Mechanism of Inhibition of Polypeptide Chain Initiation in Heat-shocked Ehrlich Cells Involves Reduction of Eukaryotic Initiation Factor 4F Activity*

Almost all living organisms studied respond to elevated temperature with a marked inhibition of overall protein synthesis but increased synthesis of a specific set of proteins, the so-called heat-shock proteins. We have prepared a cell-free protein synthesizing system (lysate) from heat-shocked Ehrlich ascites tumor cells that reflects the inhibition of protein synthesis in intact cells at elevated temperatures. We have isolated and partially purified a stimulator of the heat-shocked cell lysate from Ehrlich cells. Through four purification steps, the stimulator is chromatographically identical to eukaryotic initiation factor 4F (eIF-4F), an initia- tion factor which specifically binds mRNA cap structure. Therefore, we have tested the effects of highly purified reticulocyte eIF-4F on the heat-shocked cell lysate. Protein synthesis is strongly stimulated by ad- dition of highly purified eIF-4F. Synthesis in the heat-shocked lysate is more inhibited at high (70 mM) KC1 concentrations, than at lower concentrations, and stimulation by eIF-4F is correspondingly greater at higher KC1 concentrations, so that the rate of protein synthesis is returned to control (non-heat-shocked lysate) levels at all KC1 concentrations. Furthermore, at 70 mM KCl, in heat-shocked lysates, synthesis of the 68-kDa heat-shock protein is much less inhibited than synthesis of the bulk of non-heat-shock proteins, and eIF-4F stimulates

Almost all living organisms studied respond to elevated temperature with a marked inhibition of overall protein synthesis but increased synthesis of a specific set of proteins, the so-called heat-shock proteins. We have prepared a cell-free protein synthesizing system (lysate) from heat-shocked Ehrlich ascites tumor cells that reflects the inhibition of protein synthesis in intact cells at elevated temperatures. We have isolated and partially purified a stimulator of the heat-shocked cell lysate from Ehrlich cells. Through four purification steps, the stimulator is chromatographically identical to eukaryotic initiation factor 4F (eIF-4F), an initiation factor which specifically binds mRNA cap structure. Therefore, we have tested the effects of highly purified reticulocyte eIF-4F on the heat-shocked cell lysate. Protein synthesis is strongly stimulated by addition of highly purified eIF-4F. Synthesis in the heatshocked lysate is more inhibited at high (70 mM) KC1 concentrations, than at lower concentrations, and stimulation by eIF-4F is correspondingly greater at higher KC1 concentrations, so that the rate of protein synthesis is returned to control (non-heat-shocked lysate) levels at all KC1 concentrations. Furthermore, at 70 mM KCl, in heat-shocked lysates, synthesis of the 68-kDa heat-shock protein is much less inhibited than synthesis of the bulk of non-heat-shock proteins, and eIF-4F stimulates synthesis of 68-kDa protein to a much lesser extent than non-heat-shock proteins. Thus, addition of purified eIF-4F reverses the effects of elevated temperatures on Ehrlich cells that are reflected in lysates. Therefore, we propose that the inhibition of translation in heat-shocked Ehrlich cells is the result of inactivation of eIF-4F function.
Almost all living organisms studied respond to elevated temperatures by increasing the rate of synthesis of a small set of proteins, the so-called heat-shock proteins, through a very rapid induction of synthesis of mRNA for these proteins (1)(2)(3)(4). An equally universal, but less emphasized, response is a generalized inhibition of the overall rate of total protein synthesis a t elevated temperatures. Polypeptide chain initia-R01 CA21663-09, P30-CA11198-16, and GM 26796. The costs of * This work was supported by National Institutes of Health Grants publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ T o whom correspondence should be addressed University of Rochester Cancer Center, 601 Elmwood Avenue, Box 704, Rochester, NY 14642. tion is the major site of this inhibition in many systems (5)(6)(7)(8). Further, heat-shock mRNAs are preferentially translated relative to most non-heat-shock mRNAs at the elevated temperatures in Drosophila (9), Xenopus oocytes (lo), and HeLa cells (11). However, little is known of the mechanism by which initiation is inhibited or by which heat-shocked mRNAs are preferentially translated a t elevated temperatures. In HeLa cells, Duncan and Hershey (12) have recently reported that heat shock inhibits eIF-2,' eIF-4B, and either one of or both eIF-3 and eIF-4F. Further, they have found increased phosphorylation of eIF-2 and decreased phosphorylation of eIF-4B. Hickey and Weber (11) showed in HeLa cells that heat-shock mRNAs were translated with increasing efficiency relative to non-heat-shock mRNAs as initiation became more rate-limiting, so that inhibition of initiation, per se, can contribute to the preferential translation of heatshock mRNA, which is "more efficient." However, the data in vivo and in vitro suggest that in Drosophila there is a qualitative change in the protein synthesis apparatus such that a factor is inactivated which is required specifically for translation of non-heat-shock mRNAs and is not required for heat-shock protein mRNAs (13, 14).
Recently we reported that the level of 40 S initiation complexes is reduced in heat-shocked Ehrlich ascites tumor cells concomitant with inhibition of polypeptide chain initiation (8). We report here that cell-free protein synthesis in lysates isolated from heat-shocked cells is inhibited and that synthesis is restored by the addition of eIF-4F. Furthermore, addition of eIF-4F preferentially stimulates translation of non-heat-shock mRNAs. Eukaryotic initiation factor 4F binds specifically to mRNA cap structure and is required for efficient binding of mRNA to 40 S ribosome subunits (15).
The factor probably consists of three nonidentical subunits of 220, 45, and 24 kDa and is thought to be identical to cap binding protein I1 (16,17). It restores translation of capped mRNAs in extracts of poliovirus-infected HeLa cells. Although heat-shock mRNA is not known to be uncapped, this does a t least show that differential translation is possible by modulating eIF-4F activity.
We conclude that inactivation of eIF-4F in lysates of heatshocked cells is responsible for the decreased rate of protein synthesis in these lysates and for the preferential translation of heat-shock mRNAs. Inactivation of eIF-4F could, therefore, explain the alterations in translation in intact heat-shocked cells. 'The abbreviations used are: eIF, eukaryotic initiation factor; Mops, 3-(N-morpholino)propanesulfonic acid GEF, GDP/GTP exchange factor; hsp, heat-shock protein, SDS, sodium dodecyl sulfate.
Cell Culture-Ehrlich cells were grown in spinner suspension culture as described previously (18). Briefly, stock cultures were diluted routinely to 2 X lo5 cells/ml, and the volume was kept constant at 100 ml, in antibiotic-free minimal essential medium modified for spinner culture and supplemented with 10% (v/v) calf serum and 20 mM Mops. Under these conditions the cells grew with a doubling time of approximately 14 h. When large cultures were required, this stock was expanded.
Preparation of 37 and 44 "C Cell Lysates (19)"Suspension cultures (3 liters) at 6 X 10' cells/ml centrifuged at 500 X g for 10 min at room temperature and cells were resuspended in minimal essential medium supplemented with 1% calf serum and 20 mM Mops, pH 7.3, at a density of 3 X IO6 cells/ml. The cultures were brought rapidly to either 37 or 44 "C, as described (8), and incubated for 20 min. The cells were rapidly cooled and sedimented by centrifugation at 500 X g for 7 min and washed twice with cold phosphate-buffered saline. The final cell pellets were resuspended in an equal volume of hypotonic buffer (20 mM Mops (KOH), pH 7.5, 10 mM KCl, 2.5 mM magnesium acetate, 0.5 mM dithiothreitol, 0.1 mM EDTA) and cells were lysed with a Dounce homogenizer. The homogenates were made 180 mM KC1 and 4 mM magnesium acetate and centrifuged at 10,000 x g for 10 min at 0 "C. The supernatants were incubated at 37 "C for 30 min with ATP, GTP, an energy regenerating system, and amino acids to run ribosomes off pre-existing polyribosomes while high magnesium and KC1 prevented reinitiation. The pre-incubated preparation was gel-filtered at 0 "C through Sephadex G-25 equilibrated in 20 mM Mops (KOH), pH 7.4, 80 mM KC1, 2.5 mM magnesium acetate, 0.5 mM dithiothreitol. The excluded fractions were combined, divided into aliquots, and kept under liquid nitrogen.
Preparation of Ribosomal Salt Wash-Ehrlich cell postmitochondrial supernatant was centrifuged at 200,000 X g for 1. Activity was assayed throughout by adding 10 p1 of fractions to the heat-shocked cell-free system described below (50 p1, final volume), adjusting the assay KC1 concentration to 70 mM. Preparation of Initiation Factors-GEF and eIF-2 were prepared 50% pure from Ehrlich cells as described (20). Eukaryote initiation factors eIF-4A, eIF-4B, and eIF-4F were purified from rabbit reticulocytes as described (15).
SDS-Polyacrylamide Gel Electrophoresis-SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli (21). Equal amounts of radioactivity were applied to the gels. Gels containing protein labeled in the cell-free system (6 X lo5 cpm/lane) were autoradiographed for 5 days, while gels with proteins labeled in intact cells (lo6 cpm/lane) were autoradiographed for 3 days. Where indicated, lanes from x-ray plates were scanned at 633 nm for optical density with an Ultroscan laser densitometer (LKB).
Assay of Lysates for Protein Synthesis-Protein synthesis in the cell-free systems was followed as described (19). Briefly, 50-pl reactions contained 12.

"
filter-paper discs. The discs were boiled in 5% trichloroacetic acid, washed with 10% trichloroacetic acid, ethanol, and acetone, dried, and counted for radioactivity. Unless otherwise indicated, radioactive counts represent incorporation of ["Clleucine into protein for 30 min.

RESULTS
Characterization of Cell-free Systems-Lysates were prepared from cells incubated at 37 and 44 "C. They were preincubated to run off pre-existing polyribosomes, and endogenous amino acids were removed by gel filtration. In the cellfree protein-synthesizing system, the two lysates had similar requirements for M e (data not shown) and both incorporated [14C]leucine linearly for at least 30 min. The 37 "C lysate exhibited a KC1 optimum typical of mammalian cell-free systems (70 m M ) (22,23) and had a relatively broad optimum (Fig. 1). At 70 mM KC1 the incorporation by the 44 "C lysate was below that in the 37 "C lysate, thus mimicking the in vivo situation, in which protein synthesis was inhibited by about 90% in the cells at 44 "C. The low synthesis by the heatshocked lysate proved to be because it had an unusually low and sharp KC1 optimum (Fig. 1). Both cell-free systems were incubated at 37 "C so that this change reflects an effect of elevated temperatures on the intact cells from which the lysates were prepared. The sharp KC1 optimum indicates strong inhibition a t higher KC1 concentrations. Inhibition was much less when K' acetate was substituted for KCl, suggesting that C1-is the major inhibitory ion (Fig. 2). Elongation was ruled out as the inhibition site at higher KC1 concentrations through measurements of rate of elongation at different KC1 concentrations, as follows. The lysates were  incubated under conditions which produced similar numbers of polyribosomes and then edeine (a specific inhibitor of initiation) was added to prevent further initiation. Elongation by the preformed polyribosomes was followed at different KC1 concentrations by the addition of [14C]leucine. In both the 37 and 44 "C lysates the rate of elongation was relatively insensitive to KCl, rates being similar at 40 and 80 mM KC1 (data not shown).
Incubation of Ehrlich cells above 44 "C resulted in the partial degradation of mRNA so that lysates prepared from these cells were partially dependent on added mRNA. However, they responded to addition of KC1 in a similar manner to the lysates used for the work presented here which have no mRNA dependence (data not shown).
Purification of a Stimulator of Heat-shocked Cell-free Systems"Smal1 amounts of 37 "C lysate were strongly stimulatory when added to the 44 "C lysate. We, therefore, fractionated the 37 "C lysate to isolate the factor lacking in the 44 "C lysate, using as an assay the ability to stimulate protein synthesis in a heat-shocked cell-free system at the higher KC1 concentration (70 mM). Ribosomal salt wash that had been dialyzed against Ao.l (20 mM Mops (KOH), pH 7.6, 100 mM KC1, 0.25 mM dithiothreitol, 0.1 mM EDTA, 10% (v/v) glycerol) was found to contain a stimulator (Table I)

Purification from Ehrlich cells of a stimulator of protein synthesis in heat-shocked cell lysates
Details of purification and assay are described under "Experimental Procedures." 1 unit of activity was defined arbitrarily as an amount which stimulated heat-shocked cell-free protein synthesis at 70 mM KC1 by 100%. For consistency, only one lysate was used throughout the purification; the control incubation incorporated 5500 cpm of [14C]leucine into protein; one unit therefore stimulated to 11,000 cpm. Generally stimulation was kept below 12,000 cpm where the lysate response to stimulator was approximately linear. The concentration of KC1 that elutes stimulator from each column is shown.

Addition
Specific activity

FIG. 3. Effect of Ehrlich cell stimulator on the KC1 dependence of protein synthesis in heat-shocked cell lysate.
Lysates were incubated at 37 "C as described under "Experimental Procedures" without (0) or with (0) 6.4 pg of stimulator (purified through the phosphocellulose step). Fold stimulation is the ratio of synthesis with/without stimulator.

Effect of initiation factors on cell-free protein synthesis in heat.
shocked cell lysate Assay of the cell-free system is described under "Experimental Procedures." The active fractions pooled from DEAE-cellulose chromatography were applied to a phosphocellulose column, and bound protein was eluted stepwise. Active fractions were eluted between 0.1 and 0.3 M KC1 (Table I); however, after this step the stimulator was relatively unstable.
Effect of the Stimulator on KC1 Dependence of Protein Synthesis-The stimulator had greater activity at higher KC1 concentrations than at lower concentrations in the heatshocked lysate. In Fig. 3 the stimulation of protein synthesis (ratio of synthesis with stimulator over synthesis without stimulator) is 1.9-fold at 80 mM KC1 but only 1.2-fold at 40 mM. There is, therefore, a tendency for the stimulator to shift the KC1 optimum to higher concentrations. There was only a small effect of the stimulator on 37 "C lysates a t all KC1 concentrations used.
Effect of eIF-4F and Other Initiation Factors on the Heatshocked Cell-free System-Comparison of the chromatographic properties of the Ehrlich cell factor to those of known initiation factors showed that eIF-4F had the greatest resemblance. Eukaryotic initiation factor 4F is purified from the ribosomal salt wash, elutes from DEAE-cellulose at 0.22 M KC1, elutes from phosphocellulose between 0.15 and 0.45 M KC1, and behaves similarly to the Ehrlich cell factor in gel filtration (15). A number of initiation factors were tested for stimulatory activity in the heat-shocked system (Table  11). Only highly purified eIF-4F had any significant stimulatory effect. Other factors, eIF-2, GEF, and eIF-4B, had no stimulatory effect at any KC1 concentration; however, eIF-4A stimulated by 30%. Further, eIF-4F has greater stimulatory activity on the heat-shocked lysate a t higher KC1 concentrations ( Fig. 4B) and only minimal effect on the 37 "C lysate (Fig.  4A). The reason for the inhibition by eIF-2 and GEF is unknown; these factors have no effect on control cell lysates (data not shown).
Effect of KC1 and eIF-4F on Heat-shocked Protein mRNA Translation-The heat-shocked system was examined for the synthesis of heat-shock protein. Ehrlich cells induce at least one heat-shock protein (68 kDa); others are not detectable in the one-dimensional electrophoresis system shown in Fig. 5. The induction is prevented completely by actinomycin D (Fig.  5, lane 4). The heat-shocked cell lysate contains heat-shock mRNA so that the relative efficiency of translation at different KC1 concentrations could be studied. Densitometric scans (Fig. 6) of autoradiographs of SDS gels of [35S]methioninelabeled protein from lysates incubated at 30 and 70 mM KC1 were analyzed for synthesis of hsp 68 and actin. The results are expressed in Table I11 as the per cent of total protein synthesized. Essentially, a constant percentage of actin is synthesized under different conditions while total protein synthesis is inhibited by 70% at the higher KC1 concentration. However, hsp 68 synthesis is less inhibited at higher KC1 concentrations so that it represents a greater percentage of total protein synthesis. Addition of eIF-4F stimulated total protein synthesis and actin synthesis by 2.6-fold at 70 mM KCl; however, hsp 68 synthesis was stimulated 1.5-fold only.

DISCUSSION
We have shown that lysates prepared from cells incubated at elevated temperatures have a striking alteration in their response to KC1 when incubated in the cell-free proteinsynthesizing system a t 37 "C. At the optimum KC1 concentration (30-40 mM) there is no apparent inhibition in lysates from heat-shocked cells. However, heat-shocked cell lysates are severely inhibited by KC1 concentrations (60-70 mM) that are optimal for 37 "C lysates. The inhibition is not due to a reduction in elongation, which is independent of KC1 concentration in the range used. Premature termination is excluded hecause the pattern of proteins being synthesized is the same at different KC1 concentrations (Fig. 6) and is the same as that found with pulse-laheled whole cells. We conclude that the reduced rate of protein synthesis in 44 "C lysates a t high KC1 concentrations is caused by inhibition of initiation reflecting the inhibition in heat-shocked cells (8). Further, heatshock mRNA is preferentially translated at the high KC1 compared to low KC1 concentrations in the 44 "C lysates.
Ot,hers have observed in intact cells that heat-shock protein mRNAs are more efficiently translated than non-heat-shock . The inhibition of initiation in heat-shocked cell lysates at high KC1 concentrations can he reversed by a factor purified from Ehrlich cell ribosomal salt wash and by highly purified rahhit reticulocyte eIF-4F. The two factors have indistinguishahle chromatographic properties and identical effects on the response of heat-shocked lysates to KCI. The Ehrlich cell factor is presumably eIF-4F, and it can he concluded that an alteration of eIF-4F or an altered requirement for eIF-4F produces changes in the response to KC1 of heat-shocked lysates. However, it is unknown whether the eIF-4F added to our lysates is acting catalytically or stoichiometrically. We propose that the inhibition of initiation in heat-shocked cells is due to reduction of eIF-4F activity which results from either an alteration of eIF-4F or an altered response of the Ehrlich cell translation apparatus to eIF-4F. This conclusion is strengthened by the observation that translation of non-heat-shock mRNA is stimulated more than hsp 68 mRNA hv addition of elF-4F to cell lysates, indicating that the loss of activity of this mRNA binding factor prohahly plays a major role in the preferential translation of hsp 68 mRNA in cells at elevated temperatures. (This well-documented effect of heat shock (9-11) is in addition to the striking induction of transcription of hsp mRNAs.) Ray et al. (24) have concluded previously that eIF-4F (cap binding protein 11) is an mRNA discriminating factor. They observed that addition of eIF-4F to a Krebs ascites cell-free translation system relieved competition between large concentrations of globin and reovirus mRNAs. Therefore, it is possible that hsp 68 mRNA outcompetes other mRNAs for a limited amount of eIF-4F in our heat-shocked cell-free system so that hsp 68 synthesis is less inhibited than non-heat-shock protein synthesis. Addition of eIF-4F to the heat-shock cell lysate relieves this competition (Fig. 6). Further, synthesis of heat-shock proteins in poliovirus-infected HeLa cells is more resistant to inhibition than normal host proteins (25). Poliovirus infection results in loss of eIF-4F activity (26) (possibly by degradation of the 220-kDa subunit) which is consistent with our hypothesis that heat-shocked Ehrlich cells also have reduced eIF-4F activity. We have found no apparent inhibition of any other factor in our heat-shocked cell lysate. Duncan and Hershey (12) have reported reduced eIF-2, eIF-4B, and eIF-3 and/or eIF-4F activity. However, they have found that protein synthesis in HeLa cells grown in monolayer culture is less sensitive to inhibition by elevated temperatures than cells grown in suspension culture. Further, they used a reconstituted in vitro assay to identify changes in initiation factor activity. Therefore, differences in culturing conditions, cell lines, in vitro assay conditions, or all three may explain differences in our findings.
We have previously reported that 40 S initiation complex levels are reduced in heat-shocked Ehrlich cells (8). This is surprising because inactivation of the cap binding protein eIF-4F would, at first sight, be expected to cause inhibition of mRNA binding to 40 S initiation complexes (40 S. GTP. eIF-2 . Met-tRNA) and cause an accumulation of free 40 S initiation complexes. We now have evidence that addition of eIF-4F restores 40 S initiation complex levels in heat-shocked lysates, suggesting that, in an as yet unexplained way, eIF-4F can affect eIF-2 function.' The evidence presented here is consistent with a reduction in eIF-4F activity at higher KC1 concentrations resulting in reduced rates of initiation and preferential synthesis of 68-kDa hsp. However, it is possible that elevated temperatures have other effects on the translation apparatus which are not expressed in our Ehrlich cell-free system.