Phosphorylation of Specific Serine Residues in the PKR Activation Domain of PACT Is Essential for Its Ability to Mediate Apoptosis*

Activation of the latent protein kinase, PKR, by extracellular stresses and triggering of resultant cellular apoptosis are mediated by the protein, PACT, which itself gets phosphorylated in stressed cells. We have analyzed the underlying biochemical mechanism by carrying out alanine-scanning mutagenesis of the PKR activation domain of PACT. Among the indispensable residues identified were two serine residues, whose phosphorylation was essential for the cellular actions of PACT. Two-dimensional gel analysis, Western analysis using phosphoamino acid-specific antiserum, and in vivo 32P labeling of PACT demonstrated that constitutive phosphorylation of one of the two residues, Ser246, was required for stress-induced phosphorylation of the other, Ser287. Substitution of either of them by threonine or aspartic acid, but not alanine, was tolerated. Substitution of both residues with the phosphoserine mimetic, aspartic acid, produced a mutant PACT that, unlike the wild-type protein, caused PKR activation and apoptosis, even in unstressed cells. These results indicate that phosphorylation of specific serine residues in the activation domain of PACT is the major mode of transmission of cellular stress response to PKR.

exposed to a variety of stresses such as, withdrawal of growth factors or treatment with a low dose of actinomycin D, arsenite, thapsigargin, or peroxide (20,24,25), PACT or its murine homolog RAX, is phosphorylated and associates with PKR with increased affinity (24,25). In this study, we investigated further the mechanism of PACT-mediated activation of PKR in vivo. For this purpose, we have used an experimental system of expressing a variety of PACT mutants in human HT1080 cells, which express little endogenous PACT. They undergo apoptosis, when functional exogenous PACT is expressed and extracellular stress is applied to them. We identified two serine residues in PACT domain 3, whose phosphorylation was essential for the cellular actions of PACT. Our results indicate that constitutive phosphorylation of one of these residues was required for stress-induced phosphorylation of the other, leading to strong association of PACT with PKR and its activation.
Construction of PACT Mutants-The generation of PACT mammalian expression constructs was described previously (20). Overlap extension PCR was used to construct all PACT mutants (27). PACT⌬1 was used as a template to construct each mutant used for mammalian transfection, with the exception of S18A, S18D, S246A, and S287A, which used full-length PACT as the template. PCR fragments containing the desired PACT mutant were ligated into restriction enzyme-digested pcDNA3. A FLAG epitope tag was added at the N-terminal coding end of all PACT constructs.
Apoptosis Assays-TUNEL: HT1080 cells growing on glass coverslips in 6-well dishes were cotransfected with pcDNA3-FLAG PACT or pcDNA3-FLAG PACT mutant. At 6 h after transfection, the cells were stressed by treatment with 40 l of Lipofectamine 2000 and 50 ng/ml actinomycin D. Cells were fixed in 4% methanol-free formaldehyde 24 h after transfection. TdT-mediated dUTP nick-end labeling (TUNEL) assay using the Dead End Fluorometric TUNEL System (Promega) was performed using the manufacturer's protocol. After washing, cells were stained with primary anti-FLAG antibody and secondary anti-mouse IgG Texas Red conjugate (Molecular Probes) as described (20). The cells were mounted on glass slides in Vectashield with DAPI (4Ј-6Ј-diamidino-2-phenylindole) (Vector Laboratories), and examined under a fluorescence microscope. Quantitative Cell Survival Assay: HT1080 cells were cotransfected in 100-mm culture dishes with pcDNA3 vector or pcDNA3-PACT, and pEGFP-C1 in a 8:1 ratio using the Lipofectamine 2000 reagent (Invitrogen). At 6 h after transfection, the cells were stressed and allowed to undergo cell killing. At 48 h after transfection, the medium containing dead cells was removed, and the plate washed with phosphate-buffered saline. The population of cells remaining was collected, and 20,000 cells were monitored for GFP expression using a flow cytometer. The nonspecific background was 4 -7% from the loss of stressed GFP-expressing cells transfected with vector, and has been subtracted from all values presented in the figures and tables. Each value presented is an average of three independent experiments.
Western Blotting for Phosphorylated and Unphosphorylated PACT, PKR, and eIF2␣-HT1080 cells were transfected with FLAG-PACT or its mutants using the Lipofectamine 2000 reagent. At 23 h after transfection, cells were stressed for 1 h where indicated. Cells were lysed in RIPA buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1% Nonidet P-40, 0.2 mM phenylmethylsulfonyl fluoride, 100 units/ml aprotinin, 20% glycerol) on ice. The cell extract was used in Western blot analysis with anti-FLAG (M2) monoclonal antibody as described before (20). For the detection of phosphothreonine residues, the extract was used to immunoprecipitate FLAG-PACT or its mutant with anti-FLAG (M2) agarose. The agarose beads were washed six times with RIPA buffer, and the immunoprecipitates were analyzed by Western blotting with anti-FLAG or with a mixture of three monoclonal anti-phosphothreonine antibodies as described by the manufacturer (Biodesign International). For the detection of phosphorylated PKR or phosphorylated eIF2␣, at 47 h after transfection with PACT or mutant, cells were stressed for 1 h with 50 ng/ml actinomycin D when indicated. Cells were then lysed in RIPA buffer and extracts Western-blotted with anti-phospho-PKR, anti-PKR, anti-phospho-eIF2␣, anti-eIF2␣, or anti-FLAG (M2) antibody.
Two-dimensional Gel Analysis-HT1080 cells were transfected with 5 g of expression vector containing FLAG-PACT or its mutant using Fugene 6 reagent (Roche Applied Science). At 6-h post-transfection, the cells were washed with phosphate-buffered saline once and incubated with fresh complete medium for an additional 18 h. The cells were lysed with 1 ml of cell lysis buffer (20 mM Tris-HCl, pH 7.5, 50 mM KCl, 200 mM NaCl, 1% Triton 100, 1 mM EDTA, 100 units of aprotinin per ml, 0.2 mM phenylmethylsulfonyl fluoride, 20 mM NaF, 1 mM Na 3 VO 4 , 10 mM ␤-glycerophosphate), and PACT-immunoprecipitated using anti-FLAG M2 affinity gel. FLAG-PACT was eluted from the gel by adding 125 l of rehydration buffer (8 M urea, 4% CHAPS, 10 mM dithiothreitol, 0.5% IPG buffer 3/10). IPG strips (7 cm, pH 3-10; Amersham Biosciences) were rehydrated with rehydration buffer containing FLAG-PACT for 12 h before running isoelectric focusing (IEF). The IEF electrophoresis was performed with an IPGphor TM Isoelectric Focusing System (Amersham Biosciences) according to the instruction manual. The focusing time was 40,000 vhr. After focusing, the strips were laid on top of SDS-polyacrylamide gels for second-dimension separation. FLAG-PACT was examined by Western blotting using an anti-FLAG M2 monoclonal antibody. For alkaline phosphatase (calf intestine) (CIP, Invitrogen) treatment, after washing four times with cell lysis buffer, the immunoprecipitate was washed once with 1ϫ phosphatase buffer, divided into two 1.5 ml Eppendorf tubes, and incubated in 100 l of phosphatase buffer with or without 10 units of CIP at 37°C for 1 h. After the CIP treatment, the immunoprecipitate was washed twice with cell lysis buffer and subjected to IEF electrophoresis as indicated above.
In Vivo Phosphate Labeling-HT1080 cells were transfected in 100-mm culture dishes with 5 g of FLAG-tagged PACT or PACT mutant DNA. At 24 h after transfection, cells were passaged at a 1:10 dilution into fresh growth medium. At 48 h after transfection, media was replaced with complete medium containing 800 g/ml G418, and selection pressure applied for a minimum of 3 weeks. A pool of PACT-expressing cells was metabolically labeled with 1000 Ci/ml of [ 32 P]orthophosphate (PerkinElmer Life Sciences) for 2.5 h in phosphate-free medium. Cells were stressed for the last 30 min where indicated. Cells were washed with phosphate-buffered saline, flashfrozen on liquid nitrogen, and cell extracts prepared using modified RIPA buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 20 mM Na 4 P 2 O 7 , 2 mM Na 3 VO 4 , 0.1% SDS, 1% Triton-X 100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1ϫ protease inhibitor mixture (Sigma)). For the detection of phosphorylated PACT, the extract was used to immunoprecipitate FLAG-PACT or its mutant with anti-FLAG (M2) agarose. The agarose beads were washed six times with RIPA buffer, and the immunoprecipitates were analyzed by SDS-PAGE and autoradiography.
Coimmunoprecipitation Assay-HT1080 cells were transfected in 100-mm culture dishes with 10 g of total DNA (5 g of PKR (K296R) DNA and 5 g of FLAG-PACT or mutant DNA) using the Lipofectamine 2000 reagent (Invitrogen). At 23 h after transfection, cells were stressed. At 24 h after transfection, cells were lysed in immunoprecipitation (IP) buffer (20 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1% Triton X-100, 1 mM dithiothreitol, 100 units/ml aprotinin, 2 mM MgCl 2 , 20% glycerol). The cell extract was used to immunoprecipitate FLAG-PACT with anti-FLAG (M2) agarose as described (20), washing the beads six times with IP buffer. The immunoprecipitates were analyzed by Western blotting with anti-PKR polyclonal (Santa Cruz Biotechnology) and anti-FLAG monoclonal antibodies (Sigma) (26). (20) has established that domain 3 of PACT is essential for activating PKR (Fig. 1A). For exerting its action in vivo, this domain needs to be attached to another domain of PACT or a heterologous domain that interacts with PKR strongly. We wanted to identify the residues within domain 3 that are required for PACT activity. For this purpose, progressive deletions of ten residues were introduced from the N terminus or the C terminus of domain 3 (Fig. 1B). PACT proteins carrying these mutations were expressed in cells that were stressed by treatment with a low dose of actinomycin D, and apoptosis was measured by TUNEL assays. PACT expression and TUNEL positivity of each cell was determined by double immunofluorescence assay as described before (20,26). Deletion of ten residues from the C terminus of domain 3 (mutant C1) was tolerated, but removal of another ten residues (mutant C2) destroyed the activity. In contrast, even the first ten residues from the N terminus (mutant N1) could not be removed without losing activity (Fig. 1C). The wild-type and the mutant proteins were expressed at similar levels (Fig. 1D). Thus, residues 240 -295 of domain 3 were necessary for the apoptotic activity of PACT.

Essential Residues in PACT Domain 3-Our previous study
To identify the essential residues within the above region, we carried out alanine scanning mutagenesis using a quantitative cell survival assay. In this assay, a GFP expression vector was co-transfected with PACT or its mutants so that the transfected cells could be identified by GFP expression. After stressing by a very low dose of actinomycin D treatment, the wild-type PACT-expressing cells started undergoing apoptosis and massive cell death was obvious as dead cells floated off the monolayers. For quantifying the survival index of transfected cells expressing GFP and different PACT proteins, cells still attached to the plate were removed from each plate, and the same number of cells analyzed by FACS for GFP expression. In this quantitative assay, almost no GFP-expressing stressed cells survived when wild-type PACT was co-expressed ( Fig. 2A). The point mutants carrying substitution of a single residue with alanine had different properties: most of them were as active as the wild-type protein, whereas ten mutants lost their activity almost completely. The inactive mutants were Q243A, S246A, D260A, D262A, S265A, Q271A, S279A, S287A, G288A, and C291A. In addition, a few mutants had intermediate activities; these residues were located very close to the essential residues. The ten inactive mutant proteins were expressed at the same level as the wild type or an active mutant (244) protein (Fig. 2B). We confirmed that the active PACT mutant protein-expressing cells were undergoing apoptosis by performing TUNEL assays with selected mutants. Cells expressing the active L244A mutant protein were TUNEL-positive, whereas those expressing the S246A or the S287A proteins were TUNEL-negative (Fig. 2C).
Because we are interested in the mechanism of PACT activation by its stress-induced phosphorylation, we focused our attention to those residues in domain 3 that can potentially be phosphorylated and whose replacement by Ala produced inactive PACT mutants. There were four such mutants: Ser 246 , Ser 265 , Ser 279 , and Ser 287 ( Fig. 2A). In an in vitro binding assay for domain 3/PKR interaction, among the four mutants, mutant 287 bound PKR most strongly and mutant 246 did not bind at all (data not shown). We chose those two mutants for further studies, because they represented two different mechanistic classes. Like wild-type PACT, neither mutant caused PKR activation, as measured by its autophosphorylation, when expressed in HT1080 cells (Fig. 3). However, as expected, upon application of stress, there was strong phosphorylation of PKR in cells expressing wild type, but not mutant, PACT.
Stress-activated Phosphorylation of PACT and Its Mutants-To investigate further the role of Ser 246 and Ser 287 in PACT function, we wondered whether these two residues are the targets of phosphorylation by cellular protein kinases. This notion was supported by the properties of two other mutants, S246T and S287T (Table 1). Although Ala was not tolerated in these sites, Thr was tolerated, partially for 246  and completely for 287, suggesting that Thr, another phosphorylatable residue, can substitute for Ser at these sites. Further support came from the observation that substitution of either Ser residue with the phosphoserine mimetic Asp was fully tolerated (Table 1).
Evidence for phosphorylation of PACT on these two Ser residues was provided by biochemical analyses of wild-type PACT and mutant proteins purified from cell extracts. Two dimensional gel analyses followed by Western blotting showed that bulk of the wild-type protein (Spot 1, Fig. 4A) had a higher isoelectric point than that of the minor spot (Spot 2). Applying cellular stress caused the appearance of a new spot (Spot 3) with the lowest isoelectric point, suggesting that Spot 2 and Spot 3 could be a partially and the fully phosphorylated forms of PACT respectively. As expected, treatment with protein phosphatase caused Spots 2 and 3 to collapse into Spot 1 (Fig. 4B). The S246A mutant protein did not produce any phosphorylated species either before or after stress (Fig. 4C), whereas the S287A mutant protein contained Spot 2, but not Spot 3 even after stress (Fig. 4D). These results indicate that Ser 246 is a target of phosphorylation even in unstressed cells whereas Ser 287 gets phosphorylated after the cells are stressed. Moreover, it appears that Ser 287 cannot be phosphorylated without Ser 246 phosphorylation. Neither the constitutive nor the stress-induced phosphorylation of PACT was mediated by PKR, because both phosphorylations were observed in Pkr Ϫ/Ϫ MEF cells (Fig. 4E).
Phosphorylation of residues 246 and 287 was tested by a different approach in the experiment shown in Fig. 5; we used the S246T and the S287T mutants and a phosphothreonine-specific antibody for this purpose. Because these two mutants were functionally active in our cell survival assay (Table 1), we anticipated that the introduced threonine residues would be phosphorylated in vivo. As expected, the wild type protein was not phosphorylated in Thr residues even after stressing the cells (Fig. 5, lanes 6 and 7). The S287T mutant was Thr-phosphorylated only after stressing (Fig. 5, lanes 4 and 5) whereas the S246T mutant was constituitively Thr-phosphorylated before and after stressing (Fig. 5, lanes 2 and 3). We ensured that the transfected PACT proteins were expressed to similar levels (lower panel, Fig. 5).
Finally, stress-induced PACT phosphorylation was confirmed by its in vivo labeling with 32 P. Cell lines expressing wild type or mutant PACT were generated; wild-type and the S287A mutant proteins were expressed at similar levels whereas the S246A mutant protein was expressed more highly, probably    NOVEMBER 17, 2006 • VOLUME 281 • NUMBER 46 because the latter mutant protein was constitutively inert, and cells could tolerate its higher expression (bottom panel, Fig. 6). There was no detectable labeling of wild-type PACT or its mutants in cells that had not been stressed (top panel, Fig. 6), indicating that we could detect only stress-inducible phosphorylation under our labeling conditions. When stress was applied, strong labeling of PACT was observed only in cells expressing the wild-type protein, not the mutants (middle panel, Fig. 6). Thus, both Ser 246 and Ser 287 were required for stress-inducible phosphorylation of PACT. From the above observations, we concluded that phosphorylation of both Ser 246 and Ser 287 of PACT, one constitutive and the other stress-inducible, is necessary for its ability to activate PKR in vivo. Moreover, constitutive phosphorylation of Ser 246 was a prerequisite for inducible phosphorylation of Ser 287 .

PACT Action Regulated by Serine Phosphorylation
Properties of Asp Substitution Mutants-To further solidify the above conclusions, we generated several PACT mutants, in which Ser 246 and Ser 287 were replaced with Ala (AA) or the phosphoserine mimetic Asp (DD). As expected, the AA mutant was completely inactive, whereas the DD mutant was as active as wild-type PACT in stressed cells ( Table 2). The AD mutant was also inactive, indicating the absolute need for phosphorylation of the 246 residue. Finally, the DD mutant, unlike the wild-type protein, was highly active even in unstressed cells causing death of 86% of cells expressing it. These results were confirmed by TUNEL assay as well ( Fig. 7 and data not shown).
To probe the cause of the observed apoptosis in cells expressing the DD mutant, we examined the extent of PKR activation by measuring the levels of phosphorylation of eIF2␣, a substrate of PKR. PKR activation in vivo, as measured by eIF2␣ phosphorylation, was equally strong, before and after stress, in cells expressing the DD mutant (lanes 6 and 7, Fig. 8A). There was little eIF2␣ phosphorylation in cells expressing the AA mutant even after the application of stress (lanes 4 and 5, Fig. 8A), and the wild-type protein caused eIF2␣ phosphorylation only after the cells were stressed (lanes 2 and 3, Fig. 8A). To seek the reason behind the above observations, we wondered whether the enhanced eIF2␣ phosphorylation was caused by a stronger association between PACT and PKR. Indeed, more PKR coimmunoprecipitated with wild-type PACT after the cells were stressed (lanes 3 and 4, Fig. 8B). In contrast, when the AA mutant of PACT was expressed, only a small amount of PKR was coimmunoprecipitated even after stress ( lanes 5 and 6, Fig.  8B). Strikingly, the DD mutant of PACT coimmunoprecipitated PKR more efficiently, even without stress (lanes 7 and 8, Fig.  8B). These results demonstrated that phosphorylated wild-type PACT bound PKR more strongly in vivo, as did the DD mutant.

DISCUSSION
We have used a combination of genetic and biochemical experiments to explore the mechanism by which various extracellular stresses use PACT to activate PKR and cause apoptosis (20,24,25). We used a human cell line that did not express detectable PACT and consequently was a good recipient for testing various mutant PACT proteins in vivo expressed by transfection. Because the cell survival assay was convenient and quantitative, it was used for screening mutants of domain 3.  Three fulllength mutants were tested for inducing apoptosis with or without stress. Methods were as described in the legend to Fig. 1C.

TABLE 2 Survival of cells expressing PACT mutants, with and without stress
A quantitative cell survival assay was used to measure the survival of full-length PACT-or PACT mutant-transfected HT1080 cells by cotransfecting with a GFP expression vector (8:1) so that the transfected cells could be identified by GFP expression. Where indicated, cells were stressed. PACT/mutant-expressing cells that underwent apoptosis floated off the monolayers. Cells still attached to the plate were removed from each plate, and the same number of cells analyzed by FACS for GFP expression as a measure of cell survival. This assay was in perfect concordance with the apoptosis assay (Fig. 2). Out of the fifty-six residues needed for activity ( Fig. 1), replacement of any of ten specific residues with alanine almost completely destroyed the activity. Further structural studies will be needed for understanding why alanine substitution of these ten residues disrupted the function of domain 3. It is conceivable that the mutated Asp and Gln residues form salt bridges to stabilize its structure and Cys 291 may be involved in a disulfide linkage. In contrast, the four serine residues may be structurally required or be the targets of phosphorylation. We explored the latter possibility because of the information in the literature that PACT gets phosphorylated in stressed cells. Our results indicate that Ser 246 and Ser 287 are indeed the targets of phosphorylation. While our work was in progress, Bennett et al. (28) reported that the S18A mutant of RAX (murine homolog of PACT) was inactive in their assay system of PKR activation in vivo. Moreover, the mutant protein could interact with PKR as effectively as the wild-type protein and functioned as a dominant negative inhibitor of the latter protein. Most importantly, because the mutant protein was not labeled with 32 P in stressed cells, the authors concluded that Ser 18 was the essential target of phosphorylation for the activation of RAX, although more direct evidence for the conclusion was not provided. We examined whether similar conclusions were true for PACT and found no evidence for either the necessity of Ser 18 or its phosphorylation. The S18A mutant killed cells as effectively as the wild-type pro-tein in stressed cells (Table 1) and the S18D mutant did not cause apoptosis in unstressed cells (Fig. 7). Moreover, the S246A mutant was completely unphosphorylated, even in stressed cells, as revealed by no change in its isoelectric point (Fig. 4C). Similarly, neither S246A nor the S287A mutant was 32 P-labeled in stressed cells, although both contained unmutated S18 (Fig. 6). These results clearly showed that, in our system, Ser 18 was not essential for stress-induced phosphorylation of PACT and its resultant ability to activate PKR. The basis for the apparent discrepancy between the properties of PACT and RAX remains to be explored further.

Mutations
Our results indicate that Ser 246 and Ser 287 are the only two targets of PACT phosphorylation in vivo, because the S246A mutant protein migrated to the same place as dephosphorylated wild-type PACT (Spot 1, Fig. 4), indicating that the mutant protein isolated from stressed or unstressed cells was unphosphorylated. The two-dimensional gel analysis of wild-type and mutant PACT proteins present in stressed and unstressed cells revealed that Ser 246 was partially phosphorylated even in unstressed cells, suggesting that a constitutively active kinase phosphorylates Ser 246 . In contrast, Ser 287 was phosphorylated only after stress, suggesting that it is the target of a stress-activated protein kinase pathway. Phosphorylation of the S246T and the S287T mutants in threonine residues (Fig. 5) unequivocally proved that these two residues are indeed the targets of constitutive and induced phosphorylation respectively. In vivo radiolabeling of the stress-activated phosphoserine residue further confirmed the above conclusion (Fig. 6). The properties of the mutants uncovered a surprising aspect of phosphorylation of these two serine residues: not only was it sequential, but the order was also fixed. It appears that Ser 287 phosphorylation required prior phosphorylation of Ser 246 , because the S246A mutant produced only unphosphorylated protein even after stress, whereas the S287A mutant produced a monophosphorylated species, before and after stress (Fig. 4). However, the 246A/287D mutant was as ineffective in cell killing as the 246A/ 287A mutant (Table 2) indicating that the phosphoserine at 246 was needed for another purpose in addition to allowing the phosphorylation of Ser 287 . Substitution of the two other essential Ser residues, Ser 265 and Ser 279 ( Fig. 2A), with Asp did not produce constitutively active mutants (data not shown), indicating that their role is structural and they are not targets of phosphorylation.
Until the structure of PACT, or at least its domain 3, is solved, we can only speculate about the reasons for the observed need of most of the ten essential residues in domain 3, as identified by alanine-scanning mutagenesis. However, the results presented here provide strong evidence for the roles of Ser 246 and Ser 287 as targets of necessary phosphorylation. Our observation that residues in domain 3, but not elsewhere, regulate PACT activation, is consistent with our earlier observation that domains 1 and 2 of PACT could be functionally substituted by the corresponding dimerization domains of PKR (20). The fusion protein could induce apoptosis only in stressed cells, indicating that the stress-sensing residues were located in domain 3. Our data support a two-step phosphorylation model of PACT, one constitutive and the other stress-activated. Future biochemical analyses will be needed for identifying the respective protein kinases that FIGURE 8. eIF2␣ phosphorylation and PACT/PKR association in stressed and unstressed cells expressing PACT mutants. A, eIF2␣ phosphorylation in stressed and unstressed cells expressing full length PACT mutants. At 47 h after transfection, where indicated, cells were stressed for 1 h, and cell extracts were made. Phospho-eIF2␣ levels were analyzed by Western blotting with an antibody for phospho-eIF2␣ and total eIF2␣ levels were analyzed with an antibody for eIF2␣. B, coimmunoprecipitations of PKR with full-length wild-type PACT or its mutants were done as described under "Experimental Procedures." Lanes 1 and 2, pcDNA3 vector; lanes 3 and 4, wild-type PACT; lanes 5 and 6, S246A/S287A; lanes 7 and 8, S246D/S287D. Cells in even lanes were stressed. NOVEMBER 17, 2006 • VOLUME 281 • NUMBER 46 mediate the phosphorylation of Ser 246 and Ser 287 . Our results with various mutants, especially the DD mutant, strongly indicate that the phosphorylation of these two serine residues is both necessary and sufficient for PACT activation. Results in Fig. 8B presented a biochemical basis for understanding why phosphorylated PACT could activate PKR in vivo: it associated with PKR more efficiently. This observation is in line with that made by others using wild-type PACT (25). In contrast, the S18A inactive mutant of RAX, associated with PKR as strongly as the wild-type protein (28) indicating that it does not regulate PKR activation, a prediction supported by our experiments. However, it remains possible that different kinds of stresses can activate PACT through phosphorylation of different residues, a paradigm established for the activation of transcription factors such as NFB, IRF-3, or p53.