p38α Stabilizes Interleukin-6 mRNA via Multiple AU-rich Elements*

AU-rich elements (AREs) in the 3′-untranslated region (UTR) of unstable mRNA dictate their degradation or mediate translational repression. Cell signaling through p38α MAPK is necessary for post-transcriptional regulation of many pro-inflammatory cytokines. Here, the cis-acting elements of interleukin-6 (IL-6) 3′-UTR mRNA that required p38α signaling for mRNA stability and translation were identified. Using mouse embryonic fibroblasts (MEFs) derived from p38α+/+ and p38α–/– mice, we observed that p38α is obligatory for the IL-1-induced IL-6 biosynthesis. IL-6 mRNA stability is promoted by p38α via 3′-UTR. To understand the mechanism of cis-elements regulated by p38α at post-transcriptional level, truncation of 3′-UTR and the full-length 3′-UTR with individual AUUUA motif mutation placed in gene reporter system was employed. Mutation-based screen performed in p38α+/+ and p38α–/– mouse embryonic fibroblast cells revealed that ARE1, ARE2, and ARE5 in IL-6 3′-UTR were targeted by p38α, and truncation-based screen showed that IL-6 3′-UTR-(56–173) was targeted by p38α to stable mRNA. RNA secondary structure analysis indicated that modulated reporter gene expression was consistent with predicted secondary structure changes.

Interleukin-6 (IL-6) 2 is a multifunctional cytokine produced by lymphocytes, macrophages, fibroblasts, synovial cells, endothelial cells, glia cells, and keratinocytes (1). IL-6 expression was induced by a variety of stimuli, including interleukin-1(IL-1), tumor necrosis factor (TNF), platelet-derived growth factor, and lipopolysaccharide (LPS) (1). Evidence exists for the role of IL-6 in various diseases, including inflammation and cancer. For example, constitutive overexpression of IL-6 by synovial tissues of rheumatoid patients has been reported (2) and the role of IL-6 in human malignancy is most clearly established in multiple myeloma (3) where monoclonal antibodies directed against IL-6 enhance the effectiveness of chemotherapy in this disease. There is also evidence that IL-6 acts as an autocrine growth factor in a number of human epithelial malignancies, including renal, lung, and prostate cancer (4 -6). To clarify the mechanism involved in the abnormal expression of IL-6, it is imperative to investigate the mechanism of the IL-6 gene expression under physiological conditions.
The production of IL-6 is under transcriptional and posttranscriptional control (7). The transcriptional regulation of IL-6 expression has been thoroughly investigated. IL-1-responsive element was mapped within both the murine and human IL-6 promoter region, whereas regulation at the post-transcriptional level has not been as thoroughly investigated (8,9). Posttranscriptional mechanisms controlling pre-mRNA splicing and maturation, as well as mRNA transport, turnover, and translation, critically influence gene expression programs in mammalian cells. Central to the post-transcriptional regulatory events is the interaction of RNA with RNA-binding proteins that influence their splicing, localization, stability, and association with the translation machinery (10 -12). Many ribonucleoprotein complexes that govern mRNA stability and translation in response to various stimuli (e.g. developmental, stress-inducing, immune, and proliferative) are composed of transcripts that bear adenine and uridine (AU)-rich elements (AREs) in the 3Ј-untranslated region (-3Ј-UTR) and the RNA-binding proteins (ARE-RNA-binding proteins) (13,14). Many ARE-RNA-binding proteins have been described that regulate the stability of target mRNAs, their translation, or both processes: AU-binding factor 1, tristetraprolin, K homology splicing-regulatory protein, butyrate response factor-1, the Hu protein family (HuR, HuB, HuC, and HuD), T-cell-restricted intracellular antigen-1, and the T-cell-restricted intracellular antigen-1-related protein TIAR (15)(16)(17)(18)(19)(20)(21)(22).
In recently years, significant information has accumulated relative to the role of AREs of pro-inflammatory gene as the targets of the mitogen-activated protein kinase (MAPK) p38 pathway, and many of the effects appear to be mediated by its substrate, MAPK-activated protein kinase 2 (MK2) (23,24). It has been shown that p38 MAPK/MK2 cascade is involved in regulating mRNA stability via 3Ј-UTRs of TNF, IL-8, granulocyte macrophage-colony stimulating factor, Cox-2 and vascular epidermal growth factor mRNA (25)(26)(27). In the MK2 knockout mouse strain, LPS-induced expression of IL-6 was blocked at both protein and mRNA levels, while expression of both TNF␣ and interferon ␥ was blocked at the protein but not the mRNA level (24). For IL-1-induced IL-6 biosynthesis, the role of p38 is involved in vivo at post-transcriptional level is poorly understood.
Some p38␣ inhibitor and overexpression data indicate that p38␣ is involved in the regulation of IL-6 production at the post-transcriptional level (26). However, no cis-elements in IL-6 3Ј-UTR targeted by p38␣ have been conclusively identified. To avoid some of the potential artifacts associated with inhibitor and overexpression experiments, we have performed the analysis of IL-6 production in stable fibroblast cell lines derived from p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ embryos. We found that p38␣ is critical for the IL-1-induced IL-6 production and mRNA stability of IL-6 is promoted by p38␣ via the IL-6 3Ј-UTR. Deletion and mutation analysis identified three ARE elements that require p38␣ signaling, and IL-6 3Ј-UTR-(56 -173) is critical for p38␣ to promote mRNA stability. We have further analyzed the secondary structure of IL-6 3Ј-UTR and mutants by using a computational approach. Interestingly, the data indicated that p38␣ targeting IL-6 mRNA required the secondary structure of wild-type 3Ј-UTR.
Reagents-IL-6 enzyme-linked immunosorbent assay kit and recombinant mouse and recombinant mouse IL-1␤ and TNF␣ were purchased from R&D Systems. Lipopolysaccharide from Escherichia coli strain 0127:B8 was purchased from Sigma. Dual-Luciferase Reporter Assay System was purchased from Promega. SB203580 was from Calbiochem. Actinomycin D was from Invitrogen. Assays-on-Demand Gene Expression Products (mIL-6 and mGAPDH) and Taq Man Universal PCR Master Mix were from Applied Biosystems.
Luciferase Assay-Luciferase activity was determined using a luciferase assay system, following the manufacturer's protocol (Promega, Madison, WI) Briefly, cell monolayers in 12-well plates were removed by scraping into 200 l of reporter lysis buffer. Cells were vortexed, and cellular debris was removed by centrifugation (30 s at 12,000 ϫ g). Luciferase activity was measured using a luminometer (LMaxII 384, Molecular Devices). A Renilla luciferase reporter vector was included in every experiment for transfection efficiency control. Relative luciferase activity was determined and normalized to Renilla luciferase activity.
Northern Blot-RNA was prepared from actively growing cells by using the TRIzol reagent (Invitrogen) as specified by the manufacturer. 10 g of RNAs was separated in 1.25% Seakem gold-agarose, transferred onto a nylon membrane (GeneScreen Plus, Dupont), and cross-linked by UV treatment (Stratalinker apparatus). The membrane was hybridized in ULTRA Hyb solution (Ambion), using 32 P-labeled firefly luciferase probe (nucleotides 491-1049). The mRNA levels were normalized by 28 S RNA.
Reverse Transcription-PCR and Quantitative Real-time PCR-IL-6 mRNA expression was analyzed by reverse transcription-PCR and quantitative real-time PCR. First strand cDNA was synthesized from RNA (600 ng) using SuperScript III reverse transcriptase (Invitrogen). First strand cDNA was used for PCR with specific oligonucleotide primers for mIL-6: forward, 5Ј-ATGAAGTTCCTCTCTGCAAGAGACT-3Ј; reverse, 5Ј-CACTAGGTTTGCCGAGTAGATCTC-3Ј. Quantitative real-time PCR with primers were designed by Applied Biosystems (mIL-6, mm00446190; mGAPDH, mm99999915) using an ABI7500 thermocycler, and the gene expression was analyzed by using 7500 System SDS v1.4 software.
RNA Secondary Structure and Analysis-RNA secondary structure analysis was performed using the M-fold program (28). The folding temperature was fixed at 37°C.
Statistical Analysis-Student t tests were utilized to obtain individual p values using GraphPad Prism version 4.03 for Windows, GraphPad Software, San Diego, CA.

JOURNAL OF BIOLOGICAL CHEMISTRY 1779
To confirm that this defect was caused by p38␣ deficiency rather than genetic mutations associated with IL-1R signaling, wild-type p38␣ and kinase-dead p38 (AF) cDNAs were transiently transfected into MEFs, and IL-6 production induced by IL-1␤ measured. Results indicate that wild-type p38␣ but not p38AF restored the IL-6 production in p38␣ Ϫ/Ϫ MEFs (Fig.  1B). In contrast, overexpression of p38␣ in p38␣ ϩ/ϩ MEFs produced IL-6 in amounts similar to those of parental control cells. As expected, overexpression of kinase-dead p38AF decreased the IL-6 production in p38␣ ϩ/ϩ MEFs. Interestingly, IL-6 production in p38␣ Ϫ/Ϫ MEFs was also decreased by p38AF overexpression. These data strongly indicate that p38␣ is critical for IL-1R-induced IL-6 production.
p38␣ Stabilizes Endogenous IL-6 mRNA-Previous data have indicated that the stability of IL-6 mRNA is regulated by MK2 via an AU-rich 3Ј-UTR of the IL-6 mRNA (29). To confirm whether p38␣ signaling was required for IL-6 mRNA turnover, the stability of endogenous IL-6 mRNA transcripts were monitored by actinomycin D pulse chase. After 24 h of IL-1␤ treatment with or without SB203580 (p38␣ inhibitor), actinomycin D was added to p38␣ ϩ/ϩ MEFs, and IL-6 mRNA levels were measured at the indicated time points. The addition of SB203580 resulted in a rapid decrease in IL-6 mRNA level ( Fig.  2A). The density was measured and shown in Fig. 2B. The analysis revealed that SB203580 decreased IL-6 mRNA t1 ⁄ 2 from 3.7 to 1.7 h. Importantly, the stability of endogenous IL-6 mRNA was compared in the p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEFs. After MEFs were stimulated by IL-1␤ for 24 h, actinomycin D was added, and IL-6 mRNA levels were measured by real-time PCR. Fig. 2C shows a relative shorter half-life in p38␣ Ϫ/Ϫ MEFs (t1 ⁄ 2 ϭ 1.9 h) compare with p38␣ ϩ/ϩ cells (t1 ⁄ 2 ϭ 4.6 h) indicating that IL-6 mRNA is stabilized by p38␣.
p38␣ Targets IL-6 3Ј-UTR to Promote mRNA Stability-Because AREs in the 3Ј-UTR of many cytokine mRNAs are responsible for both translational control and mRNA stability, we examined whether the ARE-containing 3Ј-UTR of IL-6 is sufficient to confer p38-dependent stabilization to a reporter mRNA. The full length of 3Ј-UTR of IL-6 contains five AREs with no overlapping pentanucleotide AUUUA core motifs. Instead, the AUUUA motifs are scattered throughout the IL-6 3Ј-UTR where both the human and mouse IL-6 3Ј-UTR exhibit high homology to each other (30).
To clarify the functional role, the full-length 3Ј-UTR was inserted into the 3Ј-UTR of the luciferase reporter gene and transiently transfected into p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEF cells.  A, p38␣ ϩ/ϩ MEFs were seeded into 60-mm plates. After 24 h of IL-1 ␤ (1 ng/ml) stimulation with or without SB203580 (5 M), actinomycin D (10 g/ml) was added, and total RNA was harvested at indicated time points. Reverse transcription-PCR for IL-6 and GAPDH were performed as described under "Experimental Procedures." B, graphical depiction of IL-6 mRNA decay from panel A normalized to GAPDH. C, p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEF cells were stimulated by IL-1 ␤ (1 ng/ml) for 24 h. Actinomycin D (10 g/ml) was added, and the mRNA level of IL-6 was measured by quantitative real-time PCR. Data represent the average of three independent experiments.
Proximal AREs Contain Elements Targeted by p38␣-To investigate the role of individual AUUUA motifs in mediating post-transcriptional control regulated by p38␣, a series of luciferase reporter-gene constructs, containing the various regions of IL-6 3Ј-UTR and the motifs mutated from AUUUA to AGGGA, were generated (Fig. 4A) and transiently expressed in MEFs. The luciferase activity of the full-length 3Ј-UTR and all truncations was compared first in the individual cell type. For the p38␣ ϩ/ϩ MEFs, when the full-length 3Ј-UTR-(1-403) was placed in the reporter message, luciferase activity decreased more than 30% compared with luciferase alone (Fig. 4B). These data indicate that the 3Ј-UTR contains the elements that confer mRNA stability and/or translational efficiency. The pGL3-IL-6ARE-(56 -173) truncation reporter includes the two proximal AREs (ARE1 and ARE2), whereas the three distal AREs (ARE3, ARE4, and ARE5) are included in pGL3-IL-6 3Ј-UTR-(172-403). The decreased expression was lost in both pGL3-IL-6 3Ј-UTR-(56 -173) and pGL3-IL-6 3Ј-UTR-(172-403) compared with the full-length 3Ј-UTR-(1-403), suggesting that multiple control elements were operative. Whereas in the p38␣ Ϫ/Ϫ MEFs, the luciferase activity of the full-length 3Ј-UTR decreased more than 50% compared with luciferase alone and the significant decreased expression still occurred in the presence of both proximal and distal truncations. This result indi-cated that p38␣ was involved when these multiple elements mediated post-transcriptional control. When the truncation pGL3-IL-6 3Ј-UTR-(1-70), which includes none of the AREs was inserted in the reporter message, luciferase activity in both wild-type and p38␣ Ϫ/Ϫ MEFs was comparable with the control. These data indicate that the decreased expression only occurred in the presence of AUUUA motifs.
To estimate the influence of the individual AREs, AUUUA motifs were mutated individually and transiently expressed in MEFs. In p38␣ ϩ/ϩ MEFs, the luciferase activity of Mu3 increased significantly compared with full-length 3Ј-UTR (p Ͻ 0.05) but still was significantly lower than that of control (p Ͻ 0.05). For the Mu5, the luciferase activity decreased another 20% compared with full-length 3Ј-UTR. This result indicated that ARE3 included the major repressive element, whereas ARE5 bears an enhancive element. In the p38␣ Ϫ/Ϫ MEFs, the luciferase activity of Mu1-4 increased significantly compared with full-length 3Ј-UTR, whereas Mu5 was comparable.
Again, the luciferase activity of all mutants between p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEFs was compared. After the individual ARE motifs mutated, the significant difference due to the p38␣ deficiency was lost for M1, M2, and M5, whereas the luciferase activity of M3 and M4 were still significantly lower in p38␣ Ϫ/Ϫ MEFs compared with p38␣ ϩ/ϩ cells suggesting that ARE1, -2, and -5 were the elements targeted by p38␣.
Secondary structure prediction of IL-6 3Ј-UTR and mutants. Although RNA-binding proteins bind RNA in a sequencespecific manner, previous studies have indicated that RNA secondary structure plays a critical role in defining 3Ј-UTR RNA-protein interaction. Binding sites are often located in single-stranded RNA region (31). Using M-fold, the secondary structures of IL-6 3Ј-UTR and mutants were predicted (Fig. 6). For wild-type 3Ј-UTR, ARE2 and ARE5 comprised the sequence to form the external loop structures (LoopA and LoopD), whereas ARE3 and ARE4 formed the internal loop structures (LoopB and LoopC). ARE1 formed the stem adjacent to the external LoopA.
Mutation of ARE1 and ARE2 abolished external LoopA, and mutation of ARE5 abolished external LoopD, whereas mutation of ARE3 and ARE4 had a relatively minor effect on the secondary structure. Consistent with minor changes in secondary structure, mutation of ARE3 and ARE4 to luciferase reporter gene assay did not abolish the effect of p38␣. This indicates that the secondary structure of wildtype 3Ј-UTR is necessary for the p38␣ targeting.

DISCUSSION
TNF␣ remains the best demonstrated cytokine regulated by the p38␣ MAPK pathway at post-transcriptional level. The tight control of TNF␣ expression is achieved by regulating mRNA stability, translational initiation, and polyadenylation. During this process, transacting factors that bind to TNF␣ 3Ј-UTR, including AREs are essential. In contrast, limited information is available regarding the nature of IL-6 mRNA stability and regulation by MAPK p38␣. Using a p38␣-deficient cell line, strict criteria regarding signaling intermediate requirements for IL-6 expression and mRNA stability could be applied. According to previous p38␣ inhibitor experiments, we predicted that IL-6 expression was down-regulated in p38␣ Ϫ/Ϫ MEFs at posttranscriptional level. Consistent with our expectation, IL-6 mRNA decay increased in p38␣ Ϫ/Ϫ MEFs compared with p38␣ ϩ/ϩ cells.
In the present study, MEF cells treated with IL-1 resulted in phosphorylation of p38␣ to produce IL-6. Fibroblast IL-6 production was inhibited by SB203580 (26). To determine p38␣ is critical for the IL-1-induced IL-6 production by using p38␣ Ϫ/Ϫ cells, the normal expression of IL-1 receptor in p38␣ Ϫ/Ϫ cells is required. It has been reported that p38␣ Ϫ/Ϫ ES cells also gave rise to a population of IL-1R-positive cells after the in vitro differentiation process (32). The percentage of cells that express IL-1R was comparable in p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ cultures. Our data show that reconstitution of p38␣ cDNA in p38␣ Ϫ/Ϫ cells restored the IL-1␤-induced IL-6 production consistent with that the decrease of IL-6 production is due to the p38␣ deficiency rather than the defect of IL-1R. p38␣ is critical but not essential for the IL-6 production. There-p38␣ Regulates IL-6 Gene Expression at Post-transcriptional Level fore, failure to increase IL-6 production by overexpression p38␣ in wide-type cells was not unexpected. It is possible that other proteins are limiting the involvement of p38␣. For example, MAPK/extracellular signal-regulated kinase kinase kinase 3 (MEKK3), which is the potent activator of MAPKs, is essential for the IL-1-induced IL-6 production (33). However, it was somewhat surprising that p38AF decreased the IL-6 production in both p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ cells. Four members of the p38 MAPK family have been cloned: p38␣, -␤, -␥, and -␦ (34,35). p38␤ shares ϳ74% sequence identity with p38␣ and contains the TGY dual phosphorylation motif observed in p38␣ (36). p38␤ is activated by IL-1 stimulation, but 4-folder lower than p38␣ (37). In addition, MK2 is the substrate of both p38␣ and p38␤ (37). Therefore, we predict that p38AF decreased the activation of MK2 by the co-repression of p38␤.
In our present study, no significant effects of p38␣ deficiency on IL-6 proximal promoter reporter activity were demonstrated (data not shown). This result is consistent with previous reports from our group, where p38 inhibitors had minimal effects on IL-6 proximal promoter in MC3T3-E1 osteoblastic-like cells (38). Others have reported that SB203580 did not effect the activation of the IL-6 promoter by IL-1␤ in human fibroblast-like synoviocytes (26). Collectively, these results indicate that the effect of p38␣ deficiency on IL-6 gene expression results from the down-regulation of post-transcriptional level, however, we cannot rule out any role of transcriptional activation in these present studies.
SB203580 significantly reduced the mRNA stability of endogenous IL-6. The results were also confirmed in vivo by comparing half-life of IL-6 mRNA between p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEFs. To analyze the regulation of IL-6 mRNA stability, the luciferase reporter gene system was employed. We found that p38␣ stabilized the mRNA reporter construct carrying the IL-6 3Ј-UTR. All these data supported the conclusion that p38␣ signaling increases IL-6 mRNA stability via 3Ј-UTR. The degradation of endogenous IL-6 mRNA and mRNA reporter construct containing the IL-6 3Ј-UTR is only marginally affected in p38␣ ϩ/ϩ MEFs. This is probably due to the less classic ARE and perhaps weaker IL-6 ARE elements of the IL-6 3Ј-UTR compared with TNF␣ and other cytokines. TNF␣, granulocyte macrophage-colony stimulating factor, and IL-3 AREs are typical class II AREs with a core AUUUA motif cluster, which mediated rapid mRNA decay (30). The AREs of c-myc and c-fos are prototypes of class I, containing one to three scattered copies of the AUUUA motif, whereas other cytokine transcripts, including IL-2, IL-4, and IL-6 contain class I-like AREs that are non-clustered (30). These non-clustered motifs are considered weaker AREs. The differences in cells types and stimulus may also account for this minimal degradation. In the present study, we observed only modest changes with IL-6 ARE reporter activity following stimulation with IL-1␤ (data not shown). These data are consistent with the concept that the IL-6 ARE is weaker compared with other AREs or alternatively there are differences in the MEF cells lines used in these studies.
A paradigm that has been helpful in understanding the regulation of post-transcriptional control was that RNA-binding protein binds to specific cis-acting elements in target transcript and activate 3Ј-to-5Ј-exonucleolytic decay or/and influence the translational control. Some RNA-binding proteins, like tristetraprolin, AU-binding factor 1, and K homology splicing-regulatory protein, promote mRNA decay, whereas others, like members of the Hu family, prevent mRNA degradation (15-17, 19, 20). Deletion or mutation of AUUUA motifs could alter the components of mRNA-trans factor complex and affect the luciferase activity of reporter gene containing the IL-6 3Ј-UTR with mutated AUUUA motifs. Using this approach, we identified ARE1, ARE2, and ARE5 as the cis-elements regulated by p38␣ in IL-6-3Ј-UTR. Interestingly, mutation of ARE1 and ARE2 resulted in the enhanced expression due to the deficiency of p38␣, whereas mutation of ARE5 decreased the luciferase expression in the presence of p38␣. These data are consistent FIGURE 4. Reporter-gene assay identifies p38␣-dependent elements in the 3-UTR of IL-6. A, schematic representation of IL-6 3Ј-UTR deletion and ARE mutation reporter constructs. B and C, p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEFs were transiently transfected with pGL3promoter, pGL3-IL6-3Ј-UTR, or pGL3-IL-6 3Ј-UTR vectors containing the indicated deletion and ARE mutants of UTR shown in A. Renilla luciferase (pEF-RLuc) vector was co-transfected as transfection efficiency control. Significant differences in reporter expression between p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ is indicated (*, p Ͻ 0.05; **, p Ͻ 0.01). The values represent from three independent experiments, and each experiment was performed in duplicate. with the idea that p38␣ is required for the ARE1 and ARE2 to repress expression and for ARE5 to enhance expression.
Our data indicated that distinct cis-elements of 3Ј-UTR exert opposing influence on the post-transcriptional fate of IL-6 mRNA. A similar bi-functional role for the 3Ј-UTR has been described for other mRNAs, including RNase-L (39). In that study, eight AREs were identified in the RNase-L 3Ј-UTR, and deletion analysis identified AREs 7 and 8 served as a strong positive regulatory function, whereas AREs 2 and 4 are critical for the negative regulatory function. Also, HuR and AU-binding factor 1 were found to bind to p21 and cyclin D1 mRNAs in the nucleus, and independently operate opposite function in the cytoplasm depending on the physiological context (40).
RNA-binding proteins recognize RNA targets in a sequencespecific manner, and the secondary structure context of the binding site also affects the binding affinity (31). Binding sites are often located in single-stranded RNA regions. For example, HuR protein influences mRNA stability by binding to the motif NNTTNNTTT where HuR affinity correlates with the singlestrandedness of this binding motif (41,42). IL-6 3Ј-UTR-(56 -173) displays a typical "stem-loop" or hairpin structure. This structure might be critical for IL-6 3Ј-UTR-(56 -173) inducing rapid mRNA decay.
Recently studies have shown that microRNA targeting of ARE appeared to be an essential step in ARE-mediated mRNA degradation (43). The ability of microRNA to target mRNA is directed by the paring of microRNA to mRNA. It is an intriguing possibility that microRNA is implicated for p38␣ to regulate the repressive elements in IL-6 3Ј-UTR.

CONCLUSION
Cell signaling through p38␣ MAPK is necessary for posttranscriptional regulation of many pro-inflammatory cytokines. In this study, the cis-acting elements of IL-6 3Ј-UTR mRNA that required p38␣ signaling for mRNA stability and translation were identified. AUUUA motif mutation analysis using gene reporter systems performed in p38␣ ϩ/ϩ and p38␣ Ϫ/Ϫ MEF cells revealed that ARE1, ARE2, and ARE5 in IL-6 3Ј-UTR were targeted by p38␣, and IL-6 3Ј-UTR-(56 -173) is critical for p38 to promote mRNA stability. The predicted RNA secondary structure of these ARE reporter mutants was consistent with the ability to alter reporter expression data, FIGURE 6. Major structural differences in IL-6 3-UTR are observed following ARE mutations. Secondary structures of full-length (A) and mutant IL-6 3Ј-UTR (B-F) were predicted using M-fold software. Highlighted regions in B, C, and F indicate major loop change in RNA structure as predicted following ARE1, -2, and -5 mutations.