Does Prothymosin α Affect the Phosphorylation of Elongation Factor 2?*

Prothymosin α is a small, acidic, essential nuclear protein that plays a poorly defined role in the proliferation and survival of mammalian cells. Recently, Vega et al.proposed that exogenous prothymosin α can specifically increase the phosphorylation of eukaryotic elongation factor 2 (eEF-2) in extracts of NIH3T3 cells (Vega, F. V., Vidal, A., Hellman, U., Wernstedt, C., and Domı́nguez, F. (1998) J. Biol. Chem. 273, 10147–10152). Using similar lysates prepared by four methods (detergent lysis, Dounce homogenization, digitonin permeabilization, and sonication) and three preparations of prothymosin α, one of which was purified by gentle means (the native protein, and a histidine-tagged recombinant prothymosin α expressed either in bacteria or in COS cells), we failed to find a response. A reconstituted system composed of eEF-2, recombinant eEF-2 kinase, calmodulin, and calcium was also unaffected by prothymosin α. However, unlike our optimized buffer, Vega’s system included a phosphatase inhibitor, 50 mm fluoride, which when evaluated in our laboratories severely reduced phosphorylation of all species. Under these conditions, any procedure that decreases the effective fluoride concentration will relieve the inhibition and appear to activate. Our data do not support a direct relationship between the function of prothymosin α and the phosphorylation of eEF-2.

The detailed function of prothymosin ␣ remains cryptic and controversial. The protein is not, as its name intimates, a precursor for much smaller, processed, thymic hormones (1)(2)(3)(4)(5). Rather, prothymosin ␣ is exclusively mammalian (see Refs. 6 and 7, and references therein), extremely abundant (8,9), probably unfolded (10), and highly acidic throughout its length of 109 -110 amino acids (1,11,12). It localizes to the nucleus in intact form in all tissues (13)(14)(15)). An overview of the prothymosin ␣ literature strongly suggests that the protein plays an essential, but nevertheless undefined, role in the growth and survival of virtually all cells (1, 16 -19). Several laboratories have noted (i) that an increase of greater than 10-fold in the mRNA and the protein occurs when quiescent cells proliferate, (ii) that the amount of prothymosin ␣ in a tissue mirrors its inherent growth rate, and (iii) that prothymosin ␣ is expressed in a chronologically contemporaneous manner with other growth related molecules such as Myc, histone H3, and proliferating cell nuclear antigen (1, 3, 8, 20 -24). However, the most compelling evidence comes from synchronized human myeloma cells, which, when exposed to antisense oligodeoxyribonucleotides directed toward prothymosin ␣ mRNA at several locations, undergo growth arrest (25).
The particulars of prothymosin ␣'s function are contested. The amino acid sequence, which should provide clues, is essentially unique. Except for a "classical" bipartite nuclear localization signal included in many otherwise unrelated nuclear proteins, the only shared featured is a reiterated region of glutamic acid residues found also in the Xenopus chromatin remodeling protein, nucleoplasmin (26,27). Binding studies in vitro have targeted various histones as binding partners, including core histones or histones H1, H3, and H4 depending on the conditions (28), whereas an examination of a single clone of HL-60 cells permanently transfected with the prothymosin ␣ gene suggested that excess prothymosin ␣ might deplete and sequester nucleosomal histone H1 (29). In contrast, prothymosin ␣ did not bind to histone H1-containing chromatin or to mononucleosomes in related studies (30,31). Kubota et al. demonstrated the binding of prothymosin ␣ to the activation domain of the human immunodeficiency virus protein, rev, and the human T-cell leukemia virus protein, rex, in vitro (32). Since rev is involved in the export of incompletely spliced viral RNA from the nucleus, it is tempting to speculate that prothymosin ␣ interacts with a rev-like protein produced in normal cells and that its role is to assist in the export of RNA or protein from the mammalian nucleus. This view of prothymosin ␣ as a facilitator of nuclear export is entirely distinct from the previous view of the protein as a histone-binding factor involved with chromatin.
Studies performed in vivo provide yet another perspective. The Berger group has shown that prothymosin ␣ is phosphorylated in vivo on glutamic acid residues in the region homologous with nucleoplasmin (33). These acyl phosphates are unstable in vivo and energy rich (33,34). Upon cell lysis, most hydrolyze almost instantaneously, while a tiny fraction migrate to more stable positions on nearby serine or threonine residues (33,35). By observing the changes in stability of prothymosin ␣'s phosphates in vivo in cells overexpressing prothymosin ␣, in synchronized cells at different stages of the cell cycle, and in cells treated with metabolic inhibitors, the group concluded that prothymosin ␣'s phosphates become stabilized and inactive during mitosis and in transcriptionally compromised cells (36). These experiments did not identify the prothymosin ␣-requiring process, but imply that prothymosin ␣'s phosphates participate in some aspect of RNA biosynthesis. This idea is compatible with all of the binding partners noted above, but is also consistent with a catalytic function involving transfer of high energy phosphate.
Yet another unrelated function has been proposed. Vega et al. (37) assert that prothymosin ␣ stimulates Ca 2ϩ -dependent phosphorylation of eukaryotic elongation factor 2 (eEF-2) 1 in cellular extracts. eEF-2 catalyzes translocation of the ribosome along mRNA during translation. Phosphorylation, which renders eEF-2 inactive (38), could play a role in the regulation of the rate of protein synthesis (reviewed in Refs. 39 and 40). Vega et al. surmise that prothymosin ␣ exerts its effect on eEF-2 phosphorylation at mitosis when the nuclear and cytoplasmic compartments are no longer distinct. Hence, they have implicated prothymosin ␣ in the down-regulation of translation during M phase. We take issue with this proposal and with the observations on which it is based.
Here, we show that prothymosin ␣ has no effect on the ability of extracts to phosphorylate eEF-2. We found that Vega and co-workers studied the reaction under conditions of severe, nonspecific inhibition and that in our hands even in inhibited extracts, prothymosin ␣ is ineffective. Finally, we studied the phosphorylation of eEF-2 by eEF-2 kinase with purified components in the presence and absence of prothymosin ␣ and found no effect. Therefore, we conclude that prothymosin ␣ has no role to play in the regulation of eEF-2 phosphorylation.

EXPERIMENTAL PROCEDURES
Growth and Transfection of Mammalian Cells-NIH3T3 cells from the American Type Culture Collection (CRL 1658) were cultivated in Dulbecco's modified Eagle's medium containing 10% calf serum from Life Technologies, Inc., antibiotic-antimycotic, and 2 mM glutamine as described by Tao et al. (36). COS 1 cells (ATCC, CRL 1650) were cultured in the same medium containing the same additives, but with 10% fetal bovine serum (Hyclone). The cells were transiently transfected with a prothymosin ␣ gene containing six histidine codons inserted immediately upstream of the stop codon (33) using the DEAEdextran procedure (41). Cultures were maintained at 37°C in an atmosphere of 5% CO 2 . HeLa S-3 cells in liter quantities were a kind gift of Dr. Bernard Moss (NIAID, National Institutes of Health, Bethesda, MD).
Bacteria-In some experiments, Escherichia coli strain BL21(DE3) transformed with pET3a containing the prothymosin ␣ cDNA coding region similarly tagged with six histidine codons (33) was grown and induced to overexpress histidine-tagged prothymosin ␣ using a kit and instructions provided by Stratagene.
Purification of Prothymosin ␣ from HeLa Cells, Transfected COS Cells, and from Bacteria-Native prothymosin ␣ was isolated from HeLa cells as described by Trumbore et al. (33), Wang et al. (34), and Sburlati et al. (9). Briefly, washed cells were lysed in the presence of a detergent, nuclei were removed by centrifugation, and prothymosin ␣, which leaks out quantitatively, was recovered from the supernatant fluids by means of a phenol extraction. However, because such preparations contain RNA and other unknown contaminants as well as prothymosin ␣ in the aqueous phase, the protein was further purified by ion exchange chromatography on a Bioscale Q2 column run with a Biologics (Bio-Rad) medium pressure chromatography system. The Q2 column replaced the TSK-gel DEAE-5PW column used in earlier studies (33). Samples were also subjected to reverse phase chromatography on a C-18 column to remove predominantly non-macromolecular contaminants (8). Histidine-tagged prothymosin ␣ from COS cells was isolated from lysates with the aid of metal chelate chromatography on nickel nitrilotriacetic resin (Qiagen) and purified further on the aforementioned Q2 column. Procedures and reagents that might inactivate or denature globular proteins such as heat, acid, phenol, or reverse phase chromatography were rigorously eschewed.
Recombinant prothymosin ␣ tagged with six histidine residues (tagged prothymosin ␣) was obtained using the methods detailed in the Stratagene instruction manual entitled pET System Vectors and Hosts. Further purification was effected with the Bioscale Q2 and C-18 columns noted above.
The concentration of purified prothymosin ␣ was determined from the absorbance at 214 nm and validated by amino acid analysis; in lysates, the amount in arbitrary units was obtained by densitometric scanning of bands in the identical stained gel (33).
Purification of eEF-2 and eEF-2 Kinase-Eukaryotic EF-2 was puri-fied from rabbit reticulocytes essentially as described by Ryazanov and Davydova (42). DNA coding for eEF-2 kinase was obtained from a cDNA library (CLONTECH) derived from human glioma cell line T98G, subcloned into pGEX-2T (Amersham Pharmacia Biotech), and expressed in Escherichia coli, BL21(DE3)pLysS. A portion of the resultant glutathione S-transferase (GST)-linked kinase was purified as an active enzyme from the soluble fraction, whereas the majority was recovered from inclusion bodies after denaturation in urea and renaturation by dialysis. Our methods were modifications of published procedures (43). The use of the GST-linked eEF-2 kinase made it possible to distinguish phosphorylation of eEF-2 from autophosphorylation of the kinase. Without the adduct, the kinase at ϳ105 kDa and eEF-2 at ϳ100 kDa would have comigrated during gel electrophoresis. Preparation of Lysates of NIH3T3 Cells-Lysates for analysis of eEF-2 phosphorylation were prepared by four methods: detergent lysis; Dounce homogenization; sonication; and digitonin permeabilization. The preliminary steps performed in all cases were identical: NIH3T3 cells were trypsinized, suspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, recovered by centrifugation, and washed once in phosphate-buffered saline. They were then resuspended in homogenization buffer (50 mM Tris-HCl at pH 7.5, 0.1 mM EDTA, 1 mM dithiothreitol, 250 mM sucrose, and protease inhibitors: 5 g/ml soybean trypsin inhibitor, 2 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride) and stored on ice at a concentration of 40 ϫ 10 6 /ml in preparation for lysis. Detergent lysis was carried out by making the homogenization buffer 1% in Triton X-100 and incubating the sample for 10 min on ice. Nuclei and debris were removed by brief centrifugation in a Tomy Tech USA refrigerated microcentrifuge at top speed (15,000 rpm) for 5 min, and the supernatant fluids were recovered and stored frozen. Cell lysates were prepared by Dounce homogenization by subjecting the cell suspension in homogenization buffer to 10 strokes of the tight pestle. Nuclei and debris were eliminated centrifugally as noted above, and the supernatant fluids were collected and stored frozen in aliquots. Cells were disrupted by sonication by treating them in homogenization buffer with a Sonic Dismembrator 550 (Fisher Scientific) equipped with a microtip probe at a power setting of 2 with a 1-s pulse and a 1-s delay between pulses. Sonication was continued on ice until froth formed in the sample, ϳ15 s. Debris was removed by brief centrifugation. The supernatant fluids were collected and centrifuged at 100,000 ϫ g for 30 min in a Sw55Ti rotor. The soluble material was dispersed into several tubes and frozen. The gentlest procedure was permeabilization in homogenization buffer containing 40 g/ml digitonin. After incubation at 37°C for 10 min, the perforated cells and debris were removed by centrifugation and the supernatant fluids were recovered and stored frozen. In all cases, samples were desalted by two successive passes through G-50 spin columns that had previously been equilibrated with homogenization buffer without sucrose. In some cases, the lysates were freshly prepared, desalted, and used immediately without undergoing a freeze-thaw cycle. Sample Preparation and Analysis-Reactions were stopped by a combination of rapid cooling to 4°C, dilution with 5ϫ SDS electrophoresis sample buffer, and boiling for 5 min. Proteins were resolved electrophoretically in 8, 10, or 18% polyacrylamide SDS gels run in Trisglycine buffer. The gels were either prepared in the laboratory or purchased from Novex, and stained with Coomassie Brilliant Blue. We used SeeBlue prestained standards from Novex according to the en-closed directions. After destaining, gels were dried and exposed to X-AR film (Eastman Kodak Corp.) in the presence or absence of intensifying screens.

RESULTS
Phosphorylation of Sonicated Extracts-Cytoplasmic extracts from mammalian tissues, when incubated with [␥-32 P]ATP incorporate radioactivity into proteins. In the presence of Ca 2ϩ , a band at ϳ100 kDa, which can be identified as eEF-2, acquires label rapidly and dominates the electrophoretic pattern (39). Accordingly, we examined the ability of sonicated lysates of NIH3T3 cells to phosphorylate a ϳ100-kDa protein in a series of incubations extending from 1 to 20 min. When the products were resolved electrophoretically (Fig. 1), the expected band at ϳ100 kDa appeared after the shortest incubation period, and gained a modicum of additional radioactivity with increased reaction times. Although the putative eEF-2 remained prominent throughout the study, by 20 min many other radioactive proteins had appeared.
The literature states that prothymosin ␣ stimulates the rapid phosphorylation of eEF-2 (37). In order to corroborate these findings, we tested two different kinds of prothymosin ␣; the first was native prothymosin ␣ purified to homogeneity from HeLa cells, whereas the second was a recombinant human prothymosin ␣ with six histidines affixed to the carboxyl terminus. The results (Fig. 1) were inconsistent with the published record (37). Neither type of prothymosin ␣ stimulated the phosphorylation of the ϳ100-kDa band. Indeed, the incorporation of radioactivity into any band, at any period of incubation, was not enhanced by the presence of prothymosin ␣. In these experiments, the Coomassie Blue-stained gels of all samples were identical regardless of the nature of the addition or the time of incubation (data not shown). Clearly under these conditions, prothymosin ␣ appeared to be inert.
Effect of Lysate Preparation on the Phosphorylation of eEF-2-Sonicated extracts of cells prepared with separate instruments differ. Therefore, in order to test widely differing lysates, we prepared cell extracts using four methods: detergent lysis, Dounce homogenization, sonication, and permeabilization with digitonin. From inspection of Coomassie Blue-stained gels ( Fig.  2A), it is obvious that the compositions of the extracts vary and that each extract contains an array of prominent bands of similar size. It is not obvious that the entire cell complement of endogenous prothymosin ␣ is also present in these lysates ( Fig.  2A, lanes labeled Ϫ). In order to obtain sufficient prothymosin ␣ to be visible in a stained gel, it would have been necessary to load 1.5 mg of total protein, an amount ϳ100-fold higher than that in Fig. 2A. We have found that Triton lysis quantitatively releases the prothymosin ␣ in a cell into the supernatant fluids (9) and that electrophoretic analysis followed by densitometry of the extracted materials from 5 ϫ 10 6 NIH3T3 cells resulted in 710 arbitrary units of prothymosin ␣. Within an error of less than 5%, the identical amount of prothymosin ␣ was recovered from sonicated and digitonin-permeabilized extracts, but lesser amounts were obtained from Dounce-homogenized cells owing to the failure of the method to rupture all of the cells. In contrast, the exogenous prothymosin ␣ at 8 M in the assay was easily viewed as an intense, stained band near the bottom of the 18% polyacrylamide gel (lanes labeled P), with the more slowly migrating histidine-tagged prothymosin ␣ located immediately above it (lanes labeled T).
The products of the phosphorylation reactions catalyzed by these lysates in the presence of [␥-32 P]ATP are exhibited in Fig.  2B. Here, it is apparent that the intensities of the ϳ100-kDa bands of eEF-2 are a function of the composition of the lysate, but it is equally clear that neither prothymosin ␣ from HeLa cells nor its histidine-tagged recombinant counterpart affect the reaction. The phosphorylation of eEF-2 remained independent of prothymosin ␣ regardless of the quality or the quantity of cytoplasmic components in the lysate. Although there is no certainty that any of these extracts is identical to those of Vega et al. (37), it is also unlikely that the reported stimulatory effect of prothymosin ␣ was obliterated in all of them by the acquisition of inhibitors or the loss of cofactors during preparation.
Effect of Distinct Preparations of Prothymosin ␣ and ATP on the Phosphorylation of eEF-2 in Crude and Purified Systems-In a reconstituted system, it is possible to study the effect of prothymosin ␣ on the phosphorylation of eEF-2 by its specific kinase in the absence of complicating factors. Differences in the levels of eEF-2, or kinase, or phosphatases in the lysates, as well as variations in the amount of nonradioactive ATP remaining after desalting all have the potential to influence the results. To avoid these pitfalls, we examined the effect of prothymosin ␣ on the reaction using purified enzyme and substrate. We focused, first, on the reconstituted system using [␥-33 P]ATP as the source of radioactivity. In Fig. 3, it can be seen that both eEF-2 and the kinase had to be present, together with calmodulin, for phosphorylation of eEF-2 to occur (Fig. 3,  lane 3) and that the separated components were inactive (Fig.  3, lanes 1 and 2). Fig. 3 (lane 3) also revealed a band of lesser mobility and intensity, which represents autophosphorylation of the GST-linked kinase. When the effect of native prothymosin ␣ on the reaction was examined, the results (lane 6) reinforced those in Fig. 1; prothymosin ␣ did not influence the phosphorylation of eEF-2 by its kinase in the reconstituted system. The 33 P-labeled ATP was then tested in sonicated lysates. Once again, prothymosin ␣ did not improve phosphate acquisition by eEF-2 in a crude system (lanes 4 and 5), but with the less energetic [ 33 P]ATP, the broad band of radioactivity at ϳ100 kDa in Figs. 1 and 2B was resolved into two discrete bands, the more mobile of which comigrated with genuine, phosphorylated eEF-2.
Because the activity of prothymosin ␣ is unknown and cannot be assayed in vitro, there is no guarantee that functional prothymosin ␣ was used in either these or the published experiments (37). Therefore, we avoided all harsh steps in the purification by isolating histidine-tagged prothymosin ␣ from overexpressing COS cells using only methods that would not denature a sensitive globular protein. The data show that the gently handled, tagged prothymosin ␣, which is clearly visible as a band in the stained gel (Fig. 4A, lane 5), did not significantly enhance the phosphorylation of purified eEF-2 by its purified kinase (Fig. 4B, lanes 4 and 5). In the same experiment, we demonstrate that the gently treated, tagged prothymosin ␣ (Fig. 4A, lane 1) also had no effect on the ability of eEF-2 to acquire phosphate in lysates (Fig. 4B, lanes 1 and 2). Moreover, in order to prove without question that the prominent labeled band in lysates is eEF-2, we mixed the purified and crude samples after the phosphorylation reaction had taken place and analyzed them together. As shown in Fig. 4B  (lane 3), the mixing experiment clearly distinguished eEF-2 as the labeled band with increased intensity. Similarly, in the stained gel in Fig. 4A (lane 3) eEF-2 is less conspicuous but nevertheless recognizable as a darker band among a cluster near 100 kDa (see arrow). Furthermore, when the stained and labeled gels in Fig. 4 were superimposed, it became evident that phosphorylated (labeled) eEF-2 comigrated with bulk eEF-2 and that the protein does not undergo a change in mobility upon acquiring phosphate.
Effect of Fluoride on the Phosphorylation of eEF-2-Our studies, which focused on the phosphorylation of eEF-2 from the outset, were quite different from those of Vega et al. (37), which sought prothymosin ␣-induced effects on the phosphorylation state of any physiologically relevant protein. In particular, our system did not contain fluoride, presumably added as a nonspecific phosphatase inhibitor. The ability of fluoride to form complexes with magnesium and phosphate was recognized as early as 1941 (45) and reported thereafter in a rich literature encompassing the inhibitory effects of fluoride on systems such as adenylate cyclase (46,47), DNA polymerase (48), and ATPase (49), to mention only a few. We, therefore, tested sonicated lysates of NIH3T3 cells for their ability to phosphorylate eEF-2 in the optimized buffer in the presence and absence of fluoride. Our data (Fig. 5) clearly confirm the older fluoride literature; 50 mM fluoride, the amount used by Domínguez and co-workers (37), when added to the reaction mixture virtually eliminated the phosphorylation of every band in the gel (Fig. 5B), although both extracts contained an identical array of stained proteins (Fig. 5A). In order to quantify the effect of graded doses of fluoride on the phosphorylation of eEF-2, we made use of the reconstituted system. The results in Fig. 5B show that low doses of fluoride were tolerated, but that a marked inhibition occurred at doses higher than 30 mM. When the fluoride concentration was increased to 50 mM, the ability of the kinase to phosphorylate eEF-2 was virtually abolished. There are two conclusions to be drawn from these experiments. (i) Fluoride is a potent nonspecific inhibitor of the phosphorylation reactions in both the crude and purified sys- tems; and (ii) in a severely compromised system, relief of the inhibition of phosphorylation by any agent can readily be construed as activation.
Effect of Prothymosin ␣ on the Phosphorylation of eEF-2 in Lysates under Conditions used by Vega et al-To reproduce the conditions used in the previously published report (37), we replaced the optimized buffer with the fluoride-containing buffer used by Vega et al. (37), and reduced the ATP concentration by eliminating the nonradioactive ATP from the assay. When the phosphorylation reaction was examined in Vega's buffer in cell extracts prepared by detergent lysis and by sonication, there was a dramatic transformation in the nature of the radioactive products with no substantive effect on the array of stained protein (Fig. 6). There are several important points to consider. (i) Although the specific activity of the radioactive ATP was much higher, under these conditions, the autoradiograms had to be exposed for several days, rather than hours to discern a signal. (ii) The location of the major bands shifted. Despite superficial similarities with the autoradiograms in Fig.  2, namely a cluster of radioactive high molecular weight products as well as low molecular weight material of comparable intensity, the high intensity cluster did not include eEF-2. As shown on the right side of Fig. 6, which also displays the stained reconstituted system, genuine eEF-2 is substantially smaller than any of the intensely labeled large products and migrates in a region of the gel which contains no perceptible labeled protein. (iii) Reactions performed in the presence of fluoride, unlike those in its absence (Fig. 2) generated qualitatively different patterns of radioactive products depending on the nature of the lysate. And most important, (iv) the addition of native prothymosin ␣ (lanes denoted P) to the phosphorylation assay had no effect on eEF-2, which was not labeled; it had no effect on the labeling of most other proteins; and when an effect could be discerned, it was an inhibitory one (Triton lysis, lane denoted P). Furthermore, prothymosin ␣ also failed to influence the phosphorylation of eEF-2 at pH values from 6.7 to 7.5 in buffers otherwise identical to that of Vega et al. (37) whether or not 50 M ATP was added to the tracer (data not shown). The same negative results were obtained in the presence of fluoride with recombinant prothymosin ␣ and with lysates obtained by Dounce homogenization or digitonin permeabilization (data not shown). Our attempts to reproduce the published results were unsuccessful despite alterations in the buffer, the ATP concentration, the components of the lysate, and the nature of the prothymosin ␣ used as an effector. DISCUSSION Our study was prompted by the work of Domínguez and co-workers who found that the addition of prothymosin ␣ aug- mented the ability of 100,000 ϫ g supernatants prepared from NIH3T3 cells to phosphorylate eEF-2 (37). Since prothymosin ␣ is a nuclear protein (13)(14)(15), and the translation factor, eEF-2, is exclusively cytoplasmic, it was not immediately clear how and where the two proteins might interact in a physiologically relevant manner. The dilemma was resolved by considering M phase. During mitosis, cells curtail translation and accumulate phosphorylated eEF-2 (50). They also release nuclear proteins into the cytoplasm upon destruction of the nuclear membrane. Hence, conceptually, an interaction between prothymosin ␣ and eEF-2 could be envisaged. Nevertheless, we remained dubious. The proposed function required prothymosin ␣ to be active in the cytoplasm only at mitosis, whereas Wang et al. (34) and Tao et al. (36), assessing prothymosin ␣'s activity by observing its unstable glutamyl phosphates, found phosphate turnover throughout the cell cycle except during mitosis. If prothymosin ␣ were to function transiently in the cytoplasm, why was it continuously consuming high energy phosphate in the nucleus throughout most of the cell cycle? Could the known connection between prothymosin ␣ and cell proliferation be reconciled with prothymosin ␣ as an inhibitor of translation? We set out to evaluate the relationship between prothymosin ␣ and the phosphates of eEF-2 anew.
A series of suggestions has been made (37): that mitotic lysates have the ability to phosphorylate eEF-2 maximally, without the need for exogenous prothymosin ␣; that cell lysates become refractory and require exogenous prothymosin ␣ to facilitate the reaction at the termination of M phase when prothymosin ␣ is sequestered in the nucleus; and that lysates from synchronized cells exhibit intermediate, graded behavior as they progress through the cell cycle. Hence, prothymosin ␣ is thought to regulate eEF-2 phosphorylation and presumably influence translation. We do not wish to challenge these observations, but we disagree with their interpretation for two reasons.
First, we have shown that all lysates under discussion, here, contain prothymosin ␣. We found that prothymosin ␣ was recovered virtually quantitatively in sonicated lysates, and from the post-nuclear supernatant fluids of ruptured cells after Dounce homogenization and detergent lysis, methods that preserve the nucleus as a microscopically identifiable entity. Even when cells were permeabilized with digitonin, the washed cells lost virtually all of their prothymosin ␣ to the supernatant fluids with little or no apparent damage to the nuclei. We have also studied synchronized or mitotically arrested NIH3T3 cells extensively (36) and found quantitative discharge of prothymosin ␣ into the soluble fraction regardless of when in the cell cycle samples were prepared. Furthermore, several groups agree (2,51,52) that prothymosin ␣ will remain in extracts that have lost other small molecules during the process of desalting. Our point is that all of the cellular prothymosin ␣ will be recovered in the supernatant fluids of all samples. Therefore, we believe that prothymosin ␣ has no role to play. Since Vega et al. (37) achieved cell synchronization by serum starvation and since new proteins are synthesized during the first cell cycle of the recovery, there are a plethora of candidate proteins available to explain differences in eEF-2 phosphorylation.
Second, prothymosin ␣ from a quantitative point of view is unconvincing as an effector of eEF-2 kinase activity. In a rapidly growing myeloma cell, we have shown that prothymosin ␣ is about 0.02% of total protein (9). Thus, in a sonicated extract containing 20 g of protein, the maximum amount used by the Domínguez group (37) and here, there would be about 4 ng of prothymosin ␣. When dispersed into a volume of 50 l as required by the published procedure (37), the maximum prothymosin ␣ concentration would be ϳ0.08 mg/liter or 0.007 M. Yet, in Fig. 1B of Vega et al. (37), exogenous prothymosin ␣ had no effect on eEF-2 phosphorylation until the concentration approached 1 M. Hence, if synchronized cells contributed graded amounts of endogenous prothymosin ␣ to the assay and if the entire cellular complement of prothymosin ␣ were available in lysates of cells arrested in mitosis, the maximal concentration of 0.007 M prothymosin ␣ would still have been grossly suboptimal. The optimal dose of 8 M prothymosin ␣ used by these investigators to promote phosphorylation of eEF-2 in vitro exceeds the physiological amount by about 1000-fold.
Our study examined the phosphorylation of eEF-2 in crude cell lysates and in a reconstituted system consisting of purified eEF-2, recombinant enzyme, calcium, and calmodulin. In no case were we able to corroborate the previously published results (37). Because we had failed, we scrutinized every component of the system looking for contaminants that might either activate or inhibit the reaction. We showed, first, that sonicated lysates tested in a buffer optimized for eEF-2 kinase activity did not exceed the dynamic range of the assay; eEF-2 could have been phosphorylated further (Fig. 1). We also considered the lysates themselves. Lysates are a source of salt, phosphatases that attack phosphoproteins and nucleotides such as ATP, and a host of enzymes and substrates that compete for the reagents. They are poorly defined and difficult to reproduce, particularly when prepared by sonication. Although our results indicated that lysates differed considerably, both in their total protein and in the relative intensity of the radioactive phosphorylated proteins generated by the reaction, the phosphorylation of eEF-2 was not dependent on prothymosin ␣. Nor were substantive changes in the reaction or in prothymosin ␣'s effect on it introduced by replacing [ 32 P]ATP with [ 33 P]ATP, or by adopting conditions in which the sole source of the reagent was [ 32 P]ATP at a specific activity of 4500 Ci/mmol.
Since prothymosin ␣ itself is obviously a key component of the system, we used several versions: a native protein, a recombinant, tagged protein, and the tagged protein produced in transiently transfected COS cells expressing an ectopic prothymosin ␣ gene. In the last case, the protein was isolated using FIG. 6. Effect of prothymosin ␣ on the phosphorylation of eEF-2 in crude lysates in the presence of fluoride. Lysates containing 20 g of protein were obtained by Triton X-100 lysis and sonication and incubated at 37°C with [␥-32 P]ATP in the buffer used by Vega et al. (37) in the absence (-) and in the presence (P) of native prothymosin ␣. Samples, which were evaluated in 10% polyacrylamide SDS gels, are displayed in pairs as follows from left to right: radioactive products obtained with Triton-lysed cell extracts; the corresponding Coomassie Blue-stained gel; radioactive products obtained with sonicated lysates, the corresponding stained gel. Far right, the Coomassie-Blue stained electropherogram of the reconstituted system containing genuine eEF-2 (E) is displayed as a marker (lane marked C). The control and sonicated samples were analyzed in the same gel. the gentle methods of nickel chelate chromatography and column chromatography on Q2 resin without exposing it to denaturing agents such as perchloric acid, used by Vega et al. (37), phenol, used by both our laboratories (9), and reverse phase chromatography, known to inactivate many proteins. All of these preparations behaved similarly in our hands. Because none of them stimulated the phosphorylation of eEF-2, we believe that our result is correct. However, we do not think that the differences between our results and theirs reflect the integrity of prothymosin ␣; the protein is apparently unfolded (10) and consequently should be unaffected by the harsh conditions of purification. In contrast, the components of the solutions in which prothymosin ␣ was introduced into the reactions are highly suspect. It may be important to reiterate that prothymosin ␣ can be obtained as a homogeneous protein yielding a single band on a Coomassie Blue-stained gel while retaining large amounts of RNA, carbohydrate, and small molecules (9,33). Any of these might modulate the activity of eEF-2 kinase.
Finally, the buffer must be considered. When we were unsuccessful in reproducing the published data with a system optimized for eEF-2 kinase activity, we examined the reaction in the same system use by Vega et al. and found that fluoride, included at high concentration in the buffer, is inhibitory. Since the effect is nonspecific and mediated through complexes of small ions (44 -48), it was not surprising that the phosphorylation of eEF-2 by its kinase was affected.
We suspect that relief of fluoride inhibition probably played a part in the published results. In Fig. 1 of Vega et al. (37), which shows a gel of [ 32 P]proteins phosphorylated in the presence and absence of prothymosin ␣, all of the bands became more intense in the presence of prothymosin ␣, a result most consistent with a nonspecific effect. Furthermore, it is important to point out that eEF-2 is an abundant protein and that it is rapidly phosphorylated (39). Therefore, in short incubations with labeled ATP it is the most prominent labeled product. Then, when autoradiograms are not overexposed, eEF-2 will dominate the pattern regardless of whether additions to the system specifically affected eEF-2, or its kinase, or influenced phosphorylation reactions in general. The very prevalence of eEF-2 in lysates complicates the identification of specific effects. In contrast, using the reconstituted system, we had difficulty phosphorylating eEF-2 in 50 mM fluoride regardless of the presence or absence of prothymosin ␣ and found a severe increase in inhibition at doses of fluoride between 30 and 50 mM.
Our data strongly suggest that prothymosin ␣ does not modulate the phosphorylation of eEF-2. With the evidence accumulated to date, there is no reason to connect prothymosin ␣ to the phosphates of eEF-2 and every reason to seek its function elsewhere.