Regulation of Initiation Factors during Translational Repression Caused by Serum Depletion ABUNDANCE, SYNTHESIS, AND TURNOVER RATES*

During growth in unreplenished medium, the fraction of active, polysomal ribosomes progressively de- creases about %fold from 80-90% to only 20-40% due to a reduced rate of initiation. To assess whether the abundance of initiation factors could be involved in this repression of translational activity, HeLa cell cy- toplasmic lysates were resolved by two-dimensional isoelectric focusing/sodium dodecyl sulfate-polyacryl- amide gel electrophoresis, and spots corresponding to the initiation factor proteins were quantitated. The relative abundance of most of the initiation factor pro- teins only decreases by 10-40% and roughly parallels that of the ribosomes. Measurement of the rates of synthesis and turnover of the initiation factor proteins establishes that during periods of active growth, syn- thesis and degradation occur coordinately with total cell protein. As growth rate decreases, the synthesis of some initiation factor proteins, particularly eukaryotic initiation factor (e1F)-3 subunits, becomes depressed.

The activities of numerous enzymes are regulated by the protein levels in cells. Frequently, enzymes which regulate the overall flux of metabolites through a pathway exhibit high synthesis and turnover rates (7). By altering the rate of synthesis or degradation, a new cellular level can be reached rapidly. It is reasonable to postulate that changes in the abundance of one or more initiation factors may serve to regulate the overall rate of protein synthesis, and that such factors would be synthesized and degraded at relatively rapid rates. In this report, the synthesis, turnover, and abundance of over 10 of the initiation factor proteins have been surveyed to ascertain whether the metabolism of these proteins is correlated with regulated translational activity. Lysates were examined from cells exhibiting different rates of protein synthesis due to varying times of growth in unreplenished medium. The status of the factor proteins was analyzed by twodimensional gel electrophoresis of unfractionated cell lysates, relying upon our identification of a set of spots which correspond to the HeLa cell initiation factors (8).

EXPERIMENTAL PROCEDURES AND RESULTS~
Quantitution of Initiation Factor Proteins-When HeLa cells are grown without replenishing the growth medium, the rate of protein synthesis progressively declines (15, 16) due to a reduction in the initiation rate ( Table 1). The characteristics of this translational repression during serum depletion are described in detail the Miniprint Section. To determine whether the amounts of the initiation factors decrease during serum depletion and might limit the rate of initiation, their levels were measured at 24-h intervals after the serum depletion regime was begun. Quantitations were determined by fractionating 35S-labeled protein on IEF/SDS-PAGE and excising and counting those spots previously identified as corresponding to the initiation factors ( Fig. 1 and Table 2). Additional details covering the electrophoresis, quantitation, and labeling procedures are presented in the Miniprint Sec-tion. Two '%-labeling regimes were used: (i) [35S]methionine was added from the time of replating until protein was extracted; and (ii) ["SJmethionine was added for a 24-h interval prior to extraction (see "Experimental Procedures" for details). The quantitative results were indistinguishable using either labeling regimen.
The results of the quantitations show that for all of the eIF proteins, a progressive decline in amount occurs (Fig. 2). The extent of the decline for eIF-4A, eIF-3p44, and eIF-3p24 is about 50%. eIF-2P and eIF-3p40 decrease by -40%, although the decline in eIF-28 is likely caused in part by the appearance of variant forms in depleted cells (17), which were not quantitated because they were not unambiguously resolved on the autoradiograms. For the remainder of the eIF proteins, the extent of decline is only 10-30%. Such small abundance reductions are not likely to be directly responsible for inhibiting the rate of initiation by 3-4-fold, although a contributory role cannot be excluded.
The initiation factor quantitations determined by IEF/ SDS-PAGE were confirmed by immunoblotting (see "Results" in Miniprint and Fig. 3). The abundance of the ribosomal proteins likewise shows little or no variation during the serum depletion regime (see "Results" in Miniprint and Table  3). Thus, the initiation factor/ribosome ratios remains relatively constant for the initiation factors that show little abundance decline and may decrease by 50% for the factors whose abundance decreases the most.
Rates of Synthesis of the Initiation Factor Proteins-The intracellular abundance of proteins, including the initiation factor proteins, is controlled by their respective rates of synthesis and degradation. We next measured these rates to provide additional insight into how initiation factor protein levels are regulated. We note that the maintenance of relatively constant levels of eIF proteins during serum depletion does not imply that synthesis and degradation rates remain constant, but only that they are coordinately varied. In order to approach the problem of factor synthesis rates, we measured the fraction of radioactivity incorporated into specific eIF proteins compared to total proteins. Cells were pulse- Cells were plated and propagated in nonradioactive culture medium just as for the labeling analysis. After 1. 2. 3 , or 4 days of growth, protein was extracted by the IEF/SDS-PACE sample preparation method. Samples of about 30 pg were resolved by onedimensional SDS-PAGE on 100 X 150 X 0.75 mm d a h gels, electrophoretically transferred to nitrocellulose paper, and sequentially treated with anti-initiation factor protein antihodies followed by '=Ilabeled second antihody. Autoradiograms of gels treated with antibodies to eIF-Za, 28, and 2.1 @one1 A ) or eIF-4A and elF-4R (panel R ) are shown. The gel in panel A was 10% total acrylamides, 0.26% bisacrylamide and the gel in p a n e l R was 1 5 6 total acrylamides, 0.06% bisacrylamide. L o w s D l . 02. 0 3 , and 114 were extracted after 1, 2, 3, and 4 days of growth, respectively.

TABLE 3
Quantitotion of ribosomal protein9 Cells were labeled continuously from the time of replating until harvesting wth [%]methionine. Extracted protein was analyzed by two-dimensional gel electrophoresis in the presence of carrier ribosomal protein as described under "Experimental Procedures." The amount of radioactive protein applied to each gel was determined from a trichloroacetic acid precipitate of an aliquot of the gel sample. Several independently prepared samples were each analyzed two or more times by two-dimensional gel electrophoresis. and the averages are reported as the counts in the specific ribosomal protein spot divided by the input radioactivity. The ahundancies have not been corrected for either protein molecular weight or methionine content. Thus, the values are not absolute abundancies. but they may be uwd to compare the relative abundance of the same protein in different states. labeled with [%S]methionine over a 30-min interval, and radioactivity incorporated into eIF proteins was quantitated as described above. The per cent radioactivity values were then converted to molar ratios. Details of the labeling protocols and data analysis are given under "Experimental Procedures." These procedures provide a measure of the relative values for two eIF-2 subunits.

Ribosomal
. ~. . rates of synthesis of eIF proteins compared to total protein.
Relative synthesis rates were examined in serum-stimulated, exponentially growing, and serum-depleted cells as described under "Experimental Procedures" and are reported in Table 4. In serum-stimulated and exponentially growing cells, the percentage of factor molecules synthesized lies in the range of 0.05-0.10 for most factors, with eIF-3p24 and eIF-4A showing significantly higher rates of synthesis. The relative synthesis rate values are essentially indistinguishable from their steady state levels ( Table 2, Miniprint Section). Thus, these factors do not possess unusually high rates of synthesis which could allow a rapid modulation of their cellular levels.
Comparison of serum-stimulated and exponentially growing cells versus serum-depleted cells shows that the relative rates for the latter are somewhat lower for a number of factors ( Table 4). Five of the 10 factors analyzed (2a, 2& 31336, 4A, and 4B) show no significant (f25%) change; 3p44,3p47, and 3~2 2 0 decrease to 55-65% of serum-stimulated values, whereas 3p24 and 3p40 decrease to 39 and 44%, respectively. We conclude that the relative rates of synthesis of initiation factors either change very little or decrease to about half as cells become serum-depleted. The changes observed are generally comparable to the changes in abundance reported in Fig. 2. While the relative synthesis rates for eIF proteins are quite comparable in serum-stimulated versus serum-depleted cells, the absolute rates of synthesis differ 4-6-fold (Table 1 and Table 4, legend). The data indicate the absence of a feedback mechanism which would cause a specific inhibition of the relative rates of synthesis of eIF proteins when the absolute rate of protein synthesis decreases.
Pulse intervals as short as 1 min were also examined. Gels comparing a 1-min versus a 24-h labeling time (Fig. 1, Mini-

TABLE 4
Relntive rates of synthesis of the initiation factor proteins Cells were grown, labeled, and analyzed for relative rates of eIF synthesis as described under "Experimental Procedures." In this and subsequent tables multiple samples were analyzed and results averaged. Standard errors are as described (see Table 11, Ref. 8,and Fig. 1). Absolute rates of synthesis were calculated based on the rate of total protein synthesis, the fractional percentage due to initiation factor protein synthesis, and the [35S]methionine uersuS "C-amino acid labeling percentages. The total protein synthesis rate was calculated based on 3 X lo6 ribosomes/cell, of which about 80% are active in recently serum-stimulated cells, and an average elongation rate of 5 amino acids/s in serum-stimulated cells (9). Thus, the values in the table represent the maximum synthetic rates for protein synthesis which occur in serum-stimulated cells. In exponentially growing cells, the fraction of active ribosomes usually decreases to 65-75% and the polymerization rate decreases to about 3 amino acids/s (9). Thus, the rates in exponentially growing cells would be about 50% of the ones reported in the table, and the rates in serumdedeted cells would be about 20% of these. print Section) show initiation factor protein spots of approximately equal darkness. The results confirm and extend our conclusion that the rates of eIF protein synthesis mirror their accumulation and abundance even during an interval approximating a single ribosomal transit time. In addition, they provide further evidence that the steady state form of each initiation factor protein is also its primary mRNA translation product, since initiation factor protein spots after 1 min of labeling are neither of decreased intensity nor absent as might occur if some were derived by protein processing from a larger precursor.
Rates of Turnover of the Initiation Factor Proteins-Measurements of protein turnover were performed in parallel with the synthesis measurements. Following a 24-h incubation in medium with [35S]methionine, cells were washed 3 times and then incubated in nonradioactive medium containing 400 PM methionine for 24 h during which time several replicate cultures were extracted for protein analysis (see "Experimental Procedures" for details). About a 20% loss of total labeled protein occurred during the 24-h chase for serum-stimulated and exponentially growing cells, whereas a more pronounced decrease of about 70% was found in depleted cells (Fig. 4). For the serum-stimulated and exponentially growing cells, this corresponds to protein half-lives of about 40-60 h; for serum-depleted cells, the rate of loss is consistent with a halflife of 4 h for about 75% of the labeled protein. How salvage reutilization of radioactive amino acids derived from protein degradation affects the turnover rate measurements has not been determined. We then measured the loss of radioactivity in individual eIF protein spots separated by IEF/SDS-PAGE. The percentages of radioactivity remaining after a 24-h chase are reported in Table 5. The turnover rates of the initiation factor proteins are indistinguishable from the turnover of total cell proteins in all cell states examined.
Initiation Factor mRNA Utilization-Experiments described thus far suggest that the mRNAs encoding the eIF proteins form a constant percentage of the translationally active species in rapidly growing cells and may decrease for   Initiation factor protein synthesis following serum stimulation of serum factor-depleted cells Cells were grown for 4 days without replenishing the serum or medium. To some cells [36S]methionine was added for 60 min, and then protein was extracted and quantitated (column 0 h). The remaining cells were serum-stimulated by removing the old medium and replacing it with fresh medium containing 10% calf serum. Cells were labeled in complete medium for 60 min and extracted at the times after serum stimulation indicated in the some factor proteins in serum-depleted cells. In the depleted cells, less than half of the mRNA and the ribosomes are functioning, but most of these are rapidly activated upon addition of fresh serum (Table l).3 It is possible that initiation factor synthesis is greatly favored when depleted cells are serum-stimulated. This would lead to increased levels characteristic of active dividing cells. Thus, we wished to test whether the free mRNAs mobilized into polysomes by serum in serum-depleted cells are enriched in initiation factor mRNAs. Cells grown 96 h in unreplenished medium were given fresh 10% serum-containing medium and pulse-labeled at several times thereafter, and the rates of initiation factor synthesis were determined ( Table 6). The rate of initiation factor protein synthesis provides an indirect measure of the number of active mRNAs. Little change in the relative rates of initiation factor protein synthesis is detectable for most factor proteins, although all proteins are synthesized to a greater extent. The results indicate that initiation factor mRNAs are not selectively recruited from the free mRNA pool following serum stimulation. Three proteins showed enhanced synthesis in several independent analyses. eIF-3p44 R. Duncan and J. W. B. Hershey, unpublished results. synthesis increases about 30%, and eIF-4A synthesis increases about 25%. The most striking alteration is in eIF-3p24 synthesis which has been found to increase 3-&fold. The relative rate of synthesis of this protein is also the most severely depressed in serum-depleted cells ( Table 2). The change in the synthesis of eIF-3p24 is detected within 2 h after serum stimulation (pulse label from 1-2 h), and a high relative rate persists for a few hours more, finally dropping by 11.5 h to the relative rate observed in exponentially growing cells.

DISCUSSION
Cell growth in unreplenished medium is accompanied by a progressive decline in the fraction of active ribosomes. This repression of protein synthesis in attached HeLa cells occurs principally at the level of initiation, as indicated by the polysome profiles and by the fact that low levels of cycloheximide cause polysomes to reassemble. Such behavior has been reported previously for HeLa cells grown in suspension cultures (15). Translational repression in unreplenished medium is likely due in large part to depletion of serum growth factors, although this issue is not addressed here. The repression can be reversed in suspension cultures by adding insulin and epidermal growth factor to the depleted growth medium? We asked whether this decrease in the rate of initiation of protein synthesis is due to changes in the abundance of one or more initiation factors, and how the levels of eIF proteins are established and regulated.
The cellular levels of 10 initiation factor proteins were measured by two-dimensional gel fractionation of labeled lysate protein. During the 4-6-fold reduction in protein synthesis rate as cells are grown for up to 96 h in unreplenished medium, relative eIF protein levels decrease either slightly or to about half (Fig. 1). Such moderate changes in eIF abundance were confirmed by an independent method, immunoblotting (Fig. 3), for 5 proteins. It seems likely that the modest changes in factor levels are not primarily responsible for the much larger effects on overall protein synthesis rates. Of the three factors showing the greatest reduction in level, two (eIF-3p24 and eIF-4A) are the most abundant factors measured ( Table 2) and are therefore poor candidates for a regulatory role. However, eIF-4A in crude lysates has been shown to be limiting for the in vitro translation of certain viral mRNAs (21) and might therefore contribute to the repression observed. Whereas we can rule out the likelihood that overall protein synthesis rates are controlled by the levels of any one of these 10 eIF proteins, it is possible that one of the initiation factors not analyzed plays such a regulatory role. It has been proposed that eIF-4F (the cap-binding protein complex) is limiting in cells and is responsible for mRNA discrimination (22). Of the three components of eIF-4F (p220, eIF-4A, and eIF-4E), the p220 subunit appears by immunoblotting to be as abundant as the eIF-2 p r~t e i n .~ Its level was assessed during the serum depletion regimen and no change was detected (results not shown). The level of eIF-4A is very high and changes little, as has been discussed above. eIF-4E (the 26-kDa cap-binding protein) is much more likely to be limiting in absolute level. The abundance of eIF-4E is too low to quantitate accurately by cutting out gel spots. Other means of estimating its abundance indicate that eIF-4E comprises about 0.01% of the cellular protein molecules.' Whether eIF-4E levels are regulated in serum-depleted cells cannot be determined by the methods employed here.
In exponentially growing HeLa cells, the relative rates of synthesis of eIF proteins are comparable to their cellular ' R. Duncan, unpublished results. levels, and they are degraded at slow rates characteristic of the total protein population. In effect, rates of factor synthesis determine factor levels in these cells. The results refute the possibility that initiation factor proteins are members of the small subclass of proteins that are synthesized and degraded unusually rapidly, presumably to facilitate rapid modulation of intracellular concentration connected with metabolic regulation (7).
During serum depletion, the relative rates of synthesis change very little if at all for most of the eIF proteins, although their absolute rates decrease 4-6-fold. In contrast, degradation rates for both eIF and total protein increase substantially in the serum-depleted state. This means that both synthesis and degradation make important contributions to establishing factor levels in very slowly dividing cells. Most noteworthy is that synthesis and degradation of initiation factors are strictly coupled to the metabolism of total protein. This is consistent with the observation above that relative levels of the initiation factors do not change appreciably in serum-depleted cells. The coupling is also seen in serum-stimulated cells which had been previously depleted of serum (Table 4), where the relative rates of factor synthesis change very little (except for eIF-3p24). In contrast, resting chick embryo fibroblasts stimulated by insulin preferentially synthesize ribosomal proteins (23). Further work is required to determine whether initiation factor synthesis is generally coupled to total protein metabolism and whether they are always coordinately expressed with other components of the translational apparatus.
[3$linethionine presented in the min results SKtfonr specific protein are internally valid regardless of the dbsDIute abuodaacfcs OF there proteins, teins Mar also ewaluatea by sawn etting vnlabeled lysates prepared fmn CUItures 1. 2, 3 ami 4 days after replating. m e t e lysates wee reIolYed by one-dimnstoRil1 gels and probed with affinity putifted antlnodies to efF-Ze, elF-28. eIF-2r. elf-48, and eIF-B. me a w m t O f anttbodv baurrd provider a yamblottfng results ifipure 3 ) support the labeling mdlSUrCD3nts: the akUndlnce t l t s t i v e estimate of initirtion factor pmtein abundance 1111. me inunoof a l l the elF-2 =&units and eIf-&a sDw l i t t l e varidtton. and elf-4.4 ab""dance *creases about 2 fold (saasund br hand cutting and eosnting).