Further characterization of eukaryotic initiation factor 5 from rabbit reticulocytes. Immunochemical characterization and phosphorylation by casein kinase II.

Eukaryotic initiation factor (eIF)-5, isolated from rabbit reticulocyte lysates, is a monomeric protein of Mr = 58,000-62,000. Immunochemical methods were employed to identify eIF-5 in crude cell lysates. Antisera against purified denatured eIF-5 were prepared in rabbits and characterized by immunoblotting and immunoprecipitation techniques using native and denatured eIF-5 as antigens. Monospecific antibodies to denatured eIF-5 were affinity-purified using eIF-5 blotted onto aminophenylthioether paper. Rabbit reticulocytes, HeLa cells and mouse L cells were lysed directly into a denaturing buffer containing 3% sodium dodecyl sulfate. The denatured proteins were analyzed by polyacrylamide gel electrophoresis followed by immunoblotting with anti-eIF-5 antibodies. With each lysate, one major immunoreactive polypeptide was observed whose molecular weight corresponded to that of purified eIF-5 (Mr = 58,000-62,000). No degradation products or precursor forms of molecular weight higher than 62,000 were detected in any lysate. These results indicate that isolated eIF-5 is the same size as that found in crude lysates. Additional characterization of eIF-5 indicates that purified eIF-5 can be phosphorylated at serine residues in vitro by casein kinase II. Furthermore, in vitro phosphorylated eIF-5 retains full biological activity in catalyzing the joining of 60 S ribosomal subunits to a preformed 40 S ribosomal initiation complex to form an 80 S initiation complex. Based on its specific activity, we demonstrate that 1 pmol of rabbit reticulocyte eIF-5 mediates the formation of approximately 180 pmol of 80 S initiation complex under the conditions of in vitro initiation reactions.


Further Characterization of Eukaryotic Initiation Factor 5 from Rabbit Reticulocytes
Eukaryotic initiation factor (eIF)-6, isolated from rabbit reticulocyte lysates, is a monomeric protein of M, = 58,000-62,000. Immunochemical methods were employed to identify eIF-5 in crude cell lysates. Antisera against purified denatured eIF-5 were prepared in rabbits and characterized by immunoblotting and immunoprecipitation techniques using native and denatured eIF-6 as antigens. Monospecific antibodies to denatured eIF-5 were affinity-purified using eIF-6 blotted onto aminophenylthioether paper. Rabbit reticulocytes, HeLa cells and mouse L cells were lysed directly into a denaturing buffer containing 3% sodium dodecyl sulfate. The denatured proteins were analyzed by polyacrylamide gel electrophoresis followed by immunoblotting with anti-eIF-6 antibodies. With each lysate, one major immunoreactive polypeptide was observed whose molecular weight corresponded to that of purified eIF-6 (M, = 68,000-62,000). N o degradation products or precursor forms of molecular weight higher than 62,000 were detected in any lysate. These results indicate that isolated eIF-5 is the same size as that found in crude lysates. Additional characterization of eIF-5 indicates that purified eIF-6 can be phosphorylated at serine residues in vitro by casein kinase 11. Furthermore, in vitro phosphorylated eIF-6 retains full biological activity in catalyzing the joining of 60 S ribosomal subunits to a preformed 40 S ribosomal initiation complex to form an 80 S initiation complex. Based on its specific activity, we demonstrate that 1 pmol of rabbit reticulocyte eIF-6 mediates the formation of approximately 180 pmol of 80 S initiation complex under the conditions of in vitro initiation reactions.
Eukaryotic initiation factor 5 (eIF-5)l catalyzes the hydrolysis of GTP bound to a 40 S ribosomal polypeptide chain initiation complex with the simultaneous joining of a 60 S ribosomal subunit to form an 80 S ribosomal polypeptide chain initiation complex (for a review, see Refs. 1-3). In our laboratory, using an assay that directly measures the formation of an 80 S polypeptide chain initiation complex from a * This investigation was supported in part by Grants GM15399 and P30CA13330 from the National Institutes of Health and Grant NP-636 from the American Cancer Society. 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.
The abbreviations used are: eIF-2, eIF-5, eukaryotic initiation factors 2 and 5, respectively; SDS, sodium dodecyl sulfate; TBS, Trisbuffered saline; PVDF, polyvinylidene difluoride; HPLC, high performance liquid chromatography. Sankar Ghosh, Jorge Chevesich, and Umadas Maitra From the DeDartment of DevebDmental Bioloev and Cancer, Division of Biology, Albert Einstein College of Medicine, preformed 40 S initiation complex, eIF-5 was purified to apparent electrophoretic homogeneity from the ribosomal salt-wash proteins of calf liver and rabbit reticulocyte lysates (4,5). The homogeneous factor, isolated from these mammalian cell extracts, prepared in the presence of a large number of protease inhibitors, is a monomeric protein of apparent M, = 58,000-62,000. These results were in contrast to reports from several laboratories (6-9) suggesting that eIF-5 purified from rabbit reticulocyte lysates is a protein of much higher molecular weight ranging from 120,000 to 168,000 (6-9). In one report (9), eIF-5 was isolated in multiple forms which appeared to be generated by limited proteolysis. In recent review articles, the factor was listed as a protein of M , = 150,000 (10) or 120,000 (11). We have consistently observed that, although homogeneous, eIF-5 behaves as a monomeric protein of apparent M, = 58,000-62,000, less pure preparations of reticulocyte eIF-5 behave in gel filtration columns and in glycerol gradient centrifugation in buffers containing 75-100 mM KC1 as a protein of apparent M, = 140,000-160,000 (4, 5). Presumably, this is due to association of eIF-5 with other proteins, since eIF-5 activity present in such preparations was shown by (a) glycerol gradient centrifugation in buffers containing 500 mM KC1 or (b) elution of protein from gel slices following electrophoresis under denaturing conditions, to be associated with a 58,000-62,000-dalton protein (4, 5).
To obtain further evidence concerning the in vivo form of the initiation factor, we describe here the preparation and characterization of rabbit antisera specific for rabbit reticulocyte eIF-5. We have used affinity-purified anti-eIF-5 antibodies to show that the molecular size of eIF-5 in cells lysed rapidly in denaturing buffers containing high concentrations of SDS is similar to that of homogeneous eIF-5 isolated from rabbit reticulocyte lysates. Furthermore, we have carried out additional characterization of purified eIF-5 with regard to phosphorylation of the protein by purified casein kinase I1 and the effect of phosphorylation on the activity of eIF-5 in uitro. A preliminary report of this work has been presented (12 and a subunits of eIF-2, but did not react with the y subunit. amide (15%) gel electrophoresis followed by sensitive silver staining (16). The purified preparation was stored at -20 "C in a buffer containing 20 mM Tris-HC1, pH 7.5,190 mM KC1,l mM dithiothreitol, 1 mM EDTA, and 55% glycerol. The procedure for isolation of homogeneous eIF-5 from rabbit reticulocyte lysates was as described (5). In addition to homogeneous eIF-5, a partially purified "eIF-2/eIF-5 preparation," designated "CM-Sephadex eIF-5," was also isolated by chromatography of the 0.5 M KC1-wash proteins of poly-soma1 pellets obtained from rabbit reticuloycte lysates on a DEAEcellulose column followed by phosphocellulose and CM-Sephadex chromatography as described (5). Analysis of this eIF-2/eIF-5 preparation by SDS-polyacrylamide (15%) gel electrophoresis showed five major and a few minor polypeptide chains (see Fig. 1, Panel A, lane c). Three of these major polypeptides correspond to a, P, and y subunits of eIF-2 while a fourth, M , = 58,000, corresponded to eIF-5.
The specific activities of eIF-2 and eIF-5 in these partially purified preparations were about 3,300 and 130,000 units/mg protein, respectively. Homogeneous eIF-2 containing a, P, and y subunits was isolated by an adaptation of the procedure of Schreier et al. (7) except that CM-Sephadex was used instead of CM-cellulose. The specific activity of homogeneous eIF-2 is about 6,000 units/mg protein while that of homogeneous reticulocyte eIF-5 was 2.6 X lo6 units/mg protein. Definition of units for eIF-5 and eIF-2 were as described previously (5, 13). Protein was determined as described in these communications. Preparation of Antisera against Rabbit Reticulocyte eIF-5"Antisera against purified rabbit reticulocyte eIF-5 were prepared in mature rabbits following the procedure of Knudsen (18). Approximately 20 pg of eIF-5 present in a purified eIF-2/eIF-5 preparation (25% pure eIF-5 protein) was phosphorylated by [-p3'P]ATP and casein kinase I1 as described under "Experimental Procedures" ("Phosphorylation by Casein Kinase 11") and was resolved by electrophoresis in SDSpolyacrylamide (15%) gel. The proteins were then electroblotted onto nitrocellulose, which was then exposed to X-Omat AR film. To identify the 32P-labeled eIF-5 band on the nitrocellulose membrane, homogeneous eIF-5 and homogeneous eIF-2 containing a, P, and y subunits (where only the P-subunit was phosphorylated by casein kinase 11) were run in parallel lanes of the same gel. After aligning the nitrocellulose paper with the autoradiogram, a strip containing eIF-5 was excised. The strip was dried thoroughly and suspended in a minimal volume of dimethyl sulfoxide, mixed with complete Freund's adjuvant (no emulsion was formed), and injected intradermally into four sites of each of two rabbits (approximately 8 pg of eIF-5/rabbit). A booster of the same amount of antigen was given after 4 weeks and the animals bled after another 3 weeks. The presence of anti-eIF-5 antibodies in the serum was assayed by performing Western blot analysis using both homogeneous eIF-5 as well as CM-Sephadex eIF-2/eIF-5 fraction as antigens. A 1:lOO dilution of the serum from one rabbit gave a strong signal against eIF-5 and was used as a working dilution in subsequent experiments. Blood was collected from the rabbit, allowed to clot and retract, and the sera was clarified and stored in aliquots at -70 "C.
Immunobbtting of Initiation Factors-Protein samples were resolved on SDS-polyacrylamide gels according to Laemmli (19), electrophoretically transferred to either a sheet of nitrocellulose or a PVDF (polyvinylidene difluoride) membrane (20). Following transfer, the gels were stained with silver reagents (16) to verify quantitative transfer. The membrane was first rinsed with 20 mM Tris-HCI, pH 7.5, and 150 mM NaCl (TBS), and then placed in TBS containing 5% non-fat dry milk (Blotto) (21) and shaken gently for 1 h. The blot was then incubated with the primary antibody in Blotto for 2-16 h, after which it was washed three times with Blotto + 0.05% Tween 20 and then incubated for 2 h with anti-rabbit IgG coupled to alkaline phosphatase. The blot was washed again with three changes of Blotto + 0.05% Tween 20, followed by three changes of TBS. The membrane was then immersed in the color development solution (1 ml of 3% nitro blue tetrazolium and 1 ml of 1.5% 5-bromo 4-chloro 3-indolyl phosphate dissolved in 100 ml of 100 mM Tris-HC1, pH 9.5, plus 0.5 mM MgCU and color development allowed to proceed until the bands reached the desired intensity and background. Affinity Purification of Antibodies-Antibodies specific for eIF-5 or the p subunit of eIF-2 were purified by elution of the antibody from immunoblots, as described by Olmsted (22). About 60 pg of the CM-Sephadex eIF-2/eIF-5 fraction was phosphorylated by casein kinase 11 and [-,"'P]ATP and subjected to electrophoresis in a SDS-15% polyacrylamide gel. Casein kinase I1 phosphorylates both eIF-5 (this paper) and the P subunit of eIF-2 (23-28). Homogeneous eIF-5 and homogeneous eIF-2 (containing the a, @, and y subunits) were also phosphorylated by casein kinase I1 and run in parallel lanes on the same gel. Following electrophoresis, the proteins were electrotransferred onto aminophenylthioether paper in transfer buffer containing 25 mM sodium phosphate, pH 6.5. The paper was then incubated in 10% ethanolamine, 1% bovine serum albumin in 10 mM Tris-HC1, pH 8.8, for 2 h at room temperature to block all nonspecific sites and then washed with TBS. To locate the band of interest, the blot was air-dried and exposed to Kodak X-Omat AR film. Using the bands corresponding to eIF-5 and eIF-2 (8) as markers, the regions (2 mm wide) corresponding to eIF-5 and eIF-2 (8) were excised and cut into small pieces (2 X 2 mm). The pieces were then incubated with a 1:lO dilution of the appropriate antisera for 6-12 h, washed with blotting buffer (TBS + 1% bovine serum albumin + 0.05% Tween 20) and finally with a buffer (phosphate-buffered saline) containing 10 mM potassium phosphate, pH 7.4, 150 mM NaC1, and 0.5% bovine serum albumin. Antibodies were extracted from the paper pieces by stirring them with 500 p1 of 4 M guanidine hydrochloride in phosphatebuffered saline for 2 min. The extraction was repeated twice and the combined extracts dialyzed overnight against phosphate-buffered saline to allow renaturation of the eluted antibodies. Phosphorylation of eIF-5 by Casein Kinase II-Phosphorylation was carried out in reaction mixtures (20 pl) containing 20 mM Tris-HCl, pH 7.5, 100 mM KCl, 5 mM dithiothreitol, 10 mM MgCL, 150 p~ [Y-~'P]ATP (2,000-20,000 cpm/pmol), 0.5 pg of purified casein kinase 11, and appropriate amounts of the substrate eIF-2 or eIF-5 as indicated. Following incubation at 30 "C for 45 min, the reactions were terminated by the addition of 20 pl of an electrophoresis loading buffer (250 mM Tris-HC1, pH 6.8, 2% SDS, 570 mM 2-mercaptoethanol, and 0.01% bromphenol blue) and heating for 3 min at 100 "C. After standing at room temperature for 10-20 min, the reaction mixtures were electrophoresed in 15% polyacrylamide, 0.09% bisacrylamide gel at pH 8.8 in the presence of 0.1% SDS at 110 V for 3 h according to Schreier et al. (7). The gels were stained with silver reagents according to Merril et al. (16), and autoradiograms of the dried gel were obtained. Marker proteins were run in parallel lanes to determine the M, values of stained and radioactive bands.

Purity, Specific Activity, and Molecular Weight of eIF-5-
Purified eIF-5 was isolated from rabbit reticulocyte lysates by procedures similar to those described previously from this laboratory (5). Analysis of the final eIF-5 preparation by polyacrylamide gel electrophoresis in the presence of SDS followed by staining with silver reagents according to Merril et al. (16) displayed a single band of M, = 58,000 ( Fig. 1, Panel A, lane a; see also Ref. 5). When the eIF-5 preparation was subjected to (a) glycerol gradient centrifugation and (b) Sephadex G-75 gel filtration, eIF-5 activity was found to be associated with a protein of apparent M, of about 58,000-62,000 as described previously (5). Furthermore, in agreement with our earlier results (5), when purified eIF-5 was subjected to SDS-polyacrylamide gel electrophoresis and the separated proteins reconstituted, all the factor activity (approximately 10% yield) was found to be associated with a polypeptide of M, = 58,000-62,000 (data not shown in this paper; see Ref. 5 for earlier work). These results confirm earlier reports from this laboratory (5) that homogeneous eIF-5, isolated from rabbit reticulocyte lysates is a monomeric protein of M, = The specific activity of the homogeneous eIF-5 preparation was determined using an assay (5) that measured its ability 58,000-62,000. to mediate joining of a 60 S ribosomal subunit to a preformed 40 S initiation complex to yield an 80 S initiation complex (Fig. 2). Under the conditions of the in vitro assay described in Fig. 2 Phosphorylation of eIF-5 by Casein Kinase 11-A number of eukaryotic translation factors have been shown to be phosphorylated in vivo (reviewed in Refs. 10, 29, and 30; also see Ref. 31). Examples are phosphorylation of the a! and j 3 subunits of eIF-2, eIF-3, eIF-4B, eIF-4F, eIF-4E, and a number of subunits of eIF-2B (GEF). Studies on in vitro phosphorylation of initiation factors by protein kinases have shown that casein kinase 11, a cyclic AMP-independent protein kinase isolated from rabbit reticulocyte lysates, specifically phosphorylated the p subunit of eIF-2, eIF-4B, and eIF-3 (23-30).
It was therefore of interest to determine whether our homogeneous eIF-5 preparation (MI = 58,000-62,000) can be phosphorylated by purified casein kinase 11. For this purpose, purified eIF-5 preparations were incubated with [y3'P]ATP and casein kinase 11, and the reaction products were analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography (Fig. 1). An intense 32P-labeled band which comigrated with homogeneous eIF-5 preparation (Mr = 58,000-62,000) was observed (Fig. 1, Panel B, lane a). In addition to this band, another 32P-labeled band corresponding to the 25,000-dalton subunit of casein kinase I1 was also observed due to autophosphorylation of this kinase subunit (Fig. 1,  Panel B, lane d). When a homogeneous eIF-2 preparation, isolated from rabbit reticulocyte lysates, was used as a substrate for casein kinase 11, only the p-subunit of eIF-2 was phosphorylated (Fig. 1, Panel B, lane b). Incubation of a partially purified eIF-2/eIF-5 preparation (CM-SepMez eIF-2/eIF-5) by casein kinase I1 and [y3'P]ATP phosphorylated both the eIF-2p and eIF-5 (Fig. 1, Panel B, lane c). In addition to eIF-20 and eIF-5, a number of other polypeptides, present in this partially purified eIF-2/eIF-5 preparation, were also phosphorylated by casein kinase I1 (Panel B, lane c).
The amino acid residues of eIF-5 phosphorylated by casein kinase I1 were identified by subjecting the 32P-labeled eIF-5 band to acid hydrolysis in 6 N HC1, followed by high voltage electrophoresis at pH 3.5, using unlabeled phosphoserine, phosphothreonine, and phosphotyrosine as internal standards. Incorporation of 32P occurred only into serine residues of eIF-5 (Fig. 3).
The number of sites in eIF-5 phosphorylated by casein kinase I1 was determined by analyzing tryptic phosphopep- Purified eIF-5 (250 ng) was phosphorylated with 0.3 mM [-p3'P]ATP (10,400 cpm/pmol) and 90 units of casein kinase I1 in a 20-4 reaction mixture ("Experimental Procedures"). Following incubation for 45 min at 30 'C, the reaction was terminated and subjected to SDSpolyacrylamide (15%) gel electrophoresis. Following electrophoresis, the gel was silver-stained (16), dried, and the strip of gel corresponding to eIF-5 was excised. After swelling the gel in 10% methanol, the dried gel was mixed with 25 pg each of unlabeled phosphoserine, phosphotyrosine, and phosphothreonine and subjected to acid hydrolysis in 6 N HCl in a sealed ampula for 1.5 h at 110 "C. Following hydrolysis, the solution was lyophilized three times, dissolved in 20 pl of water, spotted on a Whatman 3MM paper, and subjected to electrophoresis for 2 h a t 4,000 volts at 10 "C in pyridine acetate buffer (0.5% pyridine, 5% acetic acid) at pH 3.5. The dried filter paper was first developed with 0.5% ninhydrin in acetone containing 30% glacial acetic acid to determine the position of each unlabeled phosphoamino acid. Subsequently, the filter paper was subjected to autoradiography.
tides derived from phosphorylated eIF-5 by reverse-phase HPLC columns using 0.13% heptafluorobutyric acid and increasing concentrations of acetonitrile as an eluant (Fig. 4).
Complete digestion of 32P-labeled eIF-5 by trypsin yielded two major and two minor phosphopeptides by reverse-phase HPLC column chromatography. It should be noted that if the HPLC column was developed with 0.1% trifluroacetic acid and increasing concentrations of acetonitrile, the two major peaks I1 and I11 were not resolved and eluted as a single peak (data not shown). Analysis of the tryptic peptides derived from "P-eIF-5 on SDS-18% polyacrylamide gels also indicated the presence of 2 major phosphopeptides and 2 minor phosphopeptides ranging in M , values from 3,000 to 1,500 (data not shown). These results indicated that casein kinase I1 phosphorylated multiple sites in the eIF-5 molecule. We also determined whether phosphorylation of eIF-5 by casein kinase I1 effected the function of eIF-5 in mediating joining of a 60 S ribosomal subunit to a preformed 40 S initiation complex. For this purpose, purified eIF-5 was phosphorylated with [y-32P]ATP and casein kinase 11, and the reaction mixture was subjected to Sephadex G-75 gel filtration (Fig. 5). Aliquots of each eluted fraction were assayed for 32P radioactivity as well as for eIF-5 activity. In addition, fractions containing 32P radioactivity were subjected to SDS-polyacrylamide gel electrophoresis to identify the phosphoproteins eluted. As expected, both 32P-labeled autophosphorylated casein kinase I1 and 32P-labeled eIF-5 eluted from the column in agreement with the M , value of each protein (Fig. 5 ) . Furthermore, 32P-labeled eIF-5 retained biological activity (Fig.  5). The total recovery of eIF-5 activity from the column was about 51%. Similar recovery of eIF-5 activity was observed 3ZP]ATP (20,000 cpm/pmol) in a reaction mixture (40 pl) similar to that described under "Experimental Procedures." Following incubation for 45 min at 30 "C, the protein was reduced and carboxymethylated by first adding 13 pl of a buffer containing 1.12 M 2-mercaptoethanol, 500 mM Tris-HC1, pH 8.8, 4% SDS, 40% glycerol, and 0.01% bromphenol blue and heating for 3 min at 100 "C. Subsequently, the cooled reaction mixture was treated with sodium iodoacetate (140 mM, final concentration) and incubated at 37 "C for 30 min in dark. The resulting protein sample was then resolved in a SDS-15% polyacrylamide (-0.09% bisacrylamide) gel electrophoresis and then electrotransferred to a nitrocellulose membrane a t 90 volts for 4 h at 5 "C using a Bio-Rad Transblot apparatus. The blotted eIF-5 was detected by immersing the nitrocellulose paper in a solution of 0.1% Ponceau S dye in 1% acetic acid for 1 min and removing the excess stain with water by gentle agitation or by autoradiography. The stain-free piece of nitrocellulose filter was incubated with 0.5% polyvinylpyrrolidone-40 dissolved in 100 mM acetic acid for 30 min at 37 "C, and the excess polyvinylpyrrolidone was removed by extensive washing with water (39). Subsequently, the 32P-labeled eIF-5 was digested in situ in 50 pl of a solution containing 100 mM NH~HCOI and acetonitrile (95:5, v/v) with 200 ng of L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin for 12 h at 37 "C followed by the addition of a second 200 ng of trypsin and incubation for 2 h at 37 "C. Following centrifugation, the supernatant containing peptides was lyophilized. The residue was dissolved in 50 pl of 10% 2propanol and lyophilized again. This was repeated twice to remove traces of NH4HC03. The dried residue was dissolved in 50 pl of formic acidacetic acidwater, 5:15:80, pH 3.5. An aliquot of the trypsindigested sample was treated with 150 pl of 5% acetonitrile containing 0.13% heptaflurobutyric acid and then subjected to reverse-phase HPLC. This was performed with a Waters HPLC system using Vydac 218TP54 Cl* column (0.46 X 25 cm) equilibrated with 5% acetonitrile and 0.13% heptaflurobutyric acid (40,41). The trypsin-digested sample was applied to the column and eluted at 1.5 ml/min with 0.13% heptaflurobutyric acid and increasing concentrations of acetonitrile as follows: 0-5 min, 5% acetonitrile; 5-15 min, 14.75% acetonitrile; 15-90 min, linear gradient from 14.75 to 40.25% acetonitrile; 90-100 min, linear gradient from 40.25 to 80% acetonitrile; and 100-110 min, 80% acetonitrile. Elution of radioactive peptides from the column was monitored by counting 0.75-ml aliquots from each fraction with 5 ml of ACS I1 (Amersham Corp.) in a liquid scintillation spectrometer. The recovery of radioactivity from the column was about 85%. The authors are indebted to Dr. Tim Stoller and Dr. Yves Bourbonnais of Professor Dennis Shield's laboratory of this institution for considerable help and advice in the separation of tryptic peptides by HPLC.
when 600 units of eIF-5, without treatment with ATP and casein kinase 11, was subjected to gel filtration in the same column (data not shown). These results demonstrate that phosphorylation of eIF-5 by casein kinase I1 does not affect its ability to catalyze joining of a 60 S ribosomal subunit to a preformed 40 S initiation complex.
Preparation and Characterization of Anti-eIF-5 Antibodies-A major problem for preparation of antibodies against eIF-5 was the relatively low yield of homogeneous eIF-5 obtained by purification from either calf liver extracts (4) or from rabbit reticulocyte lysates ( 5 ) . Furthermore, the possi- . The dried gel was subjected to autoradiography. In a separate experiment, the positions of elution of peak activity of several marker proteins, muscle aldolase, bacterial alkaline phosphatase, human hemoglobin type IV, and ovalbumin, were determined in the same column. In addition, the void volume (V,) was determined using blue dextran. From these results, it was calculated that eIF-5 eluted from the column as a protein of apparent M, of 55,000-60,000. bility exists that a protein synthesis initiation factor like eIF-5 may be highly conserved and hence might be poorly antigenic. Attempts were therefore made to raise antibodies against relatively large amounts of denatured eIF-5. We reasoned that denaturation of eIF-5 would help to overcome the problem of possible conservation in different species, as the denatured form would not resemble any self protein. We therefore followed the procedure of Knudsen (18) to raise anti-eIF-5 antibodies in rabbits ("Experimental Procedures"). In our case, approximately 16 pg of eIF-5, phosphorylated with casein kinase 11, was injected intradermally into four sites of each of two rabbits (approximately 8 pg/rabbit) ("EXperimental Procedures"). Following booster injections, the presence of anti-eIF-5 antibodies in the serum was assayed by performing immunoblotting analysis using purified eIF-5 as an antigen. A dilution of 1:lOO of the serum from one rabbit gave a strong signal of the presence of antibodies against eIF-5 (data not shown) and was used as the working dilution in subsequent experiments. Unexpectedly, there was a background staining which was most likely due to the anti-nitrocellulose antibodies which were also generated in the rabbit. To circumvent this problem, we affinity-purified the anti-eIF-5 antibodies from the serum as described under "Experi-  (M, = 14,300)) was run in a separate lane of the same gel and was also electrotransferred onto the PVDF membrane (not shown). Following electrophoresis, one-half of the gel (Panel A ) was stained with silver reagents according to Merril et al. (16), while the other half (Panel B ) was electrotransferred onto the PVDF membrane and probed with 1:lOO dilution of the anti-eIF-5 antibodies as described under "Experiment a l Procedures." 5 mM dithiothreitol at 30 "C for 45 min. Following incubation, one-half of the reaction mixture was heated at 65 "C for 10 min to denature the proteins. Both the heated and the unheated bodies (1:25 final dilution) for 12 h a t 4 "C). Subsequently, the fractions were then incubated with affinity-purified anti-eIF-5 antiantigen:antibody complex formed was incubated with 20 pl of 50% protein A-Sepharose CF-4B (Pharmacia) suspended in Blotto buffer for 3 h at 4 "C. The immune precipitate formed was collected by centrifugation, washed two times with 1 ml of RIPA buffer (1% Triton X-100, 1% deoxycholate, 0.1% SDS, 150 mM NaC1, 5 mM EDTA, 10 mM Tris-HCI, pH 7.2) followed by 1 ml of high salt buffer (2 M NaCI, 5 mM EDTA, 0.2% Triton X-100, 10 mM Tris-HC1, pH 8.0), and finally three times with 0.8 ml of low salt buffer (150 mM NaCl, 5 mM EDTA, 0.1% Triton-X-100, 10 mM Tris-HC1, pH 8.0). Finally, the pellet was boiled in 40 p1 of a SDS buffer (125 mM Tris-HCl, pH 6.8,280 mM 2-mercaptoethanol, 1% SDS, 10% glycerol, and 0.04% bromphenol blue) for 3 min and subjected to SDS-polyacrylamide gel electrophoresis (15% polyacrylamide and 0.09% bisacrylamide) at pH 8. Lysates of mouse L cells and HeLa cells used in this experiment were prepared by sonicating freshly grown packed cells into 2 volumes of a denaturing buffer containing 3% SDS, 125 mM Tris-HC1, pH 6.8, 280 mM 2-mercaptoethanol, and 10% glycerol as described by Meyer et al. (32). Other eIF-5 preparations used were as indicated below. Lane a, 0.5 pg of a homogeneous eIF-5 preparation from rabbit reticulocytes; lane b, 50  to noise ratios of the affinity-purified antibodies were far higher than whole sera, thereby allowing the detection of trace levels of eIF-5. Fig. 6 shows that affinity-purified anti-eIF-5 antibodies reacted strongly with homogeneous eIF-5
To determine specificity of anti-eIF-5 antibodies, both native and heat-denatured eIF-5 were separately treated with anti-eIF-5 antibodies (Fig. 7). Only the heat-denatured form of eIF-5 was immunoprecipitated with anti-eIF-5 antibodies indicating that denatured conformation of eIF-5 had the antigenic determinants (Fig. 7, compare lanes a and b). Consistent with this observation, prior incubation of a homogeneous non-denatured eIF-5 preparation with various dilutions of the affinity-purified anti-eIF-5 antibodies had no effect on the ability of the factor to mediate 80 S initiation complex formation (data not shown).
Immunoblot Analysis of Cell Lysates-The availability of anti-eIF-5 antibodies allowed us to examine eIF-5 in different species and to determine if the size of purified rabbit reticulocyte and calf liver eIF-5 is the same as in vivo. Lysis of cells directly into protein denaturing buffers containing high amounts of SDS is expected to minimize possible proteolytic degradation of initiation factor proteins. Such lysis procedures are thus considered to freeze the proteins in forms present in vivo (32,33). For analysis of size of eIF-5 in crude lysates under these conditions, freshly grown HeLa cells, mouse L cells, were lysed directly into denaturing buffers. In addition, crude rat liver and calf liver extracts were also treated with SDS denaturing buffers. Each of the freshly prepared denatured lysates were analyzed immediately by SDS-polyacrylamide gel electrophoresis and immunoblotting (Fig. 8). In all the species examined, a polypeptide (MI = 58,000) reacted with the antibody and was the major band in rabbit reticulocytes, rat and calf livers and mouse L cells. Both HeLa and mouse L cells lysates exhibited an additional immunoreactive polypeptide (M, = 62,000) which was the major species in HeLa extracts. No precursor or degradation forms of eIF-5 were detectable under these conditions (Fig. 8). These data indicated that the molecular weight of eIF-5 is unaltered during the isolation procedure.
Examples are phosphorylation of the (Y and B subunits of eIF-2, eIF-4B, eIF-4F, eIF-4E, eIF-3 and a number of subunits of eIF-2B (also designated as GEF). Hershey and his colleagues (34)(35)(36)(37) have demonstrated changes in covalent modification (phosphorylation status) of these initiation factors in response to physiological signals, e.g. heat shock, serum depletion, and serum stimulation of growing cells. The extents of phosphorylation of some of these proteins correlate with protein synthesis activity levels. However, utilizing existing initiation factor assays, it has not been possible to demonstrate that phosphorylation causes a change in functional activity of any of these proteins with the exception of phosphorylation of the a subunit of eIF-2. In this case, phosphorylation of the a subunit of eIF-2 affects the ability of this initiation factor to be catalytically re-utilized in initiation reactions (29,30).
Earlier reports from several laboratories (6-9) suggested that eIF-5, isolated from rabbit reticulocyte lysates, is a monomeric protein of M , of 120,000-168,000. Thus, when in vitro phosphorylation of eIF-5 was examined, a 120,000-168,000dalton polypeptide was shown to be phosphorylated by casein kinase I1 (15,24,25,38). However, results published previously from this laboratory (4,5) as well as those presented in this communication show that eIF-5 is a monomeric protein of M, = 58,000-62,000 indicating that the original eIF-5 preparations of reported M , values of >120,000 were not homogeneous. It is likely therefore that phosphorylation of contaminating proteins in eIF-5 preparations was observed.
Results presented in this communication demonstrate that our homogeneous eIF-5 preparations (Mr = 58,000-62,000) isolated from rabbit reticulocyte lysates can be phosphorylated at multiple serine residues by highly purified casein kinase 11. However, phosphorylation of eIF-5 by casein kinase I1 had no apparent effect on the activity of eIF-5 in joining 60 S ribosomal subunits to a 40 S initiation complex to form an 80 S initiation complex. However, the ability of casein kinase I1 to specifically label eIF-5 with 32P without loss of biological activity of the initiation factor will allow us to prepare labeled factor for mechanistic studies. It will also be of considerable interest to determine whether eIF-5 as isolated from cell extracts is endogenously phosphorylated in vivo at sites other than casein kinase I1 and if such phosphorylation/ dephosphorylations play a role in biological activity of the factor. Availability of homogeneous eIF-5 in our laboratory has now allowed us to raise antibodies against eIF-5. Characterization of these antibodies has shown that antibodies were directed against denatured conformation of eIF-5 and did not recognize the native eIF-5 protein. However, such antibodies are an important tool for structural studies, e.g. as probes for (a) screening mammalian h g t l l libraries to isolate cDNA clones, ( b ) immunoblot analysis of eIF-5 protein to investigate relative concentration and covalent modification of eIF-5 in growing cells in response to a variety of growth regulatory physiological signals. For example, in the present work, we have used these anti-eIF-5 antibodies to show that the size of eIF-5 in fresh mammalian liver extracts or in cell lysates, prepared by lysing growing mouse L cells or HeLa cells directly in 3% SDS containing denaturing buffers, is similar to the size of the factor purified in our laboratory from calf liver extracts and rabbit reticulocyte lysates. In all the species examined, a 58,000-dalton polypeptide band reacted with the anti-eIF-5 antibodies. However, HeLa cell extracts exhibited an additional immunoreactive band of M, = 62,000 which was the major species in these cell lysates (Fig. 8). The possibility exists therefore that the 62,000-dalton polypeptide may be a form present in uiuo which is very labile and proteolytically processed to the 58,000-dalton form. Further work is clearly necessary to answer this question. However, the immunoblot analysis of freshly prepared cell lysates in 3% SDS-buffers ( Fig. 8) clearly indicates that our original observation that eIF-5 is a protein of M, = 58,000-62,000 is not due to proteolytic processing of a 120,000-160,000 dalton protein during isolation of the initiation factor from cell extracts. This experiment therefore provides additional evidence that earlier reports from other laboratories (6-9) that eIF-5 was of apparent M , = 120,000-160,000, probably represents the presence of proteins unrelated to eIF-5 activity. However, the most convincing proof for the molecular size of eIF-5 will come from cloning and sequencing of the cDNA corresponding to the coding sequence of eIF-5. Using anti-eIF-5 antibodies, we are currently screening several mammalian Xgtll libraries to identify cDNA clones which encode mRNA for eIF-5.