Protein Synthesis in Yeast I. PURIFICATION AND PROPERTIES OF ELONGATION FACTOR 3 FROM SACCHAROMYCES CEREVISIAE*

Elongation factor 3 from the yeast Saccharomyces cerevisiae was purified over 230-fold from a high speed supernatant fraction. The homogeneity of the protein was shown by gel filtration and sedimentation equilibrium analysis of the native protein and by sodium dodecyl sulfate gel electrophoresis of the denatured protein. The molecular weight of the protein was esti- mated to be 125,000 by the above-mentioned methods. The protein single dif€usion

k To whom all correspondence should be addressed. cell-free system from yeast spheroplasts and have shown synthesis of intact proteins in response to exogenous messengers (2,3). Prior to this, Richter and Klink, Richter and Lipmann, Spremulli and Ravel, and Torano et al., had isolated elongation factors 1 and 2 from yeast and had shown that they carry out functions corresponding to those of elongation factors from other eukaryotic sources (4-7). Skogerson isolated analogous factors from Saccharomyces cereuisiae (8). During the isolation procedure, he and his associates identified a third protein that they found to be uniquely required by the yeast ribosomes. Skogerson and Engelhardt showed that yeast ribosomes were completely inactive in poly(U)-dependent polyphenylalanine synthesis in the absence of the third factor. Ribosomes from liver, reticulocyte and from the brine shrimp Artemia salina were not influenced by this material. This yeast protein was referred to as elongation factor 3 by Skogerson (8). The current paper describes the purification of the yeast EF3,' its amino acid composition, some of its functions, and its possible role in protein synthesis.

RESULTS
Factor 3 that is required by the yeast ribosomes for polyphenylalanine synthesis has been purified to homogeneity. Table I shows the yield in the peak fractions and the purification obtained in each step. The coincidence of the protein and the activity peak and the constant specific activity across the peak of the final Sephadex G-200 column ( Fig. 1) are taken as criteria of purity. The protein was also found to be homogeneous by sodium dodecyl sulfate-gel electrophoresis ( Fig. 2) and by sedimentation equilibrium analysis (Fig. 3).
The upward curvature of the 2 versuslog concentration graph ( Fig. 3) indicates a tendency of aggregation of this protein, although gel electrophoresis under nondenaturing conditions did not reveal this property (data not shown). The molecular weight of this protein calculated from the linear portion of the graph, and representing most of the material in the cell, was 110,OOO. Estimation of the molecular weight of the native enzyme by gel fiitration (Fig. 4) and of the denatured protein by sodium dodecyl sulfate-gel electrophoresis (Fig. 2) using molecular weight markers gave a value of 125,000, indicating that this protein consists of a single polypeptide chain. Fig. 5 shows the pattern of cross-reaction of EF3 in an Ouchterlony double-diffusion plate with the antiserum raised against the pure protein. Lack of cross-reaction of anti-EF3 with the elongation factors 1 and 2 (Fig. 5, wells 5 and 61, indicates that factor 3 is antigenically unrelated to either of the other two proteins. Table 11 shows the amino acid composition of factor 3. There is no specific or unusual feature in the amino acid composition of the protein. In comparison to the elongation factor 1, the content of the hydrophobic amino acids and the basic amino acids are relatively low and the acidic amino acids are high in factor 3. The acidic isoelectric point of 5.9 is consistent with the amino acid composition. Early experiments on a third factor requirement for polyphenylalanine synthesis by yeast ribosomes were done with partially pure EF3 that contained several contaminating proteins. Therefore, it was not possible to demonstrate whether the effect on polyphenylalanine synthesis was due to one or more components present in the preparation. With the homogeneous preparations of all three factors, it was possible to demonstrate unequivocally that, besides EF1 and EF2, the yeast ribosomes required one additional factor. With the homogeneous preparation of EF3, one could unequivocally demonstrate the dependence of the system on the addition of this protein. Fig. 6 shows dependence of the polymerization reaction on EF3 concentration and the effect of anti-EF3 antisera on the reaction. Poly(U)-dependent polyphenylalanine synthesis was completely inhibited by the immune serum. The inhibitory effect was overcome by the excess of EF3. Control serum had no effect on the system. One possible explanation for the requirement of an additional factor by the yeast ribosomes could be that this protein results from the separation of one of the two elongation factors into two components. In that case, one would expect to see cross-reaction of anti-EF3 with EF1 or EF2. EF3 should also stimulate the partial reactions catalyzed by one of the two proteins. Lack of cross-reaction of EF3 antisera with EF1 and EF2 (Fig. 5) indicates that the proteins are not of the same A possible effect of the third factor on EF1 activity could be followed by measuring the stimulation of aminoacyl-tRNA binding to the ribosome. Similarly, an effect on EF2 activity could be measured as stimulation of EF2-dependent nucleotide binding to the ribosome in the presence of the antibiotic fusidic acid, which prevents recycling of EF2 (17). In a series of elegant experiments, Skogerson and Engelhardt had shown that EF3 does not affect the function either of EF1 or EF2 to a significant extent (10). Similar results were obtained by us with purified EF3. EFl-dependent binding of aminoacyl-tRNA to the ribosome is very rapid and is not greatly affected by temperature (1). A 2-fold stimulation of Phe-tRNA binding to the ribosome by EF3 at 0 "C and only at saturating concentrations of EF1 cannot explain the requirement of this factor for polyphenylalanine synthesis by the yeast ribosomes. In fact, liver ribosomes that do not require EF3 also showed similar stimulation by this protein (10).
We attempted to study the role of EF3 in peptide bond synthesis. Although EF1-dependent binding of aminoacyl-tRNA to the ribosomal A-site is temperature-independent, the peptide bond synthesis may be facilitated at a higher temperature. Therefore, the experiment was carried out at 30 "C instead of at 0 "C. However, the analysis of dipeptide synthesis was complicated by the presence in the ribosome preparation, sufficient amount of EF1 and EF3 to carry out the complete elongation cycle. This diffkulty was overcome by including anti-EF2 in the assay along with the ribosome and the factors. As can be seen in Fig. 7, addition of EF3 or Origin. EF1 + EF3 to the ribosome resulted in increased binding of [14C]Phe-tRNA to the nitrocellulose membrane, but the extent of binding was reduced by a large factor when anti-EF3 was added in the system. We analyzed the products from both sets of experiments by alkali hydrolysis of the membranebound radioactivity (17) and by paper chromatography (18) of the total and the ethyl acetate-extractable products (19). The analyses showed the formation of large oligomers in the absence of anti-EF2. When anti-EF2 was present in the reaction mixture along with EF1 and EF3, the bulk of the nitrocellulose-bound product, upon alkali extraction and subsequent paper chromatography, appeared as free amino acid and not as dipeptide (data not shown). From this experiment, we concluded that the stimulatory effect of EF3 on EF1dependent Phe-tRNA binding to the ribosome is not due to enhanced rate of dipeptide synthesis but reflects increased binding of the aminoacyl-tRNA to the ribosome.
Elongation factors 2 and 3 both show ribosome-dependent GTP hydrolysis independent of each other. The rate of GTP hydrolysis by EF3 is 10 times faster than by EF2. However, EF2 has higher affinity for GTP ( K , is 10 p~) .
Moreover, GTPase activity of EF2 was 80% inhibited at a substrate concentration of 60 p~ (data not shown). The K,,, of EF3 for GTP is much higher (125 p~) and the reaction was not inhibited by high GTP concentrations. Whether GTP hydrolysis by EF3 plays any role in peptide elongation needs further experimentation.
Factor 3 also has strong ATPase activity. In fact, the rate of ATP hydrolysis was 3 times faster than the rate of GTP hydrolysis. The K , for ATP is 55 p~. Neither EF1 nor EF2 showed ATPase activity. Both GTPase and ATPase activities of EF3 were inhibited by the immune serum at comparable concentrations and were expressed after the addition of excess of EF3 (Figs. 8 and 9).
The process of elongation involves recycling of the factors. Elongation factor 1 forms a ternary complex with GTP and aminoacyl-tRNA that binds to the ribosome. In bacterial systems, GTP is hydrolyzed and Tu:GDP is released after delivering the aminoacyl-tRNA to the A-site of the ribosome. A low molecular weight protein EFTS functions in recycling of EFTu by the following exchange reaction: EFTu:GDP + GTP EFTu:GTP + GDP A protein analogous to bacterial EFTS has been reported to occur in eukaryotic systems (1). This protein, EFlP, required for the polymerization reaction, stimulates the binding of Phe-tRNA and the nucleotide exchange reactions with limiting amount of EFla. The protein EFlP has often been isolated in association with EFla. In yeast we have as yet no indication of the existence of EFlP or a high molecular weight form of EFla. Elongation factor 3, which is uniquely required by the yeast ribosomes, does not stimulate the nucleotide exchange reaction (data not shown). On the basis of this experiment and the results discussed earlier, we conclu(4e that EF3 does not correspond in function to EFlP. In the bacterial system, one of the ribosomal proteins, SI, has been shown to dissociate from and reassociate with the ribosome very easily. Poly(U) -dependent polyphenylalanine synthesis is stimulated by this protein and the translation of natural message is almost completely dependent on the addition of S1. Khanh et al. (21) have shown that the protein S1 binds strongly to synthetic and natural messages. The requirement of EF3 by the yeast ribosomes and not by the liver ribosomes indicates that this protein could be an easily dissociable ribosomal protein like the bacterial S1. Therefore, we tested the poly(U)-binding property of EF3 and checked the cross-reactivity of EF3 with the antiserum to the bacterial S1 (kindly performed by A. Subramanian of the Max Planck Institute, West Germany). The binding of poly(U) to EF3 was very weak and there was little cross-reaction of this protein with S1 antisera (data not shown). This indicates that yeast EF3 is very probably not a ribosomal protein analogous to the bacterial S1 in function.

DISCUSSION
The unusual requirement of a third protein factor by the yeast ribosomes for poly(U)-directed polyphenylalanine synthesis prompted us to characterize this protein and to study its function in the elongation cycle.
Although the elongation factors 1 and 2 are interchangeable with homologous factors from other eukaryotic sources, factor 3 is not required by ribosomes other than those isolated from yeast. It is possible that factor 3 is species-specific and, therefore, does not function in heterologous systems. Indeed, antiserum raised against EF3 did not affect the polymerization reaction by the liver ribosome^.^ Another possibility is that EF3 is a loosely bound ribosomal protein and is removed easily during the isolation procedure from yeast ribosomes whereas liver ribosomes and the ribosomes from other sources are saturated with an analogous protein and are unaffected by this factor. An unlikely possibility is that yeasts evolved differently and have a unique requirement for an additional soluble protein factor to carry out the elongation process.
Skogerson and Engelhardt had studied some of the partial reactions to determine which of the three elongation cycle reactions was affected by EF3. Our results with homogeneous EF3 preparation is in good agreement with the results obtained by them with impure preparations. As previously discussed, EF3 was not required for the EFl-and GTP-dependent binding of Phe-tRNA. The difference in the reaction observed by us between 0 "C and 30 "C could be accounted for by the stimulation of the polymerization reaction at 30 "C due to contaminating amounts of the other two factors present in the ribosome. The 2-fold stimulation of Phe-tRNA binding by EF3 at 0 "C with saturating concentrations of EF1 was due to increased binding of Phe-tRNA to the ribosome. No effect of EF3 was seen in the EF2-dependent binding of GDP to ribosomes in the presence or absence of fusidic acid (9). Ribosome-dependent GTPase activity of EF2 did not require the presence of EF3 although EF3 itself has a much stronger GTPase activity than EF2. Further studies are needed to determine the correlation between the high K , of EF3 for GTP, inhibition of EF2 by higher GTP concentration and GTP hydrolysis in protein synthesis, Poly(U)-dependent polyphenylalanine synthesis is stimulated by ATP. This observation was originally reported by Skogerson (8) . We have confirmed this result with the purified factors and showed that the ATP also stimulated polyphenylalanine synthesis in other eukaryotic systems. Elongation factor 3 has strong ATPase activity, dependent upon yeast ribosomes. These ribosomes were prepared by the two-phase extraction method (22) and were devoid of ATPase activity, whereas the liver ribosomes prepared by several different methods including the two-phase extraction method showed K. Chakraburtty and B. Dasmahapatra, unpublished results. strong ATPase activity and do not require the addition of EF3 for the polymerization reaction. A possible correlation of the ATP effect, the ATPase activity of EF3, and its requirement by the yeast ribosomes is under current investigation.
Translation of synthetic messages generally does not require initiation factors although polymerization must be initiated by some mechanism. Wahba et al. had shown a stimulatory effect of initiation factor 3 on poly(U)-directed polyphenylal.. anine and poly(A)-directed polylysine synthesis with Escherichia coli ribosome (23). The possibility that the newly discovered elongation factor is actually an initiation factor cannot be ruled out at this point. -s~a e stram 0-587-48 had been routinely used in this laboratory. We had also prepared ribosomes and factors from the Straln A2248 and had not noticed any variation among strams of yeast in terms of factor stabil?ty. Chromatographic mobillty or functional properties.
Growth Of Cells and Preparatlon Of Cell Extract: Saccharomyces mtesy of Prof. J u l l a n Davlesl ~n enriched media 10.5% yeast extract; 1% peptone and 2% glucose1 and were collected by continuous flow centrlfuqation, washed once wlth buffer and stored at -1OO'C.
The cells were grown to early log phase ln large fermentors lcour-These LnClude grinding in a blender wlth five volumes of glass beads for 1 mln or disrupting vlth a French press. Ribosomes seemed to be more actlve when prepared by the former method. ThlS could be due to the fact that fewer vacuoles were disrupted under milder condition releasing fewer pressure dlsrnptwn. For larqe scale preparations, we routinely uee the degrading enzymes. nowever. factor activity Seemed to be unaffected by French press. Several methods had been successfully used to rupture yeast cells.

ImYnlzatlOn of Rabblts and Purification of Antisera.
Rabblts vere injected in the toe pads with 100 Yg Of the factor In complete adluvant. lwo Subcntaneous booster8 o f 5 0 u q were admmlstered at 2 week intervals. The Ouchterlony test was performed after each bleedlng 111). Antlserm was purified by (NHII~SO~. precipitation and by DEAE cellulose chromatogqaphy (12). The IgG fraction was eluted in 20 nW NaC1, 3 0 ml4 phosphate pH 8 . 0 . This fraction was concentrated by the addrtion of equal volume of saturated ( N H~I~S O~. Dialyzed antisera were Stored in small fractions at -20nC. Control serum was prepared by identical methods from the blood withdrm m before immunization. to 0 . 5 M Lnmediately after low Speed oentrifugatlon. This procedire seemed to impmve the yleld of actlvc r~bonomes although the bulk of the elongation factore, includrng the factor 3 . appear in the supernatant fracflon after low salt ( 5 0 ml extraction of the ribosomes. Fnrther pulficatxon Of the r~bosomes was obtalned by pelleting through a 15% glycerol contaming buffer.
cause Of the Y base near the anticodon, blndrtzhtly to the BD-cellulose column and IS easily Beparated from the bulk Of the tRNA. Preparations of rRNAPhe at this stage vere further Purified by a second run on the phe-tRNAPhe from the Second run is more than 60% pure. Amnoacylation of same column after charging with radiolabeled phenylalanine. The Charged tRw.Phe was carried out under standard ConditlOns of charging (151 uslng partially purified phenylalanyl-tRNA synthetase. Charged tRmPhe after phenol extractlo" and ethanol precipitation v a~ further purifled by gel material was lyophilized and Stored at -100'C.