Bacterial Lipopolysaccharide Induces Tyrosine Phosphorylation and Activation of Mitogen-activated Protein Kinases in Macrophages*

Bacterial lipopolysaccharide (LPS) is a potent activator of antibacterial responses by macrophages. Fol- lowing LPS stimulation, the tyrosine phosphorylation of several proteins is rapidly increased in macrophages, and this event appears to mediate some re- sponses to LPS. We now report that two of these tyrosine phosphoproteins of 41 and 44 kDa are isoforms of mitogen-activated protein (MAP) kinase. Each of these proteins was reactive with anti-MAP kinase antibodies and comigrated with MAP kinase activity in fractions eluted from a Mono& anion-exchange column. Follow- ing LPS stimulation, column fractions containing the tyrosine phosphorylated forms of p41 and p44 exhib- ited increased MAP kinase activity. Inhibition of LPS-induced tyrosine phosphorylation of these proteins was accompanied by inhibition of MAP kinase activity. Additionally, induction of ~ 4 1 1 ~ 4 4 tyrosine phos- phorylation and MAP kinase activity by LPS appeared to be independent of activation of protein kinase C, even though phorbol esters also induced these re- sponses. These results demonstrate that LPS induces the tyrosine phosphorylation tyrosine phosphopro- teins antiphosphotyrosine monoclonal horseradish peroxidase (1:15,000 dilution in TBST, reprobed monoclonal sheep IgG-horseradish (1:15,000 were on by

publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
) I To whom correspondence and reprint requests should be ad- arachidonic acid metabolites (3,4). In addition, the combination of LPS and interferon-y induce macrophages to differentiate to a highly bactericidal state. However, excessive LPS stimulation of macrophages and other cells occurring during severe bacterial infections can also lead to extensive tissue damage and septic shock (5). Thus, a better understanding of the mechanisms by which macrophages are activated by LPS could provide important insight into the regulation of the host response to bacterial infection. We have investigated the early biochemical events that are triggered in macrophages by LPS. Previously, we reported that LPS increases protein tyrosine phosphorylation in murine macrophages and that this early signaling event appears to mediate some downstream macrophage responses to LPS (6). To extend these findings, we have attempted to identify the molecular components involved in the induced phosphorylation response. One such component is the macrophage cell surface protein, CD14. This protein has been shown to bind complexes consisting of LPS and serum LPS-binding protein and has been implicated in the cellular response to LPS (7). Inhibition of LPS binding to CD14 with anti-CD14 antibodies also inhibited LPS-induced tyrosine phosphorylation in human macrophages.' This observation suggests that CD14 plays a role in mediating this signaling event. Also of interest are the identities of the proteins that undergo increased tyrosine phosphorylation following LPS stimulation. Among the most prominent tyrosine phosphorylated bands in LPS-stimulated macrophages are a series of 40-45-kDa proteins (6). These molecular masses are similar to those of a family of serine/threonine protein kinases known as mitogenactivated protein (MAP) kinases (8)(9)(10). MAP kinases appear to participate in the signal transduction pathways activated by a variety of extracellular ligands. These kinases have been shown in several cell types to be rapidly phosphorylated on tyrosine residues following cellular activation, and this modification contributes to the increased enzymatic activity of these proteins. The in vitro phosphorylation and activation of the 90-kDa ribosomal S6 protein kinase (11,12) and the cjun transcription factor (13) by MAP kinases suggest that they may regulate fundamental cellular processes.
Given the growing evidence that MAP kinases are important signal transduction components, we tested whether any of the tyrosine-phosphorylated bands observed in LPS-stimulated macrophages correspond to MAP kinases and whether these enzymes become activated following LPS treatment. In this report, we show that MAP kinase activity is increased following LPS treatment, and this response appears to occur as the result of tyrosine phosphorylation of at least two S. L. Weinstein, C. H. June, and A. L. DeFranco, manuscript in preparation. 14955 different MAP kinase isozymes. Thus, MAP kinases are the first identified substrates for LPS-induced tyrosine phosphorylation.

EXPERIMENTAL PROCEDURES
Materials-Preparations of LPS were purchased from List Biological Laboratories (Campbell, CA) and diphosphoryl lipid A was purchased from Ribi Immunochemical Research (Hamilton, MT). Synthetic lipid A (diphosphoryl, Escherichia coli type), [y3'P]ATP, and the PY-20 antiphosphotyrosine monoclonal antibody were from ICN Biomedicals. The 4G10 antiphosphotyrosine monoclonal antibody was a gift from D. Morrison, B. Druker, and T. Roberts (Dana-Farber Cancer Institute). Goat antirabbit IgG and goat antimouse IgG conjugated to alkaline phosphatase were procured from Bio-Rad. Sheep antimouse IgG conjugated to horseradish peroxidase and the ECL detection kit for immunoblotting were purchased from Amersham. Herbimycin A was obtained from N. R. Lomax (Drug Synthesis and Chemical Branch, Division of Cancer Treatment, National Cancer Institute, Bethesda, MD). Compound 3 was a gift from Dr. Michael Venuti (Department of Bio-Organic Chemistry, Genentech, San Francisco, CA). Complete and Incomplete Freund's Adjuvant were from GIBCO. The bicinchoninic acid protein assay kit was purchased from Pierce Chemical (Rockford, IL). Other reagents were purchased from Sigma.
Cell Culture, Stimulation, and Lysis-The murine macrophage cell line, RAW 264.7 (American Type Culture Collection, Rockville, MD), was cultured in Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal bovine serum and 2 mM glutamine at 37 "C in a 5% CO,/air mixture. For each experimental sample, 2 X lo' cells were seeded into a 150-mm dish in 20 ml of medium. The cells were then cultured for about 18 h to allow the cell number to approximately double. Cells were stimulated by the addition of the indicated activator for 15 min. In some experiments, prior to the addition of stimulators, cells were pretreated with 10 pg/ml herbimycin A for 4 h or 10 p~ Compound 3 for 20 min. Following stimulation, cells were washed in situ with ice-cold phosphate-buffered saline containing 1 mM Na3V0,, then lysed in 2 ml of 20 mM MOPS, pH 7.2, 5 mM EGTA, 1% (w/v) Nonidet P-40,l mM dithiothreitol, 75 mM &glycerol phosphate, 1 mM Na3V04, and 1 mM phenylmethylsulfonyl fluoride for 20 min at 4 "C. The detergent-insoluble material was pelleted by centrifugation (10,000 X g, 15 min, 4 "C), and the soluble supernatant fraction was removed and stored at -80 "C.
For immunoprecipitation of MAP kinase, cultures were seeded at 4 x106 cells in a 100-mm dish in 10-ml medium. The cells were then cultured for about 18 h to allow the cell number to approximately double. Following stimulation, the cells were lysed in 0.3 ml of boiling 0.5% SDS, 10 mM Tris, pH 7.3, 1 mM dithiothreitol. Lysates were boiled an additional 5 min, and the insoluble material was removed by centrifugation (10,000 X g, 15 min, 25 "C).
The protein concentration of the macrophage lysates was determined using the bicinchoninic acid assay.
Isolation and Assay of MAP Kinases-Macrophage lysate protein (-1 mg) was loaded onto a MonoQ anion exchange column (1-ml bed volume) equilibrated in a column buffer (12.5 mM MOPS, pH 7.2,0.5 mM EGTA, 2 mM dithiothreitol, 12.5 mM P-glycerol phosphate, 7.5 mM MgClz). The column was eluted with a 20 ml linear 0-0.8 M NaCl gradient using a Pharmacia fast protein liquid chromatography system, and 250-pl fractions were collected. MAP kinase activity of MonoQ fractions was assessed by assaying the myelin basic protein (MBP) phosphorylating activity as described previously (14).
Production of MAP Kinase Antibodies-Antibodies were raised to synthetic peptides from the rat extracellular signal-regulated kinase 1 (ERK 1) sequence: subdomain 111 (erkl-111, PFEHQTYCQRTLRE-IQILLGFRHENVIGIRDILRAP-GGC), from the C terminus (erkl-CT, CGG-PFTFDMELDDLPKERLKELIFQETARFQPGAPEAP), and from the ATP binding site of the 44-kDa MAP kinase encoded by the sea star mpk gene (p44'"pk) (mpk-I, GLAYIGEGAYGMVC). New Zealand White rabbits were immunized subcutaneously at four sites with -500 pg of keyhole limpet hemocyanin-coupled peptide emulsified in complete Freund's adjuvant (1-ml final volume). Rabbits were subsequently boosted every 4 weeks intramuscularly at two sites with 500 pg of keyhole limpet hemocyanin-coupled peptide emulsified in incomplete Freund's adjuvant. Ear bleeding was performed 2 weeks after each boost. The blood was permitted to clot at 37 "C for 30 min and then incubated a t 4 "C overnight to allow the clot to contract. The antisera were collected and stored a t -20 "C. Antipeptide antibodies were affinity-purified on the appropriate peptide-agarose col-umn by eluting with 0.1 M glycine, pH 2.5, and subsequent neutralization with a saturated Tris solution. Antibody titers were estimated by standard enzyme-linked immunosorbent assay techniques. Rabbit polyclonal antibodies raised against the purified sea star p44"Pk protein were prepared as described previously (15) and affinitypurified on a ~44"~'-agarose column. The mouse monoclonal anti-MAP kinase antibody used for the immunoprecipitation experiment was purchased from Zymed (San Francisco, CA).
Electrophoresis and Immunoblotting-Protein samples were prepared for electrophoresis by mixing with a concentrated sample buffer to obtain a final concentration of 62.5 mM Tris-HC1, pH 6.8, 2% SDS, 100 mM dithiothreitol, 10% glycerol, and 0.01% bromphenol blue. Samples were then separated on 12% SDS-polyacrylamide gels using the buffer system described by Laemmli (16). Following electrophoresis, the separating gel was soaked in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol) for 5 min and then the proteins were transferred to nitrocellulose for 4 h at 0.5 A. Subsequently, the nitrocellulose membrane was blocked with Tris-buffered saline (TBS) containing 3% gelatin or 2% bovine serum albumin for 2 h at room temperature. The membrane was washed twice with TBS containing 0.05% Tween 20 (TBST) for 5 min before overnight incubation with rabbit polyclonal anti-MAP kinase antibodies or the mouse monoclonal PY-20 antiphosphotyrosine antibody (in 1% gelatin-TBST; 1:lOOO dilution) or the mouse monoclonal 4G10 antiphosphotyrosine antibody (1:3 diluted culture supernatant in TBST). The next day, the membrane was washed twice with TBST before incubation with the second antibody (goat antirabbit IgG or goat antimouse IgG coupled to alkaline phosphatase in 1% gelatin-TBST; 1:3000 dilution) for 2 h at room temperature. The membrane was rinsed with two washes of TBST, followed by one wash with TBS prior to color development with 0.5 mg/ml 5-bromo-4-chloro-3-indolyl phosphate and 0.25 mg/ml nitro blue tetrazolium in 0.1 M NaHC03, 10 mM MgCIZ, pH 9.8. The color development was continued for 5 min to 4 h to give the desired darkness, and the reaction was stopped by rinsing the membrane in a large volume of water.
Inmunoprecipitation of MAP Kinase-Cell lysates in 0.5% SDS, 10 mM Tris, pH 7.3, 1 mM dithiothreitol were diluted 1:5 with 12.5 mM Tris, pH 7.3, 187.5 mM NaCl, 1.25% deoxycholate, 1.25% Triton X-100, 0.625% Nonidet P-40, 1.25 mM EDTA, 1.25 mM EGTA, 0.25 mM Na3V04, 0.25 mM phenylmethylsulfonyl fluoride. Diluted lysates (1.5 ml) were precleared with 3 pg of affinity-purified rabbit antimouse IgG and 40 p1 of packed protein A-Sepharose beads for 30 min. The lysates were then incubated with a combination of 5 pg of a mouse monoclonal anti-MAP kinase antibody and 5 p1 of erkl-CT anti-MAP kinase polyclonal antibodies for 2 h followed by the addition of 3 pg of affinity-purified rabbit antimouse IgG for 1 h. Immune complexes were precipitated by transferring the lysates to tubes containing 40 pl of packed protein A-Sepharose beads and incubating for 1 h. All incubations were performed at 4 "C. The beads were washed once with 1 ml of wash buffer (10 mM Tris, pH 7.3,2 M NaCI, 0.1% SDS, 1% deoxycholate, 1% Triton X-100, 0.5% Nonidet P-40, 1 mM EDTA, 1 mM EGTA, 0.2 mM Na3V0,, 0.2 mM phenylmethylsulfonyl fluoride) and then three times with 1 ml of wash buffer containing 150 mM NaCI. The beads were resuspended in 40 p1 of 2 X concentrated SDS sample buffer and boiled 10 min. The supernatant fraction was resolved on a SDS-polyacrylamide gel and then transferred to a nitrocellulose membrane. The tyrosine phosphoproteins were visualized by immunoblotting with the 4G10 antiphosphotyrosine monoclonal antibody followed by sheep antimouse IgGhorseradish peroxidase antibodies (1:15,000 dilution in TBST, 1 h, 25 "C) and an enhanced chemiluminescence detection system used as directed by the manufacturer. The membrane was stripped (2% SDS, 62.5 mM Tris, pH 6.7, 100 mM 2-mercaptoethanol, 50 'c, 30 m i d and reprobed with a mouse monoclonal anti-MAP kinase antibody (1:5,000 dilution in TBST, overnight, 4 "C), followed by sheep antimouse IgG-horseradish peroxidase antibodies (1:15,000 dilution in TBST, 1 h, 25 "C). The immunoreactive proteins were then visualized on film by chemiluminescence. ment of these cells with phorbol 12-myristate 13-acetate (PMA) induces tyrosine phosphorylation of the 41-and 44-kDa proteins. To test whether any of these LPS-and PMAinduced bands correspond to MAP kinase isozymes, parallel immunoblots were probed with a panel of antibodies specific for MAP kinases (Fig. 1, 23-E). These antibodies were raised against the purified sea star MAP kinase, ~4 4 "~~) or against peptides derived from ~4 4 " '~~ or the rat ERK 1 gene product. These antibodies have been shown to detect a 41-42-kDa isoform of MAP kinase as well as a 43-44-kDa isoform in 3T3 fibroblasts (11) and in other cell types.3 In RAW 264.7 cells, these antibodies detected two isoforms of MAP kinase which comigrated with the 41-and the 44-kDa tyrosine phosphoproteins induced by LPS and PMA. In contrast, the anti-MAP kinase antibodies did not react with bands corresponding to the 42-or the 43.5-kDa tyrosine phosphoproteins induced by LPS but not PMA. Thus, RAW 264.7 cells have at least two isoforms of MAP kinase and each isoform comigrates on SDSpolyacrylamide gels with a protein that becomes tyrosinephosphorylated following treatment with LPS or PMA. Interestingly, the 41-kDa immunoreactive MAP kinase from LPS-and PMA-stimulated cells migrated slightly more slowly on the SDS-polyacrylamide gel than the corresponding protein from unstimulated cells. A similar mobility shift has been reported for a 42-kDa MAP kinase isozyme in Xenopus oocytes following progesterone treatment (17) and in plateletderived growth factor-stimulated fibroblasts (18). It has been suggested that this mobility shift may be a consequence of increased phosphorylation.

LPS-and PMA-induced Tyrosine Phosphorylated p41 and p44 Correspond
To further assess the relationship between the 41-and 44-kDa proteins detected by antiphosphotyrosine antibodies and by anti-MAP kinase antibodies, RAW 264.7 cell lysates were immunoprecipitated with a mixture of monoclonal and polyclonal anti-MAP kinase antibodies, and the resulting precipitated and unprecipitated fractions were separated on a SDSpolyacrylamide gel and immunoblotted with antiphosphotyrosine antibodies. Following anti-MAP kinase immunoprecipitation, the 44-kDa tyrosine phosphoprotein was completely depleted and the 41-kDa phosphoprotein was partially depleted from the cell lysate ( Fig. 2 A ) . This result indicates that both of these proteins were recognized by the anti-MAP kinase antibodies. Depletion of these two proteins was specific as other tyrosine phosphoproteins were not affected. More-S. L. Pelech, unpublished experiments.  over, when these blots were stripped and reprobed with anti-MAP kinase antibodies, the 41-and 44-kDa MAP kinase isoforms were found to be depleted to the same extent as the corresponding tyrosine phosphoproteins (Fig. 2B). These results suggest that the 41-and 44-kDa tyrosine phosphoproteins induced by LPS are isoforms of MAP kinases.
The anti-MAP kinase immunoprecipitated fractions were also analyzed by immunoblotting with antiphosphotyrosine antibodies (Fig. 2A). We found that the 41-and 44-kDa MAP kinase isoforms were phosphorylated on tyrosine to a greater extent after stimulation with LPS. Since similar amounts of these MAP kinases were immunoprecipitated from the unstimulated and the LPS-stimulated cell lysates (Fig. 2B), these results directly demonstrate that LPS treatment increases the tyrosine phosphorylation of the 41-and 44-kDa MAP kinases.
LPS Activates MAP Kinase Isoforms-Since LPS increases the tyrosine phosphorylation of two MAP kinase isozymes, and as tyrosine phosphorylation is thought to be a critical event activating these enzymes, we tested whether LPS alters MAP kinase activity. RAW 264.7 cell lysates were fractionated by MonoQ anion-exchange chromatography, and the resulting column fractions were analyzed for MAP kinase activity and immunoblotted with antiphosphotyrosine antibodies or anti-MAP kinase antibodies. LPS treatment resulted in a large increase in MAP kinase activity as assessed by the phosphorylation of MBP, a standard substrate of MAP kinase (Fig. 3A). The increased activity was partially resolved into two peaks by MonoQ chromatography with a main peak of activity in column fractions 30-35 and a smaller, second peak in fractions 35-38. In some other experiments the resolution into two peaks was more evident (see below).

LPS Activation of M A P Kinases in Macrophages
LPS treatment also resulted in the increased tyrosine phosphorylation of the 41-and 44-kDa proteins that correspond to isoforms of MAP kinase (Fig. 3, B and C). Induced phosphorylation of these isozymes was accompanied by changes in the migration of these proteins on the immunoblots (Fig.   3, B-G) . As was observed in Fig. 1, most of the 41-kDa MAP kinase isoform was shifted to a higher apparent molecular weight as indicated by the arrow in Fig. 3 (B-G). In addition, the 44-kDa MAP kinase isoform from LPS-stimulated cells eluted from the MonoQ column in later fractions indicating that this protein was bound more tightly to the column. Both of these phenomena are thought to be indicative of increased phosphorylation of these proteins (17)(18)(19)(20).
Comparison of the immunoblotting data and the MAP kinase activity profile revealed that column fractions with elevated activity contained both the tyrosine-phosphorylated, 41-and 44-kDa isoforms of MAP kinase. Thus, the contribution of each isozyme to a particular activity peak could not be determined. In any case, LPS increased MAP kinase activity in RAW 264.7 cells, and this effect correlated with the induced tyrosine phosphorylation of a t least two MAP kinase isoforms.
Similar increases in MAP kinase activity were observed following stimulation of RAW 264.7 cells with smooth or rough forms of LPS as well as purified bacterial or synthetic lipid A preparations (Fig. 4). Thus, different biologically active forms of LPS activate MAP kinases and this response appears to be lipid A-dependent, as is the case for almost all the effects of LPS on macrophages (4).
Herbimycin A Inhibits LPS-induced Activation of MAP Kinases-The enzymatic activity of MAP kinase isozymes is regulated in part by phosphorylation on tyrosine residues (21). Since LPS induces the tyrosine phosphorylation of the 41-and 44-kDa MAP kinase isoforms, LPS may modulate MAP kinase activity by inducing ~4 1 1~4 4 MAP kinase tyrosine phosphorylation. To test this possibility, we examined the effect of preventing LPS-induced tyrosine phosphorylation of ~4 1 1~4 4 on the MAP kinase activity. Previously, we reported that herbimycin A, a protein tyrosine kinase inhibitor (22-24), completely blocks tyrosine phosphorylation of all the proteins modulated by LPS including the 41-and 44-kDa proteins (6). Herbimycin A pretreatment also inhibited LPSinduced MAP kinase activity (Fig. 5A). In addition, the LPSinduced mobility shift of the 41-kDa immunoreactive MAP kinase isoform and the delayed elution of the 44-kDa isoform were absent in cells pretreated with herbimycin A (Fig. 5 pg/ml of wild type Salmonella minnesota LPS ( WT), rough type LPS from S. minnesota re595, synthetic E. coli type lipid A, or purified bacterial lipid A from S. minnesota (DPL) were subjected to MonoQ chromatography as described under "Experimental Procedures." The column fractions were assayed for MAP kinase activity using MBP as a substrate. Values are expressed relative to the maximal MBP phosphorylating activity in the first peak from MonoQ (fraction numbers 31-33). The relative MBP phosphorylating activity in the second MonoQ peak is also shown (fraction numbers 35-39). Similar results were obtained in two independent experiments. also induces the tyrosine phosphorylation of 41-and 44-kDa species that correspond to MAP kinase isoforms (Fig. 1). Therefore, we tested whether MAP kinase activity is elevated in PMA-stimulated RAW 264.7 cells. PMA stimulation strongly increased the amount of MAP kinase activity with two peaks of activity partially resolved by MonoQ chromatography (Fig. 6A). A large peak of activity eluted in fractions 28-33 and a second, smaller peak that resolved as a shoulder of the first peak eluted in fractions 34-40. Thus, like LPS, PMA appeared to activate a t least two isoforms of MAP kinase. Interestingly, following PMA treatment two 44-kDa MAP kinase isoforms with different MonoQ elution patterns were detected on antiphosphotyrosine and anti-MAP kinase immunoblots (Fig. 6, B-D). One of these proteins eluted from the column in the same fractions as p44 MAP kinase from unstimulated cells and appeared in lane 4 on the blots. The other protein was more strongly bound to the column and was detected in lanes 7 and 8. A similar phenomenon has been observed in nerve growth factor-stimulated PC12 cells (20) and in insulin-treated rat fibroblasts (19). It is unclear whether this observation results from the detection of two different MAP kinase isozymes of similar molecular weight or two different forms of the same isozyme. As mentioned earlier, LPS also induced some increased retention of p44 to the MonoQ column. However, LPS stimulation did not result in the appearance of two distinct 44-kDa MAP kinase species. Thus, the effects induced by LPS and PMA on the 44-kDa isoform(s) of MAP kinase were not identical in RAW 264.7 cells.
Protein Kinase Inhibitors Differentially Affect LPSand PMA-stimulated MAP Kinase Actiuation-To further inves- tigate the mechanism by which LPS and PMA activate MAP kinase isozymes, we examined the effect of herbimycin A on the response triggered by PMA. Herbimycin A treatment, which completely blocked LPS-induced tyrosine phosphorylation and activation of MAP kinase, only weakly inhibited PMA induction of these responses (Fig. 7, A, C, E, C). Since, the targets of herbimycin A action are thought to be protein tyrosine kinases, the activation of MAP kinase by LPS appears to involve a herbimycin-sensitive tyrosine kinase, whereas the PMA-induced response does not.
The results with a protein kinase C inhibitor provide further evidence that LPS-induced activation and PMA-induced activation of MAP kinase are mechanistically different. We found that pretreatment with the staurosporine analog, Compound 3 (25), inhibited PMA-induced tyrosine phosphorylation of p41 and p44 and also inhibited activation of MAP kinase (Fig. 7, A, B, D, and F). This result was expected since phorbol esters are thought to exert their effects on cells by activating protein kinase C. In contrast, these LPS-induced responses were insensitive to Compound 3. (Fig. 8). Thus, the mechanism by which LPS activates MAP kinases in RAW 264.7 cells does not appear to be dependent on protein kinase C.

DISCUSSION
LPS stimulation of macrophages results in the increased tyrosine phosphorylation of several proteins. Here, we report that two of the induced tyrosine phosphoproteins, of 41 and 44 kDa, correspond to isoforms of MAP kinase. Each of these proteins was immunoreactive with anti-MAP kinase antibodies and could be partially depleted from cell lysates by immunoprecipitation with these antibodies. In addition, both pp41 and pp44 coeluted with MAP kinase activity following MonoQ chromatography. Together, these results show that the 41-and 44-kDa proteins, whose tyrosine phosphorylation is induced by LPS, are isozymes of MAP kinase.
To date, at least four highly related MAP kinases have been described in a variety of species by biochemical, immunological, and molecular cloning data (8-10). One isoform has a mass of 42 kDa and is tyrosine-phosphorylated in response to mitogenic stimulation in a wide variety of cells. This MAP kinase isoform has been designated ~4 2 " ' '~~ and corresponds to the ERK 2 gene product (20, 26). A slightly larger MAP kinase isoform (43-44 kDa) exhibits induced tyrosine phosphorylation often in parallel with increased phosphorylation of the 42-kDa isoform and is thought to correspond to ERK 1 (10). The existence of additional MAP kinase isoforms has been inferred from the molecular cloning of a third sequencerelated cDNA (ERK 3; predicted molecular mass, 63 kDa) (20) and from the immunoblotting of a 45-kDa polypeptide (ERK 4) with anti-ERK 1 antibodies (20).3 The 41-kDa MAP kinase isoform from RAW 264.7 macrophages behaves most similarly to P~~"'"~/ERK 2. In addition to the similar molecular mass, both MAP kinases exhibit slightly decreased mobility on SDS-polyacrylamide gels, following cellular activation. Moreover, anti-~44"'~~ antibodies recognized the 41-kDa isoform more efficiently after LPS or PMA stimulation of RAW 264.7 cells. This property was previously observed with p42 MAP kinase from Xenopus oocytes (17). The 44-kDa MAP kinase isoform from RAW 264.7 cells behaves like p44 MAP kinase from PC 12 cells (20) and rat fibroblasts (19). For each of these 44-kDa isoforms, cellular stimulation leads to stronger binding of the 44-kDa protein to a MonoQ column. Although positive identification will require additional experiments, it seems likely that the two isoforms of MAP kinase activated by LPS in macrophages correspond to the two major MAP kinase isoforms seen in other cell types.
LPS-induced tyrosine phosphorylation of the two MAP kinase isoforms in RAW 264.7 macrophages was accompanied by increased MAP kinase activity. This activity could be partially resolved into two peaks by MonoQ chromatography suggesting that at least two MAP kinase isoforms were activated by LPS. Each of the column fractions with elevated MAP kinase activity contained tyrosine phosphorylated forms of the 41-and 44-kDa MAP kinase isozymes. Since tyrosine phosphorylation of MAP kinases is necessary for the activation of these proteins, both the 41-and 44-kDa isoforms are likely to contribute to the observed MAP kinase activity. The loss of MAP kinase activity which accompanied inhibition of induced tyrosine phosphorylation of these proteins is consistent with this interpretation. Moreover, the antiphosphotyrosine blots of the active column fractions did not reveal the presence of any other tyrosine phosphoproteins. This observation makes it unlikely that additional MAP kinase isozymes which were not detected by our MAP kinase antibodies were responsible for the observed MAP kinase activity. Therefore, LPS appears to modulate MAP kinase activity in RAW 264.7 cells by inducing the tyrosine phosphorylation of the 41-and 44-kDa MAP kinase isoforms. Since phosphorylation on threonine residues has been shown to be necessary for the full activation of MAP kinase (22, 27), LPS may additionally modulate MAP kinase activity by inducing increased threonine phosphorylation of these proteins. Alternatively, MAP kinases in RAW 264.7 cells may be phosphorylated on regulatory threonine residues prior to LPS stimulation.
The mechanism by which LPS treatment increases the tyrosine phosphorylation of MAP kinases is not clear. Recent evidence from several groups has indicated that MAP kinases can autophosphorylate on tyrosine residues as well as threonine residues (14, 28-30). Thus, LPS-induced tyrosine phosphorylation of MAP kinase isoforms may occur by a LPSstimulated autophosphorylation mechanism, independent of other protein tyrosine kinases. Alternatively, increased MAP kinase tyrosine phosphorylation could be a consequence of LPS-activated protein tyrosine kinases.
The ability of herbimycin A to inhibit both LPS-induced tyrosine phosphorylation and activation of the MAP kinase proteins suggests that an activated protein tyrosine kinase is necessary for these LPS responses. While the mechanism of action of herbimycin A is not completely understood, this inhibitor appears to inactivate protein tyrosine kinases by irreversibly binding to thiol groups in the affected kinases (22, 23). In addition, herbimycin A binding to these kinases promotes their degradation (22, 31). Since the amount of MAP kinases detected by immunoblotting did not change following herbimycin A treatment, these kinases may not be targets of this inhibitor. Moreover, PMA-induced MAP kinase tyrosine phosphorylation and activation was only weakly affected by herbimycin A treatment. This result demonstrates that herbimycin A does not directly inhibit MAP kinase. Thus, it seems most likely that herbimycin A inhibits an upstream protein tyrosine kinase that is necessary for the LPS-induced tyrosine phosphorylation of MAP kinases.
The effects of herbimycin A on LPS-and PMA-induced MAP kinase activity also suggest that MAP kinases may mediate some of the antibacterial responses of macrophages. First, herbimycin A, which prevented the LPS-stimulated increase in MAP kinase activity, also inhibits the release of arachidonic acid metabolites from LPS-treated RAW 264.7 macrophages (6). Arachidonic acid metabolites are potent inflammatory mediators, and their secretion by macrophages is characteristic of the activated state. In contrast to the results with LPS, herbimycin A did not strongly inhibit PMAstimulated MAP kinase activation or the release of arachidonic acid metabolites (Fig. 7. and Ref. 6). Thus, the induction of MAP kinase activity by LPS and PMA appears to be correlated with the appearance of at least some downstream macrophage responses. A further indication of the relationship between MAP kinase activity and macrophage activation was provided by the results obtained with the protein kinase C inhibitor, Compound 3, in PMA-stimulated cells. Compound 3, which blocked PMA-induced MAP kinase activation, also inhibited the release of arachidonic acid metabolites (data not shown). However, the results with RAW 264.7 macrophages pretreated with Compound 3 and then stimulated with LPS do not fit this pattern. Compound 3 treatment, which did not inhibit LPS induction of MAP kinase activity, did inhibit LPS stimulation of arachidonic acid metabolite release (data not shown). Thus, in this case, an activated macrophage response was not triggered despite the induction of MAP kinase activity by LPS. This result, however, is still consistent with the hypothesis that MAP kinases participate in LPSstimulated signal tranduction. For example, one obvious explanation is that protein kinase C, or another protein kinase that is inhibited by Compound 3, participates in the LPS signaling pathway downstream of MAP kinase activation. Alternatively, the release of arachidonic acid metabolites induced by LPS in RAW 264.7 macrophages may require the generation of two independent intracellular signals, one provided by MAP kinases and the other by protein kinase C. Clearly the experiments presented here do not prove that MAP kinases mediate macrophage responses to LPS, but the results with the protein kinase inhibitors are consistent with this hypothesis.
The precise function of MAP kinases in macrophages and in other cells is, however, not yet known. Several proteins have been found to be efficient in vitro substrates of MAP kinases, and these may be indicative of MAP kinase function in uivo. For example, phosphorylation of microtubule-associated protein 2 by MAP kinases (32, 33) may alter the cytoskeleton and could provide a molecular mechanism for the morphological changes induced by LPS in macrophages. Similarly, MAP kinases can phosphorylate and activate both the S6 ribosomal protein kinase (11, 12) and the c-jun product (13), proteins that are involved in the regulation of translation and transcription, respectively. If MAP kinases phosphorylate these targets in vivo, these kinases may contribute to the altered expression of LPS-modulated proteins. Since many of the responses triggered by LPS in macrophages depend on transcription and translation, LPS activation of MAP kinases could be a critical part of the mechanism by which LPS induces responses in macrophages. Despite the incomplete understanding of the role of MAP kinases in cells, our results suggest that these kinases could be very important targets of LPS action in macrophages and lend support to the hypothesis that induced protein tyrosine phosphorylation is part of the signal transduction pathway that mediates macrophage responses to LPS.