Adenomatous Polyposis Coli Protein Associates with C/EBP β and Increases Bacillus anthracis Edema Toxin-stimulated Gene Expression in Macrophages*

The production of cAMP from Bacillus anthracis edema toxin (ET) activates gene expression in macrophages through a complex array of signaling pathways, most of which remain poorly defined. In this study, the tumor suppressor protein adenomatous polyposis coli (APC) was found to be important for the up-regulation of previously defined ET-stimulated genes (Vegfa, Ptgs2, Arg2, Cxcl2, Sdc1, and Cebpb). A reduction in the expression of these genes after ET exposure was observed when APC was disrupted in macrophages using siRNA or in bone marrow-derived macrophages obtained from C57BL/6J-ApcMin mice, which are heterozygous for a truncated form of APC. In line with this observation, ET increased the expression of APC at the transcriptional level, leading to increased amounts of APC in the nucleus. The mechanism utilized by APC to increase ET-induced gene expression was determined to depend on the ability of APC to interact with C/EBP β, which is a transcription factor activated by cAMP. Coimmunoprecipitation experiments found that APC associated with C/EBP β and that levels of this complex increase after ET exposure. A further connection was uncovered when silencing APC was determined to reduce the ET-induced phosphorylation of C/EBP β at Thr-188. This ET-mediated phosphorylation of C/EBP β was blocked by glycogen synthase kinase 3 (GSK-3) inhibitors, suggesting that GSK-3 is involved in the activation of C/EBP β and supporting the idea of APC helping direct interactions between GSK-3 and C/EBP β. These results indicate that ET stimulates gene expression by promoting the formation of an inducible protein complex consisting of APC and C/EBP β.

The production of cAMP from Bacillus anthracis edema toxin (ET) activates gene expression in macrophages through a complex array of signaling pathways, most of which remain poorly defined. In this study, the tumor suppressor protein adenomatous polyposis coli (APC) was found to be important for the up-regulation of previously defined ET-stimulated genes (Vegfa, Ptgs2, Arg2, Cxcl2, Sdc1, and Cebpb). A reduction in the expression of these genes after ET exposure was observed when APC was disrupted in macrophages using siRNA or in bone marrow-derived macrophages obtained from C57BL/6J-Apc Min mice, which are heterozygous for a truncated form of APC. In line with this observation, ET increased the expression of APC at the transcriptional level, leading to increased amounts of APC in the nucleus. The mechanism utilized by APC to increase ETinduced gene expression was determined to depend on the ability of APC to interact with C/EBP ␤, which is a transcription factor activated by cAMP. Coimmunoprecipitation experiments found that APC associated with C/EBP ␤ and that levels of this complex increase after ET exposure. A further connection was uncovered when silencing APC was determined to reduce the ET-induced phosphorylation of C/EBP ␤ at Thr-188. This ET-mediated phosphorylation of C/EBP ␤ was blocked by glycogen synthase kinase 3 (GSK-3) inhibitors, suggesting that GSK-3 is involved in the activation of C/EBP ␤ and supporting the idea of APC helping direct interactions between GSK-3 and C/EBP ␤. These results indicate that ET stimulates gene expression by promoting the formation of an inducible protein complex consisting of APC and C/EBP ␤.
Edema toxin (ET) 2 is secreted from Bacillus anthracis as a binary bacterial toxin composed of two proteins, protective antigen (PA) and edema factor (EF) (1). PA mediates the membrane translocation and cell entry of EF, which functions as a calmodulin-dependent adenylate cyclase within the cytosol of mammalian cells (1). As a potent adenylate cyclase, EF generates supraphysiological levels of cAMP (1), which potentially hyperactivates cAMP-sensitive signaling pathways and alters intracellular signaling pathways not affected by normal levels of cAMP. Of particular interest to our studies are the mechanisms utilized by ET to alter signaling pathways and thus macrophage function. Macrophages are part of the innate immune system and are believed to be a critical target of this cAMP-generating toxin (2)(3)(4)(5). In macrophages, cAMP is known to activate immunosuppressive signaling pathways and help regulate immune responses (6,7). Therefore, the ability of pathogens to commandeer cAMP signaling is an important mechanism of bacterial pathogenesis.
The underlying cellular and molecular mechanisms utilized by cAMP to alter macrophage phenotypes are not fully known but likely involve a complex network of cellular factors. The most prominent and best defined branch of cAMP signaling is carried out through the activation of protein kinase A (PKA) and the basic region leucine zipper transcription factor CREB. When protein kinase A phosphorylates CREB at Ser-133, CREB activates transcription on certain gene promoters that possess cAMP response elements (CRE). In macrophages, one key transcriptional target of CREB is the transcription factor CCAAT/ enhancer-binding protein ␤ (C/EBP ␤) (8). C/EBP ␤ is also part of the basic region leucine zipper family of transcription factors and is essential for the induction of anti-inflammatory genes in macrophages (8). A considerable amount of evidence exists that both CREB and C/EBP ␤ interact with various factors that allow these transcription factors to cooperate with other signal transduction pathways (9 -11).
In previously published work from our group, ET-exposed macrophages were found to down-regulate ␤-catenin-stimulated transcription because of the activation of nuclear GSK-3␤ and subsequent phosphorylation of ␤-catenin (12). Because ET modifies GSK-3 and ␤-catenin activity, we hypothesized that these proteins or associated proteins could interact with a well defined cAMP responsive pathway such as the CREB-C/EBP ␤ pathway. The phosphorylation of ␤-catenin by GSK-3 depends on the formation of a protein complex stabilized by the scaffolding protein adenomatous polyposis coli (APC) (13). In this study, APC was examined to determine whether it is connected to the CREB-C/EBP ␤ pathway. APC has the potential to interact with the CREB-C/EBP ␤ pathway because APC is known to associate with both PKA and GSK-3 (14, 15), which are two kinases activated by ET and connected to the CREB-C/EBP ␤ signaling pathway (10, 12, 16 -18). Many previous studies have also shown that APC possesses both ␤-catenin-dependent and independent activities (19 -22). Here, APC was found to utilize a previously undescribed mechanism to alter ET-induced gene expression in macrophages. These results support the notion that APC is involved in the CREB-C/EBP ␤ signaling axis in macrophages through a mechanism that depends on the formation of a complex containing APC and C/EBP ␤.
Maintenance and Use of Cell Lines-RAW 264.7 and L-929 cells were obtained from the ATCC. Cells were grown in the presence of DMEM (ATCC) containing 10% fetal bovine serum (ATCC). All cell lines were used between passages 5 and 20. To produce L-929 cell-conditioned medium necessary for growth of BMDM, L-929 cells were cultured at a density of 1.25 ϫ 10 5 cells/ml. On the fourth day of growth, the conditioned media were removed, passed through a 0.2-m filter, and stored at 4°C until use.
Isolation and Culture of Bone Marrow-derived Macrophages-C57BL/6J-Apc Min heterozygous mice and C57BL/6J wild-type mice were purchased from The Jackson Laboratories (Bar Harbor, ME), and all studies were performed in accordance with the Institutional Care and Use Committee guidelines at the University of Oklahoma Health Science Center (IACUC approval date, 06/10/2009; IACUC protocol number 09 -064). To culture BMDM, mouse femurs were isolated, and bone marrow was flushed from femurs with a PBS solution containing 3% FBS. The bone marrow was centrifuged for 5 min at 500 ϫ g, and macrophage culture medium (DMEM with 10% FBS and 30% L-929 conditioned medium) was added to the resulting pellet. The bone marrow cells were counted, and equal numbers of cells were plated on bacterial-grade Petri dishes in macrophage culture medium. After 24 h, adherent bone marrow cells were discarded, and non-adherent bone marrow cells were maintained in culture for 6 days, at which point a monolayer of macrophages was apparent. Macrophages were then plated at a density of 2.5 ϫ 10 5 cells/ml in six-well tissue culture-treated plates in macrophage culture medium, and the macrophages were subjected to experimental conditions after maintaining the culture overnight.

Preparation of Protein Extracts and Immunoblot Analysis-
To extract total cellular protein from cells, the medium was removed and replaced with lysis buffer chilled to 4°C containing 1% SDS, 50 mM Tris (pH 7.4), 5 mM EDTA, protease inhibitor mixture (Sigma, catalog no. P8340), and 10 mM N-ethylmaleimide. The cells were incubated in this buffer on ice for 15 min, passed through a 22-gauge needle 10 times, and finally centrifuged for 5 min at 20,000 ϫ g to remove insoluble debris.
To obtain cytoplasmic and nuclear fractions, cells were grown in T-75 flasks and subjected to experimental conditions. The cells were then washed, scraped, and centrifuged for 5 min at 250 ϫ g. The resulting pellet was separated into cytoplasmic and nuclear fractions using the NE-PER nuclear and cytoplasmic extraction reagents according to the manufacturer's instructions (Pierce). A protease inhibitor mixture (Sigma, catalog no. P8340) and 10 mM N-ethylmaleimide were added to the extraction reagents.
Immunoblot analyses were performed by first combining protein extracts (10 to 25 g/well) with sample buffer (62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 5% ␤-mercaptoethanol, and 0.001% bromphenol blue) and heating the samples at 95°C for 7 min. The samples were resolved using 10 -12% SDSpolyacrylamide gel electrophoresis. For analysis of APC, we used either 6% SDS-PAGE or 3% SDS-agarose gel electrophoresis. Protein bands were either transferred to a PVDF membrane by electroblotting for SDS-PAGE or transferred by downward capillary transfer for 3% SDS-agarose gel electrophoresis. The membrane was blocked with 5-10% nonfat milk in wash buffer consisting of 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 0.1% Tween 20. Membranes were then probed with primary antibodies, washed, and incubated with the appropriate secondary antibodies conjugated to horseradish peroxidase. The blots were then developed with an enhanced chemiluminescent protein development system (GE Healthcare) and exposed to film. Densitometry analysis of immunoblot bands was accomplished on digitized images using ImageJ 1.37V software (Wayne Rasband, National Institutes of Health).
Reverse Transcriptase qPCR Analysis-To synthesize cDNA from total RNA, reverse transcription reactions were performed using SuperScript III (Invitrogen) according to the conditions suggested by the manufacturer. For qPCR analysis, the resulting cDNA was combined with a SYBR Green PCR master mix (ABI or SABiosciences) and gene-specific primers. Amplification was performed in an Applied Biosystems 7500 realtime PCR system, and the 2 -⌬⌬Ct method was used to determine relative changes in the mRNA of genes of interest compared with Actb mRNA levels.
siRNA Transfection-siRNA was transfected into RAW 264.7 cells with the transfection reagent HiPerFect (Qiagen) using a modified version of the manufacturer's protocol for RAW 264.7 cells. Briefly, 7.5 ϫ 10 5 cells in 200 l of DMEM with 10% FBS were plated into a well of a 24-well plate (the protocol was adjusted for different plate sizes). Immediately after plating the cells, the transfection mixture containing 6 l of HiPerFect and 50 -100 nM siRNA was prepared in 200 l of Opti-MEM. After 10 min of incubation, the transfection mixture was added to the cells, bringing the volume to 400 l. After a 6-h exposure, the medium was removed and replaced with fresh DMEM contain-ing 10% FBS. The siRNAs against APC (Santa Cruz Biotechnology, sc-29703) and C/EBP ␤ (Santa Cruz Biotechnology, sc-29862) each contain a pool of three to four target-specific siRNAs. As a negative control, cells were transfected with a scrambled sequence not targeting any known gene (Santa Cruz Biotechnology, sc-37007).
Immunoprecipitation-To immunoprecipitate protein complexes, intact cells or isolated nuclei were lysed by incubating for 5 min in lysis buffer (20 mM HEPES (pH 7.9), 350 mM NaCl, 30 mM MgCl 2 , 10% glycerol, 0.5% Nonidet P-40, 200 M DTT, protease inhibitor mixture), passing the lysates through a 22-gauge needle 10 times, and then centrifuging the lysates at 18,000 ϫ g for 5 min. After determining the protein concentration, 500 g of protein was diluted into a binding buffer (20 mM HEPES (pH 7.9), 30 mM MgCl 2 , 10% glycerol, 0.2% Nonidet P-40, 200 M DTT, and protease inhibitor mixture) to give a final volume of 300 l. One microgram of the appropriate antibodies was added to the immunoprecipitation mixture. After incubating for 2 h at 4°C, protein G-conjugated magnetic beads were added to the immunoprecipitation mixture, and the incubation was continued for an additional 30 min. The input protein was removed from the magnetic beads, and the beads were washed three times in binding buffer. The immunoprecipitated proteins were subsequently eluted in 15 l of buffer at pH 2.8.

Expression of ET-responsive Genes Is Influenced by Cellular
Levels of APC-To begin to study the role of APC in ET-intoxicated macrophages, the induction of ET-responsive genes was compared between macrophages containing wild type levels of APC and macrophages with APC levels reduced by siRNA. As shown by the immunoblot in Fig. 1A, APC siRNA-depleted cellular levels of APC, which led to a concomitant increase in ␤-catenin as predicted by the canonical Wnt signaling pathway. RT-qPCR was then used to measure the expression of a selection of previously defined (2, 5) ET-responsive genes (Vegfa, Ptgs2, Arg2, Cxcl2, Sdc1, and Cebpb) that are believed to be important in the pathogenesis of anthrax. Each ET-responsive FIGURE 1. Decreasing levels of APC with siRNA attenuates ET-stimulated gene expression. RAW 264.7 cells were transfected with siRNAs directed against APC or negative control siRNA. A, immunoblot analysis demonstrating decreased levels of APC and increased ␤-catenin levels in siRNA-treated cells. B-G, RAW 264.7 cells were exposed to ET (10 nM EF and 10 nM PA for 6 h), and RT-qPCR was used to measure levels of the following ET-inducible genes: Vegfa, Ptgs2, Arg2, Cxcl2, Sdc1, and Cebpb. H, as a control, RAW 264.7 cells were exposed to 150 M ZnSO 4 for 6 h, and RT-qPCR was used to measure levels of Mt1. Error bars indicate mean Ϯ S.E., and p values were determined by comparisons between the control and experimental groups using a two-tailed Student's t test with a 95% confidence interval. *, p Ͻ 0.01.
target gene was significantly reduced in RAW 264.7 cells with APC siRNA (Fig. 1, B-G), suggesting that APC is necessary for ET-stimulated gene expression. Examination of the induction of metallothionein 1 (Mt1) by ZnSO 4 revealed that reduced APC did not significantly impact the induction of a gene not connected to cAMP signaling (Fig. 1H).
As another approach to disrupt APC and demonstrate the connection between APC and ET-activated gene expression, experiments were performed utilizing BMDM from C57BL/6J-Apc Min mice. These mice are heterozygous for a truncated form of APC (95 kDa) resulting from a point mutation that creates a premature stop codon in the gene. As shown in Fig. 2A, BMDM from C57BL/6J-Apc Min mice express the truncated form of APC and reduced levels of full-length APC ( Fig. 2A). The levels of ␤-catenin were examined next, and BMDM cultured from C57BL/6J-Apc Min mice were found to express similar level of ␤-catenin compared with wild-type controls ( Fig. 2A), indicating that these mice are haplosufficient regarding regulating levels of cellular ␤-catenin. Next, RT-qPCR was used to examine the ET-responsive genes in C57BL/6J-Apc Min BMDM. As shown in Fig. 2, B and C, Vegfa and Ptgs2 expression was found to be decreased in BMDM from C57BL/6J-Apc Min mice compared with wild-type controls exposed to ET. Other ET-responsive genes analyzed showed trends toward reduction but did not reach statistical significance. Collectively, these data indicate that APC is critical for the efficient induction of these ETstimulated genes.
Exposure of Macrophages to ET Elevates Expression of Fulllength APC-The results shown in Figs. 1 and 2 indicate that APC is involved in ET-mediated transcriptional changes, but whether APC was directly affected through increased expres-sion was unknown. Therefore, the levels of Apc mRNA were measured in both BMDM and RAW 264.7 cells exposed to ET (10 nM EF plus 10 nM PA). As shown in Fig. 3, A and B, Apc transcript levels were found to increase between 1.5-and 1.8fold. To determine whether elevated Apc mRNA results in increased levels of the full-length APC protein in the nucleus, an immunoblot analysis was performed on extracts taken from the nucleus of RAW 264.7 cells exposed to 10 nM of ET. As shown in Fig. 3, C and D, levels of APC were found to increase in the nucleus in response to ET exposure. To verify that the elevation in APC is due to cAMP, macrophages were exposed to a membrane-permeable analog of cAMP, 6-MB-cAMP. In Fig. 3, E and F, nuclear levels of APC were elevated in cells exposed to 6-MB-cAMP. The purity of the nuclear extracts was demonstrated by marking the nuclear fraction with CREB and the cytoplasmic fraction with GAPDH (Fig. 3, C and E). These data suggest that cAMP from ET intoxication leads to heightened levels of APC within the nucleus.
APC Associates with C/EBP ␤-In the next set of experiments, the mechanism utilized by APC to alter ET-stimulated transcription was examined. Because APC was found to accumulate in the nucleus and affect a wide array of ET-stimulated genes, we hypothesized that APC might be interacting with a cAMP-responsive transcription factor. We therefore examined whether or not APC was able to associate with two major cAMP-responsive transcription factors, CREB and C/EBP ␤. To determine whether CREB or C/EBP ␤ were able to form a complex with APC, each of these transcription factors were immunoprecipitated, and interactions with APC were determined by immunoblot. As shown in Fig. 4, A and B, C/EBP ␤ was found to associate with APC, whereas interactions between FIGURE 2. Vegfa and Ptgs2 are reduced in C57BL/6J-Apc Min Mouse BMDM. BMDM were cultured from bone marrow derived from wild-type mice or MIN mice (mice heterozygous for truncated APC). A, immunoblot analyses examining levels of APC and ␤-catenin and a region of a Coomassie-stained gel showing equal loading between samples. B, BMDM were exposed for 6 h to 10 nM ET (10 nM EF and 10 nM PA), and RT-qPCR was used to measure levels of Vegfa. C, BMDM were exposed for 4 h to 10 nM ET (10 nM EF and 10 nM PA), and RT-qPCR was used to measure levels of Ptgs2. Bar graphs are representative of three experiments. Each experiment used a different culture generated from a different wild-type or MIN mouse. Error bars indicate mean Ϯ S.E., and p values were determined by comparisons between the wild-type and MIN groups using a two-tailed Student's t test with a 95% confidence interval. *, p Ͻ 0.05. CREB and APC were not detected. To confirm that APC and C/EBP ␤ were forming a complex, APC was immunoprecipitated, and C/EBP ␤ levels were examined by immunoblot analysis. As shown in Fig. 4C, the 34-kDa form of C/EBP ␤ was detected after APC was immunoprecipitated from ET-treated cells. Importantly, these results reveal that not only are C/EBP ␤ and APC forming a complex, but also that exposing macrophages to ET increases the levels of this complex (Fig. 4, A-C).
Next, associations between C/EBP ␤ and APC were examined in macrophages using other agents that can activate cAMP signaling. Macrophages were exposed to 6-MB-cAMP or stimulated with PGE 2 , which activates endogenous adenylate cyclases by binding the prostaglandin E 2 G-protein-coupled receptor. Both treatments resulted in increased levels of C/EBP ␤ and a concomitant increase in binding interactions between C/EBP ␤ and APC, as demonstrated by immunoprecipitating C/EBP ␤ and immunoblotting for APC (Fig. 4, D and E). Levels of the APC-C/EBP ␤ complex were compared by quantitative densitometry and revealed that ET increased levels of this complex by about 7-fold, whereas 6MB-cAMP and PGE 2 increased levels of the complex by about 5-and 2-fold, respectively. Collec-tively, these results reveal that interactions between C/EBP ␤ and APC increased not only after ET exposure but also after activating cAMP signaling through other methods.
ET-induced Genes Are Reduced by C/EBP ␤ siRNA-To determine whether C/EBP ␤ contributes to ET-induced changes in transcription, levels of ET-responsive genes were examined in macrophages with siRNA against C/EBP ␤. As shown by immunoblot analysis in Fig. 5A, ET exposure leads to increased levels of C/EBP ␤, and the addition of C/EBP ␤ siRNA reduces levels of C/EBP ␤ in both controls and ET-exposed macrophages. RT-qPCR was then used to measure the expression of the ET-responsive genes that were found in Fig. 1 to be sensitive to APC levels. The results indicated that these ET-responsive target genes were decreased in RAW 264.7 cells with C/EBP ␤ siRNA (Fig. 5B-F), which further suggests that interactions between C/EBP ␤ and APC are important for stimulating ET gene expression.
C/EBP ␤ Phosphorylation Is Dependent on APC-To further establish the importance of interactions between APC and C/EBP ␤, we determined whether APC levels would impact the phosphorylation state of C/EBP ␤ at Thr-188 (Thr-167 for the

. Intoxication by ET triggers an increase in nuclear levels of APC.
A, mouse BMDM were exposed for 6 h to 10 nM ET (10 nM EF and 10 nM PA), and RT-qPCR was used to measure levels of Apc. Bar graph values were averaged from BMDM obtained from three different mice. B, RAW 264.7 macrophages exposed to 10 nM ET. RT-qPCR was used to determine changes in Apc mRNA. A and B, error bars indicate mean Ϯ S.D., and p values were determined by comparisons between the conditions using a two-tailed Student's t test with a 95% confidence interval. C, immunoblot analysis of APC in nuclear extracts isolated from ET-treated RAW 246.7 cells (6 h of 10 nM ET). The nuclear and cytoplasmic fractions were marked by immunoblotting for CREB and GAPDH. D, the level of APC accumulating in the nucleus was quantified by densitometry analysis of four immunoblot analyses. E, immunoblot analysis of APC in nuclear extracts isolated from RAW 246.7 cells exposed to 1 mM 6-MB-cAMP for 5 h. The nuclear and cytoplasmic fractions were confirmed by immunoblotting for CREB and GAPDH. F, the level of APC in the nucleus was quantified by densitometry analysis of four immunoblot analyses. The error bars indicate mean Ϯ S.E., and the p values were determined by comparisons between the groups using a two-tailed Student's t test with a 95% confidence interval. *, p Ͻ 0.05.

34-kDa form)
. The phosphorylation of Thr-188 on C/EBP ␤ helps activate the ability of C/EBP ␤ to promote transcription (16,17,23,24). As shown by immunoblot analysis in Fig. 6A, the amount of C/EBP ␤ phosphorylated at Thr-188 was increased after ET treatment. The addition of APC siRNA to ET-intoxicated macrophages reduced the level of phosphorylated C/EBP ␤ to near that of the untreated control, as shown by immunoblot analysis and quantitative densitometry normalized to total C/EBP ␤ levels (Fig. 6, A and B). Examining total levels of C/EBP ␤ revealed that APC siRNA did not attenuate the ET induction of overall C/EBP ␤ levels (Fig. 6A). These data suggest that depletion of APC decreases the ability of C/EBP ␤ to become FIGURE 4. The activation of cAMP signaling in macrophages leads to increased binding between APC and C/EBP ␤. A, immunoprecipitation was performed using protein extracts taken from RAW 264.7 cells exposed for 4 h to 10 nM ET (10 nM EF and 10 nM PA). Immunoprecipitation was carried out with control rabbit IgG or rabbit IgG against C/EBP ␤, and the subsequent immunoblot analyses were probed with antibodies against APC or C/EBP ␤. B, immunoprecipitation was performed using protein extracts taken from RAW 264.7 cells exposed for 4 h to 10 nM ET (10 nM EF and 10 nM PA). Immunoprecipitation was performed with control rabbit IgG, a rabbit IgG against CREB, or rabbit IgG against C/EBP ␤, and the immunoblot analyses were probed with APC or CREB antibodies. C, immunoprecipitation was performed using protein extracts taken from RAW 264.7 cells exposed for 4 h to 10 nM ET (10 nM EF and 10 nM PA). Control rabbit IgG or antibodies against the C-terminal of APC were used in immune precipitations, and the immunoblot analyses were probed with antibodies against C/EBP ␤ or APC. D, immunoprecipitation was performed using lysates taken from RAW 264.7 cells exposed for 3 h to 1 mM 6-MB-cAMP. Immunoprecipitation was carried out with control rabbit IgG or rabbit IgG against C/EBP ␤, and the immunoblot analyses were probed with antibodies against APC or C/EBP ␤. E, immunoprecipitation was performed using lysates taken from RAW 264.7 cells exposed for 4 h to 1 M PGE 2 . Immunoprecipitation was carried out with control rabbit IgG or rabbit IgG against C/EBP ␤, and the immunoblot analyses were probed with antibodies against APC or C/EBP ␤. All experiments were repeated a minimum of three times. FIGURE 5. The expression of ET-stimulated genes is attenuated by C/EBP ␤ siRNA. RAW 264.7 cells were transfected with siRNA directed against C/EBP ␤ or negative control siRNA. A, immunoblot analysis demonstrating increased levels of C/EBP ␤ after exposure to ET (10 nM EF and 10 nM PA for 6 h) and a reduction of C/EBP ␤ after transfection with siRNA against C/EBP ␤. B-F, these cells were exposed to ET (10 nM EF and 10 nM PA for 6 h), and RT qPCR was used to measure levels of the following ET-inducible genes: Vegfa, Ptgs2, Arg2, Cxcl2, and Sdc1. Error bars indicate mean Ϯ S.E., and p values were determined by comparisons between the control and experimental groups using a two-tailed Student's t test with a 95% confidence interval. *, p Ͻ 0.01. phosphorylated at Thr-188 and thus activate ET-induced gene expression.
The Thr-188 site on C/EBP ␤ can be phosphorylated by various kinases, including GSK-3 (16,17). Previously we found that GSK-3␤ is activated in the nucleus of macrophages exposed to ET (12). Because of this observation, we hypothesized that GSK-3 could be involved in the phosphorylation of C/EBP ␤ at Thr-188 in ET-exposed macrophages. To examine this possibility, the ability of ET to induce C/EBP ␤ phosphorylation was examined in macrophages exposed to GSK-3 inhibitors. As shown in Fig. 6, C-E, the ET-induced phosphorylation of C/EBP ␤ at Thr-188 was reduced by two different GSK-3 inhibitors, LiCl and BIO. When overall levels of C/EBP ␤ were examined, the ET induction of the 38-kDa form of C/EBP ␤ was not reduced by LiCl or BIO. However, for the 34-kDa form, the ET induction was decreased by LiCl or BIO treatment, but this decrease was not as substantial as the reduction in phosphorylation of the 34-kDa form. Using quantitative densitometry, phosphorylated C/EBP ␤ was normalized to overall levels of C/EBP ␤, and the result of this analysis confirmed that the reduction in phospho-C/EBP ␤ was not a result of a decrease in overall levels of C/EBP ␤ (Fig. 6, D and E). These data suggest that GSK-3 is critical for the phosphorylation of C/EBP ␤ at Thr-188 during ET intoxication.
To further connect C/EBP ␤ to GSK-3 activity, we determined whether associations between GSK-3 and APC increase in the nucleus during ET exposure. Therefore, nuclei were isolated, and interactions between GSK-3␤ and APC were determined by immunoblotting for GSK-3␤ following APC immunoprecipitation. In Fig. 6F, little GSK-3␤ was detected binding to APC in the nuclei of untreated macrophages; however, ET exposure resulted in an increase in binding interactions FIGURE 6. The phosphorylation of C/EBP ␤ at Thr-188 is dependent on APC. A, RAW 264.7 cells were transfected with siRNAs directed against APC or negative control siRNA. The cells were then exposed to ET (10 nM EF and 10 nM PA for 6 h), and an immunoblot was performed, probing with antibodies against either C/EBP ␤ phosphorylated at Thr-188, total C/EBP ␤, or GAPDH. B, the bar graph represents the densitometry analysis of the phospho-C/EBP ␤ immunoblot analysis from ET-exposed cells (n ϭ 4). The phospho-C/EBP ␤ densitometry data were normalized to levels of total C/EBP ␤. The error bars indicate mean Ϯ S.E., and the p values were determined by comparisons between the control and experimental groups using a two-tailed Student's t test with a 95% confidence interval. *, significance between control and APC siRNA, p Ͻ 0.05. C, RAW 264.7 cells were treated with 10 nM ET for 6 h in the presence or absence of 20 mM LiCl or 5 M BIO. An immunoblot analysis was performed using antibodies recognizing either C/EBP ␤ phosphorylated at Thr-188, total C/EBP ␤, or GAPDH. D and E, the bar graph denotes the densitometry analysis of the phospho-C/EBP ␤ immunoblot analysis from ET-exposed cells with and without GSK-3 inhibitors (n ϭ 3). The phospho-C/EBP ␤ densitometry data were normalized to levels of total C/EBP ␤. The error bars indicate mean Ϯ S.E., and the p values were determined by comparisons between the control and experimental groups using a two-tailed Student's t test with a 95% confidence interval. *, p Ͻ 0.05. F, nuclei were isolated from RAW 264.7 cells exposed for 4 h to 10 nM ET (10 nM EF and 10 nM PA). To demonstrated the purity of the nuclear fraction for this immunoprecipitation, an immunoblot analysis was performed, probing for CREB to designate the nucleus and p65 NF-B to mark the cytoplasm. Protein was extracted from the nuclei, and immunoprecipitation was performed on the nuclear protein using control rabbit IgG or rabbit IgG against APC. Following the immunoprecipitation, immunoblot analyses were carried out, probing with antibodies against APC or GSK-3␤.
between APC and GSK-3␤ in the nucleus. This result demonstrates that ET is not only increasing the association between C/EBP ␤ and APC but also between APC and GSK-3␤, which provides further support for GSK-3 phosphorylating C/EBP ␤ at Thr-188 in a APC-dependent manner.

DISCUSSION
The findings from the current study suggest that APC is a critical signaling protein in macrophages and establishes an association between APC and the CREB-C/EBP ␤ signaling network in ET-intoxicated cells. A summary of the putative signaling pathway delineated in this study is shown in Fig. 7. This ET-sensitive connection appears to be through a mechanism involving interactions between APC and C/EBP ␤, which is supported by three major observations. First, reducing APC in macrophages using two methods resulted in the down-regulation of CREB-C/EBP ␤ responsive genes. Second, APC was able to form a complex with C/EBP ␤, and the amount of this complex increased with ET treatment. Third, levels of APC influenced the amount of C/EBP ␤ phosphorylated at Thr-188 after ET exposure. Together these observations provide the first evidence that APC is able to influence signaling in macrophages and suggest that APC as well as C/EBP ␤ are critical signaling proteins during ET intoxication of macrophages.
In macrophages, increases in intracellular cAMP through the activation of cellular adenylate cyclases suppress immune responses (6,7). Thus, the ET-mediated production of intracellular cAMP appears to be part of a strategy by B. anthracis to suppress host immune cells. Yet, important differences exist between the production of cAMP by cellular adenylate cyclases and that of EF. EF generates sustained supraphysiological levels of cAMP that accumulate at the perinuclear region of the cell, as opposed to typical endogenous adenylate cyclases that generate cAMP near the cell membrane (1,25). This unusual mechanism of cAMP production and cAMP localization may be necessary to overcome normal cellular processes that reverse the effects of cAMP, such as the production of cAMP phosphodiesterases. Additionally, this method of cAMP production may influence how this toxin affects macrophages and, more specifically, how cAMP-sensitive signaling pathways are impacted in macrophages. The findings in this study reveal that ET utilizes APC and C/EBP ␤ associations to fully activate gene expression. Interestingly, this signaling mechanism appears to be a general cAMP-dependent signaling mechanism in macrophages because a membrane-permeable analog of cAMP and PGE 2 are both able to increase binding between APC and C/EBP ␤ (Fig. 4, D-E).
In these studies, ET is able to increase C/EBP ␤ levels, likely through the activation of CREB and subsequent binding of CREB on the C/EBP ␤ gene promoter (8,26). The increased level of the transcription factor C/EBP ␤ could lead to enhanced transcript of ET-induced genes, and experiments in Fig. 5 support the notion that C/EBP ␤ helps promote the transcription of many ET regulated genes. Indeed, previous studies have found that some known ET responsive genes, Arg2 and Ptgs2, can be regulated at the transcriptional level by C/EBP ␤ (27,28). In addition to binding C/EBP responsive elements, C/EBP ␤ can also bind CRE sequences and up-regulating genes containing CRE sites within their promoters (29). Therefore, in addition to CREB, C/EBP ␤ appears to be critical for the full activation of ET-responsive genes.
Interestingly, Arg2 appears to be much more sensitive to siRNA against C/EBP ␤ than siRNA against APC. This result could stem from the ability of C/EBP ␤ to participate in more than one transcription activation complex. In addition to C/EBP ␤ functioning as a homodimer, C/EBP ␤ can form functional heterodimers with other basic region leucine zipper transcription factors such as ATF4 (30). Therefore, Arg2 induction may depend on two separate C/EBP ␤-containing transcription-promoting complexes, and one of these transcriptional complexes may not require APC for activation.
These studies reveal that APC, a component of Wnt signaling, is elevated by ET and is required to activate C/EBP ␤. The involvement of APC in cAMP signaling in macrophages was first hinted at when ET was found to activate GSK-3␤ in the nucleus and subsequently phosphorylate ␤-catenin (12). Yet, the ability of cAMP signaling to influence APC and other members of the Wnt signaling pathways is not unprecedented. For instance, the central member of Wnt signaling, ␤-catenin, can be phosphorylated at Ser-675 by protein kinase A, which enhances the ability of ␤-catenin to promote transcription (31). Conversely, protein kinase A can negatively regulate ␤-catenin activity when protein kinase A phosphorylates ␤-catenin at Ser-45 while in a complex with presenilin (32). Protein kinase A can also both activate and repress GSK-3 activity (12,33,34). For instance, protein kinase A can directly phosphorylate GSK-3␤ at Ser-9 and thus inhibit the activity of GSK-3␤ (33). This is in contrast to a study that found that cAMP could activate GSK-3 in a B cell line (34), which agrees with our previous results that GSK-3␤ is activated in the nucleus of ET-intoxicated macrophages (12). APC also possesses a consensus pro- FIGURE 7. Proposed model depicting a signaling mechanism used by ET in macrophages to activate gene expression. Once EF is delivered to the cytoplasm, cAMP is generated, leading to the phosphorylation of CREB at Ser-133 and CREB-induced transcription of C/EBP ␤ (8). This rise in cAMP levels also increases the expression of APC. These transcriptional events help lead to the ET-mediated increase in the complex consisting of C/EBP ␤ and APC. The rise in cellular cAMP also augments the activity of GSK-3␤ by reducing the phosphorylation at Ser-9 (12) and increases binding between GSK-3␤ and APC. GSK-3␤ phosphorylates C/EBP ␤ at Thr-188 when C/EBP ␤ binds to APC, and phosphorylated C/EBP ␤ promotes the expression of ET-inducible genes.
tein kinase A phosphorylation site near a nuclear localization signal, and mutagenesis studies suggest that phosphorylation at this site could reduce nuclear import (15,35). This is in contrast to our results in macrophages that indicate that APC accumulates in the nucleus after ET exposure. Our current studies also show that ET up-regulates APC levels at both the RNA and protein level and reveal increased binding between APC and C/EBP ␤. The ability of APC to bind C/EBP ␤ and increase the ability of C/EBP ␤ to promote transcription is another mode of interaction between cAMP signaling and members of the Wnt pathway.
Our results suggest that APC helps recruit GSK-3␤ to C/EBP ␤ in an ET-dependent manner. This is supported by the observations that C/EBP ␤ phosphorylation at Thr-188 is reduced by either decreasing APC levels or by inhibiting GSK-3 activity. Further, evidence for APC recruiting GSK-3 to C/EBP ␤ comes from the Wnt signaling pathway, where GSK-3 is known to bind APC (14). Similarly, these current studies reveal an increased amount of the GSK-3/APC complex in the nucleus of ET-intoxicated macrophages (Fig. 6F). A recent study also found that GSK-3 was able to bind C/EBP ␤, providing additional support for the idea of a complex consisting of APC, GSK-3, and C/EBP ␤ (11). Taken together, these results suggest a model where APC recruits GSK-3␤ to C/EBP ␤, leading the phosphorylation of C/EBP ␤ on Thr-188 and enhancing the ability of C/EBP ␤ to activate transcription. These results also reveal a novel role for APC in macrophages and demonstrate that APC modifies ET-induced gene expression through a mechanism involving APC associating with C/EBP ␤.