Regulation of TNF Expression by Multiple Mitogen-Activated Pathways Wei Zhu, Jocelyn S. Downey, Jun Gu, Franco Di Padova, Hermann Gram and Jiahuai Han This information is current as of September 26, 2021. J Immunol 2000; 164:6349-6358; ; doi: 10.4049/jimmunol.164.12.6349 http://www.jimmunol.org/content/164/12/6349 Downloaded from

References This article cites 57 articles, 36 of which you can access for free at: http://www.jimmunol.org/content/164/12/6349.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 26, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Regulation of TNF Expression by Multiple Mitogen-Activated Pathways1

Wei Zhu,2* Jocelyn S. Downey,2* Jun Gu,† Franco Di Padova,‡ Hermann Gram,‡ and Jiahuai Han3*

Stimulating macrophages with bacterial endotoxin (LPS) activates numerous intracellular signaling pathways that lead to the production of TNF. In this study, we show that four mitogen-activated protein (MAP) kinase pathways are activated in LPS- stimulated macrophages: the extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase/stress-activated protein kinase, p38, and Big MAP kinase (BMK)/ERK5 pathways. Although specific activation of a single MAP kinase pathway produces only a modest effect on TNF activation, activation of each MAP kinase pathway is important for full induction of the TNF . Interestingly, a dramatic induction of TNF promoter-driven was observed when all of the four MAP kinase pathways were activated simultaneously, suggesting a cooperative effect among these . Unexpectedly, cis elements known Downloaded from to be targeted by MAP kinases do not play a major role in multiple MAP kinase-induced TNF gene expression. Rather, a 40-bp sequence harboring the TATA box, is responsible for the gene up-regulation induced by MAP kinases. The proximity of the MAP kinase-responsive element to the transcriptional initiation site suggested that MAP kinases regulate the transcriptional initiation complex. Utilizing ␣-amanitin-resistant RNA II mutants with or without a C-terminal domain (CTD) deletion, we found that deleting the CTD to 31 tandem repeats (⌬31) led to >90% reduction in MAP kinase-mediated TNF production. Thus, our data demonstrate coordination of multiple MAP kinase pathways in TNF production and suggest that the CTD of RNA http://www.jimmunol.org/ polymerase II is required to execute MAP kinase signaling in TNF expression. The Journal of Immunology, 2000, 164: 6349– 6358.

umor necrosis factor (TNF or TNF-␣/cachectin) is a lian RNA pol II contains a characteristic C-terminal domain (CTD) proinflammatory cytokine that acts as a of host in its largest subunit, consisting of 52 repeats of the sequence T defense against both neoplasia and infection and is prin- Tyr-Ser-Pro-Thr-Ser-Pro-Ser (7–9). This CTD has been shown to cipally expressed in macrophages (1–4), where its secretion may interact with other RNA pol II subunits, including the TATA-bind- be increased 10,000-fold after exposure to bacterial endotoxin ing protein (8, 10), and it is known to be involved in regulating by guest on September 26, 2021 (LPS) (5). Along with numerous beneficial roles in immune reg- gene because partial deletions in the CTD modulate ulation, TNF has been implicated in the pathogenesis of both acute the regulatory properties of distinct promoters in different ways and chronic inflammatory disease (6), and therefore it is of great (11–16). In addition, the CTD is phosphorylated by cellular ki- interest to dissect the molecular mechanisms of TNF gene nases, and this appears to play a role in regulating expression. initiation and elongation of transcription (17–20). Despite the im- Eukaryotic gene expression is heavily controlled by / portance of RNA pol II and its CTD in transcriptional control, the promoter elements, which act in conjunction with the RNA poly- regulatory role that this holoenzyme plays in LPS-induced TNF merase II (pol II)4 holoenzyme to mediate transcription. Mamma- transcription has not been studied to date. Sequence analysis has revealed a number of cis elements present in the TNF promoter, including several NF-␬B-like motifs, which *Department of Immunology, The Scripps Research Institute, La Jolla, CA 92037; are thought to play a primary role in TNF gene transcription (21). †Center of Molecular Medicine, Sun Yat-Sen University of Medical Science, Guang- However, there is evidence from studies employing deletion con- ‡ zhou, China; and Novartis Pharma, Basel, Switzerland structs in the promoter to indicate that other elements are involved Received for publication June 18, 1999. Accepted for publication March 28, 2000. in TNF regulation (22, 23). In addition, AU-rich elements (ARE) The costs of publication of this article were defrayed in part by the payment of page in the 3Ј-untranslated region (3Ј-UTR) of TNF mRNA are thought charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this . to play a role in mediating mRNA stability and efficiency of trans- lation (24, 25). Therefore, TNF production is controlled at both 1 This work was supported by grants from the U.S. Public Health Service (National Institutes of Health nos. GM51470 and AI41637; to J.H.) and the China Natural transcriptional and posttranscriptional levels. Science Foundation (no. 39730140; to J.G.). W.Z. was supported by a fellowship Exposure of cells to LPS activates a number of signaling path- from the Chinese Medical Board. J.H. is an established investigator of the American ␬ Heart Association. This is publication no. 12269-IMM from the Department of Im- ways, including NF- B, phosphatidylinositol 3-kinase, and protein munology of The Scripps Research Institute. kinase C (26). Previous work from this and other laboratories has 2 W.Z. and J.S.D. contributed equally to this work. found that LPS activates the extracellular signal-regulated kinase 3 Address correspondence and reprint requests to Dr. Jiahuai Han, Department of (ERK) (27), the c-Jun N-terminal kinase (JNK) (28, 29), and the Immunology IMM-9, The Scripps Research Institute, 10550 North Torrey Pines p38 mitogen-activated protein (MAP) kinase pathways (30, 31). Road, La Jolla, CA 92037. E-mail address: [email protected] Selective inhibition of the p38 and ERK pathways inhibits LPS- 4 Abbreviations used in this paper: pol II, polymerase II; BMK, Big MAP kinase; stimulated TNF production (32, 33), and it has also been reported CTD, C-terminal domain; ARE, AU-rich elements; UTR, untranslated region; ERK, extracellular signal-regulated kinase; JNK, c-Jun N-terminal kinase; MAP, mitogen- activated protein; MEF, myocyte-enhancer factor; ATF, activating transcription fac- m, murine; h, human; pol II ⌬31, pol II mutant with a CTD containing only 31 tor; nmp, nonrelated minimal promoter; MEK, MAP/ERK kinase; M⌽, macrophage; repeats.

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00 6350 REGULATION OF TNF BY MITOGEN-ACTIVATED PROTEIN KINASES

that JNK kinases are required for translation of TNF mRNA after cedure to give the best reproducibility in transfection efficiency and results LPS induction (34). However, attempts to induce TNF with Raf I, when between different batches of plasmids. an upstream activator of ERK pathway, indicate that although Transfection ERK activation is sufficient to produce small amounts of TNF, the RAW264.7 cells were maintained in DMEM supplemented with 10% FBS, levels produced are 20 times smaller than those produced in re- 2 mM glutamine, 50 U/ml penicillin, 50 mg/ml streptomycin, and 1% sponse to LPS stimulation of cells (35–37). Studies in glial cells nonessential amino acids. With the exception of the drug-inhibition assay, have also shown that although inhibiting either the ERK pathway transfections were performed on 2 ϫ 106 cells seeded into wells in six-well using the selective drug PD98059 or the p38 pathway using dishes. However, to investigate the effects of MAP kinase inhibitors on a ϫ 7 SB203580 suppresses TNF biosynthesis, inhibiting both ERK and TNF reporter gene, 1 10 cells were transfected, and these cells were then subdivided onto six-well plates 24 h after transfection. Cells were p38 together results in almost complete abolition of TNF biosyn- transfected with 0.6 ␮g of each plasmid per well using calcium phosphate thesis (38). Therefore, it is clear that multiple signals must con- precipitation before glycerol shock (24). Empty pcDNA3 vector was used verge on TNF synthesis to elicit its full response. to normalize the amount of total DNA used in each transfection to 3 ␮g/ To define their role in TNF biosynthesis, we have studied the well. An exception to this was the transfection of pol II expression plas- mids, in which we used 6 ␮g/well for transfection because of the much regulation of the TNF promoter by various MAP kinase pathways. larger size of the plasmids. Pol II-transfected cells were selected with 2.5 We have evaluated the activation and requirement for different mg/ml ␣-amanitin 16 h after transfection. The other transfected cells were MAP kinase pathways, including the Big MAP kinase (BMK)/ processed 24 h after transfection. Some cells either were left unstimulated ERK5 pathway, in LPS responsiveness of the macrophage (M⌽) or were treated with LPS (1 ng/ml) for 8 h. The cells were washed in PBS, ␮ cell line RAW264.7. Selective activation of one, two, three, or four harvested, and resuspended in 100 l of a reporter lysis buffer (Promega, Madison, WI). Lysed cells were briefly centrifuged, and the relative MAP kinase pathways in different combinations was used to de- strength of reporter induction was calculated by measuring the luciferase Downloaded from termine the biological consequences of MAP kinase activation in activity of the supernatant by luminometer in a luciferase assay reagent terms of TNF expression. Although activation of any individual (Promega). Transfection efficiency was normalized by cotransfecting cells MAP kinase had little effect on TNF gene expression, activation of with 0.6 ␮g of an expression plasmid containing a CMV promoter-driven ␤-galactosidase reporter. ␤-Galactosidase activity was measured by using all four MAP kinase pathways produced a dramatic synergistic the chemiluminescent assay Galacto-Light (Tropix, Bedford, MA) or by effect on TNF promoter-driven gene expression. It is further shown using O-nitrophenyl-␤-D-galactopyranoside (ONPG) as follows: 20 ␮lof

that known cis elements such as AP-1 and NF-␬B are not the major lysis supernatant was added to 80 ␮l 3.5 mM ONPG solution, incubated at http://www.jimmunol.org/ target of MAP kinases, whereas the MAP kinases’ responsiveness 37°C for 30 min, and measured at 405 nm. is mediated by an element located in Ϫ43 to Ϫ1 bp immediately Reporter constructs upstream of the transcriptional initiation site. We further demon- strated that the MAP kinase-induced TNF gene expression is, at Murine TNF (mTNF) luciferase reporter was generated from a chloram- phenicol acetyltransferase (CAT) reporter described previously (40). A least in part, mediated by a regulation of RNA pol II CTD. Our 1-kb mTNF promoter was taken from Pro-CAT construct by digesting with results demonstrate a convergence of different MAP kinase path- BamHI and HindIII and was inserted into pGL2 basic using BglII and ways to one regulatory site that may play a key role in TNF HindIII sites. A progressive series of deletions was generated in the murine Ј expression. 5 -TNF promoter using PCR. A series of human TNF (hTNF) reporter

constructs were generously provided by Dr. J. S. Economou (University of by guest on September 26, 2021 California, Los Angeles, CA). 5ЈUTR of hTNF was removed by digesting Materials and Methods with BamHI and KpnI and was replaced with a synthesized oligonucleotide Preparation of recombinant linker. Deletion constructs of Ϫ75, Ϫ65, and Ϫ55 and chimeric constructs C1, C2, C3, and C4 were constructed by oligonucleotide replacement. The Escherichia coli BL21(DE3) was transformed with the vector pETM1 con- constructs were sequenced to confirm their integrity. Oligonucleotide se- taining cDNAs encoding myocyte-enhancer factor 2C (MEF2C). The quences used in PCR and oligonucleotide replacement are available upon BL21 strain of E. coli was transformed with the vector pGEX containing request. NF-␬B and AP-1 luciferase reporters containing a basic promoter cDNAs encoding ELK1 and activating 2 (ATF2). The element (TATA box) joined to a NF-␬B or AP-1 site were obtained from transformed bacteria were grown at 37°C in Luria-Bertani broth until the Stratagene (La Jolla, CA). The nonrelated minimal promoter (nmp) re- ␤ A600 was 0.5, at which time isopropyl- -D-thiogalactopyranoside was porter was constructed using the same backbone but was joined to Ϫ55 to added to a final concentration of 1 mM for 5 h. Cells were collected by Ϫ43 bp of human TNF promoter sequence. Sp1 luciferase reporter was centrifuging at 8000 ϫ g for 10 min, and the bacterial pellet was resus- constructed using the same backbone inserted with three repeats of pended in 10 ml of buffer A (30 mM NaCl, 10 mM EDTA, 20 mM Tris- Sp1 site. HCl, and 2 mM PMSF) for every 100 ml of original bacterial culture. The cell suspension was sonicated, and cellular debris was removed by centri- Other constructs ϫ fuging at 10,000 g for 30 min. Recombinant proteins were purified from The expression constructs for different MAP kinase kinase mutants were the cleared lysate using nickel-nitriloacetic acid (Ni-NTA) purification sys- cloned in pcDNA3 expression vector as described previously (41–45). The tem (Qiagen, Valencia, CA) or glutathione-Sepharose (Amersham Phar- expression constructs of ␣-amanitin-resistant RNA pol II and its CTD de- macia Biotech, Piscataway, NJ) following the manufacturer’s instruction. letion mutants were kindly provided by Hans-Peter Gerber (Genentech, South San Francisco, CA). Protein kinase assays Quantification of reporter mRNA In vitro kinase assays were conducted at 37°C for 30 min using purified immunoprecipitate as the kinase, 5 mg of kinase substrate, 250 ␮M ATP, An RNA isolation kit from Qiagen was used to isolate total RNA from and 10 ␮Ci of [␥-32P]ATP in 20 ml of kinase reaction buffer as described cultured cells according to manufacturer’s instruction. Luciferase mRNA previously (39). Reactions were terminated by the addition of Laemmli was quantified by Taqman (Perkin-Elmer, Foster City, CA). Cotransfected sample buffer. Reaction products were resolved by 12% SDS-PAGE. Phos- ␤-galactosidase was used as an internal control. phorylated proteins were visualized by autoradiography and were quanti- fied by phosphoimaging. Metabolic labeling and immunoprecipitation of transfected pol II and pol II ⌬31 Plasmid preparation Pol II or pol II ⌬31 was cotransfected into 5 ϫ 107 RAW cells with or All plasmid DNA used in transfection experiments was prepared using without the expression plasmids of MAP/ERK kinase 1 (MEK1)(E),

CsCl2-gradient ultracentrifugation. Possible LPS and other bacterial sugar/ MEK5(D), MKK6(E), and MKK7(D) using calcium phosphate precipita- lipid contamination was subsequently removed with Endotoxin Removal tion. Sixteen hours after transfection, these cells were selected with 2.5 Affinity Resin (Associates of Cape Cod, Falmouth, MA). Because differ- ␮g/ml ␣-amanitin for 3 days, after which time surviving cells were replated ences in transfection efficiency may result if DNA quality is not well con- onto 10-cm dishes. Cells were metabolically labeled as previously de- trolled, we have evaluated several methods and have determined this pro- scribed (46). Briefly, the ATP pool of cells will be labeled using The Journal of Immunology 6351

been reported by any other stimulus. This may be important and may be one possible reason for the phlogistic effects of LPS in initiating innate immune responses. Although different signal path- ways may be responsible for different cellular changes initiated by LPS, they may also ultimately converge onto one cellular reaction, such as TNF production. Therefore, we performed the experiments described below to address the role of these MAP kinase pathways in regulating TNF gene expression. Activation of each of the four MAP kinase pathways is required for the full induction of TNF gene FIGURE 1. Activation of multiple MAP kinases in LPS-stimulated Because of the availability of specific inhibitors for the p38 and RAW264.7 cells. RAW cells were stimulated with LPS (1 ng/ml) for dif- ERK pathways, inhibition of these two pathways has been shown ferent periods of time and then the cells were harvested. BMK1, ERK2, to inhibit TNF production in several cell systems (32, 33, 38). Here p38, and JNK1 were immunoprecipitated from cell lysates using specific we used specific inhibitors of the ERK and p38 pathways to test polyclonal Abs. Immunokinase assays were performed using the immuno- precipitates as kinase. Recombinant His-MEF2C, GST-ELK1(307–428), whether a luciferase reporter gene driven by a 1-kb human TNF GST-ATF2(1–109), and GST-ATF2(1–109) were used as substrate for promoter can mimic the responsiveness of the endogenous TNF BMK1, ERK2, p38, and JNK1, respectively. The kinase reactions were gene to MAP kinases in RAW264.7 cells. RAW cells were tran- stopped by SDS-sample buffer, and the reaction products were resolved by siently transfected with TNF reporter gene and then were replaced Downloaded from SDS-PAGE. Phosphorylated proteins were visualized by autoradiography. onto 6-well plates. Twenty-four hours after transfection, the cells Comparable results were obtained in two experiments. were pretreated with different concentrations of SB203580 or U0126 for 30 min and then were stimulated with LPS (1 ng/ml) for 8 h before a luciferase assay was performed on the lysed cell ex- [32P]orthophosphate (1 mCi/ml for 2 h), and recombinant RNA pol II pro- tract. As shown in Fig. 2A, the inhibitor of p38␣␤, SB203580,

teins were immunoprecipated using anti-hemagglutinin (anti-HA) Ab (Co- inhibited LPS-induced TNF reporter gene expression in a dose- http://www.jimmunol.org/ vance, Richmond, CA). SDS-PAGE was performed on the immunopre- ϳ ␮ dependent manner to give an IC50 of 18 M. U0126, an inhibitor cipitates, and the dried gel was exposed on a phosphoimaging cassette for of MEK1/2 (47), also inhibited TNF reporter gene expression 2 days. ϭ ␮ (IC50 18 M). These IC50 are similar to those for endogenous Western blotting TNF in this cell line (data not shown). A further inhibition of the Cells were rapidly chilled on ice, washed twice with ice-cold washing reporter synthesis was observed when both inhibitors were present (Fig. 2A), the same as previously reported for endogenous TNF (38). buffer (10 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 1 mM Na3VO4), and then lysed in 250 ␮l (per 106 cells) lysis buffer (20 mM Tris-HCl (pH 7.5), To test the importance of each of the MAP kinase pathways in 120 mM NaCl, 10% glycerol, 1 mM Na3VO4, 1 mM EDTA, 1% Triton TNF induction, especially the MAP kinase pathways that do not X-100, 1% deoxycholate, 0.1% SDS, and 1 mM phenylmethylsulfonyl flu- have a specific drug inhibitor, another approach was adopted in by guest on September 26, 2021 oride). The proteins were separated by SDS-PAGE and transferred to a nitrocellulose membrane. mAb H14 (Covance) was used to detect phos- which dominant negative MAP kinase kinase molecules were used phorylation of pol II CTD. The H14 Ab was developed against a purified to suppress individual MAP kinase pathways and to examine their phosphorylated form of RNA pol II and recognizes the phosphorylation of effect on TNF reporter gene synthesis. p 5 in the heptapeptide repeat YSPTS PS at the C-terminal domain. Normal activation of MAP kinase kinase family members is achieved by the dual phosphorylation of two serine/threonine res- Results idues located between kinase domains VII and VIII in the proteins. ⌽ Stimulation of M with LPS can lead to activation of multiple Substituting these residues with alanine (A) produces a dominant- MAP kinase pathways negative form of the molecules. Dominant negative forms of Previous studies by others and us have revealed that LPS activates MEK1, MEK5, MKK6, and MKK4, the MAP kinase kinases for each of the ERK, JNK, and p38 MAP kinase pathways in M⌽ (27, the ERK, BMK/ERK5, p38, and JNK/stress-activated protein ki- 28, 31). By using immunokinase assays and Western blotting, we nase pathways, respectively, were created and have been used in now know that LPS can also activate BMK1/ERK5 in RAW264.7 previous experiments by us (41). RAW cells were then cotrans- cells, a murine macrophage cell line. Fig. 1 shows the kinetics of fected with a TNF reporter gene, and MEK1(A), MEK5(A), the activation of the four different MAP kinases: ERK2, JNK1, p38 MKK4(A), or MKK6(A), and LPS stimulation was applied for 8 h, (p38␣), and BMK1. Activation of all MAP kinases can be detected 1 day after transfection. As shown in Fig. 2B, suppressing any one within 5 min of LPS stimulation; however, the time at which the of the MAP kinase pathways led to a reduction in LPS-induced activation reaches a maximum is different: both ERK2 and BMK1 TNF reporter gene expression. Because a great excess of an indi- were maximally activated in 5 min, whereas maximal activation of vidual MKK(A) is needed to competitively inhibit endogenous JNK1 and p38 did not occur until 15 and 30 min, respectively. In MKKs, we have estimated that this approach may inhibit an en- addition, along with the differences in the time taken to activate dogenous MAP kinase by less than 50% (data not shown). This each of the MAP kinases, there were significant differences in the finding agreed with the inhibition observed in Fig. 2B. Simulta- intensity of activation, such that although p38, ERK2, and JNK1 neously expressing all MKK(A) led to a stronger reduction in the were strongly activated (ϳ10-, 6-, and 9-fold, respectively), reporter gene expression (Fig. 2B), which may be due to the ad- BMK1 demonstrated only an ϳ3-fold level of induction. Further- ditive effect of the partial inhibition of each MKK. Because the more, studies by several investigators have shown that LPS-me- endogenous level of expression of different MEK/MKKs is differ- diated MAP kinase activation is not confined to the activation of a ent and because the competitive inhibitory effect of different single member MAP kinase in each pathway: ERK1 and ERK2 MKKs may also be different, the inhibition data presented in Fig. (27), JNK1 and JNK2 (28), p38, p38␤, p38␥, and p38␦ (our un- 2B may not represent the relative contribution of the different MAP published results) are all activated by LPS stimulation in M⌽. kinase pathways. It is clear that all four MAP kinase pathways are Activation of all known MAP kinases by a single stimulus has not involved in TNF gene expression induced by LPS. 6352 REGULATION OF TNF BY MITOGEN-ACTIVATED PROTEIN KINASES Downloaded from http://www.jimmunol.org/

FIGURE 2. Requirement of MAP kinase pathways in LPS-induced TNF reporter gene expression. A, RAW cells were transfected with hTNF reporter, and the cells were replated into 6-well plates to ensure equal FIGURE 3. Individual and combination effects of different MAP kinase transfection in each sample. LPS (1 ng/ml) stimulation was applied 24 h pathways on TNF reporter gene expression. hTNF or mTNF reporter with after transfection. Samples that were pretreated for 30 min with a different different dominant-active MKKs or their combinations were transfected dose of SB203580 or U0126 or both inhibitors are as indicated. Luciferase into RAW cells. The empty vector pcDNA3 was used to normalize the amount of total DNA used in the transfection. CMV-␤-galactosidase was

activity was measured 8 h after LPS stimulation. B, RAW cells were co- by guest on September 26, 2021 ␤ transfected with hTNF reporter and MEK1(A), MKK4(A), MEK5(A), cotransfected, and -galactosidase activity was used to normalize the trans- MKK6(A), or empty vector pcDNA3. LPS stimulation was applied 24 h fection efficiency. Luciferase activity was used to calculate the induction. later after transfection for 8 h. Luciferase activity was measured. Cotrans- Comparable results were obtained in three experiments. fected ␤-galactosidase was used to normalize the transfection efficiency. The fold of induction was calculated by dividing the luciferase value for the LPS-treated sample with that from the sample without LPS treatment. Comparable results were obtained in two experiments. together resulted in the highest activation of the TNF promoter. Simultaneous activation of ERK, BMK, and p38 pathways led to the highest expression of TNF reporter in comparison with the Cooperative effect of different MAP kinase pathways on TNF other three MAP kinase combinations, and activation of all four gene expression MAP kinase pathways had the highest induction of TNF reporter In contrast to substituting the dual phosphorylation sites with ala- gene expression. Therefore, it is clear that all MAP kinase path- nine, substituting the residues with either glutamic acid (E) or as- ways participate in regulating TNF gene expression, with the p38 partic acid (D) produces dominant-active forms of MAP kinase pathway seeming to contribute to a greater extent than other path- kinases (41). Because all MAP kinase pathways are involved in ways. In addition, activation of multiple pathways induced the LPS-induced TNF gene transcription, we sought to determine the TNF promoter to a much greater extent than the sum of the induction role of each MAP kinase and the combinatorial effect of these produced by individual MAP kinases, and, as such, the combinatorial MAP kinases in the gene expression of TNF. Sole activation of a effect of these MAP kinase pathways cannot be explained by an ad- given MAP kinase pathway was achieved by transiently express- ditive effect. For example, the ERK, BMK, and p38 pathways to- ing dominant-active mutants of different MKKs, and the level of gether led to an ϳ75-fold induction of hTNF reporter gene expres- induction of TNF transcription was evaluated using a luciferase sion, whereas the additive effect of these three MAP kinase pathways reporter driven by either the mouse or human TNF promoters. is only ϳ5-fold (1.1 ϫ 1.3 ϫ 2.8). Thus, MAP kinase pathways seem Almost the same activation profiles of hTNF and mTNF reporter to act through a coordinated mechanism to produce the high expres- were observed (Fig. 3), suggesting that MAP kinases regu- sion of TNF observed in M⌽ stimulated with LPS. late the TNF promoter via a conserved mechanism. When individ- To assess whether the MAP kinases acted at the level of tran- ual MAP kinase pathways were activated, the folds of induction of scriptional activation, the relative amount of mRNA produced the human TNF reporter expression by the ERK, BMK, p38, and from the luciferase reporter was measured when the four MAP JNK activation were 1.1, 1.3, 2.8, and 1.3, respectively, indicating kinase pathways were activated, using an RT-PCR-based instru- that the p38 pathway produced the strongest effect on TNF pro- ment, Taqman (Perkin-Elmer). After normalization of the trans- moter activation (Fig. 3). When two different MAP kinase path- fection efficiency using cotransfected ␤-galactosidase mRNA, we ways were activated simultaneously, the BMK and p38 pathways detected an ϳ30-fold induction of luciferase mRNA in comparison The Journal of Immunology 6353 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 4. The known cis elements in TNF promoter are not the major sites regulated by MAP kinases. A, The hTNF promoter point mutation series was cotransfected with or without dominant-active mutants of MEK1, MEK5, MKK6, and MKK7 as described in Fig. 3. B, The NF-␬B or AP-1 luciferase reporter with different combinations of dominant MKKs was transfected into RAW cells as described in Fig. 3. The induction of the reporters was determined, and comparable results were obtained in three experiments.

FIGURE 5. The induction of a deletion series of 5Ј-hTNF-luciferase reporters by the four MAP kinases. A series of deletions were made in the 5Ј-hTNF promoter, and these deletion mutants were cotransfected with or without dominant-active mutant of MEK1, MEK5, MKK6, and MKK7 as described in Fig. 3. The induction was determined by luciferase expression, and each experiment was performed three times. 6354 REGULATION OF TNF BY MITOGEN-ACTIVATED PROTEIN KINASES Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 6. The Ϫ43 to Ϫ1 TNF promoter is required for TNF transcription mediated by the four MAP kinases. A, Luciferase reporter genes were constructed by fusing the Ϫ55 to Ϫ1 human TNF (Ϫ55 hTNF) and Ϫ55 to Ϫ43 TNF promoter fragments to a nmp sequence. These constructs comprised various elements of the Ϫ55 region of the promoters including the TNF TATA box (f), the TATA box from the nmp (Ⅺ), the sequence from surrounding TNF DNA (u), and nmp DNA (solid line). B, RAW cells were then transfected with cDNA from different combinations of dominant MKKs, along with luciferase reporters containing the Ϫ55 hTNF promoter, the nmp promoter, or a TNF-nmp chimeric promoter, as described in Fig. 3. Reporter induction was determined. Comparable results were obtained in three experiments. with that of the control (data not shown). We also measured the tion (21–23), we examined whether MAP kinases require these half-life of reporter mRNA with or without induction. The lucif- sites to function. Reporter constructs with specific site mutations erase mRNA is very stable and its half-life (ϳ240 min) was not were obtained from Dr. J. S. Economou and were coexpressed with altered by LPS stimulation (data not shown). Therefore, the reg- dominant-active MKKs. Mutations in Egr-1, cAMP response ele- ulatory effect of MAP kinases observed in our reporter gene ex- ment, NF-␬B, AP-1, or AP-2 produced no significant effect on periments probably occurs at a transcriptional level. However, MAP kinase-mediated TNF reporter gene expression, indicating these data do not deny additional posttranscriptional processes in that these sites are not targeted by MAP kinases (Fig. 4A). Two cis the regulation of endogenous TNF by MAP kinases. reporter constructs, NF-␬B and AP-1 reporter, were used to di- Because the hTNF reporter contained ϳ106 bp of the 5Ј-UTR of rectly test whether these cis elements were regulated by MAP ki- TNF mRNA, we removed the hTNF 5Ј-UTR from the reporter nases. As shown in Fig. 4B, only a modest induction was achieved construct and compared hTNF reporter with or without the 5Ј-UTR when the MAP kinases were activated. when different MAP kinases were activated. In agreement with the To localize MAP kinase-responsive elements in the TNF pro- data achieved using the murine TNF reporter (which lacked the moter, RAW cells were transiently transfected with a series of 5Ј-UTR from TNF), no differences were observed (data not luciferase reporter constructs containing progressive deletions in shown), suggesting that MAP kinase activation does not act on the the 5Ј-TNF promoter sequence. Deleting sequence upstream of 5Ј-UTR of TNF. position Ϫ55 did not reduce MAP kinase-mediated gene expres- sion (Fig. 5). However, deletions to and beyond position Ϫ43 re- Mapping the cis element targeted by the MAP kinases duced the basal expression of the reporter to a level below the Because there are several transcription factor binding sites that detection limit of the luciferase assay and prevented the determi- have been suggested to be involved in LPS-induced TNF produc- nation of the relative induction of the reporter gene. Thus, the role The Journal of Immunology 6355

FIGURE 7. CTD of RNA pol II is involved in LPS-mediated TNF tran- Downloaded from scription. TNF or pGL2 reporters were cotransfected with or without ␣-amanitin-resistant pol II possessing either the full-length CTD (pol II) or a shortened CTD with 31 repeats (pol II ⌬31). The cells were grown for 16 h to allow the transfected pol II gene to be expressed, and ␣-amanitin (2.5 ␮g/ml) was added at this point to inhibit endogenous pol II. Seventy- two hours later, cells were stimulated with LPS (1 ng/ml) for 8 h, after

which time luciferase activity was measured. Comparable results were ob- http://www.jimmunol.org/ tained in two experiments.

FIGURE 8. CTD of RNA pol II is involved in MAP kinase-mediated of the cis element within Ϫ55 to Ϫ43 bp (which contains a Sp1- TNF transcription. TNF or pGL2 reporter was cotransfected with or with- ␣ like sequence) in MAP kinase-mediated gene induction was not out -amanitin-resistant pol II that contained the full-length CTD (pol II) ⌬ resolved from these experiments. To overcome the problem of low (A) or the shortened CTD with 31 repeats (pol II 31) (B) and dominant- active mutants of MEK1, MEK5, MKK6, and MKK7 into RAW cells as expression, we constructed a reporter containing the Ϫ55 to Ϫ43 Ј described in Fig. 3. The cells were grown for 16 h to allow for expression sequence placed 5 to an unrelated sequence derived from cis re- of the transfected pol II gene. ␣-amanitin (2.5 ␮g/ml) was added at this porter backbone plasmid (termed nmp). The nmp had the same point to inhibit endogenous pol II. Luciferase activity was measured 72 h by guest on September 26, 2021 length as Ϫ55 reporter. When these two constructs were com- later. Comparable results were obtained in two experiments. pared, nmp was much less responsive to MAP kinases than the Ϫ55 TNF reporter was (Fig. 6), suggesting the importance of the sequence harbored Ϫ43 to Ϫ1 in the TNF promoter in responding transcription, it is possible that RNA pol II complex is responsible to multiple MAP kinase activation. To further map the sequence for MAP kinase-mediated up-regulation of TNF synthesis; and be- within the Ϫ43 element controlling MAP kinase-mediated TNF cause phosphorylation of the CTD of RNA pol II has been impli- gene transcription, we created a series of chimeric promoter ele- cated in transcriptional regulation, we thought that a possible role ments and evaluated the effect of the sequence within this element of the CTD in MAP kinase-induced TNF gene activation should be on its responsiveness to MAP kinases. C1 was created by replacing examined. Transient transfections utilizing ␣-amanitin-resistant the sequence between the Sp1 site and TATA box of Ϫ55 with CTD deletion mutants of pol II’s largest subunit have been suc- nmp sequence. C2 was created by further replacing the C1 TATA cessfully used to examine the function of CTD (11). We adopted box with the nmp TATA box. C3 contained the Sp1 and TATA this established method to study the possible functional relations box sequence from Ϫ55 and the rest from nmp. C4 is a hybrid of between CTD and MAP kinase activation in TNF gene expression. Ϫ55 and nmp joined 3Ј to the TATA box. As shown in Fig. 6, To demonstrate the involvement of RNA pol II in LPS-induced swapping the Ϫ43 sequence with a nonresponsive sequence de- TNF expression, pol II- and pol II ⌬31- (a pol II mutant with a rived from a cis reporter backbone plasmid did not change the CTD containing only 31 repeats) transfected cells cotransfected profiles in response to different combinations of MAP kinases. with TNF reporter genes were stimulated with LPS (1 ng/ml). However, an ϳ2-fold general reduction was observed for the chi- Cells were cultured in the presence of 2.5 ␮g/ml ␣-amanitin, meras in comparison with native sequence. Analysis of the results which inhibited Ͼ95% endogenous pol II activity in RAW cells presented in Fig. 6 did not further narrow down the location of the (data not shown). The expression levels of the full-length pol II but did confirm the importance of Ϫ43 sequence and pol II ⌬31 were shown to be comparable by Western blotting in response to MAP kinases and suggested that the integrity of using anti-HA Ab (data not shown). The basal expression levels of endogenous gene is also important. Taken together, our data indi- the TNF-reporter, mediated either by the full-length pol II or by cated that Ϫ43 to Ϫ1 is the region in TNF promoter that is reg- pol II ⌬31, were similar (Fig. 7, lanes 1 and 3). Measuring the ulated by MAP kinases. luciferase activity produced by the reporter promoters demon- strated that LPS caused a significant induction of the TNF reporter C-terminal domain of RNA pol II is involved in MAP in full-length pol II-transfected cells (Fig. 7, lanes 1 and 2). How- kinase-mediated TNF transcription ever, the induction of the TNF reporter in cells containing the RNA Because Ϫ43 to Ϫ1 encompasses the TATA box and pol II ⌬31 showed a reduced level of induction (around one-third the RNA pol II complex assembles around this region to initiate of the full-length RNA pol II; Fig. 7). These data suggested that the 6356 REGULATION OF TNF BY MITOGEN-ACTIVATED PROTEIN KINASES

compared with full-length pol II. We compared the activity of pol II and pol II ⌬31 on the pGL2 reporter when the dominant-active MKKs were expressed. Activation of the four MAP kinase path- ways modestly up-regulated the reporter expression regardless of whether full-length pol II or pol II ⌬31 was used (Fig. 8, A and B, lanes 7–12), indicating that the pol II CTD may be involved in the selective enhancement of transcription. Collectively, our data sug- gest that the CTD of RNA pol II participates in LPS-induced TNF gene activation and that MAP kinase pathways are involved in the regulation of pol II. Having ascertained this, we further investigated the involvement of the RNA pol II CTD by examining whether activation of the four MAP kinase pathways would induce increased levels of phos- phorylation of RNA polymerase II mutants. To achieve this, RAW cells transfected with the RNA pol II constructs along with either pcDNA3 vector or dominant-active mutants of MEK1, MEK5, MKK6, and MKK7 were metabolically labeled using [32P]ortho- phosphate. The recombinant pol II proteins were then immunopre-

cipitated using anti-HA Ab, resolved on SDS-PAGE gels, and an- Downloaded from alyzed by phosphoimaging. Activation of the pol II-transfected cells with the four MAP kinases showed enhanced pol II phos- phorylation compared with that of unstimulated cells (Fig. 9A). In contrast, the pol II ⌬31 mutant showed little evidence of increased phosphorylation associated with MAP kinase activation. There-

fore, it is evident that the CTD of RNA pol II is targeted for http://www.jimmunol.org/ phosphorylation after activation of the four MAP kinase pathways. To confirm that this CTD phosphorylation occurs in response to LPS stimulation of RAW cells, we measured CTD phosphoryla- tion before and after LPS stimulation in the presence or absence of SB203580 and/or U0126. A mAb, H14 (Covance), specifically targeting the heptapeptide repeat YSPTSpPS intrinsic to the CTD FIGURE 9. Phosphorylation of pol II in response to activation of MAP of pol II, containing a phosphoserine at position 5 was then used ⌬ kinases or LPS stimulation. A, Pol II and pol II 31 were cotransfected in Western analysis. LPS stimulation produced an ϳ2-fold in-

with either pcDNA3 vector or dominant-active mutants of MEK1, MEK5, by guest on September 26, 2021 crease in CTD phosphorylation (Fig. 9B, top panel). This LPS- MKK6, and MKK7 into RAW cells. The cells were selected with ␣-aman- itin for 72 h, and then were metabolically labeled with [32P]orthophosphate induced CTD phosphorylation was completely inhibited by treat- for 2 h. Recombinant pol II proteins were immunoprecipitated with anti- ing the cells with SB203580 or U0126 (Fig. 9B, two middle HA Ab. These samples were resolved by SDS-PAGE and exposed on panels). Because SB203580 and U0126 had an additive effect on phosphoimaging cassettes. B, RAW cells were stimulated with 1 ng/ml LPS-induced TNF reporter gene expression (Fig. 2) and because LPS in presence or absence of SB203580 and/or U0126 for different pe- we did not observe an additive effect of these two drugs on CTD riods of time. The cells were lysed, and equal protein from different sam- phosphorylation (Fig. 9B), CTD may not be the only target regu- ples were separated on 6% SDS-PAGE. The proteins were transferred to lated by p38 or ERK. It is also possible that the measurement of nitrocellulose membrane, and the phosphorylated RNA pol II was detected CTD phosphorylation by Western blotting is not sensitive enough p using anti-YSPTS PS mAb H14. Normal loading of protein in each lane to show the combined effect of these two inhibitors. Nevertheless, was visualized by staining the membrane with Ponceau S (data not shown). our data assert that the p38 and ERK pathways are required for LPS-induced CTD phosphorylation. The observed pre-existence of CTD phosphorylation in unstimulated cells is consistent with re- CTD of RNA pol II may participate in TNF promoter activation by ports in other cell lines (48). Because there are more than 50 hep- LPS. The reporter vector pGL2 promoter (Promega, Madison, WI) tapeptide repeats in CTD and because Western analysis cannot containing a 195-bp SV40 minimal promoter sequence is almost distinguish the position of phosphorylation, we do not know whether LPS unresponsive (Fig. 7, lanes 5 and 6). The CTD deletion did LPS-induced phosphorylation differs from the pre-existing not alter the expression profile of pGL2 in the presence or absence phosphorylation. of LPS (Fig. 7, lanes 5–8), suggesting that RNA pol II promotion of SV40 minimal promoter differs from that of TNF. To determine whether CTD-dependent TNF induction by LPS is mediated by Discussion MAP kinases, we cotransfected pol II or pol II ⌬31 with or without Transcriptional activation of the TNF gene in response to LPS the four active MKKs/MEKs. As in the observation using endog- stimulation is controlled by various transcription factors in con- enous pol II, individual MAP kinase pathways did not lead to junction with the RNA pol II complex. By studying the mechanism significant induction of the TNF reporter gene when ␣-amanitin- through which MAP kinases operate in TNF transcription, we have resistant pol II was employed (Figs. 3 and 8A, lanes 1–5). Signif- found that multiple MAP kinase pathways coordinate with each icant induction of the TNF reporter by the four MAP kinase path- other to regulate the gene. The cis elements found in the TNF ways was observed when ␣-amanitin-resistant full-length pol II promoter, such as Egr-1, cAMP response element-, ␬B-, AP-1-, was used (Fig. 8A, lanes 1 and 5). Induction of the TNF reporter and AP-2-like motifs, are not the targets of these MAP kinase by the four MAP kinase pathways was much lower when pol II pathways, whereas the TATA box and its flanking sequence is ⌬31 was used (Fig. 8B, lanes 1 and 5), with ϳ90% reduction required for MAP kinase-mediated TNF gene induction. This The Journal of Immunology 6357

MAP kinase-responsive site represents the region around which positive function in initiation or elongation of specific gene tran- the RNA pol II complex assembles, and we have presented evi- scripts. Because multiple MAP kinase pathways have to be acti- dence to demonstrate that the CTD of RNA pol II participates in vated to achieve a significant level of TNF transcription (Fig. 3), MAP kinase-induced TNF gene transcription. the CTD has to be modified by multiple MAP kinase pathways at There are a number of transcription factors, such as c-Jun, once or the other component(s) in the pol II complex need to be MEF2A/2C, ELK-1, serum response factor accessory protein-1, modified at the same time. This is a complicated situation and there ATF1, ATF2, and cAMP response element binding protein, that could be many possible mechanisms by which MAP kinases reg- have been suggested to be regulated by MAP kinase pathways ulate the TNF gene. One possibility is that the CTD is modified by (49). It was predicted that MAP kinases would act via the tran- multiple MAP kinases and/or their downstream kinases, leading to scription factors mentioned above and that they would directly or a specific state of CTD phosphorylation that is optimal for an LPS- indirectly regulate TNF gene transcription. It was a surprise to find inducible gene such as TNF. Different phosphorylation states of that MAP kinase-mediated TNF gene induction is largely depen- CTD can be generated by different kinases and their combinations dent on a 43-bp sequence surrounding the TATA box rather than because CTD of mammalian pol II has hundreds of , on the rest of the TNF promoter. These data suggested that im- threonines, and tyrosines that have the potential to be selectively portant physiological substrates of MAP kinases in TNF expres- phosphorylated by these kinases. These phosphorylatable amino sion remain unidentified. It is interesting that the MAP kinase- acids can provide up to 1036 phosphorylation states that may pro- responsive element seems to possess some level of redundancy in vide a basis for the selective and quantitative control of gene ex- that replacing different regions of the element with unrelated se- pression in eukaryotic cells. As for the cooperative action of MAP quences did not abolish its function. The importance of this ele- kinases in TNF gene expression, an initial point from which we Downloaded from ment is clear because completely replacing this sequence with an can understand the mechanism can be found by identifying the unrelated sequence leads to almost nonresponsiveness. The precise functional substrate(s) of these MAP kinases. determinant harbored within this 43 bp was not revealed by our TNF biosynthesis is controlled at multiple levels. The ARE in experimental approaches. Sequence comparison between different the 3Ј-UTR is known to play a pivotal role in TNF mRNA stability species shows that the primary sequence in this region of the TNF and translational regulation (24, 25). The JNK pathway had been

promoter is poorly conserved outside of the TATA box, although shown to be involved in the stability of some ARE-bearing http://www.jimmunol.org/ careful comparison of this region among 16 reported TNF promot- mRNAs (56) and to be required for TNF translation. The studies ers does reveal a very similar AT content (50–60%) and a C-rich based on the specific inhibition of p38 by SB203580 suggested that cluster in the nonconserved region between different species. the primary site of p38 action is TNF translation (24, 57). Our Therefore, it is possible that MAP kinase responsiveness is depen- unpublished results show that the 3Ј-UTR in the TNF reporter dent on a secondary structural feature, possibly associated with the further up-regulates gene expression in response to the activation AT content or C-rich cluster of the sequence. This structural fea- of MAP kinases, supporting the notion that posttranscription reg- ture may not necessarily be required for selective protein binding, ulation is indeed an important site targeted by MAP kinase. How- but it may be required for specific remodeling. The ever, the data presented here add TNF transcription to the list of mechanism by which this sequence determines gene regulation MAP kinase regulatory sites. Thus, MAP kinase pathways act on by guest on September 26, 2021 awaits further investigation. multiple levels for the full regulation of TNF. Several observations suggested that the four MAP kinase path- Although MAP kinase pathways are important, other signal ways converge onto one regulatory site. First of all, the MAP ki- transduction pathways may also play important roles. Studies have nases exhibited effects over and above those which could be at- shown that the upstream sequence of the TNF promoter also con- tributed to an additive mechanism, indicating that one regulatory tributes to LPS-induced TNF gene expression (21–23). Coordina- mechanism may be the ultimate target for all four MAP kinase tion of MAP kinases with other pathways must exist, and indeed, pathways. Second, a region can be localized in the TNF promoter we have observed activation of the NF-␬B pathway by inhibitory that is responsive to the activation of the four MAP kinase path- ␬B protein kinase-2 to further up-regulate MAP kinase-induced ways. Third, deleting the CTD of RNA pol II almost abolished TNF gene expression. TNF induction mediated by the four MAP kinases. The coopera- It is believed that requires a set of gen- tive effect of these MAP kinase pathways is probably mediated by eral transcription factors to be assembled with RNA poly II to form targeting a single substrate such as the RNA pol II holoenzyme at a transcription initiation complex, and the CTD of pol II has been different sites or subunits. The CTD of pol II is a potential target reported to specifically interact with some of these factors (8, 10). of MAP kinases because there are a number of potential MAP Our finding that the CTD of RNA pol II is involved in multiple kinase targeting sites there. Phosphorylation of pol II CTD has MAP kinase-mediated TNF gene transcriptions indicates that been intensively studied and, although its role in controlling pol II MAP kinase pathways may regulate the RNA pol II complex. The remains unresolved, a number of protein kinases are able to phos- robust induction of the TNF reporter upon activation of the four phorylate the CTD on different sites in vitro (50–54). Experiments MAP kinase pathways suggests that study of the regulation of TNF in yeast suggest the involvement of several protein kinases in CTD gene transcription could provide a model system for the function phosphorylation because different CTD phosphorylation events of the CTD in gene regulation. Our data provide an important role have been implicated in different stages of pol II activity (including for the general transcriptional machinery in the control of cytokine elongation) and RNA polymerase release (17–20). Immunofluo- synthesis. A future challenge for our study is to define each of the rescence studies and protein DNA cross-linking assays suggest signal transduction pathways in regulating promoters as well as the that the phosphorylation state of CTD may vary depending on RNA pol II complex. which gene is being transcribed (20, 55). It is possible that different are involved in modifying the CTD to specifically tran- scribe selected genes. Negative regulation of gene transcription has Acknowledgments been observed upon CTD phosphorylated by Cdc2 or Cdk7 (17). We thank Dr. J. S. Economou (University of California, Los Angeles, CA) It would be very interesting to address whether MAP kinases play and Hans-Peter Gerber (Genentech, South San Francisco, CA) for plasmids a role in CTD phosphorylation, and if this phosphorylation has a and J. V. Kuhns for excellent secretarial assistance. 6358 REGULATION OF TNF BY MITOGEN-ACTIVATED PROTEIN KINASES

References 31. Han, J., J.-D. Lee, L. Bibbs, and R. J. Ulevitch. 1994. A MAP kinase targeted by endotoxin and hyperosmolarity in mammalian cells. Science 265:808. 1. Beutler, B., and A. Cerami. 1988. Tumor necrosis, cachexia, shock, and inflam- 32. Lee, J. C., J. T. Laydon, P. C. McDonnell, T. F. Gallagher, S. Kumar, D. Green, mation: a common mediator. Annu. Rev. Biochem. 57:505. D. McNulty, M. J. Blumenthal, R. J. Heyes, S. W. Landvatter, et al. 1994. Iden- 2. Ziegler, E. J. 1988. Tumor necrosis factor in humans. N. Engl. J. Med. 318:1533. tification and characterization of a novel protein kinase involved in the regulation 3. Waage, A., A. Halstensen, and T. Espevik. 1987. Association between tumour of inflammatory cytokine biosynthesis. Nature 372:739. necrosis factor in serum and fatal outcome in patients with meningococcal dis- 33. Dong, Z., X. Qi, and I. J. Fidler. 1993. Tyrosine phosphorylation of mitogen- ease. Lancet 1:355. activated protein kinases is necessary for activation of murine macrophages by 4. Old, L. J. 1987. Tumor necrosis factor: another chapter in the long history of natural and synthetic bacterial products. J. Exp. Med. 177:1071. endotoxin. Nature 330:602. 34. Swantek, J. L., M. H. Cobb, and T. D. Geppert. 1997. Jun N-terminal kinase/ 5. Beutler, B., N. Krochin, I. W. Milsark, C. Luedke, and A. Cerami. 1986. Control stress-activated protein kinase (JNK/SAPK) is required for lipopolysaccharide of cachectin (tumor necrosis factor) synthesis: mechanisms of endotoxin resis- stimulation of tumor necrosis factor ␣ (TNF-␣) translation: glucocorticoids in- tance. Science 232:978. hibit TNF-␣ translation by blocking JNK/SAPK. Mol. Cell. Biol. 17:6274. 6. Beutler, B., and A. Cerami. 1986. Cachectin and tumour necrosis factor as two 35. Reimann, T., D. Buscher, R. A. Hipskind, S. Krautwald, M. Lohmann-Matthes, sides of the same biological coin. Nature 320:584. and M. Baccarini. 1994. Lipopolysaccharide induces activation of the Raf-1/ 7. Bartolomei, M. S., N. F. Halden, C. R. Cullen, and J. L. Corden. 1988. Genetic MAP kinase pathway: a putative role for Raf-1 in the induction of the IL-1␤ and analysis of the repetitive carboxyl-terminal domain of the largest subunit of the TNF-␣ genes. J. Immunol. 153:5740. mouse RNA polymerase II. Mol. Cell. Biol. 8:330. 36. Liu, M. K., P. Herrera-Velit, R. W. Brownsey, and N. E. Reiner. 1994. CD14- 8. Corden, J. L. 1990. Tails of RNA polymerase II. Trends Biochem. Sci. 15:383. dependent activation of protein kinase C and mitogen-activated protein kinases 9. Drapkin, R., and D. Reinberg. 1994. The multifunctional TFIIH complex and (p42 and p44) in human monocytes treated with bacterial lipopolysaccharide. transcriptional control. Trends Biochem. Sci. 19:504. J. Immunol. 153:2642. 10. Thompson, C. M., A. J. Koleske, D. M. Chao, and R. A. Young. 1993. A mul- 37. Hambleton, J., M. McMahon, and A. L. DeFranco. 1995. Activation of Raf-1 and tisubunit complex associated with the RNA polymerase II CTD and TATA- mitogen-activated protein kinase in murine macrophages partially mimics lipopo- binding protein in yeast. Cell 73:1361. lysaccharide-induced signaling events. J. Exp. Med. 182:147. 11. Gerber, H. P., M. Hagmann, K. Seipel, O. Georgiev, M. A. West, Y. Litingtung,

38. Bhat, N. R., P. Zhang, J. C. Lee, and E. L. Hogan. 1998. Extracellular signal- Downloaded from W. Schaffner, and J. L. Corden. 1995. RNA polymerase II C-terminal domain regulated kinase and p38 subgroups of mitogen-activated protein kinases regulate required for enhancer-driven transcription. Nature 374:660. inducible nitric oxide synthase and tumor necrosis factor-␣ gene expression in 12. Okamoto, H., C. T. Sheline, J. L. Corden, K. A. Jones, and B. M. Peterlin. 1996. endotoxin-stimulated primary glial cultures. J. Neurosci. 18:1633. Trans-activation by human immunodeficiency virus Tat protein requires the C- 39. Jiang, Y., C. Chen, Z. Li, W. Guo, J. A. Gegner, S. Lin, and J. Han. 1996. terminal domain of RNA polymerase II. Proc. Natl. Acad. Sci. USA 93:11575. Characterization of the structure and function of a new mitogen-activated protein 13. West, M. L., and J. L. Corden. 1995. Construction and analysis of yeast RNA kinase (p38␤). J. Biol. Chem. 271:17920. polymerase II CTD deletion and substitution mutations. Genetics 140:1223. 14. Patturajan, M., R. J. Schulte, B. M. Sefton, R. Berezney, M. Vincent, 40. Han, J., G. Huez, and B. Beutler. 1991. Interactive effects of the tumor necrosis factor promoter and 3Ј-untranslated regions. J. Immunol. 146:1843. O. Bensaude, S. L. Warren, and J. L. Corden. 1998. Growth-related changes in http://www.jimmunol.org/ phosphorylation of yeast RNA polymerase II. J. Biol. Chem. 273:4689. 41. Huang, S., Y. Jiang, Z. Li, E. Nishida, P. Mathias, S. Lin, R. J. Ulevitch, 15. Liu, Z., H. Hsu, D. V. Goeddel, and M. Karin. 1996. Dissection of TNF receptor G. R. Nemerow, and J. Han. 1997. Apoptosis signaling pathway in T cells is com- 1 effector functions: JNK activation is not linked to apoptosis while NF-␬B ac- posed of ICE/Ced-3 family proteases and MAP kinase kinase 6b. Immunity 6:739. tivation prevents cell death. Cell 87:565. 42. Han, J., X. Wang, Y. Jiang, R. J. Ulevitch, and S. Lin. 1996. Identification and 16. Meisels, E., O. Gileadi, and J. L. Corden. 1995. Partial truncation of the yeast characterization of a predominant isoform of human MKK3. FEBS Lett. 403:19. RNA polymerase II carboxyl-terminal domain preferentially reduces expression 43. New, L., Y. Jiang, M. Zhao, K. Liu, W. Zhu, L. J. Flood, Y. Kato, G. C. Parry, of glycolytic genes. J. Biol. Chem. 270:31255. and J. Han. 1998. PRAK, a novel protein kinase regulated by the p38 MAP 17. Gebara, M. M., M. H. Sayre, and J. L. Corden. 1997. Phosphorylation of the kinase. EMBO J. 17:3372. carboxy-terminal repeat domain in RNA polymerase II by cyclin-dependent ki- 44. Wang, Y., S. Huang, V. P. Sah, J. J. Ross, J. H. Brown, J. Han, and K. R. Chien. nases is sufficient to inhibit transcription. J. Cell. Biochem. 64:390. 1998. Cardiac muscle cell hypertrophy and apoptosis induced by distinct mem- 18. Peterson, S. R., A. Dvir, C. W. Anderson, and W. S. Dynan. 1992. DNA binding bers of the p38 mitogen-activated protein kinase family. J. Biol. Chem. 273:2161.

provides a signal for phosphorylation of the RNA polymerase II heptapeptide 45. Wang, Y., B. Su, V. P. Sah, J. H. Brown, J. Han, and K. R. Chien. 1998. Cardiac by guest on September 26, 2021 repeats. Genes Dev. 6:426. hypertrophy induced by mitogen-activated protein kinase kinase 7, a specific 19. Kang, M. E., and M. E. Dahmus. 1993. RNA IIA and IIO have activator for c-Jun NH2-terminal kinase in ventricular muscle cells. J. Biol. Chem. distinct roles during transcription from the TATA-less murine dihydrofolate re- 273:5423. ductase promoter. J. Biol. Chem. 268:25033. 46. Han, J., Y. Jiang, Z. Li, V. V. Kravchenko, and R. T. Ulevitch. 1997. Activation 20. O’Brien, T., S. Hardin, A. Greenleaf, and J. T. Lis. 1994. Phosphorylation of RNA of the transcription factor MEF2C by the MAP kinase p38 in inflammation. polymerase II C-terminal domain and transcriptional elongation. Nature 370:75. Nature 386:296. 21. Shakhov, A. N., M. A. Collart, P. Vassalli, S. A. Nedospasov, and 47. Favata, M. F., K. Y. Horiuchi, E. J. Manos, A. J. Daulerio, D. A. Stradley, C. V. Jongeneel. 1990. ␬B-type enhancers are involved in lipopolysaccharide- W. S. Feeser, D. E. Van Dyk, W. J. Pitts, R. A. Earl, F. Hobbs, et al. 1998. mediated transcriptional activation of the tumor necrosis factor ␣ gene in primary Identification of a novel inhibitor of mitogen-activated protein kinase kinase. macrophages. J. Exp. Med. 171:35. J. Biol. Chem. 273:18623. 22. Rhoades, K. L., S. H. Golub, and J. S. Economou. 1992. The regulation of the 48. Kim, E., L. Du, D. B. Bregman, and S. L. Warren. 1997. Splicing factors asso- ␣ human tumor necrosis factor promoter region in macrophage, T cell, and B cell ciate with hyperphosphorylated RNA polymerase II in the absence of pre-mRNA. lines. J. Biol. Chem. 267:22102. J. Cell Biol. 136:19. 23. Yao, J., N. Mackman, T. S. Edgington, and S. Fan. 1997. Lipopolysaccharide in- 49. New, L., and J. Han. 1998. The p38 MAP kinase pathway and its function. Trends ␣ duction of the tumour necrosis factor- promoter in human monocytic cells: regu- Cardiovasc. Med. 8:220. ␬ lation by Egr-1, c-Jun, and NF- B transcription factors. J. Biol. Chem. 272:17795. 50. Lee, J. M., and A. L. Greenleaf. 1997. Modulation of RNA polymerase II elon- 24. Han, J., T. Brown, and B. Beutler. 1990. Endotoxin-responsive sequences control gation efficiency by C-terminal heptapeptide repeat domain kinase I. J. Biol. cachectin/tumor necrosis factor biosynthesis at the translational level. J. Exp. Chem. 272:10990. Med. 171:465. 51. Feaver, W. J., O. Gileadi, Y. Li, and R. D. Kornberg. 1991. CTD kinase asso- 25. Shaw, G., and R. Kamen. 1986. A conserved AU sequence from the 3Ј untrans- ciated with yeast RNA polymerase II initiation factor ␤. Cell 67:1223. lated region of GM-CSF mRNA mediates selective mRNA degradation. Cell 46:659. 52. Garcia-Martinez, L. F., G. Mavankal, J. M. Neveu, W. S. Lane, D. Ivanov, and 26. DeFranco, A. L., M. T. Crowley, A. Finn, J. Hambleton, and S. L. Weinstein. R. B. Gaynor. 1997. Purification of a Tat-associated kinase reveals a TFIIH 1998. The role of tyrosine kinases and map kinases in LPS-induced signaling. complex that modulates HIV-1 transcription. EMBO J. 16:2836. Prog. Clin. Biol. Res. 397:119. 53. Stone, N., and D. Reinberg. 1992. Protein kinases from Aspergillus nidulans that 27. Weinstein, S. L., J. S. Sanghera, K. Lemke, A. L. DeFranco, and S. L. Pelech. 1992. phosphorylate the carboxyl-terminal domain of the largest subunit of RNA poly- Bacterial lipopolysaccharide induces tyrosine phosphorylation and activation of mi- merase II. J. Biol. Chem. 267:6353. togen-activated protein kinases in macrophages. J. Biol. Chem. 267:14955. 54. Serizawa, H., R. C. Conaway, and J. W. Conaway. 1992. A carboxyl-terminal- 28. Hambleton, J., S. L. Weinstein, L. Lem, and A. L. DeFranco. 1996. Activation of domain kinase associated with RNA polymerase II transcription factor ␦ from rat c-Jun N-terminal kinase in bacterial lipopolysaccharide-stimulated macrophages. liver. Proc. Natl. Acad. Sci. USA 89:7476. Proc. Natl. Acad. Sci. USA 93:2774. 55. Weeks, J. R., S. E. Hardin, J. Shen, J. M. Lee, and A. L. Greenleaf. 1993. 29. Sanghera, J. S., S. L. Weinstein, M. Aluwalia, J. Girn, and S. L. Pelech. 1996. Locus-specific variation in phosphorylation state of RNA polymerase II in vivo: Activation of multiple proline-directed kinases by bacterial lipopolysaccharide in correlations with gene activity and transcript processing. Genes Dev. 7:2329. murine macrophages. J. Immunol. 156:4457. 56. Ming, X. F., M. Kaiser, and C. Moroni. 1998. c-jun N-terminal kinase is involved in 30. Han, J., J. D. Lee, P. S. Tobias, and R. J. Ulevitch. 1993. Endotoxin induces rapid AUUUA-mediated interleukin-3 mRNA turnover in mast cells. EMBO J. 17:6039. protein tyrosine phosphorylation in 70Z/3 cells expressing CD14. J. Biol. Chem. 57. Baldassare, J. J., Y. Bi, and C. J. Bellone. 1999. The role of p38 mitogen-acti- 268:25009. vated protein kinase in IL-1␤ transcription. J. Immunol. 162:5367.