Tristetraprolin (TTP)-14-3-3 Complex Formation Protects TTP from Dephosphorylation by Protein Phosphatase 2a and Stabilizes Tumor Necrosis Factor-α mRNA*

Tumor necrosis factor (TNF)-α is a major cytokine produced by alveolar macrophages in response to pathogen-associated molecular patterns such as lipopolysaccharide. TNF-α secretion is regulated at both transcriptional and post-transcriptional levels. Post-transcriptional regulation occurs by modulation of TNF-α mRNA stability via the binding of tristetraprolin (TTP) to the adenosine/uridine-rich elements found in the 3′-untranslated region of the TNF-α transcript. Phosphorylation plays important roles in modulating mRNA stability, because activation of p38 MAPK by lipopolysaccharide stabilizes TNF-α mRNA. We hypothesized that the protein phosphatase 2A (PP2A) regulates this signaling pathway. Our results show that inhibition of PP2A by okadaic acid or small interference RNA significantly enhanced the stability of TNF-α mRNA. This result was associated with increased phosphorylation of p38 MAPK and MAPK-activated kinase 2 (MK-2). PP2A inhibition increased TTP phosphorylation and enhanced complex formation with chaperone protein 14-3-3. TTP physically interacted with PP2A in transfected mammalian cells. A functional consequence of TTP-14-3-3 complex formation appeared to be protection of TTP from dephosphorylation by inhibition of the binding of PP2A to phosphorylated TTP. Mutation of the MK-2 phosphorylation sites of TTP did not influence TNF-α adenosine/uridine-rich element binding and did not alter the increased TNF-α 3′-untranslated region-dependent luciferase activity induced by PP2A-small interference RNA silencing. Our data indicate that, although phosphorylation stabilizes TNF-α mRNA, PP2A regulates the mRNA stability by modulating the phosphorylation state of members of the p38/MK-2/TTP pathway.

Tumor necrosis factor-␣ (TNF-␣) 2 is a pro-inflammatory cytokine that influences a broad range of immunological pro-cesses, including autoimmune diseases, rheumatoid arthritis, septic shock, and acute respiratory distress syndrome. In these disease states, macrophages are widely recognized as cells that play a central role in the regulation of immune and inflammatory activities. In response to a variety of stimuli, for example LPS, macrophages generate pro-inflammatory cytokines such as TNF-␣, IL-1␤, IL-6, IL-12, and oxidants. TNF-␣ is an important alveolar macrophage secretory product, and depletion of alveolar macrophages reduces TNF-␣ production, thereby neutralizing and attenuating LPS-induced lung inflammation (1). Therefore, understanding all potential regulatory pathways of TNF-␣ expression is an important goal.
ARE-associated RNA-binding proteins are classified as stabilizing or destabilizing proteins depending on their effect on mRNA stability. Destabilizing proteins include tristetraprolin (TTP) (7)(8)(9), BRF1 (10), and AUF1 (11)(12)(13), which can directly bind to the AREs to promote mRNA decay. In contrast, stabilizing proteins such as HuR (14 -16) inhibit mRNA decay. The mitogen-induced immediate-early gene product, TTP, is a prototypical member of zinc finger proteins that contain two CCCH zinc fingers and three tetraproline (PPPP) motifs (17,18). Involvement of TTP in the post-transcriptional regulation of TNF-␣ mRNA was suggested by the increased TNF-␣ secretion from TTP-deficient mouse macrophages (19,20). Lai et al. (8) demonstrated that integrity of both zinc fingers was required for TTP to directly bind to the TNF-␣ ARE to promote mRNA decay. However, zinc fingers alone were not sufficient to mediate TNF-␣ mRNA degradation, because truncation of either the carboxyl or amino ends of TTP caused a loss TTP function (21). Both ends of the TTP zinc fingers contain a cluster of phosphorylation sites (22), yet the effects of phosphorylation on the mRNA binding and destabilizing activity of TTP remain controversial (22)(23)(24)(25).
Phosphorylation of TTP is regulated by p38 MAPK/MK-2 pathway (22,26,27). Activation by MK-2 results in a complex formation with chaperone protein, 14-3-3 (22,28). This association subsequently effects nuclear export of TTP (28,29) and prevents TTP-associated mRNA from being targeted to the degradative machinery (30). Regulation of dephosphorylation of TTP, and its upstream pathway, is incompletely understood. PP2A is a serine-threonine protein phosphatase involved in the regulation of signal transduction, cell growth, and apoptosis (31)(32)(33)(34). The core heterodimer of PP2A consists of a 36-kDa catalytic subunit (PP2A-C) and a 65-kDa regulatory subunit (PR65 or PP2A-A). This heterodimer interacts with a variable third subunit (PP2A-B isoforms) that influences cellular substrate specificity (35). The role of PP2A in the cellular response to inflammatory stimuli that causes cytokine expression has been increasingly investigated. In monocytes, inhibition of PP2A resulted in increased JNK activity and subsequent enhanced IL-1␤ secretion (36). However, the role of PP2A in the regulation of LPS-induced TNF-␣ production remains unknown.
In the present study, we tested the hypothesis that PP2A is involved in the post-transcriptional regulation of TNF-␣. Our data showed that inhibition of PP2A increased TNF-␣ mRNA stability. This effect was accompanied by increased p38 and MK-2 activity as well as increased phosphorylation of TTP. Interestingly, binding of TTP to 14-3-3 appeared to protect phosphorylated TTP from access by PP2A, thereby preventing its dephosphorylation.
PP2A Phosphatase Assay-A non-radioactive, malachite green-based immunoprecipitation assay kit (Upstate Biotechnology) was used to measure PP2A activity. MHS cells were either pretreated with OA or siRNA as described. Total cellular proteins were then extracted in Triton X-100 lysis buffer containing 50 mM Tris-HCl, pH 7.5, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 1% Triton X-100, and 0.5% Nonidet P-40 with no phosphatase inhibitors. Subunit C (catalytic) of PP2A was immunoprecipitated by anti-PP2A antibody. The precipitates were washed twice with lysis buffer and once with phosphatase assay buffer (50 mM Tris-HCl, pH 7.0, 0.1 mM CaCl 2 ). The pellets were resuspended in assay buffer and incubated with 750 M phosphopeptide for 15 min at 30°C. Reactions (25 l) were then transferred to a microtiter plate and incubated with 100 l of malachite green reagent. Color was developed for 5 min, and changes in absorbance were measured at 650 nm in a Spectra-MAX 250 (Molecular Devices plate reader).
Luciferase Assay-MHS cells were transiently transfected with pMT2-luc-UTR using Lipofectamine 2000 (Invitrogen). Two days after transfection, cells were either pretreated with or without 1 M OA for 1 h, washed once with cell culture medium, and then stimulated with 200 ng/ml LPS. For PP2A-siRNA silencing experiments, 0.3 ϫ 10 6 cells plated in 12-well plates were first transfected with control siRNA or mouse PP2A-siRNA for 48 h and then cotransfected with pMT2-luc-UTR and siRNA for another 24 h. Cells harvested at different LPS treatment times were lysed in Reporter lysis buffer (Promega, Madison, WI), and luciferase activity was measured using a Berthold Autolumat Plus LB 953 system with the protocol of a 2-s measurement delay followed by a 10-s measurement read.
p38 MAPK and MAPKAPK2 Kinase Assay-4 ϫ 10 6 MHS cells were lysed in 500 l of lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). 500 g of cell lysates was incubated with 20 l of immobilized phospho-p38 antibody overnight at 4°C. Beads were washed twice with lysis buffer and once with 1ϫ kinase buffer and then incubated with 8 g of recombinant ATF-2 (Cell Signaling) in 40 l of kinase reaction buffer containing 25 mM Tris-HCl, pH 7.5, 5 mM ␤-glycerophosphate, 2 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 200 M ATP, and 8 Ci of [␥-32 P]ATP for 30 min at 30°C. The reaction mixture was then boiled in an equal volume of 2ϫ Laemmli buffer and subjected to 10% SDS-PAGE. Gels were dried and exposed at Ϫ80°C for 5 h. MK-2 kinase assay was performed using a non-radioactive immunoprecipitation kinase assay kit from Upstate. Briefly, 1 mg of cell lysates was incubated with 10 l of anti-MK-2-agarose conjugate at 4°C for 1.5 h. Immunoprecipitated MK-2 was then incubated with 0.5 g of recombinant Hsp27 in 50 l of kinase buffer containing 100 M ATP for 45 min at 30°C. Onefifth of the sample was subjected to SDS-PAGE followed by immunoblotting with the phospho-Hsp27 antibody and a secondary antibody for enhanced chemiluminescence (Amersham Biosciences).

PP2A Inhibition by OA or siRNA Increased LPS-induced TNF-␣ Secretion in MHS Cells-To investigate the role of PP2A
in regulating LPS-induced TNF-␣ secretion from alveolar mac-rophages OA was used to examine the effect of phosphatase inhibition on TNF-␣ production. MHS cells representative of mouse alveolar macrophages were treated with varying concentrations of OA. The catalytic unit of PP2A was immunoprecipitated from MHS cell lysates and subjected to an in vitro phosphatase assay. OA inhibited PP2A specific activity in a dosedependent manner with nearly 90% inhibition of phosphatase activity achieved at 1 M OA (Fig. 1A). As a result, subsequent studies used a standard protocol of 1-h pretreatment with 1 M OA followed by culture medium change prior to LPS stimulation. This OA dose achieved PP2A inhibition with only modest affects on PP1-specific activity (data not shown).
MHS cells were pretreated with OA and then stimulated with 200 or 500 ng/ml LPS for 2, 4, and 6 h. OA treatment significantly augmented TNF-␣ secretion. At 2-and 4-h time points (Fig. 1B), the amount of TNF-␣ secreted into culture supernatants approximately doubled in the OA-treated group. By 6 h of stimulation, the augmentation of TNF-␣ produced in the presence of OA was less but remained significant (Fig. 1B). Although stimulation with 500 ng/ml LPS increased the absolute value of TNF-␣, the relative amount of OA-mediated augmentation, while significant, was not further enhanced. Therefore, 200 ng/ml LPS were used in all subsequent experiments. Moreover, OA treatment alone did not change the basal level of TNF-␣ secretion, so that the enhancing effects were only observed following LPS stimulation.
To demonstrate that the augmented TNF-␣ secretion by OA was specifically attributable to PP2A inhibition instead of PP1, silencing of PP2A and PP1 was achieved by transfecting MHS cells with either PP2A-or PP1-siRNA (Fig. 2). Western blot showed that the endogenous expression of PP2A and PP1 in MHS cells were both reduced by siRNA transfection (Fig. 2A). Similar to OA treatment, PP2A-siRNA silencing significantly increased TNF-␣ secretion after LPS stimulation (Fig. 2B). Compared with control siRNA-transfected cells, the amount of secreted TNF-␣ was nearly doubled in the PP2A-silencing group at all time points (Fig. 2B). Of note, no increase in TNF-␣ secretion was observed following PP1 silencing at any time (Fig.  2B). As expected on the basis of transfecting less confluent cell cultures with siRNA, the overall level of TNF-␣ secretion was slightly less with control siRNA transfection (Fig. 2B) as compared with non-transfected cells (see Fig. 1). Thus, attributing the OA-induced augmentation of TNF-␣ secretion to PP2A inhibition was corroborated by PP2A-siRNA silencing experiments. When feasible, siRNA was used as a complimentary strategy to confirm results from OA treatment studies.
Because recent studies suggested that PP2A positively regulated I kappa kinase (IK) and subsequently activated the NF-B pathway (40), we surmised the effect of OA was largely independent of transcription. Instead we hypothesized that post-transcriptional regulation was being affected.
PP2A Inhibition Prevents Decay of TNF-␣ mRNA-Posttranscriptional regulation of TNF-␣ involves the mRNA stability of its transcript. To examine the role of PP2A in the regulation of the TNF-␣ mRNA stability, MHS cells pretreated with   FEBRUARY 9, 2007 • VOLUME 282 • NUMBER 6 or without OA were stimulated with LPS for 2 h and then treated with 5 g/ml actinomycin D to block further transcription. Total RNA was extracted at the indicated time intervals. Northern blot analysis revealed that the decay of TNF-␣ mRNA was significantly prevented by OA pretreatment (Fig. 3). This result suggested that the effect of OA treatment on TNF-␣ secretion was due to the stabilization of mRNA transcript.

Post-transcriptional Regulation of TNF-␣ by PP2A
PP2A Inhibition by OA or PP2A-siRNA Increased the TNF-␣ 3Ј-UTR-dependent Luciferase-To further characterize the effects of PP2A inhibition on TNF-␣ mRNA stability, a luciferase reporter construct that incorporated the 3Ј-UTR of TNF-␣ (pMT2-luc-UTR) was transiently transfected into MHS cells and used to assess mRNA stability of transcribed luciferase. Transfected cells were pretreated with or without OA followed by LPS stimulation, and luciferase activity was measured at different time points. In response to OA, 2-, 15-, and 11-fold increases in luciferase activity were observed at the 2-, 4-, and 6-h time points, respectively (p Ͻ 0.01 for all time points, Fig. 4A, left panel).
Similar to OA treatment, knock-down of endogenous PP2A bysiRNAalsosubstantiallyincreased(ϳ10-fold)TNF-␣3Ј-UTRdependent luciferase activity at all time points tested (Fig. 4A, right panel). Because siRNA transfections were performed on less confluent cell cultures than used in OA studies, luciferase activity was normalized to micrograms of protein. As a result, no effect was seen using control siRNA as compared with non-transfected cells. Taken together, these results indicate that the unstable elements conferred by the 3Ј-UTR of TNF-␣ mRNA were suppressed by the inhibition of PP2A in MHS cells. It has been demonstrated that activation of p38 MAPK stabilizes TNF-␣ mRNA (26,27). To determine whether the effect of PP2A on TNF-␣ mRNA stability was mediated via the p38 pathway, inhibitors of p38, JNK, MEK1/2, and NF-B were added 30 min after stimulation with LPS to transfected cells that had been pretreated with OA. Results showed that luciferase activity was significantly inhibited only by the p38-specific inhibitor SB 203580 (5 M) (Fig. 4B). These data provided additional support to the hypothesis that PP2A regulated TNF-␣ mRNA stability via the p38 pathway.
PP2A Inhibition by OA or PP2A-siRNA Increased Phosphorylation of p38 MAPK/MK-2-To further support this observation, MHS cells pretreated with or without OA were   Cell lysates were assayed for phospho-p38 and its downstream kinase, phospho-MK-2, by Western blot. OA pretreatment increased the LPS-induced phosphorylation status of both p38 MAPK (Fig. 5A, upper panel) and MK-2 (Fig. 5B, upper panel).
To corroborate these data with increased kinase activity, p38 MAPK and MK-2 activities were measured by in vitro kinase assays. Autoradiography analysis for the p38 kinase substrate (phospho-ATF2, Fig. 5A, bottom panel) and immunoblotting for the MK-2 substrate (phospho-Hsp27, Fig. 5B, bottom panel) confirmed increased p38 and MK-2 activity following OA pretreatment. Similarly, inhibition of PP2A by siRNA showed increased phosphostatus of p38 MAPK (Fig. 5A, middle  panel) and MK-2 (Fig. 5B, middle  panel). These data further confirmed the specific involvement of PP2A in the p38 MAPK/MK-2 posttranscriptional pathway.
PP2A Inhibition Increased Phosphorylation of TTP-RNA-binding protein TTP is a substrate for MK-2 (22,27) and has been implicated in the 3Ј-UTR-mediated degradation of TNF-␣ (19,20). MK-2 phosphorylates TTP at two major serine sites, Ser-52 and Ser-178, both of which serve as 14-3-3 binding sites (22,28,30). To measure MK-2-induced phosphorylation of TTP, a phospho-specific antibody was raised against Ser-178. The antibody recognized endogenous TTP in lysates of LPS-activated Raw264.7 macrophages (Fig. 6A, lane 2), whereas treatment of the lysate with alkaline phosphatase abolished the signal (lane 3). To confirm specificity of this antibody, COS7 cells were transiently transfected with TTP constructs. The novel, phospho-specific antibody recognized both human and mouse TTP (Fig. 6B, lanes 2 and 3) but failed to recognize TTP that was mutated at Ser-178 (lanes 5 and 6). After knocking down MK-2 in COS7 cells by siRNA, the phospho-(Ser-178)-TTP signal was reduced by ϳ50%, both in the absence and presence of the p38-MAPK activator anisomycin (Fig. 6C). Consistent with prior studies (22,30), this result confirmed that MK-2 was required for phosphorylation of TTP at Ser-178.
To examine the effects of PP2A on TTP phosphorylation, MHS cells pretreated with or without OA were stimulated with LPS for 1 and 2 h. Phosphorylated TTP was pulled down by GST-14-3-3 beads in vitro and analyzed using the phospho-(Ser-178)-TTP antibody. The amount of total TTP in the cell lysates was detected with an affinitypurified TTP antibody (CARP-3). LPS induced TTP expression with its phosphorylation being apparent by 1 h (Fig. 6D, upper  panel). At 2 h, a second, lower mobility band was induced, suggesting an enhanced phosphorylation state. OA strongly . 500 g of cell lysates treated with or without recombinant PP2A for 30 min at 30°C was incubated with immobilized phospho-p38 antibody overnight at 4°C. Beads were washed and subjected to in vitro kinase assay using ATF-2 as the substrate, and the phosphorylated ATF-2 was visualized by autoradiography (bottom panel). B, MHS cell lysates were examined for phospho-MK-2 by immunoblotting (upper and middle panels). MK-2 kinase activity was measured by a non-radioactive immunoprecipitation method using recombinant Hsp27 as MK-2 substrate. Phospho-Hsp27 was visualized by immunoblotting using enhanced chemiluminescence (bottom panel). Results shown are representative of three independent experiments.
increased the phosphorylated bands, whereas incubation with recombinant PP2A significantly decreased phospho-(Ser-178)-TTP (Fig. 6D, upper panel). p38 activation is known to increase TTP expression by preventing its proteolytic degradation via the proteasome (41) or by activating its expression at the transcriptional level. Because OA treatment increased p38 activity (Fig. 5A), it was crucial to determine the total amount of TTP expression in these experiments. As expected, increased total TTP expression was seen in MHS cells after OA treatment (Fig.  6D, bottom panel). However, a substantial amount of OA-augmented TTP protein was the phosphorylated isoform, because densitometry analysis showed that total TTP expression increased ϳ2.5-fold after OA treatment, whereas the phosphorylated isoform (Ser-178 of TTP) increased ϳ4.2-fold.
Binding of 14-3-3 and PP2A to TTP showed an inverse correlation when tested on non-phosphorylatable mutations of TTP (Fig. 7A). Thus, we hypothesized that 14-3-3 association with TTP may prevent its dephosphorylation by PP2A. To test this, HEK293 cells were co-transfected with wild-type TTP and gradually decreasing amounts of Myc-tagged 14-3-3␤ construct. Western blotting confirmed equal TTP expression but a decreasing amount of 14-3-3 in the cell lysates (Fig. 7B, upper  panel). These same blots were stripped and examined for phospho-(Ser-178)-TTP. In this case, a decreasing amount of phospho-TTP was detected in the cells that had less expression of 14-3-3 (Fig. 7B, bottom panel). Finally, knocking down PP2A by FIGURE 6. Establishing specificity of phospho-(Ser-178)-specific TTP antibody. A, cells were stimulated for 4 h with LPS (10 ng/ml), and cytoplasmic lysates were prepared using RNA immunoprecipitation buffer. A portion of the lysate prepared in the absence of phosphatase inhibitors was treated for 20 min at 37°C with 1 unit/ml alkaline phosphatase (lane 3). B, COS7 cells were transfected with human and mouse TTP cDNAs, and additional TTP constructs in which serine 52 and serine 178 were substituted by alanine. C, COS7 cells were transfected over a period of 4 days with either a control siRNA or siRNA targeting MK-2, and co-transfected with TTP. Cells were cultured for 48 h in the absence of FCS and treated with control medium or 10 g/ml anisomycin for 30 min prior to lysis. Knock-down efficiency was monitored using an MK-2 antibody. Serine 82 of Hsp27 is a known target site of MK-2. D, phosphatase inhibition increased LPS-induced phosphorylation of TTP. GST-14-3-3 fusion proteins were expressed in E. coli HB101 cells by isopropyl-␤-D-thiogalactoside induction. Bacterial lysates were incubated with glutathione-Sepharose 4B beads (Amersham Biosciences) for 1 h at 4°C and washed three times with phosphate-buffered saline. GST alone or GST-14-3-3 fusion protein attached to glutathione-Sepharose beads were added to the lysates and incubated for 2 h at 4°C. Proteins pulled down by GST-14-3-3 beads were analyzed with a phospho-(Ser-178)-specific TTP antibody. Total TTP in the lysates was detected with non-phospho-TTP (CARP-3) antibody. siRNA in this model system increased the signal of phospho-(Ser-178)-TTP by ϳ50%, whether or not 14-3-3 was overexpressed (Fig. 7C, lanes 2 and 4). Also, whether PP2A was knocked down or not, the absence of 14-3-3 overexpression correlated with a diminished phospho-TTP signal (lanes 3 and 4). These data provided further evidence that TTP-14-3-3 complex formation appeared to prevent TTP dephosphorylation by PP2A.
PP2A Interacts with and Dephosphorylates TTP from LPS-stimulated MHS Cells-Although these data using an overexpression system suggested direct regulation of TTP phosphorylation state by PP2A, an indirect effect could not be excluded. Nor was it certain this biology could be extrapolated to immune-relevant cells such as macrophages. Therefore, to further determine whether endogenous PP2A could directly mediate TTP dephosphorylation, we first immunoprecipitated PP2A from unstimulated MHS cells. This intraperitoneal product was then incubated (30°C, from 5 to 60 min) with cell lysates recovered from MHS cells stimulated with LPS for 4 h (Fig. 8). Lysates were then subjected to Western blot analysis for the phosphorylated and dephosphorylated isoforms of TTP on the basis of molecular weights of the protein. Dephosphorylation appeared at 15 min and longer incubation with immunoprecipitated, endogenous PP2A resulted in complete dephosphorylation of TTP (Fig. 8, upper  panel). To further detect the interaction of TTP and PP2A, the beads were washed and then subjected to reducing Laemmli buffer, boiling, and immunoblotting with anti-TTP (Carp-3) antibody. As shown in Fig.  8 (middle panel), the presence of TTP on anti-PP2A-IgG-beads correlated with decreased TTP phosphorylation (lanes 4 -7). As an important control, no dephosphorylation of TTP and no TTP-PP2A interaction were observed with the immunoprecipitated product using an isotype-matched, irrelevant antibody (lane 1).
Effects of TTP Phosphorylation on TNF-␣ ARE Binding-To determine the effect of TTP phosphorylation on mRNA binding, RNA gel-shift assays were performed. Cytoplasmic extracts from LPS-stimulated MHS cells were incubated with a TNF-␣ ARE probe without the addition of supershifting antibodies (Fig. 9A). Of the five major bands resolved by non-denaturing gel electrophoresis, LPS induced one band designated in the prior literature as C4 (lanes 2 and 3) (27). Pretreatment with OA strongly increased this C4 band signal (lanes 4 and 5) by ϳ3.3-fold as compared with non-treated samples (lanes 2-3).
We therefore used an RNA-immunoprecipitation method to examine the ability of wild-type TTP (mTTP-wt) and Ser to Ala double mutation TTP (TTP-S52A/S178A) to co-immunoprecipitate a reporter mRNA containing the TNF-␣-ARE (Fig. 9C). Globin-ARE (TNF) mRNA was effectively precipitated by both constructs with nearly identical efficiency. The TTP-m1,2 expressed possesses an altered zinc finger section of TTP such that it cannot bind TNF-␣ mRNA and was used as a negative control to demonstrate the specificity of this experiment (lane 3). These data suggested that TTP binding to the TNF-␣ ARE may be independent of the two MK-2-dependent phosphorylation sites (22,28,30).
Ser-52/178 3 Ala-52/178 Mutation of TTP Did Not Disrupt the Increased Luciferase Activity Induced by PP2A-siRNA Silencing-Because Ser-52/178 to Ala-52/178 mutation did not change the binding of TTP to TNF-␣ ARE, we tested the effects of this mutation on luciferase reporter gene expression and whether knocking down PP2A altered this expression. Compared with vector control, co-transfection of reporter gene (pMT2-luc-UTR) with either wild-type TTP (mTTP-wt) or TTP-S52A/S178A strongly inhibited the luciferase gene expression (data not shown). However, higher luciferase activity was observed for wild-type TTP than the mutated TTP (Fig. 10, left panel compared with right panel). This finding is consistent with prior reports of MK-2 phosphorylating TTP to stabilize TNF-␣ mRNA (21,30,42). Knock-down of PP2A increased the luciferase activity of both constructs, although less dramatically with the mutated TTP (Fig. 10). Nevertheless, it was surprising that PP2A silencing increased the luciferase activity of the mutated TTP to any degree as the two MK-2 phosphorylation sites in this construct have been mutated. It may be that the increased p38/MK-2 activity caused by PP2A silencing was responsible for the increased luciferase activity of the mutated TTP. Alternatively, other phosphorylation sites on TTP, independent of MK-2, that regulate TNF-␣ mRNA stability could be additional targets by PP2A.  . Effects of TTP phosphorylation on TNF-␣ ARE binding. A, MHS cells were pretreated with or without 1 M OA for 1 h followed by LPS stimulation for 0, 1, and 2 h. Cytoplasmic extracts were incubated with a radiolabeled RNA probe containing TNF-␣ ARE for 15 min at room temperature, followed by RNase T1 treatment. Reaction samples were separated by 6% non-denaturing PAGE, and the ARE-binding bands were visualized by autoradiography (labeled as the C4 band). B, prior to separation by non-denaturing PAGE, supershift antibodies were added to the reactions: normal rabbit IgG (lanes 1-3), rabbit anti-14-3-3 (lanes 4 -8), normal goat IgG (lanes 9 -11), and goat anti-TTP (lanes [12][13][14]. Results shown are representative of five independent experiments. C, binding of TTP to the ARE of TNF-␣ by RNA immunoprecipitation. COS7 cells were transfected with a ␤-globin reporter gene containing the ARE of TNF-␣ together with vector alone, TTP-wt, TTP-m1,2 mutated in the RNA-binding zinc finger domain, or TTP-AA. MycHis-tagged TTP was immunoprecipitated using anti-Myc antibody, and RNA was isolated from both the lysate (input) and the immunoprecipitated material. Globin-ARE (TNF) mRNA was detected by Northern blot, TTP protein was visualized by Western blot analysis using anti-Myc antibody. Results shown are representative of two independent experiments.

DISCUSSION
Our laboratory has been interested in identifying the role of endogenous proteins that negatively modulate inflammatory cell signaling pathways with the hopes of identifying additional therapeutic targets. A number of inflammation-triggered disease states affecting critically ill patients (e.g. sepsis and acute respiratory distress syndrome) are initiated by pathogen-mediated activation of cells of the innate immune system. Because of this key role played by cells derived from the mononuclear/ macrophage lineage, we have focused our current studies on this cell type. A canonical pathogen-associated molecular pattern that triggers cellular activation is the Gram-negative bacterial cell wall product, LPS. LPS is known to activate a myriad of pathways essential to the subsequent expression of a number of gene products. These genes include cytokines (e.g. TNF-␣), chemokines, and adhesion molecules that mediate pathological responses in inflammatory disease states (43,44). As a result, a number of studies have examined the transcriptional and post-transcriptional mechanisms regulating LPS-induced TNF-␣ expression.
Post-transcriptional regulation of TNF-␣ is mediated by binding of TTP to the AU-rich elements in the 3Ј-untranslated region of TNF-␣ mRNA. The p38 MAPK has been shown to increase TNF-␣ expression via ARE-mediated mRNA stability (26,27). This action of p38 is mediated through MK-2, which phosphorylates TTP in vitro and in vivo (22,30,45,46). It has been established that regulation of phosphorylated proteins results from a balance between the activity of specific kinases and phosphatases. The current studies aimed to elucidate the role of PP2A in the post-transcriptional regulation of TNF-␣. We observed that modulating PP2A activity greatly enhanced LPS-induced TNF-␣ mRNA stability. These data provide the first evidence for the involvement of PP2A in the post-transcriptional regulation of TNF-␣ production. Although previous results have shown that TTP phosphorylation modulates TNF-␣ expression (21,30), our understanding of the molecular mechanism by which phospho-TTP controls inflammatory gene expression remains incomplete.
Previous studies have established that binding of phosphorylated TTP to 14-3-3 (22, 28) mediates a variety of cellular functions. For example, the TTP-14-3-3 interaction was reported to be responsible for the nuclear exportation of TTP (29,47). This observation suggests subcellular localization of the complex was an important contributor to function. This concept of localizationdependent function was supported by the additional observation that TTP-14-3-3 interaction reduced the entry of TTP into the stress granule after arsenite treatment in COS7 cells (30). In the current study, we have identified a potential additional and novel function of TTP-14-3-3 association. Our results demonstrate that interaction with 14-3-3 protects phosphorylated TTP from access by the catalytic unit of PP2A thus preventing its dephosphorylation.
We hypothesize that this protective role of TTP-14-3-3 complex formation against PP2A dephosphorylation is crucial for stabilizing cytokine messenger RNA after stimulus exposure. Our experiments suggest that LPS-induced p38 MAPK/MK-2 phosphorylation is quick and transient. Phosphorylation peaks at 30 min and is followed by subsequent, rapid dephosphorylation. In contrast, TTP phosphorylation usually persists for several hours (27,30). Thus, it appears that 14-3-3 binding may be critical for maintaining a prolonged phosphorylation state of TTP once upstream kinases have been deactivated. Subsequently, via an unknown mechanism, 14-3-3 is dissociated, leaving the unbound TTP to be accessed by PP2A for dephosphorylation, thereby promoting mRNA decay. A recent study reported that phosphorylation of 14-3-3 by IK␤ and protein kinase C␦ resulted in the dissociation of 14-3-3 from the ␤4GalT1 mRNA thereby enhancing mRNA stability (48). These reports support our hypothesis that dissociation of 14-3-3 from its binding to TTP might be dependent on the phosphorylation state of 14-3-3. Such a function could reflect another layer to the complexity of TTP-mediated mRNA stability.
It has been established that binding of TTP to 14-3-3 stabilizes TNF-␣ mRNA. This is at least in part p38/MK-2-dependent, because mutation of two key serine residues affected luciferase expression mediated by the TNF-␣ 3Ј-UTR. Importantly, in the current studies, this mutated TTP had reduced binding to 14-3-3 and showed increased association with PP2A. Thus, the enhanced mRNA decay observed with the Ser-mutated TTP appears consistent with our hypothesis. However, these results cannot exclude other phosphorylation sites as being critically involved in the stability of mRNA transcript. Consistent with our observations, recent MK-2/TTP function studies (21) as well as additional studies employing the MK-2 knockout mice (42) have suggested that alternative pathways other than p38/MK-2 could modify TTP to affect its function. An indirect effect of p38/MK-2 activation caused by PP2A inhibition could play a role in the increased stability of TNF-␣ mRNA; however, direct dephosphorylation of TTP by PP2A was substantiated by the present data. These two regulatory mechanisms (p38/MK-2-dependent versus independent pathways) may not be mutually exclusive. Instead the regulatory mechanisms may act at different stages of TTP activation/phosphorylation. For an example, p38 and MK-2 act immediately upon cellular activation to phosphorylate TTP and stabilize mRNA. In contrast, PP2A may function at a later stage to dephosphorylate TTP thereby affecting mRNA stability.
The current studies also examined the effect of TTP phosphorylation on its binding to mRNA. Other studies exploring this question have reported varying results (22)(23)(24)(25). Carballo et al. reported that phosphorylated TTP bound less to a granulocyte macrophage-colony stimulating factor ARE probe than dephosphorylated TTP. This finding suggested that the increased stability of granulocyte macrophage-colony stimulating factor mRNA after LPS stimulation was caused by the release of TTP from the transcript (23). In a similar manner, Hitti et al. (42) reported that a lower phosphorylation state of TTP correlated to increased binding to a TNF-␣ ARE probe. In contrast to these observations, we did not detect a difference between wild-type TTP and the Ser-mutated TTP (S52A/ S178A) in their ability to bind to a globin-TNF␣-ARE reporter mRNA. This was corroborated by RNA gel-shift studies that showed an increased C4 band after OA treatment that was supershifted by both TTP and 14-3-3 antibodies. Thus, phosphorylation of TTP, at least at the MK-2-dependent sites, did not appear to disrupt mRNA binding. These RNA-binding data are consistent with reports showing the effect of phosphorylation of other RNA-binding proteins to their target AU-rich elements such as the binding of BRF1 to an IL-3 ARE (49).
Because TTP is considered a destabilizing RNA-binding protein, it would have been predicted that increased mRNA stability should have been associated with less RNA binding by TTP. Our current data suggest there may be additional effects mediated by PP2A regulation of the phosphorylation state if TTP. We speculate that TTP may regulate mRNA stability via the interaction with additional regulatory proteins. The presence of such putative regulatory proteins that could comprise a particular RNA-binding complex may influence the ultimate state of a mature mRNA transcript. In support of this notion, 14-3-3 binding appears to prevent TTP-bound RNA from moving to the degradation machinery, thereby increasing mRNA stability (30). One possible explanation is that 14-3-3 binding is antagonistic to the interaction of TTP with RNA degradation machinery. TTP-14-3-3 complex formation could impair the localization of the transcript to either the exosome, a multiprotein complex of 3Ј-to-5Ј exoribonucleases (50,51), or the decapping complex possessing the 5Ј-to-3Ј exoribonuclease Xrn1 (52,53). Therefore, TTP-RNA binding alone may not be the only indicator of mRNA processing. Identification of additional putative proteins involved in the ARE-binding complex requires further study. Further insight into the role of these proteins in modifying mRNA stability will be necessary. Finally, understanding how this function is influenced by their phosphorylation state and regulated by phosphatases such as PP2A is necessary to increase our understanding of the molecular regulation of cytokine gene expression.