Initiation Factor Activity Associated with Free 40 S Subunits from Rat Liver and Rabbit Reticulocytes

Abstract At least one initiation factor, required for the translation of rabbit globin messenger RNA in a partially purified cell-free system, is found to be associated with free 40 S ribosome subunits in cell extracts from rat liver and rabbit reticulocytes. The free 40 S ribosomes carrying initiation factor activity are devoid of messenger RNA activity. The initiation factor activity is dissociated from the 40 S particles by high salt concentration and then is found in the ribosome-free upper region of sucrose density gradients. The resulting salt-washed free 40 S subunits are still active in protein synthesis if supplemented with initiation factors. Our results imply that the free 40 S ribosomes in mammalian cells are intermediates in the ribosome cycle of protein synthesis and may represent at least in part true initiation complexes prior to mRNA binding.


Initiation
Factor Activity Associated with Free 40 S Subunits from Rat Liver and Rabbit Reticulocytes (Received for publication, January 31, 1974) INGEGERD C. SUNDKVIST, WALLACE L. MCKEEHAN, MAX H. SCHREIER, AND THEOPHIL STAEHELIN* From the Base1 Institute for Immunology, CH-4068 Basel, Switzerland SUMMARY At least one initiation factor, required for the translation of rabbit globin messenger RNA in a partially purified cell-free system, is found to be associated with free 40 S rlbosome subunits in cell extra&s from rat liver and rabbit reticulocytes. The free 40 S rlbosomes carrying initiation factor activity are devoid of messenger RNA activity. The initiation factor activity is dissociated from the 40 S particles by high salt concentration and then is found in the ribosome-free upper region of sucrose density gradients. The resulting saltwashed free 40 S subunits are still active in protein sjllthesis if supplemented with initiation factors. Our results imply that the free 40 S ribosomes in mammalian cells are intermediates in the ribosome cycle of protein synthesis and may represent at least in part true initiation complexes prior to mRNA binding.
In both bacterial and mammalian cells initiation factors for protein synthesis are usually extracted by 0.5 to 1 M KC1 or NH&l from crude total ribosome preparations. For Eschetihm coli it has been well established that in the process of polypeptide chain initiation the initiation factors interact with the small ribosome subunits, fMet-tRNA and presumably mRNA (for review see Refs. l-3). Furthermore, free 30 S bacterial ribosome subunits, or proteins derived from them by high salt treatment, can be used as the only source of initiation factors in bacterial systems (4-6).
The quantitative distribution of the individual initiation factors between ribosomes and the high speed supernatant cell fractions in eucaryotes may well differ from one species to another or even from one tissue to another. In immature red blood cells initiation factors required for protein synthesis in reconstituted systems have been obtained so far from crude ribosomes only (7-9). Thii does not exclude that at least some factor(s) may also occur free in the cell sap. The only initiation factor characterized and purified from Artemia sdina embryos (10) was found exclusively in the high speed supernatant of cell extracts. The apparent functional equivalent of the Artemia factor was isolated from high speed supematants from rat liver (ll-13), rat muscle (12), and mouse Krebs ascites cells (12). * To whom correspondence and reprint requests should be sent.
In cells active in protein synthesis one would expect some or all of the initiation factors to be associated at least temporarily with ribosomes during initiation complex formation. Whether one or more initiation factors have such a high affinity to ribosomes that, under physiological ionic conditions, their existence in the high speed supernatant is practically excluded is an open question at present. Furthermore, initiation factors obtained from the ribosome fraction in mammalian cells have been extracted only from the whole ribosome population. So far nothing is known about the distribution of initation factors within the ribosome population.
In this study, however, we show for both rat liver and rabbit reticulocytes that initiation factor activity required for the cellfree translation of rabbit globm mRNA is found associated with free 40 S ribosome subunits at low KC1 concentrations in these cell extracts. The initiation factor activity is removed from the subunits by treatment with high salt. No endogenous mRNA seems to be associated with free 40 S subunits.
The polysomes were converted into single ribosomes by a "run-off" in an in vitro protein-synthesizing system, using a pH 5 enzyme preparation from rat liver as a source for all the components needed for polypeptide chain elongation and termination. After incubation, KC1 was added to a final concentration of 0.4 M to dissociate the single ribosomes into subunits. The subunits were separated by sucrose density gradient centrifugation and the 40 and 60 S peak fractions were concentrated by centrifugation to pellet the subunits. The ribosome subunits were resuspended in 0.1 M KCl, 4 mM MgCls, 0.02 M Tria-HCl (pH 7.6), 1 my dithiothreitol, and stored at -80" until used.
Rat Liver pH 6 Enzymes-Rat liver pH 5 enzymes were prepared according to Falvey and Staehelin (14).
In Vitro Protein Synthesis-The system is basically that described by Schreier and Stsehelin (9). Incubation mixtures, 0.1 ml, contained free 40 S ribosome subunits present in pH 5 enzyme preparations of either rat liver or rabbit reticulocyte high speed supernatante, or polysome-derived 40 S subunits. In either case polysome-derived 60 S subunits were added. A standard assay contained 0.36 AZ60 units of 60 S subunits and 0.14 A160 units of 40 S subunits or, as indicated for each experiment, 25 ~1 of rat liver pH 5 enzymes, 2 pg of rabbit globin mRNA, and the following reactants per ml: 20 pmoles of creatine phosphate, 3.75 units of 6513 mole of dithiothreitol, 30 nmoles of each L-amino acid and 0.75 &i of [14C]leucine (final specific activity, 23 mCi per mmole). Incubation wae for 40 min at 30". The samples were processed for hot trichloroacetic acid (lo%)-insoluble radioactivity measurements as described previously (9). One thousand counts per min correspond to 26 pmoles of leucine.
If exogenous initiation factors were used for control experiments, they were prepared as an "A-fraction" from rabbit reticulocyte ribosomes as described previously (9).
Density Gradient Centrijugation-Density gradient centrifugation of pH 5 enzymes for analysis or preparative use of free 40 S subunits were done in sucrose or glycerol gradients.
Centrifugation was at 50,600 rpm at 4" for 2.5 hours without using the brake in a Beckman SW 50.1 rotor. The absorption was monitored continuously as previously described (9) and the preparative gradients were collected into 0.25-ml fractions.

RESULTS
CorreZatiun between Znitiatwn Factor Activity and Presence of Free 40 S Subunits in pH 6 Enzyme Preparations-A highly efficient mammalian system for the in vitro translation of exogenous rabbit globin mRNA has been reported previously (9). This partly purified cell-free system was shown to be dependent on the addition of initiation factors and messenger RNA. However, we noticed that the system also was able to promote at low, but significant, efficiency de norm synthesis of rabbit globin in the absence of added initiation factors. This synthesis was found to be inversely related to the length of centrifugation time used to prepare the high speed supernatant for the pH 5 enzyme preparation. These rat liver pH 5 enzymes contain all the components required for amino acid activation, polypeptide elongation, and termination in the cell-free system. Sucrose gradient analysis of different pH 5 enzyme preparations revealed a striking correlation between the amount of 40 S subunits ( Fig.  1) and their ability to promote de nova protein synthesis in the presence of polysome-derived 40 and 60 S subunits and rabbit globin mRNA (Table I). It should be noted that the synthetic activities of the pH 5 enzyme Preparations A, B, and C without added initiation factors were about 22, 7, and l%, respectively, compared to similar incubation mixtures supplemented with saturating amounts of partially purified rabbit reticulocyte initiation factors (results not shown).
Free .&I S Subunits Present in pH 6 Enzymes Participate in Globin Synthesis-The results described above do not prove that the free 40 S subunits present in rat liver pH 5 enzymes have initiation factor activity or can participate in protein synthesis. Therefore, we used the pH 5 enzyme Preparations A, B, and C as the exclusive source of both initiation factors and 40 S ribosome subunits in the cell-free system. Increasing amounts of polysome-derived 60 S subunits were added to incubation mixtures containing globin mRNA, pH 5 enzymes (A, B, or C) and the other components required for protein synthesis (Fig. 2). The results show that the rat liver 40 S subunits present in the pH 5 enzyme preparations actually participate in globin synthesis together with the added mouse liver 60 S subunits. The activity of pH 5 enzyme Preparation A in the absence of added 60 S subunits is due to the small amount of 60 S subunits present Fifty microliters of pH 5 enzyme preparation were analyzed on each gradient. in this pH 5 enzyme preparation (Fig. 1A). From Curves A and B in Fig. 2 we calculate that maximum activity was reached at a 60:40 S Azso ratio of about 2.8 which corresponds to a molar ratio of slightly more than 1. In the linear range of Curue A of Fig. 2 FIQ. 3. Distribution of initiation factor activity of ratliver pH 5 enzymes centrifuged either in low salt or high salt glycerol density gradients.
Two 1.5-ml samples of liver pH 5 enzymes (Fig. 1A) were adjusted to the ionic conditions of preparative glycerol gradients containing either low salt (0.03 M KCl, 0.002 M MgClr) or high salt conditions (0.5 M KCl, and 0.003 M MgCls). The samples were layered onto the 3.5-ml convex gradients (see "Methods and Materials") and centrifuged as described under "Methods and Materials." After centrifugation the optical density was recorded and 0.25-ml fractions were collected. Ten microliters of each fraction was assayed for initiation factor activity with a complete globin-synthesizing system containing polysome-derived subunits and an absolutely ribosome-free pH 5 enzyme preparation.
The results from both gradients are plotted into the same figure. Initiation factor activity measured as protein synthesizing activity: low salt gradient, O---0 and high salt gradient, l ---0 ; 1000 cpm = 26 pmoles of leucine incorporated. tion were assayed in a complete globin synthesizing system lacking only initiation factors but containing polysome-derived 40 and 60 S subunits. The results are shown in Fig. 3. Due to overloading, the 40 S absorbance peak is only visible as a shoulder. For convenience the results from both gradients were plotted into the same figure. The initiation factor activity coincides with the position of the 40 S subunits in the gradient containing 0.03 M KCl. However, in the gradient containing 0.5 M KC1 all the initiation factor activity is found on the top of the gradient. The sample had been adjusted to 0.5 M KC1 before layering on the high salt gradient. Since all the assays for globin synthesis now contained 40 S-free pH 5 enzymes, we conclude from these experiments that at least one initiation factor occurs at low ionic strength exclusively on the 40 S subunits, while others might be free or subunit-bound, or both. As expected, 0.5 M KC1 completely dissociates factor activity from the subunits. Protein Synthesis Promoted by Free .JO S Subunits is mRNA-Dependent-In Fig. 4 we show the mRNA dependence of protein synthesis by free 40 S ribosome subunits in rat liver pH 5 enzymes supplemented with an equimolar amount of polysomederived 60 S subunits. Each assay contained about 2 pmoles each of free 40 S subunits and polysome-derived 60 S subunits. Globin messenger RNA was titrated into the system and at saturation about 2.5 pmoles of globin were synthesized per pmole of ribosomes. The strong dependence on, and very high response to exogenous mRNA indicate that the majority of the free 40 S subunits isolated with pH 5 enzymes is at a functional stage of the ribosome cycle prior to mRNA binding.

Experimenls with pH 6 Enzyme Preparations
Obtained from Rabbit Reticulocytez-In order to substantiate and generalize our findings with free 40 S subunits from rat liver, we did similar experiments with pH 5 enzyme preparations obtained from high speed supernatants of rabbit reticulocytes. Whereas from rat liver only the upper three-fourths of the high speed supernatant The gradients were centrifuged and collected as previously described (Fig. 3). Twenty microliters of each fraction were assayed for initiation factor activity with a complete globin-synthesizing system containing an absolutely ribosome-free pH 5 enzyme preparation. The low salt gradient also was assayed with an incubation mixture without globin messenger. A, low salt gradient. Initiation factor activity in the presence (O---0 ) and absence (A.. . * *A) of messenger. B, high salt gradient. Initiation factor activity in the presence of messenger (O---0 ); 1000 cpm = 26 pmoles of leucine incorporated.
per centrifuge tube was used for pH 5 enzyme preparation, from rabbit reticulocytes the total supernatant was used. Accordingly, considerably more 40 S subunits as well as 60 S subunits and 80 S ribosomes were present in the reticulocyte pH 5 enzyme preparations. Fig. 5 shows preparative glycerol gradients of reticulocyte pH 5 enzymes in 0.03 M KC1 (Panel A) and 0.5 M KC1 (Panel B). Again we assayed each gradient fraction for initiation factor activity in the cell-free protein synthesizing system in the presence of polysome-derived 60 and 40 S mouse liver subunits, globm mRNA, and rat liver pH 5 enzyme preparation free of 40 S ribosome subunits. The results are identical with those obtained with rat liver pH 5 enzymes in Fig. 3. With 0.03 M KC1 in the gradient all the initiation factor activity coincides with the absorbance of the 40 S subunits. No initiation factor activity is found with the 80 S monosomes and most likely none with the 60 S subunits (Fig. 5A, open circles). Exposure to 0.5 Y KC1 dissociates all the initiation factor activity from the subunits (Fig. 5B).
Each fraction of the low salt gradient also was assayed for The same high salt glycerol gradient as shown in Fig. 5B was used. Ten microliters of each gradient fraction were tested for protein synthesis in the cell-free system containing ribosome-free pH 5 enzymes, polysome-derived 60 S subunits, globin mRNA, and the other components required; initiation factors were supplied by 10 ~1 of pooled Fractions 16 and 17; 1000 cpm = 26 pmoles of leucine incorporated. messenger activity using the same system as above but with the omission of globin mRNA. The result demonstrates that in reticulocytes, too, no endogenous mRNA is associated with the free 40 S subunits (Fig. 5A, open t&mgZes).
In the following experiment we tested the activity of the 0.5 M KCl-washed 40 S subunits present in the glycerol gradient shown in Fig. 5B, when supplemented with the washed off initiation factors (gradient Fractions 16 and 17), polysome-derived 60 S subunits, and globin mRNA. The result is shown in Fig. 6. The washed 40 S subunits recovered in the glycerol gradient were indeed active in protein synthesis in the presence of initiation factors recovered from the top of the same gradient. The relatively high incorporation in the absence of 40 S subunits was due to contamination of polysome-derived 40 S subunits in the polysome-derived 60 S subunit preparation used.

DISCUSSION
In this study we show for two mammalian cell types of different species the free 40 S ribosome subunit as the carrier of initiation factor activity. Since 40 S-free pH 5 enzymes cannot promote protein synthesis in our system, they must lack at least one component essential for chain initiation. Thus, one or more initiation factor(s) are exclusively bound to the 40 S subunit at the low ionic strength used here.
By in vi&o reconstitution from polysome-derived 40 S subunits and purified initiation factors we have demonstrated a class of "40 S" subunits carrying the large initiation factor IF-E& (16). IF-E8 could indeed be purified from isolated free 40 S subunits of mouse Krebs II ascites cells and rabbit reticulocytes.2 This factor seems to be bound exclusively to ribosomes under our conditions of cell disruption and ribosome isolation, whereas other initiation factors are also detectable in the ribosome-free supernatant.s The results presented here give strong support to the interpretation of Ayuso-Parilla et al. (17)  7. Dependence on the KC1 concentration of initiation _. factor extraction from guinea pig liver microsomes.
Seven guinea pig livers weighing about 20 g each were homogenized in a low ionic strength buffer as described previously (24) The final volume was 80 ml which were divided into four equal fractions.
In three of them the KC1 concentration was raised to 0.3, 0.5, and 0.7 M, respectivelv. followed bv 20 min stirrine at 0'. The four fractions were centkfuged for 3 hours at 60,000 Grn in the IEC A321 angle rotor. The supernatants were fractionated by ammonium sulfate.
The proteins precipitated between 25 and 65% saturation were dissolved in 3 ml each of 0.12 M KCl, 0.02 M Tris-HCl (pH 7.6), 0.2 rnM EDTA, 1 mM dithiothreitol, and 5% glycerol and dialyzed against the same buffer. They then were tested for initiation factor activitv in the cell-free svstem described under "Methods and Mater&. some or possibly all of the extra protein they found associated with free 40 S subunits obtained from Ehrlich ascites cells might be due to initiation factors. Very few, if any, free 40 S subunits have been shown to carry mRNA (18-21) but a significant amount seem to have bound initiator . To what extent free subunits serve as carrier of initiation factors not directly engaged in initiation cannot be determined for the moment.
This problem may be a question of semantics depending on the definition of individual steps of initiation complex formation.
As it has been shown, however, that the initiation of protein synthesis occurs on ribosomes drawn from the pool of free or so-called native subunits in the mammalian cell (22,23), we would suggest that free 40 S subunits represent at least in part true intermediates in a stage prior to messenger binding in the process of chain initiation carrying specifically the factors required for their formation.
Further evidence for this conclusion will be published elsewhere.2 Another problem of importance is the stability of initiation factor-ribosome complexes. The conditions used for the isolation procedure may determine whether a given initiation factor wilI be mainly bound to ribosomes and therefore called a "ribosome factor" or dissociated into the ribosome free supernatant and be called a "supernatant factor" or "cytosol factor." In order to illustrate this problem, which may be a serious source of confusion between diierent investigators, we extracted protein from a liver microsome fraction with different concentrations of KC1 and tested the protein fractions for initiation factor activity in our cell-free system. The results are shown in Fig. 7. Even at 0.1 M KC1 significant amounts of the most tightly bound initiation factor(s) were extracted. The extraction was complete at 0.5 M KCl. To what extent the hydrostatic pressure during centrifugation helped dissociate initiation factors from ribosomes at the lower KC1 concentrations is not known.
However, this experiment at least suggests that even the initiation factor with the highest affinity to ribosomes might be partly free under the ionic conditions of protein synthesis. Although diierent organisms, tissues, or developmental stages may show differences in the distribution of homologous initiation factors between ribosomes and supernatant under the same fractionation procedure, any classification of initiation factors as either "ribosome factor" or "supernatant factor" must be strictly operational and will be dependent on the experimental conditions. i: 3.