p38 Mitogen-activated Protein Kinase Stabilizes mRNAs That Contain Cyclooxygenase-2 and Tumor Necrosis Factor AU-rich Elements by Inhibiting Deadenylation*

AU-rich elements (AREs) in 3′-untranslated regions of mRNAs confer instability. They target mRNAs for rapid deadenylation and degradation and may enhance decapping. The p38 MAPK pathway stabilizes many otherwise unstable ARE-containing mRNAs encoding proteins involved in inflammation; however, the mRNA decay step(s) regulated by the signaling pathway are unknown. To investigate whether it regulates deadenylation or the decay of the mRNA body, we used a tetracycline-regulated β-globin mRNA reporter system to transcribe pulses of mRNA of uniform length. We measured on Northern gels the migration of reporter mRNAs isolated from cells transfected only with reporter plasmid or co-transfected with an active mutant of MAPK kinase-6, and treated either with or without the p38 MAPK inhibitor SB 203580. Differences in migration were shown by RNase H mapping with oligo(dT) to be due to poly(A) shortening. Insertion of an ARE into the β-globin reporter mRNA promoted rapid deadenylation and decay of hypo-adenylated reporter mRNA. p38 MAPK activation inhibited the deadenylation of reporter mRNAs containing either the cyclooxygenase-2 or tumor necrosis factor AREs. The regulation of deadenylation by p38 MAPK was found to be specific because deadenylation of the β-globin reporter mRNA either lacking an ARE or containing the c-Myc 3′-untranslated region (which is not p38 MAPK-responsive) was unaffected by p38 MAPK. It was concluded that the p38 MAPK pathway predominantly regulates deadenylation, rather than decay of the mRNA body, and this provides an explanation for why p38 MAPK regulates mRNA stability in some situations and translation in others.

The control of mRNA stability is fundamental in regulating gene expression. This is reinforced by the fact that mRNA stability can be modulated by cell signaling events (1)(2)(3). The default mRNA decay pathway in eukaryotes begins by shortening of the poly(A) tail (4 -7). A mammalian poly(A)-specific ribonuclease has been identified (8,9), and a human homologue of the major cytoplasmic deadenylase in yeast, Ccr4p, has been cloned (10). In mammalian cell extracts deadenylation is followed by decay of the mRNA body, which occurs mainly in the 3Ј 3 5Ј direction due to the action of a large exonuclease complex or exosome (11)(12)(13); the 5Ј cap structure of the mRNA then being removed by a scavenger activity (11). In yeast, mRNA decay initiates with poly(A) shortening, but the mRNA body is then degraded either by 3Ј 3 5Ј decay involving the exosome (14) or by decapping involving Dcp1p (15,16) and Dcp2 (17) followed by 5Ј 3 3Ј decay by the exoribonuclease Xrn1p (18). Homologues of the yeast proteins Dcp1p and Dcp2p have been found in humans (19,20), suggesting that following deadenylation 5Ј 3 3Ј decay of the body could also be a significant pathway in mammalian cells. However, despite the identification of specific steps in the mRNA decay, little is known about which process is regulated by particular cell signaling pathways.
The best characterized p38 MAPK-regulated mRNAs contain AU-rich elements (AREs), consisting of multiple, frequently overlapping copies of the AUUUA motif. An important property of the ARE is its ability to direct instability of an mRNA. A wide range of AREs destabilize mRNAs in vivo by promoting rapid deadenylation (4,24,25). In vitro, AREs target the body of synthetic mRNAs for 3Ј 3 5Ј decay by the exosome (12,13). AREs also increase decapping rates in vitro (26); however, it is unclear if this is performed by a Dcp1-like or scavenger activity (20).
Activation of p38 MAPK stabilizes mRNA by blocking the ability of AREs to destabilize mRNA (27,28). Despite the identification of numerous ARE-binding proteins (ARE-BPs), it is unclear which (if any) provides a link between p38 MAPK and the ARE. One ARE-BP, tristetraprolin (TTP), is destabilizing and directs deadenylation (29). It is an in vitro substrate of MAPKAPK-2 (30); however, TTP is not necessarily involved in the basic stabilization mechanism because it is not expressed in HeLa cells in which regulation of reporter mRNAs by p38 MAPK has been studied. 2 HuR, which is expressed in HeLa cells (31), also binds the p38 MAPK-responsive COX-2 and TNF AREs (31)(32)(33)(34). HuR is thought to stabilize mRNA by inhibiting the decay of hypo-adenylated intermediates (35,36). In order to identify the link between mRNA stability and p38 MAPK, it is necessary to ascertain which specific mRNA decay steps are regulated by the signaling pathway. To do this we transcribed pulses of reporter mRNA in cells containing active or inactive p38 MAPK and followed poly(A) tail length and decay of the mRNA by Northern blotting. We show that p38 MAPK stabilizes mRNA by blocking ARE-directed deadenylation. Plasmids-pTet-BBB-COX-2 ARE was prepared as described previously (27) by inserting a 123-nt sequence that spans the main regulatory ARE of the human COX-2 3Ј-UTR into the unique BglII site in the ␤-globin 3Ј-UTR in pTet-BBB (a kind gift of A.-B. Shyu, Houston Medical School, Houston, TX). pTet-BBB-c-Myc 3Ј-UTR was constructed previously (27). Tet-BBB-TNF ARE was constructed by inserting a 44-nt oligonucleotide spanning the human TNF ARE into the BglII site in the ␤-globin 3Ј-UTR of pTet-BBB as described previously (37). The sequences inserted into the BglII site of pTet-BBB are shown in Fig. 5.

Materials-General
Cell Culture-HeLa tet-off cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and G418 (100 ng ml Ϫ1 ). Cells were maintained at 37°C in the presence of 5% CO 2 .
Transfection-100,000 cells were seeded per well in 6-well plates and cultured for 24 h. Cells were washed with phosphate-buffered saline, and 500 l of serum-containing medium was added. 100 ng of mRNA reporter plasmid with or without 100 ng of pCMV-MKK6E DNA (a gift of J. Han) and pBluescript DNA (enough to give a total of 1 g of DNA) were combined with 5 l of Superfect in 100 l (final volume) of serum-free Dulbecco's modified Eagle's medium. The mixture was incubated for 10 min at room temperature and overlaid onto the cells. Cells were incubated at 37°C in the presence of 5% CO 2 for 3 h and washed with phosphate-buffered saline, and culture medium containing tetracycline was added.
Transcriptional Pulse-This was performed as described previously (38) with some modifications. Briefly, pulses were produced by culturing HeLa tet-off cells in a low concentration of tetracycline (50 ng ml Ϫ1 unless stated otherwise) to block transcription of reporter mRNA for 24 h following transfection. The cells were washed with 1ϫ phosphatebuffered saline, and culture medium (lacking the antibiotic) was replaced. Cells were cultured for 3 h to allow transcription to proceed, and then tetracycline at a high concentration (500 ng ml Ϫ1 ) was added to switch off transcription. Cells were then treated with p38 MAPK inhibitor or left untreated and then harvested and lysed for RNA isolation at appropriate time points.
RNA Isolation and Northern Blotting-Total RNA was isolated from cells using a kit. 10 g of RNA together with 1 g of RNA size markers was electrophoresed on denaturing gels (2% agarose and 0.41 M formaldehyde). RNA was capillary-transferred to a nylon membrane and fixed by UV cross-linking. Prehybridization (2 h) and hybridization (overnight) with ␣-32 P-labeled riboprobes were performed at 65°C in Ultrahyb. Blots were washed three times for 30 min at 65°C with 5ϫ SSC and 0.1% SDS, 1ϫ SSC and 0.1% SDS, and 0.1ϫ SSC and 0.1% SDS. Signals were visualized and quantified on a Fuji (Tokyo, Japan) FLA-2000 PhosphorImager. RNase H Mapping-10 g of RNA was mixed with 0.3 g of oligo(dT) in distilled H 2 O (final volume 9.8 l) and incubated at 65°C for 5 min and then allowed to cool slowly to room temperature. 0.8 units of RNase H, 1.2 l of 10ϫ REact 2 buffer, and 40 units of RNasin were added, and the mixture was incubated at 37°C for 30 min. RNA was precipitated using ethanol and sodium acetate and resuspended in distilled H 2 O.
Determination of Poly(A) Tail Length-This was calculated by plotting a calibration curve of log (RNA marker length) against migration on the Northern gel. The difference in migration between BBB-COX-2 ARE and GAPDH was measured, added to a fixed value of GAPDH migration (measured from the lane adjacent to the RNA markers), and plotted on the calibration curve. This allowed for correction of any slight differences in migration due to distortions in mobility during electrophoresis. BBB-COX-2 ARE mRNA length was then determined from the calibration curve. Poly(A) tail length was calculated by subtracting the theoretical length of poly(A) Ϫ BBB-COX-2 ARE mRNA (712 nt) from the measured values. All measurements of mRNA migration were taken from the peak band intensity determined by PhosphorImager analysis.

A Transcriptional Pulse Strategy to Generate mRNA of Uniform Poly(A) Tail
Length-To investigate whether p38 MAPK inhibits deadenylation or blocks the decay of the mRNA body of ARE-containing mRNAs, we used a tetracycline-regulated transcriptional pulse-chase mRNA reporter system (38). This system comprises HeLa cells stably transfected with a tetracycline-sensitive transcription factor (HeLa tet-off). The cells are transiently transfected with a plasmid containing a genomic rabbit ␤-globin gene which is under the control of the tet operon. The reporter mRNA consisted of the ␤-globin 5Ј-UTR, ␤-globin open reading frame, and ␤-globin 3Ј-UTR (BBB mRNA) without or with different AREs inserted into its 3Ј-UTR. This system allows pulses of mRNA with uniform poly(A) tail lengths to be transcribed, enabling deadenylation rates to be measured, and the decay of deadenylated intermediates to be followed.
To check that the stability of mRNA produced using this pulse-chase approach was regulated by p38 MAPK, HeLa tetoff cells were transiently transfected with pTet-BBB-COX-2 ARE reporter plasmid together with or without a plasmid expressing MKK6E, an active mutant of the p38 MAPK upstream activator (MAPK kinase-6). Cells were cultured overnight in the presence of a low concentration (25 ng ml Ϫ1 ) of tetracycline to inhibit transcription. Tetracycline was removed for 3 h to allow a pulse of transcription to occur and then added back at a high concentration (500 ng ml Ϫ1 ) to re-impose transcriptional blockade. Cells were harvested at various time points; RNA was isolated, and mRNA levels in the lysates were measured by ribonuclease protection assay for ␤-globin and GAPDH (as a loading control).
In the absence of p38 MAPK activation, no reporter mRNA was detected after overnight incubation of cells in 25 ng ml Ϫ1 tetracycline, but the transcript accumulated over 3 h following withdrawal of the antibiotic (Fig. 1). The re-addition of 500 ng ml Ϫ1 tetracycline was followed by rapid decay of the reporter mRNA, with a slight lag of around 1 h (Fig. 1). In cells cotransfected with the MKK6E expression plasmid, a small amount of BBB-COX-2 ARE mRNA was detected after overnight incubation with 25 ng ml Ϫ1 tetracycline, and strong accumulation of this mRNA was observed after tetracycline withdrawal. Under these conditions the reporter mRNA level remained constant following re-addition of 500 ng ml Ϫ1 tetracycline (Fig. 1). These observations confirm that a transcriptional pulse can be generated in HeLa tet-off cells and that the stability of this reporter mRNA produced in this way is controlled by the p38 MAPK pathway.
Inhibition of p38 MAPK Increases the Rate of Deadenylation of BBB-COX-2 ARE Reporter mRNA-To examine the effect of p38 MAPK activation on the deadenylation rate of the ARE reporter mRNA HeLa tet-off cells were co-transfected with pTet-BBB-COX-2 ARE and the MKK6E expression plasmid. Cells were cultured overnight in the presence of tetracycline (50 ng ml Ϫ1 ) to block transcription. The antibiotic was removed for 3 h and then re-added at a higher concentration (500 ng ml Ϫ1 ) to generate a pulse of transcription. The cells were then either left untreated, or SB 203580 (1 M) was added to block p38 MAPK activity. The cells were then harvested and lysed at different times, and RNA was isolated and examined by Northern blotting to measure the deadenylation rates of the reporter mRNA in the presence or absence of p38 MAPK activity.
Immediately after transcriptional blockade, BBB-COX-2 ARE mRNA isolated from cells containing active p38 MAPK was ϳ900 nt long (Fig. 2, A and B). The parental poly(A) Ϫ BBB transcript is 589 nt (6), and BBB-COX-2 ARE mRNA contains a 123-nt insert spanning the COX-2 ARE, giving a theoretical length of 712 nt for the poly(A) Ϫ BBB-COX-2 ARE transcript. Thus immediately after the transcriptional pulse ARE reporter mRNA had an estimated poly(A) tail length of ϳ200 nt. In the absence or presence of SB 203580, the poly(A) tail was shortened to ϳ50 nt, followed by rapid degradation of the mRNA body, with no further decay intermediates detected (Fig. 2). The rates of decay of the hypo-adenylated transcript appeared very similar in the absence or presence of the inhibitor (Fig.  2C); however, rates of deadenylation were markedly different (Fig. 2D). In the absence of SB 203580, it took ϳ4 h for the mRNA to become hypo-adenylated and begin rapid decay, whereas in the presence of SB 203580 deadenylation was essentially completed by 2 h.
Similar results were also obtained for cells that had not been transfected with the MKK6E expression plasmid. In these, the BBB-COX-2 ARE mRNA signal was much weaker, but a long exposure showed that deadenylation was rapid, and following an initial lag phase the mRNA also decayed rapidly (Fig. 3A). The decay of reporter mRNA following the lag phase was slightly faster than in cells transfected with pCMV-MKK6E and treated with SB 203580 (Fig. 3B). Because deadenylated BBB-COX-2 ARE mRNA decayed more slowly in cells transfected with the MKK6E expression plasmid than in cells lacking it, p38 MAPK activation may make a small contribution to the overall stabilization by slowing the decay of the mRNA body.
Differences p38 MAPK Inhibits ARE-directed Deadenylation differences in mobility of BBB-COX-2 ARE mRNA (Figs. 2 and 3) were due to differences in poly(A) tail length, the RNA was treated with RNase H and oligo(dT) prior to electrophoresis to remove the poly(A) tails of the mRNA. RNase H/oligo(dT) treatment of RNA from cells transfected with or without pCMV-MKK6E and isolated immediately after the transcriptional pulse caused the reporter mRNA to increase in mobility on the gel (Fig. 4). Reporter mRNA that was treated with RNase H and oligo(dT) and which now lacked a poly(A) tail migrated to the same extent (Fig. 4). This confirms that the differences in mobility seen in Figs. 2 and 3 are due to different poly(A) tail lengths and not to shortening of the mRNA from the 5Ј end. The mobility of the ARE reporter mRNA isolated from pCMV-MKK6E-transfected cells 6 h after tetracycline addition also increased slightly following RNase H/oligo(dT) treatment. This confirms that the most deadenylated intermediates of BBB-COX-2 ARE mRNA (Fig. 2D) still possess a short poly(A) tail.
Deadenylation of BBB mRNA That Lacks an ARE Is Not Regulated by p38 MAPK-It was necessary to check that the regulation of deadenylation by p38 MAPK was specific for mRNAs that are stabilized by p38 MAPK. To do this the effect of p38 MAPK activation on the deadenylation of the parental BBB reporter mRNA was examined. In cells transfected with pTet-BBB, either with or without pCMV-MKK6E, the reporter mRNA was deadenylated at an identical rate (Fig. 6A), showing that p38 MAPK specifically inhibits ARE-directed deadenylation and not deadenylation in general. The BBB mRNA underwent very slow shortening of the poly(A) tail, and no decay was detected over the 12-h time course (Fig. 6A). Examination of BBB mRNA at later times showed that no further deadenylation occurred after 12 h (data not shown). In experiments in which the time course was extended to 24 h following tetracycline addition, no decay of the mRNA was discernible (data not shown). The marked difference in stability between deadenylated BBB mRNA and deadenylated ARE-containing reporter mRNA (Figs. 2 and 3) strongly suggests that in addition to directing rapid deadenylation, the ARE also promotes decay of the mRNA body. The BBB mRNA levels increased about 2-fold upon p38 activation (Fig. 6A). This is likely to arise from a slightly elevated rate of transcription upon activation of the p38 MAPK pathway.
Deadenylation Directed by the c-Myc ARE Is Not Regulated by p38 MAPK-To check that p38 MAPK only regulated deadenylation of mRNAs that contain p38 MAPK-sensitive AREs, we investigated the effect of activating the kinase on deadenylation directed by the c-Myc ARE. The c-Myc 3Ј-UTR contains four dispersed AUUUA motifs and several long stretches of   (37), and c-Myc 3Ј-UTRs (C) (27), that were inserted into the unique BglII site in the 3Ј-UTR of pTet-BBB are shown. Note that although the c-Myc 3Ј-UTR contains an internal polyadenylation signal, the mobility of pTet-BBB c-Myc 3Ј-UTR mRNA indicated that the entire insert and ␤-globin 3Ј-UTR was transcribed in the HeLa tet-off cells.
p38 MAPK Inhibits ARE-directed Deadenylation uridylates (Fig. 5C). When inserted into the ␤-globin 3Ј-UTR it causes instability but does not respond to p38 MAPK activation (27). As shown previously, BBB-c-Myc 3Ј-UTR mRNA decayed rapidly following blockade of transcription, and its stability was unchanged in cells co-transfected with pCMV-MKK6E (Fig. 6B). As for BBB-COX-2 ARE mRNA (Fig. 2), decay only commenced once the BBB-c-Myc mRNA was deadenylated. The smaller shift in mobility upon deadenylation seen here compared with that seen in Figs. 2 and 3 is due to the larger size of this reporter mRNA because of the insertion of 442 nt spanning the entire c-Myc 3Ј-UTR. Deadenylation of this mRNA was rapid and was not regulated by p38 MAPK (Fig. 6B), showing that p38 MAPK-regulated deadenylation is specific for only some AREs. As seen for BBB mRNA, the levels of the c-Myc reporter mRNA increased upon p38 MAPK activation (Fig. 6B), again possibly due to some increase in transcription.
p38 MAPK Inhibits Deadenylation Directed by the TNF ARE-Similarly to the ARE of COX-2, that of TNF directs instability and is also regulated by p38 MAPK (37). To see if p38 MAPK also regulates deadenylation directed by the TNF ARE, pTet-BBB-TNF ARE (containing 44 nt of the human TNF ARE (Fig. 5)) was transfected into cells, with or without pCMV-MKK6E. This ARE reporter mRNA underwent rapid poly(A) shortening and decay in the absence of p38 MAPK activation, with maximal deadenylation being reached by 90 min (Fig. 7, A  and B). In the presence of p38 MAPK activity deadenylation of the reporter mRNA was slower, taking about 4 h to complete (Fig. 7, A and C) after which it decayed. The kinetics of BBB-TNF ARE mRNA decay were almost identical to those of BBB-COX-2 ARE mRNA (Fig. 7C), and as before, decay was more pronounced after an initial period of poly(A) shortening. Thus p38 MAPK not only regulates deadenylation directed by the COX-2 ARE but may provide a more general mechanism for mRNA stabilization.

DISCUSSION
The aim of this work was to establish whether p38 MAPK regulates the stability of ARE-containing mRNAs by controlling deadenylation or the decay of the mRNA body. We found that p38 MAPK stabilizes reporter mRNAs containing either the COX-2 or TNF AREs by inhibiting deadenylation. The results are consistent with the observation in LPS-treated RAW 264.7 cells that the mobility of TNF mRNA increased slightly upon inhibition of p38 MAPK (37). Deadenylation of COX-2 mRNA is also stimulated by dexamethasone (39) that is now known to inhibit p38 MAPK.
It is generally believed that the major default mRNA decay pathway in eukaryotes begins with shortening of the poly(A) tail (40). c-Fos mRNA is a good example for which the poly(A) tail is shortened from about 200 to ϳ20 nt and the mRNA is then rapidly degraded in the absence of any other detectable intermediates (6). The initial poly(A) tail length of the COX-2 mRNA reporter in this study was about 200 nt. The poly(A) tail underwent shortening to ϳ50 nt before decay commenced. This is demonstrated by the lag phases that occurred before the onset of mRNA decay: 4 h in cells transfected with MKK6E and 1 h in MKK6E-transfected cells that were treated with p38 MAPK inhibitor. The extended lag phase in cells transfected with MKK6E shows that the inhibition of deadenylation makes the major contribution to stabilization of the reporter mRNA by p38 MAPK.
Following deadenylation, BBB-COX-2 ARE mRNA decayed at the same rate in cells transfected with MKK6E in the absence or presence of SB 203580. This is consistent with p38 MAPK regulating only deadenylation and not the degradation of the body of the transcript. By comparison, the deadenylated

p38 MAPK Inhibits ARE-directed Deadenylation
reporter mRNA decayed more rapidly in cells that lacked pCMV-MKK6E. However, cells containing pCMV-MKK6E were subjected to a continuous high level of p38 MAPK activation which in itself may affect the general degradative mechanisms. Nevertheless, we cannot exclude the possibility that p38 MAPK may partly stabilize mRNA by inhibiting decay of the body. In contrast, there was a very large difference in stability between the deadenylated forms of reporter mRNAs that contained or lacked an ARE. This shows that unlike p38 MAPK, which mainly regulates deadenylation, the ARE regulates both deadenylation and decay of the mRNA body in vivo. This is consistent with the observation that AREs accelerate 3Ј 3 5Ј decay of the body of the mRNA by the exosome in vitro (12).
We note that although it was possible to quantify accurately reporter mRNA levels in cells lacking MKK6E, it was much more difficult to follow deadenylation of BBB-COX-2 ARE mRNA. In these cells the reporter mRNA signal was weak and diffuse. The diffuse bands suggested that deadenylation in these cells might be asynchronous (i.e. occurring at different rates for different transcripts), producing a range of poly(A) tail lengths. This was in contrast to the discrete bands of mRNA from cells transfected with pCMV-MKK6E that were indicative of a simultaneous (or synchronous) process. However, the population distributions determined by PhosphorImager analysis were identical in cells transfected with or without MKK6E, showing that in both cases deadenylation was synchronous.
It is unclear which enzyme is responsible for poly(A) shortening in mammalian cells. A poly(A)-specific ribonuclease has been purified from calf thymus (8) and has very similar enzymatic properties to the activity in HeLa cells (9). In yeast there are two deadenylase complexes, Pan2p/Pan3p (41,42) and Ccr4p/Pop2p/Notp (43). Ccr4p is now thought to be the major deadenylase in yeast (43), and a human homologue has been cloned (10). It is possible that p38 MAPK could directly regulate the deadenylase; alternatively, the kinase pathway could target poly(A)-binding protein (PABP), the protein which is thought to stabilize mRNA by protecting the poly(A) tail. Recently, PABP has been shown to be phosphorylated by MAP-KAPK-2 in vitro (44). However, it is not obvious how either of these mechanisms would confer specificity for AREs. It has been suggested that PABP, in addition to binding poly(A) RNA, can bind the GM-CSF ARE; however, the specificity of the latter interaction is not known (44).
It seems more likely that specificity is achieved by ARE-BPs. In electrophoretic mobility shift assays using cell extracts, the ARE-BP HuR binds both the COX-2 and TNF AREs with high affinity and specificity (31)(32)(33)(34). The COX-2 ARE also binds AUF2 (and probably AUF1) in electrophoretic mobility shift assay of HeLa cell extracts (32). Both of these ARE-BPs bind the TNF ARE tightly in recombinant form (32,45). HuR overexpression stabilizes ARE-containing mRNAs but appears to do so by stabilizing deadenylated mRNA intermediates (35,36). Immunodepletion of the Xenopus orthologue of HuR from oocyte extracts had no effect on the ARE-directed deadenylation of an exogenous RNA in vitro (46). Recently, TTP has been shown to promote ARE-directed deadenylation in a cell-free system (29). However, in cells lacking TTP, such as the HeLa cells used in this study, it is unknown which factor is responsible for targeting ARE-containing mRNAs for rapid deadenylation or for inhibition of deadenylation by p38 MAPK.
It has been suggested that apart from destabilizing mRNAs, AREs such as that in the TNF 3Ј-UTR also act as cis-acting repressors of translation (47). Early work showed that inhibition of p38 MAPK in lipopolysaccharide (LPS)-treated RAW 264.7 cells had little effect on TNF mRNA levels but resulted in dissociation of TNF mRNA from polysomes and its association with single ribosomes (48). This suggested a role for the p38 MAPK pathway in translational initiation. These findings were substantiated by the fact that TNF mRNA can be induced by LPS in MAPKAPK-2-null mice but is inefficiently translated (49).
However, in LPS-treated human monocytes, inhibition of p38 MAPK results in a commensurate decrease in both TNF mRNA and protein levels (1). p38 MAPK also regulates TNF mRNA stability in both human monocytes (50) and RAW 264.7 cells (37). Despite no apparent effect on TNF mRNA stability in macrophages from the MAPKAPK-2-null mice, IL-6 mRNA was found be destabilized (49).
The presence of a long poly(A) tail on an mRNA greatly enhances its translation. The poly(A) tail interacts with the 5Ј cap through protein-protein interactions to synergistically increase translation (51). The translation initiation complex eIF-4F binds the 5Ј cap and interacts with PABP on the poly(A) tail, the mRNA forming a closed loop (52). By maintaining the mRNA in a closed loop, the poly(A) tail is thought to promote translation by allowing ribosomes to circulate more efficiently around the circular mRNA. Thus, regulation of deadenylation by p38 MAPK provides a mechanism to control both mRNA stability and translation.