Association of Mitogen-Activated Protein Kinase Pathways with Gingival Epithelial Cell Responses to Porphyromonas gingivalis Infection

ABSTRACT Mitogen-activated protein (MAP) kinase pathways are key factors in host signaling events and can also play important roles in the internalization of pathogenic bacteria by host cells.Porphyromonas gingivalis, a periodontal pathogen, can efficiently invade human gingival epithelial cells (GECs). In this study, we examined the activation of MAP kinase pathways in GECs infected with P. gingivalis. c-Jun N-terminal kinase (JNK) was activated after 5 min of infection with P. gingivalis, whereas noninvasiveStreptococcus gordonii did not have a significant effect on JNK activation. In contrast, extracellular signal-regulated kinase (ERK) 1/2 was downregulated in a dose-dependent manner by P. gingivalis, but not by S. gordonii, after a 15-min exposure. Nonmetabolically active P. gingivaliscells were unable to modulate MAP kinase activity. U0126, a specific inhibitor of MEK1/2 (ERK1/2 kinase), and toxin B, a specific inhibitor of Rho family GTPases, had no effect on P. gingivalis invasion. Genistein, a tyrosine protein kinase inhibitor, blocked uptake of P. gingivalis. The transcriptional regulator NF-κB was not activated by P. gingivalis. These results suggest that P. gingivalis can selectively target components of the MAP kinase pathways. ERK1/2, while not involved in P. gingivalisinvasion of GECs, may be downregulated by internalized P. gingivalis. Activation of JNK is associated with the invasive process of P. gingivalis.

Porphyromonas gingivalis is a major etiologic agent in severe forms of periodontitis, a chronic inflammatory condition that leads to destruction of the periodontal tissues and eventual exfoliation of the teeth (44). Both in vivo (39,40) and in vitro (15,22,23,31,41), P. gingivalis can adhere to and invade epithelial cells, properties that facilitate retention in the oral environment and may contribute to immune evasion and tissue destruction. In primary cultures of gingival epithelial cells (GECs), invasion is a rapid event that is complete within 15 min, and large numbers of viable P. gingivalis cells accumulate in the perinuclear region (2). Host cell cytoskeletal rearrangements that accommodate invasion involve both microfilament and microtubule activity (23). The major fimbriae of P. gingivalis initiate the invasive process through binding to specific receptors on epithelial cell surfaces (31,48). Disruption of eukaryotic cell signaling pathways accompanies invasion. P. gingivalis induces a transient increase in epithelial cell cytosolic Ca 2ϩ , as a result of release of Ca 2ϩ ions from thapsigarginsensitive intracellular stores (20). In addition, epithelial cells induce P. gingivalis to secrete a novel set of proteins (37), one of which has homology to phosphoserine phosphatase enzymes (6) and thus could potentially interfere with eukaryotic information flow. One phenotypic outcome of P. gingivalis subversion of epithelial cell signaling is the inhibition of transcription and secretion of the neutrophil chemokine interleukin-8 (IL-8) (11,19,26). Moreover, P. gingivalis can also antagonize IL-8 secretion after stimulation of GECs by common plaque commensals (11), a process that could dampen the immune response in the periodontal area.
The mitogen-activated protein (MAP) kinases are central to many host cell signaling pathways. In addition to the mitogenic response to growth factors, from which their name derives, MAP kinases are involved in cytokine responses, cytoskeletal reorganization, and stress responses, with activity often funneling through nuclear transcription factors (38). The MAP kinase group includes the following three serine-threonine kinases: extracellular signal-regulated kinases (ERK) and the stress-activated protein kinases c-Jun N-terminal kinase (JNK) and p38 MAP kinase (12). MAP kinases are usually activated by phosphorylation of tyrosine and threonine residues by MAP kinase kinases (MEKs) (3,7). There are several MEKs with specific targets in the MAP kinase pathways that can thus provide a link between surface receptor-small G protein, Ras/ Rho cascades, and the MAP kinase signaling circuitry (24,27,28).
The invasive process of pathogenic bacteria is frequently associated with MAP kinase signaling activity. For example, infection of epithelial cell lines with Listeria monocytogenes, Salmonella enterica (serovar Typhimurium), or enteropathogenic Escherichia coli (EPEC) induces the activation of ERK1/2, JNK, and p38 MAP kinases (5,10,18,46,47), while invasive Neisseria gonorrhoeae can activate JNK specifically (29). In contrast, Yersinia species downregulate the activity of ERK1/2, JNK, and p38 MAP kinases in both macrophages and epithelial cells (33,36). This is effected through the YopJ protein of Yersinia pseudotuberculosis that is delivered into the host cell through the type III secretion machine and binds to MEKs, thus blocking phosphorylation and subsequent activation (33,34).
The involvement of MAP kinases in P. gingivalis invasion of GECs has not yet been demonstrated. In this study, we investigated the activity of MAP kinases in primary cultures of GECs after infection with P. gingivalis and correlated these activities with other components of the signaling pathways and with bacterial invasion. The results show that P. gingivalis downregulates ERK1/2 but induces the phosphorylation of JNK. The downregulation of ERK1/2 is associated with inhibition of the NF-B pathway. These results further suggest that JNK activation may be required for P. gingivalis invasion.
Culture of GECs. Primary cultures of human GECs were generated as described previously (22). Briefly, healthy gingival tissue was collected from patients undergoing surgery for removal of impacted third molars. Surface epithelium was separated by overnight incubation with 0.4% dispase (B-M Biochemicals, Indianapolis, Ind.). Single epithelial cells were recovered by centrifugation after digestion with 0.05% trypsin and 0.53 mM EDTA. Cells were cultured as monolayers in serum-free keratinocyte growth medium (Clonetics, San Diego, Calif.) at 37°C in 5% CO 2 . For invasion assays, cells were seeded in 24-well culture plates. To obtain lysates for protein kinase and NF-B assays, GECs were seeded in 75-cm 2 culture flasks. Cells were used at 80 to 90% confluence for all experiments and reacted with bacteria at a multiplicity of infection (MOI) of between 10 and 1,000. This ratio overlaps the range reported in vivo for the association of P. gingivalis with buccal epithelial cells (39).
Invasion assay. Invasion of GECs by bacteria was quantitated by the standard antibiotic protection assay, modified for P. gingivalis (23). In brief, P. gingivalis cells, at an MOI of 100, were incubated with GECs for 90 min at 37°C in epithelial cell culture medium. After a washing with PBS, the remaining external bacteria were killed with metronidazole (200 g/ml) and gentamicin (300 g/ml) for 60 min. The monolayers were washed and lysed with sterile distilled water, and intracellular bacteria were enumerated by culture on blood agar supplemented with hemin and menadione. U0126 and genistein (or solvent control) were added for 60 min, and toxin B was added for 16 h, prior to infection and maintained throughout the infection period. BAPTA/AM (or solvent control) was preincubated with GECs for 20 min, and GECs were washed and incubated for 20 min in fresh medium before exposure to P. gingivalis. Controls for direct effects of inhibitors on bacterial viability were included in all experiments.
Preparation of cell lysates for Western blotting and immunoprecipitation. After appropriate treatments, GECs were washed three times with ice-cold PBS and solubilized in lysis buffer (20 mM HEPES, pH 7.4; 2 mM EGTA; 50 mM ␤-glycerophosphate; 1 mM dithiothreitol [DTT]; 1% Triton X-100; 10% glycerol; 1 mM phenylmethylsulfonyl fluoride; 1 mM Na 3 VO 4 ; 10 g of aprotinin/ml; 10 g of leupeptin/ml) for 15 min on ice. The soluble fraction was collected by centrifugation at 16,000 ϫ g for 15 min at 4°C, and the protein concentration was determined by the Bio-Rad protein assay.
Western blotting of ERKs. Cell extract (10 g of protein) was denatured in sodium dodecyl sulfate (SDS) sample buffer, resolved by SDS-10% polyacrylamide gel electrophoresis (PAGE) and electrotransferred to a nitrocellulose membrane (Hybond ECL; Amersham Pharmacia, Piscataway, N.J.). The membrane was blocked with 3% bovine serum albumin in TBS-T (20 mM Tris-HCl, 150 mM NaCl, 0.05% Tween 20) for 1 h at room temperature and then incubated in primary antibody (1:5,000) to phosphorylated ERK1/2 overnight at 4°C. After a washing, horseradish peroxidase-linked donkey anti-rabbit immunoglobulin (Amersham), diluted 1:15,000, was added to the membrane, and the mixture was incubated for 1 h. Bands were visualized by an enhanced chemiluminescence detection system (Amersham). The membrane was then stripped with 2% SDS-62.5 mM Tris-HCl (pH 6.7)-100 mM ␤-mercaptoethanol for 30 min at 50°C and reprobed for nonphosphorylated total ERK1/2. ERK1/2 activity was determined by the ratio of phosphorylated to nonphosphorylated protein by NIH Image analysis.
JNK and p38 in vitro kinase assay. For the JNK and p38 in vitro kinase assay, a 300-g portion of protein from each cell lysate was immunoprecipitated with 0.6 g of anti-JNK1 or 1.0 g of anti-p38 polyclonal antibody. After continuous mixing overnight at 4°C, 20 l of 1:1 slurry of protein A-Sepharose CL4B was added and reacted for an additional 1 h. The immune complexes were pelleted at 4°C by centrifugation at 16,000 ϫ g for 30 s and washed twice each with (i) lysis buffer, (ii) LiCl wash buffer (500 mM LiCl; 100 mM Tris-HCl, pH 7.6; 0.1% Triton X-100; 1 mM DTT), and (iii) assay buffer (20 mM morpholinepropanesulfonic acid, pH 7.2; 2 mM EGTA; 10 mM MgCl 2 ; 1 mM DTT; 0.1% Triton X-100; 0.1 mM Na 3 VO 4 ). The pellets were resuspended to 50 l in kinase assay buffer containing 25 M ATP, 10 Ci of [␥-32 P]ATP (3,000 Ci/mmol; NEN), and 2 g of substrate c-Jun for JNK or ATF-2 for p38. After incubation for 20 min at 30°C, the reactions were terminated by the addition of 25 l of 3ϫ SDS sample buffer and boiling for 5 min. Samples were separated by SDS-12% PAGE. The gels were dried and subjected to autoradiography, and 32 P incorporation into c-Jun or ATF-2 was quantitated by NIH Image analysis.
EMSA. Activation of NF-B was assessed by an electrophoretic mobility shift assay (EMSA) with nuclear protein extracts. After bacterial exposure for 1 h at 37°C, GECs were scraped into 10 mM HEPES-1.5 mM MgCl 2 -10 mM KCl, and all subsequent procedures were performed at 4°C. Cells were pelleted by centrifugation and resuspended in the same buffer containing 0.1% Nonidet P-40. Nuclei were pelleted and suspended in 20 mM HEPES-1.5 mM MgCl 2 -0.42 M NaCl-0.2 mM EDTA-25% glycerol. After centrifugation, supernatants containing crude nuclear protein extracts were diluted in 20 mM HEPES-0.05 M KCl-0.2 M EDTA-20% glycerol, and protein concentrations were determined by a Bio-Rad protein assay. Samples with 20 g of nuclear protein extract, 2 g of poly(dI-dC) ⅐ poly(dI-dC) (Amersham), and 5 ϫ 10 5 to 1 ϫ 10 6 cpm of 32 P-end-labeled synthetic oligonucleotide were incubated at 25°C for 20 min and electrophoresed on 4% native polyacrylamide gels. The target DNA probe was 5Ј-GCCATTGGGGATTTCCTCTTT-3Ј in which the NF-B consensus binding sequence is underlined. Gels were dried, and the protein-DNA complexes were visualized by autoradiography.

P. gingivalis downregulates ERK1/2 in GECs.
The effect of P. gingivalis on ERK1/2 phosphorylation in GECs was studied at various time points and over a range of MOIs. Interestingly, Western blotting with an antibody that recognizes the dually phosphorylated ERK1/2 forms revealed the presence of activated ERK1/2 in control unstimulated primary GECs (Fig. 1). Exposure of GECs to P. gingivalis for 15 min reduced activation of ERK1/2 at MOIs of 10, 100, or 1,000 (Fig. 1). The ratio of phosphorylated to total ERK1/2, as determined by quantitative densitometry, confirmed that P. gingivalis could inhibit ERK1/2 activation by up to 90% at an MOI of 1,000. After 60 min of infection, activation of ERK1/2 began to recover in an inversely dose-dependent manner, with the largest rebound occurring in cells infected at an MOI of 10. After 5 min of exposure to P. gingivalis, inhibition of ERK1/2 activation was only observed at an MOI of 10. Higher numbers of P. gingivalis did not reduce ERK1/2 activity after this time period. These data suggest that GECs may be able to respond to P. gingivalis infection by more than one mechanism. High numbers of P. gingivalis may transiently stimulate ERK1/2, but this effect is quickly overcome by a simultaneous and longer-lasting specific downregulation that can also occur at lower bacterial numbers. The downregulation of ERK1/2 by P. gingivalis was statistically significant and was not a nonspecific response of these GECs to the presence of bacteria since the adherent, noninvasive oral organism, S. gordonii, did not reduce ERK1/2 activity but rather transiently increased phosphorylated ERK1/2 levels (Fig. 2). P. gingivalis activates JNK but not p38 in GECs. To examine the phosphorylation and activation of p38 and JNK MAP kinases in GECs in response to P. gingivalis infection, we utilized a sensitive in vitro kinase assay. We were unable to detect p38 activation in either control or bacterially stimulated cells (not shown). In contrast, active JNK was detected in unstimulated GECs and induced further by P. gingivalis at MOIs of 10 to 1,000 (Fig. 3). At an MOI of 10, activation was transient, appearing after 5 min and returning to baseline levels after 15 min. At an MOI of 100, activation was sustained through at least 60 min, and at an MOI of 1,000 activation declined after 60 min. Activation of JNK by P. gingivalis was statistically significant, whereas noninvasive S. gordonii cells did not induce significant activation of JNK (Fig. 4).
Metabolically active P. gingivalis is required for the modulation of MAP kinase activation. GEC MAP kinase responses to P. gingivalis infection could be the result of the stress of extracellular bacteria on the cell surface or could be associated with intracellular invasion. To begin to distinguish between these possibilities, P. gingivalis cells were heat inactivated at 80°C for 30 min, and MAP kinase responses were examined (Fig. 5). Heat-inactivated, and thus noninvasive (2, 23), P. gingivalis cells were unable to stimulate JNK activity or to downregulate ERK1/2 activity. Identical results (data not shown) were obtained with azide-treated (50 mM NaN 3 for 3 h) P. gingivalis cells that are also unable to invade GECs (23). Although further study is required, these data are consistent with the concept that the activation of JNK and inactivation of ERK1/2 is associated with P. gingivalis invasive process.
Effects of signaling inhibitors on P. gingivalis invasion. Since the data indicated that MAP kinase activity is associated with P. gingivalis invasion, we examined the effects of various inhibitors of these signaling pathways on P. gingivalis invasion (Table 1). U0126, a specific inhibitor of MEK1/2 (ERK1/2 kinase), neither increased nor decreased P. gingivalis internalization within GECs. Thus, invasion of GEC would appear to be independent of ERK1/2 signaling and downregulation of ERK1/2 by P. gingivalis may be a property of internalized bacteria. No effect on invasion was observed with toxin B from Clostridium difficile, a specific inhibitor of Rho family GTPases (Rho, Rac, and Cdc42) that activate predominantly JNK and p38 (25). Toxin B also did not block activation of JNK by P. gingivalis (not shown). These results argue that P. gingivalis acts on a step subsequent to GTPase activation to achieve stimulation of JNK. Genistein, a tyrosine protein kinase inhibitor, reduced P. gingivalis invasion by more than 90%. Hence, certain protein phosphorylations are required for optimal invasion, a result that is consistent with a role for JNK in the invasive process. Corroboration of the involvement of JNK in FIG. 1. Representative blots of ERK1 and ERK2 (p44 and p42, respectively) activation in GECs. GECs were lysed after infection with P. gingivalis over a range of MOIs (10 to 1,000) or after stimulation with phorbol myristate acetate (1 M) at 37°C for the time periods indicated. Samples were immunoblotted with antiphosphorylated-ERK1/2 antibodies (upper panel), and the blot was then stripped and reprobed with antibody to ERK1/2 (middle panel). The lower panel denotes the level of phosphorylation quantitated by NIH Image analysis and is presented as the intensity of phosphorylated ERK1/2 relative to total ERK1/2. Activity in control uninfected cells was set to 100%. To determine the effect of P. gingivalis on NF-B activation, the ability of nuclear extracts to complex with DNA containing the NF-B consensus binding sequence was assessed by an EMSA (Fig. 6). Whereas S. gordonii induced nuclear translocation, P. gingivalis did not activate NF-B. Thus, with regard to the NF-B pathway, P. gingivalis activity could be mediated through ERK1/2 or by the same mechanism that disrupts ERK1/2 activity.

DISCUSSION
The pathogenesis of bacterially induced periodontal diseases is complex and involves both host and microbial components. P. gingivalis is an aggressive periodontal pathogen that pos-sesses numerous virulence factors with the potential to impinge upon host tissue integrity and immune function. Included among these is the ability to invade both epithelial and endothelial cells (13,14,23). Invasion of primary epithelial cells induces calcium ion fluxes and cytoskeletal rearrangements and results in the downregulation of IL-8 secretion (2,11,20). However, the full extent to which epithelial cell signal transduction pathways are disrupted, and phenotypic properties altered, by P. gingivalis is uncertain. MAP kinase pathways have been demonstrated to play an important role in bacterial internalization and modulation of cytokine responses. Invasive L. monocytogenes, S. enterica serovar Typhimurium, and EPEC induce activation of ERK1/2, JNK, and p38 MAP kinases in epithelial cells (5,10,18,46,47). In contrast, P. gingivalis downregulated ERK1/2 activity in a dose-dependent manner after 15 min of infection. This effect is more reminiscent of yersiniae that block MAP kinase activation (albeit JNK and p38 along with ERK) through type III secretion-mediated delivery of the intracellular effector molecule YopJ (33,34,35,36). Although P. gingivalis does not possess the apparatus of a classical type III protein secretion system, a functionally equivalent pathway may be present as the organism secretes a novel set of proteins when in an epithelial cell environment (37). One of these proteins has homology to phosphatase enzymes (6), which could, therefore, play a role in the downregulation of ERK1/2. Alternatively, since YopJ exerts its activity through a cysteine protease action (34), the secreted cysteine proteases of P. gingivalis (9) could similarly disrupt MAP kinase. The inhibition of ERK1/2 kinase activity by P. gingivalis in primary cultures of GECs would appear to differ from the situation observed in the KB oral epidermal cell line, in which P. gingivalis induces tyrosine phosphorylation of a host cell protein that is similar in size to ERK1/2 (42). This may reflect distinct uptake strategies adopted by the organism in different cell types. Indeed, the physical location of P. gingivalis within GECs is in the cytoplasm, predominantly in the perinuclear area, whereas in KB cells the organisms remain in a membranebound vacuole (2,22,23,31,41).
P. gingivalis rapidly (5 min) activated JNK, whereas it had no effect on p38. These properties are similar to those of pathogenic strains of N. gonorrhoeae that activate JNK specifically after contact with epithelial cells. P. gingivalis, however, possesses the additional property of inhibition of ERK1/2. Although there are numerous opportunities for interconnectivity among MAP kinase components, the major constituents can be insulated from one another (28,49). Thus, it would appear that P. gingivalis is capable of selectively activating one MAP kinase pathway and downregulating another, a phenomenon thus far unique to this oral organism. The components of P. gingivalis responsible for JNK activation remain to be deter-mined. Enterobacterial lipopolysaccharide (LPS) has been shown to stimulate JNK in a variety of cell types via pathways that include Toll-like receptors (TLRs) and CD14 (4,16). However, P. gingivalis LPS differs structurally from enterobacterial LPS (32), and binding of P. gingivalis LPS to both CD14 and TLRs is different from that of enterobacterial LPS (8,17). Furthermore, since the GECs are cultured in serum-free media, and thus without a source of LPS-binding protein (LPB) or soluble CD14, it is unlikely that LPS is in the effector molecule. Invasive salmonellae activate JNK through the type III secreted effectors SopB and SopE. SopB is an inositol phosphate phosphatase, whereas SopE activates a host cell inositol phosphate phosphatase (50). Whether P. gingivalis cells possess similar activity is currently under investigation.
The time course of modulation of MAP kinase activity in GECs by P. gingivalis is concurrent with intracellular invasion by the organism that is essentially complete after 15 min (2), suggesting an association with the mechanism of invasion. Alternatively, MAP kinase responses could be effected by adherent extracellular bacteria. However, as heat-and azide-inactivated P. gingivalis cells, which can adhere to GECs but not invade (2,23), were unable to disrupt control of the MAP kinases, the MAP kinase responses are more likely to be associated with the invasive process or with internalized bacteria. Indeed, specific inhibition of ERK with U0126 did not affect invasion levels, indicating that downregulation of ERK1/2 may occur subsequent to P. gingivalis internalization. Another implication of this result is that, similar to EPEC (10), P. gingivalis uses nontraditional mechanisms for the actin rearrangements required for epithelial cell membrane penetration. In contrast to the ERK inhibitor, a broad-spectrum inhibitor of tyrosine kinases, genistein, reduced P. gingivalis invasion, a result consistent with the involvement of JNK in the invasive process. The finding that stimulation of JNK precedes ERK1/2 suppression is also consistent with this concept. Activation of JNK, however, bypasses GTPase activation since toxin B did not reduce invasion or prevent JNK phosphorylation. Specific inhibition of JNK activity, however, would be required to confirm the role of JNK.
Invasion of GECs by P. gingivalis results in the transcriptional downregulation of IL-8 expression (11). The IL-8 gene can be controlled by the transcriptional activator NF-B (1). Activation of NF-B requires phosphorylation and subsequent degradation of the cytoplasmic inhibitor IB. Removal of IB from NF-B allows NF-B to translocate into the nucleus, where it recognizes specific motifs to initiate transcription (1). In this manner, the sequential phosphorylations mediated through MAP kinase pathways can converge upon the NF-B pathway. P. gingivalis did not induce nuclear translocation of active NF-B in GECs, a result consistent with reduced IL-8 expression. Since P. gingivalis can both stimulate JNK and suppress ERK1/2 activity, ERK-mediated effects would appear to have the greater influence on NF-B activation in this P. gingivalis-GEC model system. It is also possible that P. gingivalis can directly impinge upon NF-B activation, as has been reported for the Yersinia YopJ protein (33,43), and nonpathogenic Salmonella strains (30). In addition to the regulation of IL-8 and other immune effectors, NF-B controls the transcription of genes involved in growth and development, cell adhesion, and cell survival (45). Interference with NF-B ac-FIG. 6. EMSA for NF-B activation in nuclear extracts of GECs infected with P. gingivalis or S. gordonii at an MOI of 100 for 60 min. A nonspecific shifted band is always observed in treated and untreated samples and is not considered to have transactivating potential. The shifted band of higher molecular weight responds to activation conditions and represents NF-B complexes. a Reduction in invasion in the presence of inhibitors compared to a control without inhibitor. Values are the means Ϯ the SDs (n ϭ 3).
b Chelation of intracellular calcium is a positive control for inhibition of P. gingivalis invasion (20).

VOL. 69, 2001
MAP KINASE AND GEC RESPONSE TO P. GINGIVALIS INFECTION tivation by P. gingivalis could, therefore, have significant implications for the health status of the cells in the gingival compartment. Epithelial cells comprise an interactive interface with colonizing bacteria and can generate and transmit signals that activate the immune response (21). The outcome of the molecular cross talk between bacteria and host epithelial cells has, therefore, important implications for health and disease. A pattern is emerging from the accumulating literature that epithelial cell responses are species or even strain specific. In the case of P. gingivalis, the results obtained in this and previous studies suggest a model whereby the invasive process of P. gingivalis is associated with activation of JNK. Internalized P. gingivalis cells then downregulate ERK1/2 activity that, in turn, prevents the activation of NF-B and the loss of IL-8 secretion. The disruption of the ratios of active MAP kinase components will also have a number of consequences for the physiologic properties of the host cell related to proliferation, differentiation, and apoptosis. The ability to locate intracellularly, manipulate signal transduction, and paralyze components of the innate host defense may contribute to both the success of P. gingivalis in the periodontal environment and its pathogenic activities.