Interferon-α-induced Expression of Phospholipid Scramblase 1 through STAT1 Requires the Sequential Activation of Protein Kinase Cδ and JNK*

Phospholipid scramblase 1 (PLSCR1), a calcium-binding protein that either inserts into the plasma membrane or binds to genomic DNA in the nucleus, has been shown to contribute to the cell proliferation, differentiation, and apoptosis as well as antiviral activity of interferon (IFN). The expression of PLSCR1 protein is also known to be markedly increased in response to IFN and to some differentiation inducing agents such as all-trans retinoic acid, but the precise mechanisms of this response remain to be investigated. In this study, we show that the protein kinase Cδ (PKCδ)-specific inhibitor rottlerin and the dominant negative mutant of PKCδ significantly antagonized IFN-induced PLSCR1 expression. The influence of PKCδ on IFN-mediated induction of PLSCR1 was dependent upon the phosphorylation of STAT1 at Ser-727. Furthermore, PKCδ-mediated activation of STAT1 required the activation of JNK, as the inhibition of JNK activity by its specific inhibitor or transfection of its dominant negative mutant suppressed both serine phosphorylation of STAT1 and PLSCR1 expression but not the activation of PKCδ. In conclusion, our results suggest that the induction of PLSCR1 transcription through STAT1 depends upon sequential activation of PKCδ and JNK.

sine phosphorylation of PLSCR1 by c-Src occurs in response to epidermal growth factor, resulting in the association of phosphorylated PLSCR1 with Shc and the activated epidermal growth factor receptor complex (6). PLSCR1 was also reported to be a substrate of protein kinase C␦ (PKC␦), although this could not be confirmed in our recent experiments with recombinant PKC␦ and apoptosis induction by Fas ligation in Jurkat cells (7). When it fails to be palmitoylated, PLSCR1 can be imported into the nucleus where it binds to genomic DNA, suggesting a potential role for this protein in gene transcription (8,9).
These biochemical studies strongly indicate that PLSCR1 exerts wide biological effects. Whereas the role of PLSCR1 in remodeling plasma membrane phospholipids and in cell surface exposure of phosphatidylserine, which was originally identified (10,11), remains a topic of great controversy, there is increasing evidence to suggest a direct role of this protein in regulating cell proliferation and terminal differentiation. As an example, proliferation and terminal differentiation of myeloid precursor cells in response to selective growth factors are impaired in PLSCR1 Ϫ/Ϫ mice (12). Furthermore, the suppression of PLSCR1 expression by small interfering RNA and antisense RNA inhibits the all-trans-retinoic acid (ATRA) and/or phorbol 12-myristate 13-acetate (PMA)-induced leukemic cell differentiation (7,13).
However, the mechanisms of regulation of PLSCR1 expression remain largely unknown thus far. Recently, we reported that PKC␦ mediates ATRA and PMA-induced PLSCR1 expression (7). PLSCR1 has also been shown to be among the most potently activated of the interferon (IFNs)-stimulated genes, suggesting that it possibly participates in cellular response to this important cytokine (14,15). As recently reported, gene deletion of PLSCR1 or suppression of PLSCR1 expression by small interfering RNA in wild type cells was found to markedly attenuate the antiviral activity of IFNs, which appeared to be related to diminished expression of a subset of IFN-stimulated genes with known antiviral activities in cells failing to up-regulate PLSCR1 (16). However, the mechanism by which IFN induces PLSCR1 expression also remains unresolved. Because IFNs was also shown to activate PKC␦ (17), in the present study we investigated whether and how activated PKC␦ contributes to the transcriptional activation of PLSCR1 by IFN.
Plasmids and Transfection -Plasmid pcDNA3-PKC-DN␦ carrying the dominant negative (DN) fragment of PKC␦ and its empty vector pcDNA3 were generous gifts from Dr. Jae-Won Soh (Inha University, Incheon, Korea) (19). pcDNA3-DN-JNK1 was kindly provided by Dr. Young-Mi Ham (Seoul National University, Seoul, Korea) (20). pRc/ CMV-STAT1 and pRc/CMV-STAT1-S727A were generously provided by Dr. James Darnell (Rockefeller University, New York) (21). These plasmids were transfected into H1080 and U3A cells using Polyfect FIGURE 1. The PKC␦ inhibitor rottlerin suppresses IFN induction of PLSCR1. A, U937 cells were treated with IFN␣2a (3000 IU/ml) for various times as indicated. PLSCR1 mRNA and protein were detected, respectively, by real-time quantitative RT-PCR (top) and Western blot (bottom) with ␤-actin as internal controls. Increased -fold (means Ϯ S.D.) for PLSCR1 mRNA in three repeated experiments with each triplicates and for PLSCR1 protein in three experiments are expressed. B and C, U937 cells were treated with 3000 IU/ml IFN␣2a for various times as indicated (B) or with 3000 IU/ml IFN␣2a in the presence or absence of rottlerin (10 M) for 4 h (C), and phosphorylation of PKC␦ on Ser-643 was detected by Western blot. Total PKC␦ was used as an equal loading control. The -fold change (means Ϯ S.D. from three independent tests) for phosphorylated PKC␦/total PKC␦ against untreated cells is shown. *, p ϭ 0.013, compared with column 1; &, p ϭ 0.025, compared with column 3; #, p ϭ 0.071, compared with column 1. D, PKC␦ activity, in the same samples as shown in C, was measured in an in vitro kinase assay using histone H1 as substrates. U937 cells treated with 20 nM PMA for 1 h as a positive control. The -fold change against control cells is shown as means Ϯ S.D. from three independent tests. *, p ϭ 0.040, compared with column 2; &, p ϭ 0.038, compared with column 4; #, p ϭ 0.048, compared with column 2. E, U937 cells were treated with IFN␣2a (3000 IU/ml) for various times as shown. Then cell lysates were immunoprecipitated with anti-PKC␦ antibody. The immunoprecipitates were analyzed by Western blots respectively against the indicated antibodies. Input came from untreated U937 cells. transfection reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. Expression of transfected cDNAs was confirmed by Western blotting.
Real-time Quantitative RT-PCR for PLSCR1 mRNA-Total RNA was isolated by Trizol kit (Invitrogen) and treated with DNase (Promega, Madison, WI). Complementary DNA was synthesized by using the cDNA synthesis kit according to manufacturer's instructions (Applied Biosystems, Foster City, CA), and fluorescence real-time RT-PCR was performed as we described previously (7) with specific primers for PLSCR1 (sense strand, 5Ј-CTGACTTCTGAGAAGGTTGC-3Ј; antisense strand, 5Ј-GAATGCTGTCGGTGGATACTG-3Ј) and for ␤-actin (sense strand, 5Ј-CATCCTCACCCTGAAGTACCC-3Ј; antisense strand, 5Ј-AGCCTGGATAGCAACGTACATG-3Ј). The -fold change is shown as means Ϯ S.D. in three independent samples with the same treatment.
Luciferase Reporter Assays -pRL-SV40 vector and human PLSCR1 promotor-primed (HPPP)-luciferase plasmid, which was a pGL3-basicluciferase reporter vector (Promega) cloned with a 4.18-kb DNA fragment consisting of the 5Ј flanking region (Ϫ1 to Ϫ4120) and the first 60 bp of the first exon of the human PLSCR1 gene (GenBank TM AF153715)    DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 as described previously (14), were transfected into HT1080 cells with or without dominant negative JNK1 (DN-JNK1) or dominant negative PKC␦ (DN-␦) using the Polyfect reagent (Qiagen) following the manufacturer's protocol. Twenty-four hours after transfection, triplicate cultures were treated with IFN␣2a in the presence or absence of rottlerin (10 M) or Gö6976 (5 M) for 18 h. Luciferase activity was measured using a Dual-Luciferase reporter assay system (Promega) following the manufacturer's protocol. The measured luciferase activity was normalized against pRL-SV40 Renilla luciferase activity for each sample, and luciferase activity was expressed as -fold over empty pGL3 vector luciferase activity normalized by pRL-SV40 Renilla activity.

Signaling Pathway for Regulation of PLSCR1 Expression
PKC␦ Activity Assays-U937 cells were lysed in phosphorylation lysis buffer (1% Triton X-100, 150 mM NaCl, 200 M sodium orthovanadate, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, 1 mM EDTA, 50 mM Hepes, 1.5 mM magnesium chloride, 10% glycerol, and protease inhibitor mixture). Cell lysates were immunoprecipitated with anti-PKC␦ antibody, and immunoprecipitates were washed three times with phosphorylation lysis buffer and two times with kinase buffer (25 mM Tris-HCl (pH 7.4), 5 mM MgCl 2 , 0.5 mM EDTA, 1 mM dithiothreitol, 20 g of phosphatidylserine, and 20 M ATP) and resuspended in 30 l of kinase buffer containing 5 g of histone H1 as an exogenous substrate, to which 1-2 Ci of [␥-32 P]ATP was added. The reaction was incubated for 15-30 min at room temperature and terminated by the addition of 2ϫ SDS sample buffer. Proteins were analyzed by SDS-PAGE, and phosphorylated histone H1 was detected by autoradiography (22).
Western Blot-Cells were harvested and lysed with ice-cold phosphorylation lysis buffer as described above and an equal volume of 2ϫSDS sample buffer. Cell lysates were loaded on 10% SDS-polyacrylamide gel and electrophoretically transferred to nitrocellulose membrane (Amersham Biosciences). After being blocked with 5% nonfat milk in Trisbuffered saline, the membranes were incubated with monoclonal antihuman PLSCR1 4D2 antibody (14) and other indicated antibodies followed by horseradish peroxidase-linked secondary antibodies. Detection was performed by chemiluminescence phototope-horseradish peroxidase kit according to the manufacturer's instructions (Cell Signaling). Blots were stripped and reprobed with mouse monoclonal anti-␤-actin antibody to ascertain equal loading of protein. As necessary, the signal intensities of proteins tested were normalized against the indicated internal control using a densitometer (SmartView, version 5.0, software from Furi, Shanghai, China), and the -fold change was expressed compared with untreated cells.
Statistical Analysis -Student's t test was used to compare the difference between two different groups. A value of p Ͻ 0.05 was considered to be statistically significant.

RESULTS
As reported previously (14,15), IFN␣2a highly induced PLSCR1 mRNA and protein in leukemic U937 cells as determined by real-time quantitative RT-PCR and Western blotting, respectively (Fig. 1A). Prior to PLSCR1 expression, IFN␣2a treatment rapidly and time-dependently resulted in an increase in PKC␦ phosphorylation in intact cells. As shown in Fig. 1B, increased phosphorylated PKC␦ began to appear at 5 min after IFN treatment and then became more significant. To determine the possible involvement of PKC␦ in IFN induction of PLSCR1, leukemic U937 cells were treated with IFN-␣2a in the absence and presence of the PKC␦ inhibitor, rottlerin. Pretreatment with rottlerin for 1 h effectively inhibited IFN activation of PKC␦ as evidenced by PKC␦ phosphorylation (Fig. 1C) and the ability of immunoprecipitated PKC␦ to phosphorylate histone H1 (Fig. 1D). It was noteworthy that in addition to PKC␦ several other kinases, including p38, ERK, and Cdc2/ Cdk1, could not be detected in such immunoprecipitates (Fig. 1E), indicating the specificity for in vitro kinase assays of PKC␦. More intriguingly, pretreatment with rottlerin prevented IFN induction of PLSCR1 HT1080 cells were preincubated with or without rottlerin for 1 h (left) or transfected with PKC-DN␦ for 24 h (right). Cells were subsequently incubated with 3000 IU/ml IFN␣2a for 2 h, and phosphoserine 727 STAT1 (Ser-STAT1) and total STAT1 were detected with ␤-actin as a control. The ⌬ symbol denotes the nonspecific band. The altered -fold of Ser-STAT1/STAT1 against untreated cells is shown as means Ϯ S.D. from three independent experiments. *, p Ͻ 0.05, compared with column 1; #, p Ͻ 0.05, compared with column 3; &, p Ͼ 0.05, compared with column 1. FIGURE 6. Kinetics of the phosphorylation of ERK, p38, JNK, and compared with the induction of PLSCR1 in response to IFN. U937 cells were incubated with IFN␣2a at 3000 IU/ml for different times as indicated. Total cell lysates were analyzed by Western blot with specific antibodies as shown. Levels of total ERK, p38, JNK, and c-Jun are shown beneath the phosphorylated forms of these proteins. mRNA and protein (Fig. 1F, left panel). In contrast, Gö6976, an inhibitor of conventional PKCs, was ineffective to IFN-induced PLSCR1 expression (Fig. 1F, right panel).
As reported previously, rottlerin directly uncouples mitochondrial respiration from oxidative phosphorylation and thus reduces cellular ATP levels that could block any number of ATP-dependent processes (23). On the other hand, rottlerin has also been shown to inhibit other kinases (24,25). Therefore, the data gathered from using rottlerin should be evaluated cautiously. Considering this, we also tested the effect of a dominant negative mutant of PKC␦ (DN␦), which selectively inhibited action of PKC␦ (19). For the convenience of transfection, the human fibrosarcoma cell line HT1080 was used. First, a plasmid containing the PLSCR1 promoter linked to luciferase cDNA (HPPP-luc) or the empty vector was transfected into HT1080 cells. As seen in Fig. 2A, IFN treatment resulted in an increase in the promoter activity, which was significantly suppressed by rottlerin but not by Gö6976 (data not shown). In agreement with the effects of rottlerin, transient transfection of DN␦ inhibited IFN induction of PLSCR1 mRNA/protein (Fig. 2B) and the PLSCR1 promoter-reporter construct (Fig. 2C). All of these findings indicate that IFN activation of PKC␦ precedes to and contributes to the subsequent induction of PLSCR1. Additionally, rottlerin and DN␦ also slightly abrogated the basal level of PLSCR protein and its promoter activity, suggesting that basal activity of PKC␦ possibly contributes to the constitutive expression of PLSCR1.
Because PKC␦ is known to phosphorylate STAT1 on Ser-727 in response to IFN, STAT1, Ser-727 phosphorylation, and PLSCR1 induction were monitored in HT1080 in comparison with its derivative, the STAT1-null U3A cell line (Fig. 3A) (18). In response to IFN, STAT1 Ser-727 phosphorylation increased at 30 min of treatment (lower, Fig.  3B), which appeared after activation of PKC␦ (Fig. 1B), whereas the induction of PLSCR1 mRNA and protein was not seen until after 3 and 6 h of treatment, respectively (Fig. 3B). Similar results were obtained in U937 cells (data not shown). Consistent with our previous report (14), however, PLSCR1 was not induced by IFN in the STAT1-deficient U3A cells (Fig. 3A).
We further determined the role of STAT1 Ser-727 phosphorylation in IFN-induced PLSCR1 expression. For this purpose, wild type STAT1 and mutant STAT1-S727A were transfected into U3A cells. Western blots showed the effectiveness of transfection of wild type STAT1 and mutant STAT1-S727A, the former but not the latter being phosphorylated on Ser-727 (Fig. 4). The results showed that ectopic expression of STAT1 but not STAT1-S727A increased PLSCR1 expression in U3A cells (Fig. 4).
In the next phase of analysis, we extended our investigation to the potential relationship between PKC␦ and STAT1 in the induction of PLSCR1. Toward this end, HT1080 cells were treated with IFN␣2a in combination with either rottlerin or DN␦ transfection (Fig. 5). The results showed that both rottlerin and DN␦ transfection effectively inhibited IFN␣2a-stimulated phosphorylation on STAT1 at Ser-727, suggesting that PKC␦ contributes, directly or indirectly, to IFN␣2ainduced STAT1 phosphorylation.
Previous studies (26,27) showed that IFN treatment leads to activation of several serine-threonine kinases including mitogen-activated protein kinases (MAPKs) besides PKC␦, although the mechanism(s) is incompletely understood. Indeed, several members of the MAPK family including ERK1/2, p38 and JNK, as well as the JNK substrate transcription factor c-Jun, were rapidly but transiently phosphorylated upon IFN␣2a treatment (Fig. 6). Thus, we sought to determine whether any of these MAPKs contribute to PLSCR1 expression in response to IFN␣2a treatment. U937 cells were pretreated for 1 h with MAPK inhibitors U0126 (28), SB203580 (29), and SP600125 (30). The results showed that SP600125, which inhibited phosphorylation of c-Jun by JNK, blocked IFN␣2a stimulation of STAT1 Ser-727 phosphorylation and PLSCR1 expression (Fig. 7A), but U0126 and SB203580 failed to do so. These two inhibitors diminished phosphorylation of ERK1/2 and p38, respectively, after IFN␣2a treatment (Fig. 7, B and C). These results indicate that JNK, but not ERK1/2 or p38, contributes to STAT1 Ser-727 phosphorylation and PLSCR1 expression induced by IFN␣2a. Consistent with this fact, DN-JNK, a dominant negative mutant JNK, also reduced phosphorylation of STAT1 Ser-727 induced by IFN␣2a and blocked IFN- ␣2a-induced expression of both the PLSCR1 promoter construct and endogenous PLSCR1 in HT1080 cells (Fig. 8).
Finally, we sought to determine whether the effects of PKC␦ and JNK on STAT1 Ser-727 phosphorylation are independent or in the same signaling pathway. Here we found that inhibition of PKC␦ by rottlerin reduced IFN␣2a-induced activation of JNK in U937 cells, as determined by reduced phosphorylation of c-Jun. However, inhibition of JNK with SP600125 had no effect on the IFN stimulation of PKC␦ activation (Fig.  9). In agreement with the effects of these inhibitors, transfection of DN␦ could inhibit the phosphorylation of JNK and c-Jun, whereas DN-JNK failed to affect the activation of PKC␦ in HT1080 cells (Fig. 10). These results indicate that PKC␦ is upstream of and required for JNK activation and that both kinases are involved in IFN signaling to the PLSCR1 gene.

DISCUSSION
Recently we reported that the activation of PKC␦ by phosphorylation mediates ATRA and PMA-induced PLSCR1 expression in leukemic cells (7). In this work, we continued to show that the inhibition of PKC␦ activation by the chemical inhibitor rottlerin and transfection of the dominant negative mutant of PKC␦ significantly inhibited IFN␣2a-induced expression of PLSCR1 mRNA and protein, whereas such inhibition was not observed for Gö6976, an inhibitor of the conventional PKC. These data, combined with our previous report demonstrating that transfection of the catalytic fragment of PKC␦ directly up-regulates PLSCR1 expression (7), strongly suggest that PKC␦ is a common mediator for PLSCR1 expression induced by IFN␣2a, ATRA, or PMA. More recently, PLSCR1 expression has been shown to mediate the antiviral activity of IFNs (16). Our previous studies also showed that induced expression of PLSCR1 contributes to ATRA/PMA-induced leukemic cell differentiation (7). Furthermore, IFN␣2a and ATRA act synergistically in the induction of cell differentiation and growth inhibition (22). Therefore, it is of interest that IFN␣2a and ATRA both up-regulate PLSCR1 expression through PKC␦.
It is well known that IFN-␣/␤ binding to the type I IFN receptor initiates a tyrosine phosphorylation cascade involving JAK kinases (JAK1 and TYK2), resulting in the heterodimerization of tyrosine-phosphorylated STAT1 and 2. These phosphorylated STAT proteins are transported to the nucleus and, in a complex with IRF9, bind to IFNstimulated response elements and activate the transcription of many  different genes (31)(32)(33). However, an additional phosphorylation event on STAT1 Ser-727 is required for a full IFN-stimulated transcriptional response (34). Previously, IFN-␣2a was reported to activate expression of the human PLSCR1 gene at the transcriptional level through a single IFN-stimulated response element located in the first exon (14). Accordingly, IFN-␣2a failed to induce PLSCR1 in STAT1-negative U3A cells.
Here we showed that complementation with wild type STAT1, but not STAT1 mutant S727A, almost completely restored IFN-␣2a-induced PLSCR1 expression. These findings support an essential role for Ser-727 of STAT1 in IFN-␣2a-induced PLSCR1 expression. Prior studies demonstrating that PKC␦ can phosphorylate STAT1 on Ser-727 indicated that a role for PKC␦ in PLSCR1 expression is likely (35). It has been proposed that PKC␦ is activated during engagement of the Type I IFN receptor and that such an activation of PKC␦ appears critical for phosphorylation of STAT1 on Ser-727 (17). DeVries et al. (36) also showed that PKC␦ activates STAT1 through phosphorylation of STAT1 on Ser-727 in etoposide-induced apoptosis. We have shown here that the PKC␦ inhibitor rottlerin or a dominant negative mutant of PKC␦ significantly antagonized IFN␣2a-induced STAT1 phosphorylation on Ser-727. Therefore, our findings demonstrate a role for PKC␦ in the Ser-727 phosphorylation of STAT1 required for IFN induction of PLSCR1.
How PKC␦ activation leads to Ser-727 phosphorylation of STAT1 is still unresolved. IFN␣ treatment results in phosphorylation of many different signaling proteins, including insulin receptor substrate (IRS)phosphatidylinositol 3-kinase (37,38), Crk (39), and MAPKs (40,41). A recent report demonstrated that genotoxic agents such as arabinofuranosylcytosine-induced JNK activation is attenuated by treatment with rottlerin, expression of a kinase-inactive PKC␦ mutant, and downregulation of PKC␦ by small interfering RNA, proposing that PKC␦ is required, in part, for induction of JNK (42). We have demonstrated here that a JNK inhibitor, but not inhibitors of ERK1/2 and p38, blocked both STAT1 serine727 phosphorylation and PLSCR1 expression following IFN treatment. More importantly, transfection of a dominant negative mutant of JNK also reduced IFN-initiated Ser-727 phosphorylation of STAT1 and PLSCR1 expression. Our results indicated that JNK contributes to both STAT1 Ser-727 phosphorylation and an IFN transcriptional response as monitored by expression of PLSCR1. However, inhibition of JNK did not suppress IFN␣2a-induced phosphorylation and activation of PKC␦. Therefore, JNK activation is downstream of PKC␦, whereas both of these kinases are required for maximal STAT1 Ser-727 phosphorylation. In summary, PKC␦ inhibition suppressed IFN-initiated JNK activation, STAT1 Ser-727 phosphorylation, and expression of PLSCR1. Transcriptional induction of PLSCR1 through STAT1 that is induced by IFN␣2a is dependent upon sequential activation of both PKC␦ and JNK kinase.