Aeromonas hydrophila Cytotoxic Enterotoxin Activates Mitogen-activated Protein Kinases and Induces Apoptosis in Murine Macrophages and Human Intestinal Epithelial Cells*

A cytotoxic enterotoxin (Act) of Aeromonas hydrophila possesses several biological activities, induces an inflammatory response in the host, and causes apoptosis of murine macrophages. In this study, we utilized five target cell types (a murine macrophage cell line (RAW 264.7), bone marrow-derived transformed macrophages, murine peritoneal macrophages, and two human intestinal epithelial cell lines (T84 and HT-29)) to investigate the effect of Act on mitogen-activated protein kinase (MAPK) pathways and mechanisms leading to apoptosis. As demonstrated by immunoprecipitation/kinase assays or Western blot analysis, Act activated stress-associated p38, c-Jun NH2-terminal kinase (JNK), and extracellular signal-regulated kinase 1/2 (ERK1/2) in these cells. Act also induced phosphorylation of upstream MAPK factors (MAPK kinase 3/6 (MKK3/6), MKK4, and MAP/ERK kinase 1 (MEK1)) and downstream effectors (MAPK-activated protein kinase-2, activating transcription factor-2, and c-Jun). Act evoked cell membrane blebbing, caspase 3-cleavage, and activation of caspases 8 and 9 in these cells. In macrophages that do not express functional tumor necrosis factor receptors, apoptosis and caspase activities were significantly decreased. Immunoblotting of host whole cell lysates revealed Act-induced up-regulation of apoptosis-related proteins, including the mitochondrial proteins cytochrome c and apoptosis-inducing factor. However, mitochondrial membrane depolarization was not detected in response to Act. Taken together, the data demonstrated for the first time Act-induced activation of MAPK signaling and classical caspase-associated apoptosis in macrophages and intestinal epithelial cells. Given the importance of MAPK pathways and apoptosis in inflammation-associated diseases, this study provided new insights into the mechanism of action of Act on host cells.

A cytotoxic enterotoxin (Act) of Aeromonas hydrophila possesses several biological activities, induces an inflammatory response in the host, and causes apoptosis of murine macrophages. In this study, we utilized five target cell types (a murine macrophage cell line (RAW 264.7), bone marrow-derived transformed macrophages, murine peritoneal macrophages, and two human intestinal epithelial cell lines (T84 and HT-29)) to investigate the effect of Act on mitogen-activated protein kinase (MAPK) pathways and mechanisms leading to apoptosis. As demonstrated by immunoprecipitation/kinase assays or Western blot analysis, Act activated stress-associated p38, c-Jun NH 2 -terminal kinase (JNK), and extracellular signal-regulated kinase 1/2 (ERK1/2) in these cells. Act also induced phosphorylation of upstream MAPK factors (MAPK kinase 3/6 (MKK3/6), MKK4, and MAP/ERK kinase 1 (MEK1)) and downstream effectors (MAPK-activated protein kinase-2, activating transcription factor-2, and c-Jun). Act evoked cell membrane blebbing, caspase 3-cleavage, and activation of caspases 8 and 9 in these cells. In macrophages that do not express functional tumor necrosis factor receptors, apoptosis and caspase activities were significantly decreased. Immunoblotting of host whole cell lysates revealed Act-induced up-regulation of apoptosis-related proteins, including the mitochondrial proteins cytochrome c and apoptosis-inducing factor. However, mitochondrial membrane depolarization was not detected in response to Act. Taken together, the data demonstrated for the first time Act-induced activation of MAPK signaling and classical caspase-associated apoptosis in macrophages and intestinal epithelial cells. Given the importance of MAPK pathways and apoptosis in inflammation-associated diseases, this study provided new insights into the mechanism of action of Act on host cells.
Aeromonas species (spp.) are significant human pathogens that have been isolated from freshwater, salt water, and a variety of foods, and are frequently isolated from patients with diarrhea (1)(2)(3). Among Aeromonas spp., Aeromonas hydrophila is most commonly associated with human infections, leading to intestinal and non-intestinal diseases (4). Furthermore, increased resistance of this organism to antibiotics and chlorination in water presents a significant threat to public health (2,3,(5)(6)(7)(8)(9)(10)(11)(12). As a result, the Environmental Protection Agency has placed this organism on the "Candidate Contaminant List," and the monitoring of United States water supplies for this organism began in 2002 (4).
A. hydrophila produces an array of virulence factors, one of the most significant of which is a 52-kDa cytotoxic enterotoxin (Act) 1 (13). Our laboratory demonstrated that Act possessed several biological activities, including hemolysis, cytotoxicity, enterotoxicity, and lethality to mice, and induced acute inflammatory responses both in vitro and in vivo (9). Additional studies by our laboratory revealed that Act significantly altered intracellular signaling in murine RAW 264.7 macrophages, which could lead to an inflammatory response in host cells (14 -16). The pattern of Act-induced gene expression changes, based on our recent microarray studies, strongly suggested that Act might activate mitogen-activated protein kinase (MAPK) pathways (16).
MAPKs are activated by phosphorylation on both a threonine and tyrosine residue, which is accomplished by dual phosphorylation enzymes, called MAP/ERK kinases (MEKs) or MAPK kinases (MKKs). MEKs are themselves activated via phosphorylation by MEK kinases (MEKKs) or MAPK kinase kinases. There are four major MAPK cascades that lead to altered gene expression: ERK1/2, JNK, p38 kinase, and ERK5. ERK1 and ERK2 are activated via phosphorylation by MEK1 and MEK2, respectively (17).
The two MAPK pathways most frequently associated with immune responses are JNK and p38 kinases. The JNK signaling pathway is a stress-activated MAPK pathway involved in the regulation of cell proliferation and apoptosis (18). The p38 MAPK signaling pathway is associated with cell growth, differentiation, and death, and is activated by a variety of cytokines, including interleukin-1 (IL-1) and tumor necrosis factor-␣ (TNF-␣) (17,19). Three MAPKK proteins (MKK3, MKK4, and MKK6) can phosphorylate and activate p38 kinase (17). The activation of p38 leads to the production of several cytokines, including IL-1␤, TNF-␣, IL-6, and IL-8, as well as cyclooxygenase-2 and inducible nitric-oxide synthase enzyme, and phosphorylation of phospholipase A 2 (PLA 2 ) (20 -25). The ERK5 pathway is the most recently described MAPK pathway and is primarily associated with response to stress (26,27). ERK5 is activated via MEK5, which is itself activated by MEKK3 (28).
In the current study, we investigated the effects of Act on MAPK signaling in RAW 264.7 cells, murine peritoneal macrophages, and human intestinal epithelial cells (T84 cells) to determine the role of MAPK signaling in host cells and ultimately dissect the definitive role of signaling molecules that lead to the disease state during Aeromonas infection.
Because we previously demonstrated that Act caused apoptosis of RAW 264.7 murine macrophages (16), we also investigated mechanisms associated with Act-induced apoptosis in murine peritoneal (primary) macrophages, TNF receptor (TNFR) knockout (KO) murine bone marrow-derived and transformed macrophages (29), and human intestinal epithelial cells. Taken together, the data clearly demonstrated for the first time that Act-induced host cell signaling involves activation of ERK1/2, p38, and JNK pathways. Furthermore, we delineated the mechanisms of apoptosis in Act-treated host cells.

EXPERIMENTAL PROCEDURES
Cell Culture-The murine macrophage (RAW 264.7) and human intestinal epithelial (T84 and HT-29) cell lines were purchased from the American Type Culture Collection (Manassas, VA). WT, double KO murine bone marrow-derived and transformed macrophages (deficient in both TNFR-1 and TNFR-2), and TNFR1/TNFR2 single KO macrophages were graciously provided by Dr. Aggarwal (29). RAW 264.7 and HT-29 cells were cultured at 37°C and 5% CO 2 in Dulbecco's minimal essential medium (Life Technologies, Inc., Gaithersburg, MD) containing 4.5 g of glucose/liter, 10% fetal bovine serum, 2 mM L-glutamine, and antibiotics, penicillin (100 units/ml) and streptomycin (0.1 mg/ml). T84 cells were cultured similarly, but in Dulbecco's minimal essential medium/F-12 (Invitrogen) and 5% fetal bovine serum. WT and TNFR KO macrophages were cultured in RPMI (Invitrogen) and 5% fetal bovine serum. For each experiment, 5 ϫ 10 5 cells/ml were plated in 35-mm dishes and allowed to attach overnight. The medium was removed, and fresh medium containing the stimulant (6 ng/ml (unless otherwise stated) lipopolysaccharide (LPS)-free Act) was added (14). After each time point, cells were lysed using 100 l/well of 1ϫ Cell Lysis Buffer (Cell Signaling Technology, Beverly, MA), sonicated twice for 15 s, and centrifuged at 10,000 ϫ g for 10 min.
Peritoneal Macrophage Isolation-Adult 25-30 g Swiss Webster female mice (6 -8 weeks old) were purchased from Taconic Farm (Germantown, CA). Mice were housed in the specific pathogen-free animal facility at the University of Texas Medical Branch, Galveston, and given free access to food and water prior to experiments. A 1-ml solution of 3% thioglycolate was injected into the peritoneal cavity of mice to elicit macrophage migration into the cavity. Mice were allowed to rest for 4 days and then sacrificed by an overdose of isoflurane followed by cervical dislocation. A partial midline incision was made to remove the dermis and expose the parietal peritoneum lining. Peritoneal exudate cells were harvested by flushing the peritoneal cavity with 7 ml of RPMI medium (Invitrogen) using a 10-ml syringe fitted with a 21-gauge, 1 1 ⁄2-inch needle. Extracted cells were counted, centrifuged, and plated at 1 ϫ 10 8 cells in 35-mm tissue culture dishes containing RPMI supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Cells were allowed to adhere for 1 h, and then non-adherent cells were washed from the cultures. Complete medium was added to the cultures and Act (12 or 20 ng/ml) was added. Cells were incubated for the appropriate times before being lysed for subsequent Western blot analysis.
Western Blot Analysis-The antibody for ␤-tubulin was purchased from Santa Cruz Biochemical Corp. (Santa Cruz, CA). All other antibodies and Jurkat T cell control extracts were purchased from Cell Signaling Technology, and Western blot analysis was performed by established procedures (30) with slight modifications according to specifications of the antibody manufacturer. Briefly, equal amounts of total protein were loaded and separated on SDS, 10% polyacrylamide gels and then transferred to nitrocellulose membranes. Membranes were blocked with 5% milk and washed in 1ϫ Tween (0.1%), Tris-buffered saline three times for 5 min each. Primary antibodies diluted 1:1000 in 5% milk or bovine serum albumin (prepared in 1ϫ Tween (0.1%), Tris-buffered saline) were allowed to incubate overnight at 4°C. After washing, horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology) was diluted 1:2000 in 5% milk and applied to membranes. Subsequently, membranes were washed and a chemiluminescence substrate (Pierce, Rockford, IL) was applied and allowed to incubate at room temperature for 5 min.
Immunoprecipitation/Kinase Assay-Kinase assay kits for p38 and JNK were purchased from Cell Signaling Technology, and the procedure was performed as described by the manufacturer. Whole cell lysates were incubated with immobilized phospho-p38 antibody or c-Jun fusion protein overnight at 4°C. Following microcentrifugation and washing, the pellet was suspended in kinase buffer supplemented with 200 M ATP and 2 g of ATF-2 fusion protein (for p38 kinase activity) or 100 M ATP (for JNK activity) and incubated for 30 min at 30°C. The reaction was terminated with SDS sample loading buffer and the samples subjected to Western blot analysis.
Gel Shift Analysis-Gel shift assays were performed as previously described (14,16). Briefly, consensus oligonucleotide for CCAAAT/enhancer-binding protein-␤ (C/EBP-␤) was labeled using T4 polynucleotide kinase (Promega, Madison, WI) according to the manufacturer. Next DNA binding reaction mixtures were assembled. We used unlabeled AP-1 consensus oligonucleotide as a specific competitor, unlabeled AP-2 or SP-1 consensus oligonucleotides as nonspecific competitors (Promega), and nuclear extracts from specific time points after Act treatment of RAW cells. Nuclear extracts were prepared using NE-PER kit (Pierce) as described by the manufacturer. The reaction mixtures were incubated at room temperature for 10 min, followed by addition of 20,000 cpm of 32 P-labeled transcription factor consensus oligonucleotide, and incubation at room temperature for 20 min. Subsequently, 1 l of gel loading 10ϫ buffer (Promega) was added to each reaction mixture and samples (5 to 20 g) were loaded on a nondenaturing 4% polyacrylamide gel. The gel was pre-run in 0.5ϫ Tris borate-EDTA buffer for 30 min at 100 V before loading the samples. After completion of the run, the gel was transferred to Whatman 3MM paper, dried at 80°C for 5 h, and exposed to x-ray film overnight to 48 h.
Laser Scanning Confocal Microscopy-Cells (RAW 264.7 and T84) were grown in cover glass chamber slides (Nalge Nunc International, Rochester, NY) and observed on a Zeiss 510 UV meta confocal microscope (Carl Zeiss, Inc., Thornwood, NY). To view membrane blebbing, Act was added (20 ng/ml), and images were taken every 30 s for 6 h. To examine mitochondrial membrane potential, Act-treated RAW 264.7 cells were stained with 10 M MitoTracker red CM-H 2 XRos dye (Molecular Probes, Eugene, OR) for 30 min at 37°C, followed by washing with phosphate-buffered saline (PBS). All observations were made using a plan-neofluar ϫ40/0.85 objective lens with an electronic zoom of 2. The Nomarski differential interference contrast images were stored on an optical disk and analyzed with Zeiss LSM software.
Flow Cytometry-For detection of apoptosis, variously stimulated HT-29 cells (1 ϫ 10 6 ) were stained with annexin V-conjugated to fluorescein isothiocyanate (FITC) and propidium iodide (PI) by using the annexin-V-FITC staining kit (BioVision Inc.) according to the manufacturer's instruction. Briefly, cells were suspended in 500 l of annexin V binding buffer and incubated with 5 l of annexin V and 5 l of PI for 5 min at room temperature in the dark. At least 20,000 cells were acquired in a FACSCalibur flow cytometer using CellQuest 3.0.1 software (BD Biosciences). Percentages of cells undergoing apoptosis were determined by dual-color analysis. This staining allowed us to distinguish 3 subsets of cells: viable cells (annexin V negative and PI negative), early apoptotic cells (annexin V positive and PI negative), and late apoptotic or necrotic cells (annexin V positive and PI positive). Immediately after staining, the cells were analyzed on a flow cytometer using 488-nm excitation and a 525-nm band pass filter for FITC and a 620-nm filter for PI detection.
Apoptotic Protein Screens-Whole lysate protein samples (350 g) from Act-treated (20 ng/ml) RAW 264.7 cells were analyzed by Kinetworks TM KAPS-1.0 Apoptosis Screens (Kinexus Bioinformatics Corp., Vancouver, British Columbia, Canada) as previously described (view the Kinexus website for details and publication links). The Kinetworks TM analysis involved resolution of a single lysate sample by SDS-PAGE and subsequent immunoblotting with panels of up to 3 primary antibodies per channel in a 20-lane Immunetics Multiblotter. The an-tibody mixtures were carefully selected to avoid overlapping crossreactivity with target proteins. Normalized trace quantity units (CPM) were arbitrary based on the intensity of ECL fluorescence detection for target immunoreactive proteins recorded with a Fluor-S MultiImager and quantified using Quantity One Software (Bio-Rad).
Isolation of Cytosolic and Mitochondrial Lysates-RAW 264.7 cells, treated with 20 ng/ml Act for 0, 2, 4, and 6 h, were washed with PBS and centrifuged at 1,000 ϫ g for 5 min. Cell pellets were washed twice with PBS and suspended in ice-cold buffer containing 20 mM HEPES (pH 7.5), 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 20 g/ml aprotinin, 10 g/ml pepstatin A, and 250 mM sucrose. Cells were subsequently incubated on ice for 15 min and then homogenized by 15 passages though a 22-gauge needle. Homogenates were centrifuged at 1,000 ϫ g for 5 min at 4°C, and the resulting supernatants were centrifuged twice at 12,000 ϫ g for 15 min at 4°C. Supernatants (cytosolic fractions) were subsequently subjected to Western blot analysis. All of the experiments were performed at least in triplicate and representative data are presented.

Act Induces Phosphorylation/Activation of p38 Kinase in RAW 264.7 Cells, Murine Peritoneal Macrophages, and Human
Intestinal Epithelial (T84) Cells-RAW 264.7 cells, thioglycolate-induced murine peritoneal macrophages, or T84 cells were treated with 6, 12, or 20 ng/ml of Act for various time points and whole cell lysates were subjected to electrophoresis and Western blot analysis. LPS-treated RAW 264.7 cells (100 ng/ml for 30 min) were used as a positive control in these experiments and other MAPK Western blot analyses. LPS has been shown to induce ERK, p38, and JNK pathways in RAW 264.7 cells (32). Subsequent blots were probed with antibodies specific for the phosphorylated form of p38 (p-p38) or total p38␣. As shown in Fig. 1A, Act (6 ng/ml) induced an increase in a 43-kDa sized band, consistent with the phosphorylated form of p38, by 1 h. This increase was observed up to 8 h but disappeared by 12 h. Western blot analysis on samples from RAW 264.7 cells treated FIG. 1. Act-induced phosphorylation of p38 in RAW 264.7 cells, murine peritoneal macrophages, and T84 cells, and increased activity of p38 kinase in Act-treated RAW 264.7 cells. Cells were stimulated with Act for the indicated times, lysed, and Western blot analysis was performed with anti-phospho-p38 (p-p38) antibody. Blots were probed with anti-p38␣ (total p38␣) antibody or anti-p38␦ (total p38␦) antibody as a control for protein loading. Panels A and B, Western blot analysis of Act-treated (6 ng/ml) RAW 264.7 cells. Panels C and D, Western blot analysis of Act-treated (20 ng/ml) murine peritoneal macrophages. Panels E and F, Western blot analysis of Act-treated (12 ng/ml) T84 cells. Panel G, lysates from Act-treated (12 ng/ml) RAW 264.7 cells were immunoprecipitated with immobilized p-p38, kinase assays were performed in the presence of ATP and the p38 substrate ATF-2, and Western blot analysis was performed with anti-phospho-ATF-2 (p-ATF-2) antibody.
with a higher dose (12 ng/ml) of Act (for up to 8 h) resulted in similar results, with the earliest induction of p38 phosphorylation observed at 30 min (data not shown). Increased phosphorylation of p38 observed upon Act treatment was not because of an increase in the protein concentration of total p38, as evidenced by similarly sized bands for each time point when blots were probed with an antibody that recognized both the phosphorylated and non-phosphorylated forms of p38␣ (Fig. 1B). Act-treated murine peritoneal macrophages exhibited a similar pattern of p38 phosphorylation for both concentrations of Act that were used (12 and 20 ng/ml): Act induced phosphorylation of p38 by 1 h (Fig. 1C), but no increase was observed for total p38␣ (Fig. 1D). Likewise, p38 phosphorylation was observed for Act-treated (12 ng/ml) T84 cells by 1 h (Fig. 1E), with no increase in total protein concentration (Fig. 1F). Because no bands were observed for total p38␣ in T84 cells (data not shown), we used an antibody specific for p38␦ (Fig. 1F). Antibodies for p-p38 recognized both phosphorylated isoforms (p38␣ and p38␦), but it is likely that the phosphorylated form observed at 1 h in T84 cells was not the p38␣ isoform.
To demonstrate that increased phosphorylation of p38 in Act-treated cells was biologically significant, we examined p38 kinase activity in Act-treated RAW cells. Whole cell lysates were immunoprecipitated using an antibody specific for the phosphorylated form of p38, and kinase assays were performed using ATF-2, a known substrate of p38. The resulting reactions were then subjected to electrophoresis and the blots were probed with an antibody specific for the phosphorylated form of ATF-2. As shown in Fig. 1G, a 70-kDa sized band consistent with phosphorylated ATF-2 was observed by 30 min (earliest time point tested), which increased ϳ3.5-fold by 1 h. All immunoprecipitated samples contained the same amount of protein, and equal concentrations of ATP and ATF-2 substrate were used. Western blot analysis using an antibody specific for total ATF-2 demonstrated bands of similar size and intensity for each time point and served as a positive control (data not shown).
Act Induces Phosphorylation/Activation of JNK in RAW 264.7 Cells, Murine Peritoneal Macrophages, and Human Intestinal Epithelial (T84) Cells-Whole cell lysates from Acttreated (12 ng/ml) RAW 264.7 cells, murine peritoneal macrophages, or T84 cells were subjected to electrophoresis and Western blot analysis using antibodies specific for either the phosphorylated form of JNK or total JNK protein. As shown in Fig. 2A, two bands consistent with phospho-JNK were apparent by 30 min for Act-treated RAW 264.7 cells and continued up to 3 h (the last time point tested). Blots were probed with total JNK as an internal control to ensure equal loading of protein (Fig. 2B). Likewise, JNK phosphorylation was observed for Act-treated murine peritoneal cells (Fig. 2C), which was apparent by 1 h and continued to at least 8 h (last time point tested). Act induction of phosphorylated JNK was not because of an increase in total protein concentration (Fig. 2D). JNK phosphorylation was also detected in Act-treated T84 cells, although the earliest time point for phosphorylation was 2 h (Fig. 2E). JNK phosphorylation in T84 cells occurred in the absence of total protein up-regulation (Fig. 2F).
To further verify activation of JNK, we performed immuno- Act-induced phosphorylation of JNK in RAW 264.7 cells, murine peritoneal macrophages, and T84 cells, and increased activity of JNK kinase in Act-treated RAW 264.7 cells. Cells were stimulated with Act for the indicated times, lysed, and Western blot analysis was performed with antiphospho-JNK (p-JNK) antibody. Blots were probed with anti-JNK (total JNK) antibody as a control for protein loading. Panels A and B, Western blot analysis of Act-treated (12 ng/ml) RAW 264.7 cells. Panels C and D, Western blot analysis of Act-treated (12 ng/ml) murine peritoneal macrophages. Panels E and F, Western blot analysis of Act-treated (12 ng/ml) T84 cells. Panel G, lysates from Act-treated (12 ng/ml) RAW 264.7 cells were immunoprecipitated with immobilized c-Jun fusion protein, kinase assays were performed in the presence of ATP, and Western blot analysis was performed with anti-phospho-c-Jun antibody.
precipitation/kinase assays on whole cell lysates from Acttreated (12 ng/ml) RAW 264.7 cells. Phosphorylated JNK was immunoprecipitated with substrate (c-Jun)-conjugated beads, kinase assays were performed, and the resulting reactions were subjected to electrophoresis and Western blot analysis using an antibody specific for phosphorylated c-Jun. As shown in Fig. 2G, JNK was indeed activated, because phosphorylated c-Jun was detected by 30 min with a ϳ3-fold increase by 2 h (last time point tested). As controls, equal concentrations of whole cells lysates were used, and Western blot analysis was performed using an antibody that recognized both the phosphorylated and non-phosphorylated forms of c-Jun (data not shown).
Act Induces Phosphorylation of MEK3/6 and MEK4 (but Not MEK7) in RAW 264.7 Cells and Murine Peritoneal Macrophages-Because MEK3/6 and MEK7 are known to phosphorylate p38 and JNK, respectively, and MEK4 has been shown to activate both the p38 and JNK pathways simultaneously, we examined the phosphorylation status of all three upstream kinases in Act-treated RAW 264.7 cells and murine peritoneal macrophages. Whole cell lysates from Act-treated (12 ng/ml) RAW 264.7 cells or murine peritoneal macrophages were subjected to electrophoresis and Western blot analysis using antibodies specific for phosphorylated or total (phosphorylated and non-phosphorylated) MEK3/6 ( Fig. 3). As shown in Fig. 3A, Act induced phosphorylation of MKK3/6 by 1 h in RAW 264.7 cells, which is consistent with the hypothesis that MKK3/6 is responsible for p38 phosphorylation at 1 h in Act-treated macrophages (Fig. 1A). All time points contained similar amounts of total MKK3/6 (Fig. 3B). Likewise, total MKK3/6 concentrations were similar for the untreated control and Act-treated (20 ng/ml) murine peritoneal macrophage samples (Fig. 3D), whereas phosphorylated MKK3/6 was up-regulated by 1 h (Fig. 3C).
Western blot analysis of whole cell lysates from Act-treated (12 ng/ml) RAW 264.7 cells also demonstrated MKK4 phosphorylation by 1 h (Fig. 3E) with no change in total protein concentration (Fig. 3F). Similar results were obtained for Acttreated (20 ng/ml) murine peritoneal macrophages (Fig. 3, G and H). However, Western blot analysis of Act-treated RAW 264.7 cells (12 ng/ml) did not demonstrate any increase in phosphorylated MKK7 (data not shown). Neither was MKK7 phosphorylated in Act-treated (12 or 20 ng/ml) murine peritoneal macrophages (data not shown). On the contrary, LPS at the tested dose (100 ng/ml) and time point (30 min) increased MKK7 levels as expected (data not shown).
Act Induces Phosphorylation of ATF-2, c-Jun, and p90RSK in RAW 264.7 Cells and Murine Peritoneal Macrophages and MAPKAPk-2 in Murine Peritoneal Macrophages-Because Act induced activation of p38 and JNK kinases and their upstream effectors, MKK3/6 and MKK4, we next evaluated the phosphorylation status of downstream factors known to be involved in MAPK signaling (Figs. 4 and 5). ATF-2, which could be phosphorylated by p38 or JNK, was indeed phosphorylated in Acttreated (12 ng/ml) RAW 264.7 cells by 30 min (Fig. 4A) and in murine peritoneal macrophages by 2 h (Fig. 4C). The observed increase in phosphorylation of ATF-2 was not because of an increase in overall protein concentration, as shown in Fig. 4, B and D. We also examined the phosphorylation status of MAP-KAPk-2, which is known to be phosphorylated by p38. As shown in Fig. 4E, Act treatment (12 ng/ml) of RAW 264.7 cells did not result in phosphorylation of MAPKAPk-2. Nor was an increase in protein production observed in response to toxin FIG. 3. Act-induced phosphorylation of MKK3/6 and MKK4 in RAW 264.7 cells and murine peritoneal macrophages. Cells were stimulated with Act for the indicated times, lysed, and Western blot analysis was performed with anti-phospho-MKK3/6 (p-MKK3/6) or p-MKK4 antibody. Blots were probed with anti-MKK3/6 (total MKK3/6) or total MKK4 antibody as a control for protein loading. Panels A and B, Western blot analysis of Act-treated (12 ng/ml) RAW 264.7 cells. Panels C and D, Western blot analysis of Act-treated (12 ng/ml) murine peritoneal macrophages. Panels E and F, Western blot analysis of Act-treated (12 ng/ml) RAW 264.7 cells. Panels G and H, Western blot analysis of Act-treated (20 ng/ml) murine peritoneal macrophages. treatment (Fig. 4F). However, MAPKAPk-2 was phosphorylated in Act-treated (20 ng/ml) murine peritoneal macrophages (Fig. 4G), which was not a result of increased protein production (Fig. 4H). An extra band of ϳ50 kDa in size was observed for all samples used (Fig. 4, E-H), which most likely represented hyperphosphorylation of MAPKAPk-2.
Because we demonstrated Act-induced activation of JNK in this study and of the downstream transcription factor AP-1 in RAW 264.7 cells previously (16), we next investigated whether or not c-Jun, the prototypical subunit of AP-1, was phosphorylated in Act-treated macrophages. As shown by Western blot analysis in Fig. 5A, Act treatment (12 ng/ml) of RAW 264.7 cells did lead to the phosphorylation of c-Jun, which was apparent by 2 h. Because protein levels of c-Jun, as well as other AP-1 subunits, are known to be altered by activation of all three major MAPK pathways, we used ␤-tubulin as the internal control for protein loading (Fig. 5B). Phosphorylation of c-Jun was also observed for Act-treated (12 ng/ml) murine peritoneal macrophages as early as 30 min (Fig. 5C) with equal loading of protein (Fig. 5D), suggesting that AP-1 activation occurs in these cells and perhaps earlier than for RAW 264.7 cells.
Because we previously demonstrated Act-induced activation of the transcription factors cyclic AMP-response element-binding protein and nuclear factor-B (NF-B) in RAW 264.7 cells (14,15), and p90RSK has been shown to activate cyclic AMPresponse element-binding protein and NF-B (33-35), we investigated phosphorylation of p90RSK in Act-treated RAW 264.7 cells and murine peritoneal macrophages. As shown in Fig. 5E, Act treatment (12 ng/ml) of RAW 264.7 cells led to the phosphorylation of p90RSK by 30 min, as determined by West-ern blot analysis. The increase in phosphorylated p90RSK was not due merely to an increase in total protein concentration, as determined by Western blot analysis using an antibody that recognized both the phosphorylated and non-phosphorylated forms of p90RSK (Fig. 5F). Act (12 ng/ml) also caused phosphorylation of p90RSK in murine peritoneal macrophages (2 h, Fig.  5G), although induction was later than for RAW 264.7 cells. Total p90RSK protein concentration in murine peritoneal macrophages was unchanged by the toxin treatment (Fig. 5H).
Act Induces Phosphorylation of ERK1/2 in RAW 264.7 Cells, Murine Peritoneal Macrophages, and Human Intestinal Epithelial (T84) Cells-Because p90RSK is activated by ERK kinase family members, we investigated the phosphorylation status of ERK1/2, the upstream ERK1/2 activator MEK-1, and ERK5 in toxin (12 ng/ml)-treated RAW 264.7 cells and murine peritoneal macrophages treated with 20 ng/ml Act. As shown by Western blot analysis, ERK1/2 phosphorylation was increased in response to Act, with up-regulation occurring as early as 15 min and more strongly by 30 min (Fig. 6A). MEK1/2 was also phosphorylated in response to Act in RAW 264.7 cells (Fig. 6B), although induction was detected later (1 h).
Act also caused phosphorylation of both ERK1/2 and MEK1/2 in murine peritoneal macrophages (Fig. 6, C and D, respectively), which was detected by 1 h (earliest time point tested). Total protein levels of ERK1/2 and MEK1 were unchanged in response to Act treatment in both RAW 264.7 cells and murine peritoneal macrophages (data not shown). Additionally, we investigated the phosphorylation status of ERK1/2 in Act-treated T84 cells (12 ng/ml), and as shown in Fig. 6E, an increase in ERK1/2 phosphorylation was observed by 2 h. No increase in FIG. 4. Act-induced phosphorylation of ATF-2 in RAW 264.7 cells and murine peritoneal macrophages and MAPKAPk-2 in murine peritoneal macrophages. Cells were stimulated with Act for the indicated times, lysed, and Western blot analysis was performed with anti-phospho-ATF-2 (p-ATF-2) or p-MAPKAPk-2 antibody. Blots were probed with anti-ATF-2 (total ATF-2) or total MAPKAPK-2 antibody as a control for protein loading. Panels A and B, Western blot analysis of Act-treated (12 ng/ml) RAW 264.7 cells. Panels C and D, Western blot analysis of Act-treated (12 ng/ml) murine peritoneal macrophages. Panels E and F, Western blot analysis of Acttreated (12 ng/ml) RAW 264.7 cells. Panels G and H, Western blot analysis of Acttreated (20 ng/ml) murine peritoneal macrophages.
Act Causes Activation of the Transcription Factor C/EBP-␤ in RAW 264.7 Cells-Our laboratory previously demonstrated Act-induced up-regulation of the macrophage-specific transcription factor C/EBP-␤ in RAW 264.7 cells using microarrays and real-time reverse transcriptase-PCR (16). Additionally, Act activated ERK1/2 in both RAW 264.7 cells (Fig. 6A) and murine peritoneal macrophages (Fig. 6C), which have been shown to activate C/EBP-␤ (36,37). Act has also been shown to induce the production of IL-1␤, IL-6, macrophage inflammatory protein-1␣, TNF-␣, and inducible nitric-oxide synthase (14 -16), all of which are transcription targets of C/EBP-␤ (38). We therefore performed gel shift assays to detect C/EBP-␤ activation in RAW 264.7 cells. Fig. 7 clearly demonstrated that Act (6 ng/ml) caused nuclear translocation of a protein capable of binding a Cells were stimulated with Act for the indicated times, lysed, and Western blot analysis was performed with anti-phospho-ERK1/2 (p-ERK1/2) antibody or antiphospho-MEK1 (p-MEK1) antibody. Blots were probed with anti-ERK1/2 or anti-MEK1 antibody as a control for protein loading (not shown). Panels A and B, Western blot analysis of Act-treated (12 ng/ml) RAW 264.7 cells. Panels C and D, Western blot analysis of Act-treated (20 ng/ml) murine peritoneal macrophages. Panel E, Western blot analysis of Acttreated (12 ng/ml) T84 cells.

radiolabeled C/EBP-␤ binding sequence. The binding was detectable by 30 min (lane 4) and strongly evident by 4 h (lane 6).
As a positive control for C/EBP-␤ activation, nuclear extracts from LPS-treated macrophages (2 h) were loaded in lane 1. We performed competition assays using unlabeled C/EBP-␤ (specific) and SP-1 or AP-2 (nonspecific) consensus oligonucleotides. As shown in Fig. 7, addition of the specific competitor abrogated binding (lane 9), whereas the nonspecific competitor did not (not shown), indicating that the band observed was indeed specific for C/EBP-␤.
Act Induces Apoptotic Blebbing of Murine Macrophages and Human Intestinal Epithelial (T84) Cells-In addition to the induction of inflammatory cascades, MAPKs have also been shown to mediate apoptosis (39,40). Recently, our laboratory demonstrated for the first time Act-induced apoptosis of RAW 264.7 cells (16). As further evidence that apoptosis is the mechanism of cell death, we visualized Act-treated RAW 264.7 cells by confocal microscopy. Live cells were grown on microscopy slides, treated with 20 ng/ml Act, and photographed every 15 s. As shown in Fig. 8A, cells were healthy before toxin treatment, and blebbing occurred by 2.5 h (Fig. 8B) and continued past 4 h (Fig. 8C). We also examined cellular morphology in Act-treated (20 ng/ml) RAW 264.7 cells by electron microscopy. As shown in Fig. 8D, Act treatment resulted in classical apoptosis morphology that was apparent by 3 h and even more pronounced by 6 h (data not shown). For instance, Act induced condensation of chromatin starting from the periphery of the nuclei, consistent with apoptosis-associated cleavage of DNA, caused cells to shrink and to become denser, and induced blebbing of fragments of the cytoplasm (Fig. 8D). Untreated cells were used as a negative control and were healthy (data not shown).
Because Act caused extensive intestinal damage in vivo, we also examined cell morphology of Act-treated T84 cells by confocal microscopy. Blebbing of peripheral cells began at ϳ2 h post-treatment (25 ng/ml of Act) and was clearly apparent by 5 h (Fig. 8F), suggesting that Act caused apoptosis of T84 cells. Fig. 8E depicts the morphology of untreated T84 cells.
To determine whether Act did indeed cause apoptosis of human intestinal epithelial cells, HT-29 cells were cultured with Act (25 ng/ml) or cycloheximide (100 g/ml) for 6 h and assayed for annexin V-FITC and PI staining by flow cytometry. As shown in Fig. 8, less than 5% of untreated HT-29 cells were apoptotic (panel G, lower right quadrant). However, HT-29 cells treated with 25 ng/ml Act for 6 h exhibited marked apoptosis (34.0%, quadrant 4, Fig. 8H), which was greater than that observed for cycloheximide-treated positive control cells (23.5%, quadrant 4, Fig. 8I). Considered with membrane blebbing of T84 cells in response to Act (Fig. 8F), the data demonstrated, for the first time, Act-induced apoptosis of intestinal epithelial cells, using two different cell lines.
Act Induces Activation of Caspases in Murine Macrophages and Human Intestinal Epithelial (T84) Cells-To determine whether Act-induced apoptosis of macrophages and intestinal epithelial cells was caspase-dependent, we performed Western blot analysis on whole cell lysates from Act-treated RAW 264.7 cells (20 ng/ml), murine peritoneal macrophages (12 ng/ml), and T84 cells (20 ng/ml). As shown in Fig. 9A, a faint band representing the cleaved/activated form of caspase 3 was apparent by 2 h, with a much stronger band appearing by 4 h and continuing to at least 6 h. Likewise, the cleaved product of caspase 3 was induced by Act treatment of murine peritoneal macrophages (Fig. 9B), although activation occurred later, beginning at 4 h and stronger by 8 h. Act treatment of T84 cells also resulted in cleavage of caspase 3, as evidenced by the lower-molecular weight band clearly visible by 4 h (Fig. 9C). Extracts from untreated Jurkat T cells were used as negative controls, and extracts from cytochrome c-treated Jurkat T cells were used as positive controls (Fig. 9). Blots were probed with anti-␤-tubulin antibody to control for protein loading (Fig. 9D). Taken together, the data suggested that Act did indeed induce caspase-dependent apoptosis of macrophages and intestinal epithelial cells.
To determine which major apoptotic pathway (intrinsic/mitochondrial or extrinsic/death receptor) might be activated by Act, we performed activity assays for caspase 9 (intrinsic-associated) and caspase 8 (extrinsic-associated) in whole cell lysates (80 g) from Act-treated (20 ng/ml) RAW 264.7 cells (0, 2, 4, and 6 h). As shown, Act induced caspase 9 activation in RAW 264.7 cells by 2 h, which increased with time (Fig. 9E). This suggested that Act might activate the mitochondrial-associated pathway of cell death via release of cytochrome c. To determine whether the death receptor pathway might also contribute to Act-induced apoptosis, we performed activity assays for caspase 8 in Act-treated RAW 264.7 cells. Surprisingly, Act also induced caspase 8 activation by 2 h (Fig. 9E). Whole cell lysates (100 g) from camptothecin-treated Jurkat T cells were used as a positive control.
Absence of the TNF Receptor Reduces Act-induced Apoptosis and Caspase Activities of Murine Macrophages-Because binding of TNF to its receptor (TNFR) could lead to cleavage of caspase 8, we utilized single and double knockout cells (bone marrow-derived murine macrophages), in which one or both of the two isoforms of the TNF receptor (TNFR-1 and TNFR-2) were deleted (29), to examine Act-induced apoptosis. WT or double KO cells were treated with 20 ng/ml Act for 0, 2, or 4 h, stained with annexin V-FITC and PI, and examined by flow cytometry. As depicted graphically in Fig. 10A, Act caused apoptosis of WT macrophages by 2 h (3.6%), which increased by 4 h (16.1%), compared with untreated controls. However, Acttreated KO macrophages exhibited no increase in apoptosis by 2 h and only a marginal increase (3.4%) by 4 h (Fig. 10A). These results strongly suggested that Act-induced apoptosis of host cells involved engagement of TNFR-1 and/or TNFR-2 and subsequent activation of caspase 8.
To determine whether TNF-␣ was involved in Act-induced activation of caspases, activity assays for caspases 3, 8, and 9  3-9). Binding reactions were also performed in the absence of nuclear extract and served as a negative control (lane 2). Nuclear extract from RAW cells treated for 4 h with Act was mixed with unlabeled C/EBP-␤ consensus oligonucleotide (ϳ50-fold excess) before adding the labeled oligonucleotide (lane 9). The gel was dried and subjected to autoradiography.
were performed on whole cell lysates (200 g) from Act-treated (20 ng/ml) WT and TNFR double KO macrophages (0, 2.5, and 4 h). As shown in Fig. 10B, Act induced activation of all three caspases in WT macrophages by 2.5 h, and activity was greatly increased by 4 h. However, activation of all 3 caspases was significantly diminished in Act-treated TNFR double KO macrophages (Fig. 10B).
To determine which TNF receptor contributed to caspase activation, caspase activity assays were performed on whole cell lysates from Act-treated WT, double KO, and single KO (TNFR-1 or TNFR-2) macrophages (0 and 6 h). As shown in Fig.  10C, Act induced activation of all three caspases in WT macrophages at 6 h. Activation of all three caspases in TNFR-2 single KO macrophages was similar to that of WT cells (Fig. 10C), indicating that TNFR-2 was not required for Act-mediated caspase activation. However, caspase activation was nearly absent in TNFR-1 single KO and double KO macrophages. Considered together, the results suggested that TNF-␣ binding to TNFR-1 contributed significantly to Act-mediated apoptosis and that apoptosis was mediated via activation of caspases 3, 8, and 9.
Act Up-regulates Apoptosis-associated Proteins in RAW 264.7 Cells-To characterize the host cell apoptotic response to Act, we undertook an un-biased proteomics based approach to discover which proteins might mediate Act-induced apoptosis. RAW 264.7 cells were chosen for this study, because Act-induced apoptosis was originally discovered using these cells (16) and because the large volumes of lysates that were needed (350 l (1 mg/ml) per treatment) could be obtained from this trans-formed cell line with little difficulty. RAW 264.7 cells were treated with PBS or Act (20 ng/ml) for 6 h and whole cell lysates were applied to Kinetworks TM Apoptosis Screens to analyze the expression profiles of 25 apoptosis-associated proteins. We compared normalized CPM, arbitrarily based on the fluorescence intensity values, and considered a 25% increase or decrease in expression to be significant. Of the 25 proteins screened (data not shown), five were significantly up-regulated and one down-regulated by Act at 6 h (Figs. 11, A-C). Immunoblots including these six proteins are shown in Fig. 11, A (0 h) and B (6 h), for comparison, and the data are graphically represented along with CPM values and percentage increases/ decreases in Fig. 11C.
Act Causes Mitochondrial Release of Cytochrome c and AIF in RAW 264.7 Cells-To determine whether Act-induced increases in cytochrome c and AIF were the result of release from mitochondria, we separated mitochondrial and cytosolic fractions from Act-treated (20 ng/ml) RAW 264.7 cells and performed Western blot analysis. Blots were probed with anticytochrome c or anti-AIF antibodies. As a positive control, we used whole cell lysates from cytochrome c-treated Jurkat T cells. As shown in Fig. 11D, Act increased cytosolic levels of cytochrome c and AIF by 2 h, suggesting that Act-induced cell death involved the mitochondrial pathway of apoptosis. Blots were probed with anti-p90RSK to control for equal loading of protein (Fig. 11D).
Mitochondrial Membrane Depolarization Does Not Precede Act-mediated Cell Death in RAW 264.7 Cells-To determine whether Act-induced apoptosis of RAW 264.7 cells involved depolarization of mitochondria, we stained Act-treated cells with the membrane potential sensor dye JC-1 and acquired cells by flow cytometry. As shown in Fig. 12, A and B, there was no shift from JC-1 aggregates (indicative of healthy, respiring cells) to JC-1 monomers (representative of cells with depolarized mitochondrial membranes) in RAW 264.7 cells treated with 20 ng/ml Act (upper right quadrant). As a positive control, cells were treated with the mitochondrial respiratory chain decoupler, carbonyl cyanide m-chlorophenylhydrazone (CCCP) (250 M for 30 min), which did result in a shift toward JC-1 monomers (Fig. 12C, upper right  quadrant).
We also examined size and morphology of cell populations, based on forward and size scattering and compared that with JC-1 staining. Based on forward and side scattering measurements, Act treatment resulted in 2 distinct populations of cells (data not shown): cells of normal, healthy size and morphology, which we colored blue, and cells that were shrunken, colored in orange. We then gated the cells, based on these two cell populations (blue and orange), and analyzed the cells for JC-1 staining, so that we could compare cell size/morphology with mitochondrial membrane potential Cycloheximide-treated (100 g/ml) cells (5 h) were used as a positive control (not shown). As shown, absence of TNFR-1 and TNFR-2 resulted in a significant decrease in the percentage of Act-induced apoptosis at 2 and 4 h when compared with WT cells. Panels B and C, whole cell lysates from WT, TNFR double KO, TNFR1 or TNFR2 single KO macrophages treated with 20 ng/ml Act were assayed for caspases 3, 8, and 9 activities, and a representative experiment is shown graphically. Absorbance is displayed on the y axis, and treatments are shown on the x axis. Whole cell lysates from camptothecin-treated Jurkat T cells were used as a positive control. As shown, caspases 3, 8, and 9 were activated in WT macrophages by 2.5, 4 (panel B), and 6 h (panel C). However, as shown in panel B, activity for each caspase was greatly reduced in TNFR double KO cells. Caspase activation (3, 8, and 9) was greatly reduced for TNFR1 single KO macrophages, but not for TNFR2 single KO macrophages (panel C). Asterisk (panel A) denotes statistically significant difference in apoptosis between WT and KO cells as determined by Student's t test (p value Յ 0.05). 20 ng/ml Act for the indicated times, cytosolic fractions were isolated, and Western blot analysis was performed with anti-cytochrome c or anti-AIF antibody. Blots were probed with anti-p90RSK antibody as a control for protein loading. Whole cell lysates from cytochrome c-treated Jurkat T cells were used as a positive control (PC). As shown, both cytochrome c and AIF were present in higher amounts in cytosolic fractions by 2 h. (Fig. 12). Whereas Act treatment did not result in mitochondrial membrane depolarization, there was a substantial decrease in cell size (blue 3 orange), indicative of cell death (Fig. 12, A and B).
To confirm the results obtained using JC-1, we viewed Acttreated (20 ng/ml) RAW 264.7 cells by confocal microscopy after staining with MitoTracker red CM-H 2 XRos. MitoTracker red does not fluoresce until it enters an actively respiring cell, where it is subsequently oxidized and then sequestered in the mitochondria. As shown in Fig. 12E, Act treatment did not result in depolarization of mitochondrial membranes, which was comparable with the untreated control (Fig. 12D). As expected, cells treated with the mitochondrial decoupler CCCP did experience loss in mitochondrial membrane potential, re- sulting in very low levels of red fluorescence (Fig. 12F). Actively respiring mitochondria, based on strong red fluorescence, were apparent even at 6 h in Act-treated cells that were clearly undergoing cell death, as evidenced by cell shrinkage and membrane blebbing (Fig. 12H) compared with PBS-treated (Fig.  12G) or CCCP-treated RAW 264.7 cells (Fig. 12I), which appeared normal. Considered with JC-1 staining, the data clearly demonstrated that mitochondrial membrane depolarization did not precede Act-induced cell death. DISCUSSION MAPK cascades have been shown to be intimately involved in cell proliferation, differentiation, apoptosis, and inflammation, and are pivotal during immune responses (41). Furthermore, the importance of MAPK signaling cascades in a variety of human diseases has led to the development of anti-inflammatory drugs that target various MAPK enzymes (42)(43)(44)(45)(46)(47)(48)(49)(50)(51)(52)(53). We therefore investigated the effect of Act on MAPKs, in which we demonstrated for the first time Act-induced activation of the p38 (Fig. 1), JNK (Fig. 2), and ERK1/2 (Fig. 6) signaling pathways in a murine macrophage cell line (RAW 264.7), murine peritoneal macrophages, and a human intestinal epithelial cell line (T84).
Because phosphorylation of p38 and JNK is mediated by MKK3/6 and MKK7, respectively, and MKK4 phosphorylates both p38 and JNK, we evaluated activation of these kinases in both RAW 264.7 cells and murine peritoneal macrophages. Whereas Act induced phosphorylation of MKK3/6 and MKK4 (Fig. 3), it did not phosphorylate MKK7 in these cells (data not shown). We therefore believe that Act-induced JNK activation occurs via MKK4. However, activation of p38 might result from phosphorylation by either MKK3/6 or MKK4. It is likely that p38 and JNK act coordinately to induce inflammatory responses observed in Act-treated host cells.
To further dissect Act-induced activation of p38 and JNK signaling cascades, we investigated factors known to lie downstream of both kinases: ATF-2, MAPKAPk-2, and c-Jun. ATF-2, which can be phosphorylated/activated by either p38 or JNK, is a transcription factor that trans-activates genes containing cAMP response-like elements in their promoters, such as inducible nitric-oxide synthase (54) and TNF-␣ (55,56). Act treatment did cause phosphorylation of ATF-2 in RAW 264.7 cells and murine peritoneal macrophages (Fig. 4, A and C), which is consistent with the activation of p38 and JNK signaling cascades. Likewise, c-Jun was phosphorylated in response to Act in both RAW 264.7 cells and murine peritoneal macrophages (Fig. 5, A and C), which strongly suggested that Act-induced AP-1 activation resulted from activation of the JNK pathway. The potential contribution of other MAPKs, such as p38, to Act-induced AP-1 activation requires further study.
Phosphorylation of MAPKAPk-2, which can regulate cytokine translation via phosphorylation of AU-rich elements in their mRNAs (57), was observed for murine peritoneal macrophages ( Fig. 4G) but was not obvious in RAW 264.7 cells (Fig.  4E). MAPKAPk-2 deficiency severely reduces the biosynthesis of TNF-␣ in response to LPS (58), but because Act induces TNF-␣ production in RAW 264.7 cells (14), the function of MAPKAPk-2 in response to Act might be redundant.
p90RSK has been shown to activate the transcription factors cyclic AMP-response element-binding protein and NF-B (33)(34)(35), which are both activated by Act in murine RAW 264.7 macrophages (14,15). We now demonstrated phosphorylation of p90RSK in Act-treated RAW 264.7 cells and murine peritoneal macrophages (Fig. 5, E and G). It is possible that Actinduced phosphorylation of p90RSK in murine macrophages was mediated by ERK1/2, a known activator of p90RSK, because Act also induced phosphorylation of ERK1/2 (Fig. 6, A and C) and the ERK1 activator, MEK1 (Fig. 6, B and D). It is unclear why phosphorylation of MEK1, which is known to phosphorylate ERK1, was detected later than phosphorylation of ERK1/2. It is possible that the sensitivities of the antibodies differed. Alternatively, because signaling cascades are often amplified as they progress downstream, it is plausible that the levels of activated MEK1 were too low to be detected as early as 15 min. Considered together, the data clearly demonstrated Act-induced activation of all three major MAPK pathways (ERK1/2, p38, and JNK) in murine macrophages and human intestinal epithelial cells, which likely contributes to inflammatory host cell responses to the toxin.
MAPKs are responsible for regulating a variety of cellular processes, including apoptosis. It is therefore possible that activation of MAPKs plays a role in Act-induced host cell death. As shown in Fig. 8, Act induced classic hallmarks of apoptosis in both macrophages and intestinal epithelial cells. An interesting observation was that Act-induced blebbing and cell death of T84 cells occurred later than for RAW 264.7 cells and proceeded from the periphery of the epithelial clump toward the center. We confirmed apoptosis as the mechanism of Actinduced intestinal epithelial cell death by annexin V-FITC and PI staining of Act-treated HT-29 cells (Fig. 8, G-I). By utilizing a different epithelial cell line (HT-29 cells in lieu of T84 cells) for this study, we also demonstrated that Act-induced cell death of epithelial cells was not peculiar to the T84 cell line. Whereas Act is known to cause severe intestinal damage in vivo, this study demonstrated for the first time that a potential cause for intestinal damage induced by A. hydrophila infection is likely Act-mediated apoptosis.
Act induced caspase 3-cleavage in RAW 264.7 cells by 2 h (Fig. 9A). This correlated with previous findings (that Act caused presentation of phosphatidylserine to the outer membrane layer between 2 and 4 h, DNA laddering by 4 h (16), and membrane blebbing by 2.5 h (Fig. 8B)), and strongly suggested that Act-induced apoptosis was mediated, at least in part, by caspase 3-cleavage. Act also induced caspase 3-cleavage in murine primary macrophages by 4 h (Fig. 9B). The later induction was because of the fact that we used a lower dose of the toxin (12 ng/ml for murine primary macrophages compared with 20 ng/ml for RAW 264.7 cells). Indeed, Western blot analysis of murine peritoneal macrophages treated with 20 ng/ml Act demonstrated caspase 3-cleavage, which occurred between 1 and 3 h (data not shown). Finally, examination of Act-treated (20 ng/ml) T84 cells also revealed caspase 3-cleavage by 4 h (Fig. 9C), which was consistent with Act-induced membrane blebbing by 5 h (Fig. 8F).
We also demonstrated caspase 9 activation in Act-treated RAW 264.7 cells (Fig. 9E), which was expected, because Act has been shown to cause production of ROS and activation of DNA repair machinery (15). Caspase 8, which is activated in response to death receptor engagement, was also activated by Act in RAW 264.7 cells (Fig. 9E). It is therefore likely that Act induces apoptosis via activation of both the death receptormediated and mitochondrial pathways. TNFR-1 (29) is a strong candidate as the death receptor responsible for caspase 8 activation in response to toxin treatment, because Act-induced apoptosis was inhibited in TNFR double KO macrophages (Fig.  10A). Additionally, Act-induced activation of caspases 3 and 9 were also greatly reduced in TNFR-1 and TNFR double KO cells compared with WT macrophages (Fig. 10, B and C). Caspase activation was not abrogated in TNFR-2 KO cells, which was expected because TNFR-1 contains death domains, whereas TNFR-2 does not (59), and binding of TNF-␣ to TNFR-2 has been shown to be insufficient for mediating cell death (60,61). The results strongly suggested that TNF-␣ was largely responsible for Act-induced apoptosis of host cells, which was mediated via engagement of TNFR-1. It is also possible that Act-induced activation of MAPKs involves TNF-␣. Both TNFR-1 and -2 are required for TNF-␣-induced activation of p38, JNK, and ERK (29). We are currently investigating whether or not Act-induced activation of these kinases is mediated by TNF-␣ using these TNFR KO macrophages.
Based on a proteomics approach, five apoptotic proteins (cytochrome c, AIF, CAS, PARP, and PERP) were up-regulated and one (superoxide dismutase) was down-regulated by Act in RAW 264.7 cell lysates (Fig. 11, A-C). Cytochrome c and AIF are both mitochondrial proteins that are transported to the cytosol during apoptosis, which occurs upon depolarization of the mitochondrial membrane. We confirmed Act-induced cytochrome c and AIF release from mitochondria by Western blot analysis of cytosolic fractions (Fig. 11D). However, we did not detect depolarization of mitochondria in RAW 264.7 cells (Fig.  12). These results were unexpected, because mitochondrial membrane depolarization typically precedes or is concurrent with release of mitochondrial proteins (62)(63)(64)(65)(66)(67). It is possible that Act caused greater fragility of mitochondria compared with controls, leading to greater levels of cytochrome c and AIF in the cytosol of treated cells upon sample manipulation. Alternatively, Act-induced release of cytochrome c and AIF might have occurred via a mechanism that does not depend upon interruption of mitochondrial respiration. There have been several reports of cytochrome c release prior to a reduction in mitochondrial membrane potential (68 -75), and future studies will be aimed at determining how this might occur in host cells in response to Act.
CAS, which was up-regulated 284% by Act in RAW 264.7 cells (Fig. 11, A-C), is a nuclear transport protein that is highly expressed in proliferating cells and is thought to mediate the decision of whether cells proliferate or undergo apoptosis (76,77). CAS might play a role in TNF-␣-mediated apoptosis, because attenuation of CAS has been shown to reduce the susceptibility of MCF-7 breast cancer cells to TNF-␣-induced apoptosis, most likely by contributing to the nuclear transport of apoptosis-associated proteins (76). It is plausible that CAS contributes to Act-induced apoptosis, possibly in response to host cell TNF-␣ production, especially given the fact that Actinduced apoptosis and caspase activation were reduced in TNFR-1 single KO and TNFR-1 and -2 double KO macrophages (Fig. 10).
PARP, which was up-regulated 61% by Act by 6 h compared with untreated controls (Fig. 11, A-C), is rapidly activated by DNA damage and facilitates DNA repair under conditions of moderate cellular stress (78). However, PARP promotes apoptosis upon extensive DNA damage, which has recently been suggested to involve depletion of NAD ϩ , opening of a mitochondrial translocation channel, and subsequent release of AIF into the cytosol (79 -81). It is possible that Act-induced increase of PARP in RAW 264.7 cells contributes to Act-mediated apoptosis and will be further investigated.
PERP was also up-regulated (35%) by Act in RAW 264.7 cells (Fig. 11, A-C). PERP is a p53 target that is highly up-regulated during p53-mediated apoptosis (82,83). Up-regulation of PERP by Act correlates with previous studies that demonstrated Actinduced ROS production and induction of the DNA repair enzyme Ref-1 (15), and with activation of PARP (Fig. 11, A-C). Interestingly, the antioxidant enzyme superoxide dismutase was down-regulated 26% by Act (Fig. 11, A-C). Induction of PARP and PERP and down-regulation of superoxide dismutase by Act strongly suggested that Act-mediated apoptosis was caused by the response of the cell to DNA damage. However, given the absence of mitochondrial membrane depolarization in Act-treated macrophages (Fig. 12) and inhibition of caspase activation in TNFR-1 KO macrophages (Fig. 10), it will be interesting to determine how these proteins might contribute to Act-induced apoptosis.
Whereas the host receptor through which Act might signal has yet to be elucidated, the data presented here strongly suggest that Act binds a host cell plasma membrane receptor or is endocytosed and activates an intracellular factor. Activation of MAPKs likely contributes to the inflammatory cascade that ensues upon exposure to Act. The results of this study provided new insights into the mechanism of action of Act and yielded likely avenues through which Act might mediate the activation of transcription factors and subsequent production of inflammatory effectors and apoptosis in macrophages. Additionally, we have demonstrated for the first time Act-induced intracellular signaling and apoptosis of human intestinal epithelial cells.