mengovirus RNA is exactly according to their molar ratio, and there no longer is competition.
The Interaction between mRNA and eIF-2"The present results support the concept that a direct and specific interaction between eIF-2 and mRNA occurs during protein synthesis and contributes significantly to translational control. A critical role for the affinity of mRNA for eIF-2 in protein synthesis is also suggested by the observation that dsRNA, an agent known to inactivate eIF-2 (Kaempfer and Kaufman, 1973;Kaempfer, 1974;Clemens et al., 1975), fails to inhibit the translation of Mengovirus RNA in conditions where translation of globin or mouse ascites cell mRNA is blocked cornpletely; direct RNA-binding studies reveal that &RNA cornpetes with mRNA for eIF-2, binding this factor more strongly than globin mRNA, but more weakly than Mengovirus RNA (Rosen et al., 1981).
Evidence for a specific interaction between eIF-2 and mRNA is furnished by a number of observations. Binding of all mRNA species examined, including species lacking the 5'terminal cap or 3"terminal poly(A) moieties, is fist order in eIF-2, pointing to the existence of a high affinity binding site in mRNA (Kaempfer et al., 1978a. By contrast, RNA species not serving as mRNA, such as negative strand RNA (Kaempfer et al., 1978a), tRNA (Barrieux and Rosenfeld, 1977; and rRNA (Barrieux and Rosenfeld, 1977) bind much more weakly. eIF-2 forms an equimolar complex with globin mRNA, possessing an apparent Kd of less than lo-' M at physiological salt concentrations, at a site that involves neither the poly(A) tail nor the 3'untranslated sequence of this mRNA . Analogs of the cap that inhibit translation also inhibit the binding of mRNA to eIF-2, as well as ternary complex formation between eIF-2, Met-tRNAf, and GTP, suggesting an interaction of eIF-2 with the modified 5'-end of mRNA (Kaempfer et al., 1978b). However, several types of results indicate that binding of eIF-2 to mRNA is primarily at an internal sequence and secondarily through the cap (Kaempfer et al., 1978b. Direct evidence for the existence of specific binding sites for eIF-2 in mRNA comes from nucleotide sequence analysis. eIF-2 binds specifically to the 5'-end of STNV RNA, a 1,239-nucleotide-long, noncapped messenger, but not to the 5'-ends of fragments of internal origin . The factor does not bind detectably to the 32-nucleotide 5"terminal fragment ending in the AUG initiation codon, or shorter ones, but does bind specifically to the 44-nucleotide 5"terminal fragment that contains the ribosome binding site. These experiments reveal not only that eIF-2 binds to a specific site in STNV RNA, but also that this site overlaps closely with the 40 S ribosome binding site . Studies of the interaction between Mengovirus RNA and eIF-2 lead to a similar conclusion. Upon digestion with ribonuclease TI, the RNA protected in either 40 S or 80 S initiation complexes yields four unique, large oligonucleotides, 17 to 28 bases in length, a l l located between the poly(C) tract and the 3'-end.' Fingerprint analysis of the Mengovirus RNA sequences protected by eIF-2 against ribonuclease TI digestion yields oligonucleotides that are identical with those protected in either 40 S or 80 S initiation complexes.' Since Mengovirus RNA is approximately 7,500 nucleotides long, these results demonstrate high specificity in the recognition of this RNA by eIF-2.
Considering that in the studies with STNV and Mengovirus RNA the interaction between eIF-2 and mRNA was examined in the absence of Met-tRNAf, ribosomes or other components of the protein-synthesizing machinery, the striking resemblance between the binding site for isolated eIF-2 on the one hand, and for the 40 S ribosomal subunit in the presence of all initiation factors on the other, supports the concept that during initiation of translation, binding of a ribosome to mRNA is guided to a significant extent by eIF-2. The findings reported here, showing that Mengovirus and globin mRNA compete in initiation of translation as they do in direct binding to eIF-2 are in harmony with this concept. Competition between mRNA species most likely is for eIF-2 molecules located in 40 S/Met-tRNAf/eIF-2 complexes (see Di Segni et al., 1979).
The implication of the present findings is that the affinity of a given mRNA species for eIF-2 directly influences its translation, relative to that of other mRNA species.