C/EBPβ and YY1 bind and interact with Smad3 to modulate lipopolysaccharide‐induced amelotin gene transcription in mouse gingival epithelial cells

Junctional epithelium (JE) develops from reduced enamel epithelium during tooth formation and is critical for the maintenance of healthy periodontal tissue through ensuring appropriate immune responses and the rapid turnover of gingival epithelial cells. We have previously shown a relationship between inflammatory cytokines and expression of JE‐specific genes, such as amelotin (AMTN), in gingival epithelial cells. Here, we elucidated the effects of Porphyromonas gingivalis‐derived lipopolysaccharide (Pg LPS) on Amtn gene transcription and the interaction of transcription factors. To determine the molecular basis of transcriptional regulation of the Amtn gene by Pg LPS, we performed real‐time PCR and carried out luciferase assays using a mouse Amtn gene promoter linked to a luciferase reporter gene in mouse gingival epithelial GE1 cells. Gel mobility shift and chromatin immunoprecipitation assays were performed to identify response elements bound to LPS‐induced transcription factors. Next, we analyzed protein levels of the LPS‐induced transcription factors and the interaction of transcription factors by western blotting and immunoprecipitation. LPS increased Amtn mRNA levels and elevated luciferase activities of constructs containing regions between −116 and −238 of the mouse Amtn gene promoter. CCAAT/enhancer‐binding protein (C/EBP) 1–, C/EBP2– and Ying Yang 1 (YY1)–nuclear protein complexes were increased by LPS treatment. Furthermore, we identified LPS‐modulated interactions with C/EBPβ, YY1 and Smad3. These results demonstrate that Pg LPS regulates Amtn gene transcription via binding of C/EBPβ–Smad3 and YY1–Smad3 complexes to C/EBP1, C/EBP2 and YY1 response elements in the mouse Amtn gene promoter.

Junctional epithelium (JE) develops from reduced enamel epithelium during tooth formation and is critical for the maintenance of healthy periodontal tissue through ensuring appropriate immune responses and the rapid turnover of gingival epithelial cells. We have previously shown a relationship between inflammatory cytokines and expression of JE-specific genes, such as amelotin (AMTN), in gingival epithelial cells. Here, we elucidated the effects of Porphyromonas gingivalis-derived lipopolysaccharide (PgLPS) on Amtn gene transcription and the interaction of transcription factors. To determine the molecular basis of transcriptional regulation of the Amtn gene by PgLPS, we performed real-time PCR and carried out luciferase assays using a mouse Amtn gene promoter linked to a luciferase reporter gene in mouse gingival epithelial GE1 cells. Gel mobility shift and chromatin immunoprecipitation assays were performed to identify response elements bound to LPS-induced transcription factors. Next, we analyzed protein levels of the LPS-induced transcription factors and the interaction of transcription factors by western blotting and immunoprecipitation. LPS increased Amtn mRNA levels and elevated luciferase activities of constructs containing regions between À116 and À238 of the mouse Amtn gene promoter. CCAAT/enhancer-binding protein (C/EBP) 1-, C/EBP2-and Ying Yang 1 (YY1)-nuclear protein complexes were increased by LPS treatment. Furthermore, we identified LPS-modulated interactions with C/EBPb, YY1 and Smad3. These results demonstrate that PgLPS regulates Amtn gene transcription via binding of C/EBPb-Smad3 and YY1-Smad3 complexes to C/EBP1, C/EBP2 and YY1 response elements in the mouse Amtn gene promoter.
Junctional epithelium (JE) is developed from reduced enamel epithelium during tooth formation [1,2] and persists on the tooth surface, including on mature enamel and cementum with gingival recession. Internal basal laminae of JE and tooth surfaces are bound to each other by hemidesmosomes. The seal of periodontal tissue prevents the entry of bacteria and products such as lipopolysaccharide (LPS), subsequently produced proinflammatory cytokines and proteolytic enzymes [3]. Consequently, it is widely accepted that the function of JE is critical for the maintenance of healthy periodontal tissue by an appropriate immune response and the rapid turnover of gingival epithelial cells [4].
Amtn genes from mouse and human display high homologous exon-intron structure and were expressed from loci on chromosomes 5 and 4, respectively, which have a close proximity to enamelin (ENAM), ameloblastin (AMBN) and the small, integrin-binding ligand Nlinked glycoprotein (SIBLING) gene family [15]. The AMTN protein is enriched in glutamine, leucine, proline and threonine (52% of total) and contains a conserved protein kinase (casein kinase 2) phosphorylation site [16]. Amtn transcripts demonstrated to have two variants, variant 1 (V1) and variant 2 (V2). V1 represented 70.8% of Amtn transcripts and displayed the structure known in rodents, whereas V2 represented 29.2% and exhibited the non-mammalian tetrapod structure [17]. Differences of the function of AMTN protein due to Amtn gene structure are unclear, though AMTN protein has a direct influence on biomineralization by promoting hydoroxyapatite (HA) mineralization [18] and is associated with amelogenesis imperfecta [19].
AMTN and ODAM are expressed not only by mature ameloblasts at amelogenesis but also in JE at an erupted tooth [16,20,21]. AMTN knockout mice showed hypomature enamel [22], and AMTN overexpression mice have irregular enamel structures and thin enamel compared with wild-type mice [23]; however, both mice exhibited no pronounced abnormality of periodontium. Whereas ODAM-deficient mice showed gingival recession with aging, delay in regeneration following gingivectomy and a tendency towards decrease of AMTN expression in JE, no irregular enamel structures were observed [24]. It has been stated that AMTN, ODAM and FDC-SP form complexes at JE of erupted tooth surfaces [25,26]. AMTN and FDC-SP localize at internal basal lamina of JE, whereas ODAM is evenly distributed in JE [27]. Therefore, it is important to understand the mechanism of proper maintenance of JE structure by investigation the regulatory mechanism of Amtn gene transcription in inflammation.
Porphyromonas gingivalis (Pg) is a major periodontopathic bacterium [28]. LPS is one of the multiple virulence factors in the Gram-negative bacterium and therefore might contribute to the initiation and progression of periodontal diseases [29,30]. In our previous studies, tumor necrosis factor-a (TNF-a) and interleukin-1b (IL-1b) stimulate human Amtn gene transcription via CCAAT enhancer-binding protein (C/ EBP) and Ying Yang 1 (YY1) elements [7,8,31]. The relationship between C/EBPs and inflammation has been reported [32] and YY1 is also a transcription factor associated with inflammation and cancer [33]. However, the regulatory mechanism of Amtn gene transcription by PgLPS, which can induce TNF-a and IL-1b in inflamed gingiva, remains to be elucidated.
Here, we report that PgLPS-induced physical interactions between C/EBPb, YY1 and Smad3 upregulate Amtn gene transcription via mitogen-activated protein kinase kinase 1/2 (MEK1/2), phosphatidylinositol 3kinase (PI3-K) and Smad2/3 pathways in mouse gingival epithelial cells. These findings raise the possibility of an association between induced C/EBPb and YY1 in inflammation and a constitutive interaction with Smad3 in gingival epithelial cells.

Cell cultures
A mouse-derived gingival epithelial cell line (RCB1709; GE1) was purchased from Riken BRC (Ibaraki, Japan) [35]; cells were cultured at 33°C in 5% CO 2 /95% air in SFM-101 containing 1% FCS, 10 ngÁmL À1 EGF and 1% penicillin and streptomycin. Cells were grown to confluence in 60 mm cell culture dishes for RNA extraction; the media were then changed to no serum and EGF SFM-101 medium for 24 h. Then, cells were incubated in this medium with or without PgLPS (0.1 lgÁmL À1 ) for 0, 12 and 24 h. For the extraction of nuclear protein and DNA-protein complex, 100 mm cell culture dishes were used.

Real-time PCR
Total RNA was isolated using Isogen II from the GE1 cell cultures, and was used as a template for cDNA synthesis with an ExScript RT reagent kit (TaKaRa). Quantitative real-time PCR was performed using the following primer sets: mouse Amtn forward, 5 0 -CTGTCAACCAGGGAAC-CACT-3 0 ; mouse Amtn reverse, 5 0 -TGTGATGCGGTT-TAGCTGAG-3 0 ; and mouse glyceraldehyde 3-phosphate dehydrogenase (gapdh) forward, 5 0 -TGAAGGGGTGGTT-GATGG-3 0 ; mouse gapdh reverse, 5 0 -A AATGGT-GAAGGTCGGTGTG-3 0 . SYBR Premix Ex Taw was used in a TP800 thermal cycler dice real-time system (TaKaRa). Primers were purchased from Sigma-Aldrich Japan. To investigate the signaling pathways in the transcriptional regulation of the Amtn gene by PgLPS, we used protein kinase inhibitors, such as KT5720, U0126, LY249002 and SB525334. After the GE1 cells were at confluence, the cells were deprived of FCS for 12 h, and first treated with KT5720 (100 nM), U0126 (5 lM), LY249002 (10 lM) and SB525334 (1 lM) for 30 min, and then incubated with or without PgLPS (0.1 lgÁmL À1 ) for 24 h before harvesting. The amplification reactions were performed in 25 lL of the final reaction mixture containing: 2 9 SYBR Advantage qPCR Premix or SYBR Premix EX Taq (12.5 lL); 10 lM forward and reverse primers (final concentration was 0.2 lM); 50 ng (5 lL) cDNA for Amtn and 20 ng (2 lL) cDNA for gapdh. To reduce variability between replicates, PCR premixes that contained all reagents except for cDNA, were prepared and aliquot into 0.2 mL Hi-8-tubes (TaKaRa). The thermal cycling condition was 10 s at 95°C and 40 cycles of 5 s, 95°C and 30 s, 60°C. Post-PCR melting curves confirmed the specificity of singletarget amplification. The initial amount of RNA was quantified using a standard curve, and fold expressions of Amtn relative to gapdh were determined in triplicate. To visualize and confirm control levels of Amtn mRNA levels and appropriate amplification, running gel was performed using 8 lL PCR products. The product size is 208 bp according to design of primers.

Gel mobility shift assay
Nuclear protein extracts from GE1 cells were incubated for 12 and 24 h with PgLPS (0.1 lgÁmL À1 ) in SFM-101 without serum. Cy5-labeled double-stranded oligonucleotide encompassing the inverted CCAAT, C/EBP1, C/EBP2 and YY1 elements in the mouse Amtn gene promoter were prepared. Their complementary oligonucleotide and Cy5labeled oligonucleotide were purchased (Sigma-Aldrich Japan; Table 1), and they were annealed under optimal conditions. The following steps were carried out as described in a recent study [7,8,31]. Briefly, nuclear proteins (20 lg) and 0.1 pM Cy5-labeled double-stranded oligonucleotide were combined and incubated for 20 min at room temperature (RT) in the binding buffer; then, the DNA-nuclear protein complexes were separated by electrophoresis in 5% acrylamide gels. Imaging of gels was analyzed using a Typhoon Trio+ Variable Mode Imager (GE Healthcare). For competition experiments, 40-fold molar unlabeled oligonucleotides of inverted CCAAT, C/EBP1, C/EBP2 and YY1 were used [7]. Supershift experiments were carried out using the appropriate unconjugated normal rabbit anti-IgG antibody (sc-2027, Santa Cruz, Biotechnology), and rabbit polyclonal antibodies to C/EBPb (sc-746; Santa Cruz Table 1. Sequences of oligonucleotides for gel mobility shift assay.

Response element
Including Location in the mouse AMTN promotor

Chromatin immunoprecipitation assay
To support the results of gel mobility shift assay, indicating that C/EBPs, YY1 and Smad3 bind specifically to their response elements in the mouse Amtn gene promoter, chromatin immunoprecipitation (ChIP) assays were performed using GE1 cells. Confluent GE1 cells were treated for 12 and 24 h with PgLPS (0.1 lgÁmL À1 ) in SFM-101 without serum. Prior to cell harvest, cell fixation with 160 lL formaldehyde to cross-link the DNA-protein complexes was performed. Extraction of DNA including response elements was conducted as described previously [8]. Two micrograms of the same antibodies as for the gel mobility shift assay was used. The purified DNA was subjected to PCR amplification for inverted CCAAT, C/EBP1, C/EBP2 and YY1 within the mouse Amtn gene promoter using primer sets (Table 2). For the PCR procedure, KAPA Taq TM EXtra HotStart was utilized.

Immunoprecipitation and western blotting
For regulation of Amtn gene transcription by PgLPS via YY1 and Smad3, immunoprecipitation was carried out to examine whether YY1 and Smad3 binding responded independently or interactively to their own elements in the mouse Amtn gene promoter. Firstly, to nuclear proteins (15 lg) in NP-40 solution (0.5% NP-40, 10 mM Tris (pH 7.5), 0.5 M NaCl, 0.2 mM PMSF in ddH 2 O) was added 0.25 lg of unconjugated normal rabbit anti-IgG antibody (sc-2027, Santa Cruz Biotechnology) together with 20 lL of appropriate suspended protein A/G PLUS-Agarose Immunoprecipitation Regent, and they were then incubated at 4°C for 30 min with gentle rotation. After centrifugation of pellet beads, to the transferred supernatant was added 10 lL of 10% BSA and 1 lg antibodies (same as ChIP assay), and they were incubated at 4°C for 1 h with gentle rotation. To nuclear protein-antibody complexes was added 20 lL of appropriate protein A/G PLUS-Agarose conjugate suspension, and they were mixed and incubated at 4°C with gentle rotation overnight. The pellet was collected by centrifugation at 30 000 g for 30 s at 4°C; then, the supernatant was carefully removed. The flow-through lysates were used for detection of histone deacetylase-1 (HDAC1) expression as controls. After washing the pellet with NP-40 solution twice, pellet was resuspended in 1 mL of 10 mM Tris/HCl buffer (pH 7.5). The final washed pellet without supernatant was resuspended in 15 lL of electrophoresis sample buffer (S3401; Sigma-Aldrich Japan), then boiled for 5 min, and used for the western blotting procedure. The nuclear proteins that were precipitated with IgG or Smad3 were separated on 12% SDS/PAGE and transferred to a Hybond 0.2-lm poly (vinylidene difluoride) membrane. The membrane was incubated at 4°C overnight with anti-YY1 (1 : 500) or anti-C/ EBPb (1 : 1000) antibodies. Anti-rabbit and anti-mouse IgG conjugated with horseradish peroxidase were used as the secondary antibodies. ECL Prime Western Blotting Detection Reagent was used to detect immunoreactivities; the imaging was visualized with an ImageQuant LAS4000 Mini system (GE Healthcare Japan, Tokyo, Japan). Anti-YY1 (1 : 500), anti-C/EBPb (1 : 1000), anti-Smad3 (1 : 3000), anti-pSmad3 (1 : 1000; anti-rabbit monoclonal antibody, ab52903, Abcam) and anti-HDAC1 (1 : 1000; anti-mouse monoclonal antibody, sc-81598, Santa Cruz Biotechnology) antibodies were used for determining (C) Regulatory elements in the proximal mouse Amtn gene promoter. Sequences of mouse Amtn gene promoter (À290 to +60) and locations of response elements, such as inverted CCAAT box (À66 to À62), SBE1 (À82 to À79), C/EBP1 (À113 to À103), C/EBP2 (À157 to À144), which partially overlaps SBE2 (À159 to À156), and YY1 (À221 to À209) are shown in the proximal promoter region of the mouse Amtn gene.
whether PgLPS regulated these transcriptional factors at protein levels. Densitometric data were provided using IMAGEJ software (National Institutes of Health, Bethesda, MD, USA), according to the supplier's description. Briefly, each area of a band was detected by the software and values were calculated compared with HDAC1 expression (Table 3).

Immunocytochemistry
To examine protein expression of C/EBPb and YY1 in GE1 cells, immunofluorescence was carried out. After plating 30 lL of GE1 cells from confluence in 75 cm 2 flask on culture slides (Falcon Ò CultureSlides, Corning Ò BioCoat TM , Hampton, NH, USA), the plates with cells were incubated at 33°C for 12 h. After treatment with PgLPS for 12 or 24 h, the cells were rinsed with PBS twice and fixed by 4% paraformaldehyde for 15 min at RT. The fixed cells were rinsed by PBS twice and permeabilized with 0.5% Triton X-100 in PBS for 10 min at RT. After removing permeabilization solution, there followed blocking with 1% bovine serum albumin in PBS for 20 min at RT. The cells were treated with 100 lL of primary antibodies diluted with buffer (1% BSA in PBS) for 1 h at RT followed by incubation with secondary antibodies for 1 h at RT. For the primary antibodies, we used C/EBPb (1 : 50, sc-746; Santa Cruz Biotechnology), YY1 (ChIP Grade; ab38422; Abcam) rabbit polyclonal antibodies and CK19 (1 : 100, ab7755, Abcam) mouse monoclonal antibody. For detection signaling for fluorescence, we used goat anti-mouse IgG H&L (Alexa Fluor Ò 488, 1 : 200, ab150113, Abcam) and goat anti-rabbit IgG H&L (Alexa Fluor Ò 647, 1 : 200, ab150079, Abcam). Mounting and nuclear staining with 4 0 ,6-diamidino-2-phenylindole (DAPI) were performed (Fluoroshield mounting medium with DAPI, ab104139, Abcam).

Statistical analysis
Triplicate samples for real-time PCR were analyzed for each experiment and were replicated to ensure consistency of the responses to PgLPS. Significant differences between the control and treatments were determined using multiple comparison calibration (the Tukey-Kramer method) for real-time PCR and one-way analysis of variance (one-way ANOVA).

Upregulation of Amtn mRNA levels by P. gingivalis LPS in GE1 cells
To determine the Amtn mRNA levels after treatment by PgLPS (0.1 lgÁmL À1 ), we performed real-time PCR using total RNA from GE1 cells. Amtn mRNA levels were significantly increased by PgLPS at 24 h (Fig. 1A). We confirmed increases of Amtn mRNA levels by appropriate amplification with visible bands in 2% agarose gel. To determine the signaling pathways for the regulation of Amtn gene transcription by PgLPS in GE1 cells, we used protein kinase inhibitors. Amtn mRNA levels were significantly increased by PgLPS at 24 h. U0126 (a MEK1/2 inhibitor), LY249002 (a PI3-K inhibitor) and SB525334 (an ALK5 inhibitor) almost completely abolished the effects of PgLPS, whereas KT5720 (a PKA inhibitor) had no effect (Fig. 1C). These results suggested that the induction of Amtn mRNA levels by LPS were mediated through the MEK1/2-, PI3-K-and ALK5involved Smad3 signaling pathways.
To determine the signaling pathways mediating the PgLPS effects on Amtn gene transcription, kinase inhibitors were used as in the real-time PCR analyses. PgLPS-induced À238AMTN promoter activity was inhibited by U0126, LY249002 and SB525334, but no effect was observed with KT5720, indicating an involvement of MEK1/2, PI3-K/Akt and ALK5 in mediating the effects on Amtn gene transcription (Fig. 2B).

Identification of response elements regulated by PgLPS in GE1 cells
To determine whether PgLPS altered nuclear proteins binding to the C/EBP1, C/EBP2 and YY1 elements, Cy5-labeled double-stranded oligonucleotides were incubated with equal amounts (20 lg) of nuclear proteins extracted from GE1 cells that were either not treated (Control) or treated with PgLPS (0.1 lgÁmL À1 ) for 12 and 24 h. After stimulation by PgLPS, C/EBP1-protein complexes (Fig. 4A, lanes 5 and 6) and YY1-protein complexes (Fig. 4A, lanes 11 and 12) were increased at 12 and 24 h, and C/EBP2-protein complexes (Fig. 4A, lane 9) were increased at 24 h in GE1 cells. The nuclear protein binding to the inverted CCAAT sequence was not increased by PgLPS treatment (Fig. 4A, lanes 1-3). To confirm the DNA-protein complexes represent specific interactions, competition gel shifts were carried out in which 40fold molar excess of C/EBP1, C/EBP2 and YY1 double-stranded oligonucleotides reduced the amount of C/EBP1-, C/EBP2-and YY1-protein complex formation (Fig. 4B). Interestingly, non-labeled 40-fold molar excess of C/EBP1, C/EBP2 and YY1 oligonucleotides except for CCAAT competed with C/EBP1-, C/EBP2-and YY1-protein complexes. These results suggested that common transcription factors bind to the three elements and/or they interact with each other.
To support the results of transient transfection assays using several kinase inhibitors, we investigated the effects of kinase inhibitors on the DNA-protein complex formation with or without stimulation by PgLPS. Increases in the C/EBP1-protein complex formation induced by PgLPS (24 h) were abolished by U0126 (Fig. 5A, lanes 5 and 6), LY249002 (Fig. 5A, lanes 7 and 8) and SB525334 (Fig. 5A, lanes 9 and 10), but not by KT5720 (Fig. 5A, lanes 3 and 4). For increases in the C/EBP2-and YY1-protein complex formation, similar effects of U0126, LY249002 and SB525334 were observed (Fig. 5B,C, lanes 5-10). These results were consistent with the results of transient transfection assays using MEK1/2, PI3-K and ALK5 inhibitors (Fig. 2).
To further clarify the interaction between specific transcription factors, such as C/EBPb, YY1 and Smad3, in the mouse Amtn gene promoter in vivo, chromatin immunoprecipitation assays (ChIP) assays were carried out. PgLPS induced C/EBPb binding to C/EBP1, C/EBP2 and YY1, and the YY1 binding to C/EBP1, C/EBP2 and YY1. Interestingly, Smad3 binding to C/EBP1, C/EBP2 and YY1 was increased by PgLPS in a time-dependent manner. The fragmented DNA-protein complexes were used for positive control (input) without using specific antibodies for immunoprecipitation (Fig. 6).

Increase of protein levels of LPS-induced transcription factors in GE1 cells
We next investigated effects of PgLPS on the protein levels of transcription factors, such as C/EBPb, YY1, Smad3 and pSmad3 which bind to C/EBP1, C/EBP2 and YY1 elements in the mouse Amtn gene promotor. PgLPS increased C/EBPb (1.29-fold) and YY1 (1.40-fold) protein levels at 12 h and pSmad3 protein levels at 12 (2.67-fold) and 24 h (2.81-fold), whereas Smad3 protein levels were not consistently changed by PgLPS ( Fig. 7 and Table 3).
To elucidate the localization and protein expression of C/EBPb and YY1, immunocytochemistry was carried out. C/EBPb and YY1 protein levels in the nucleus were increased by treatment with PgLPS for 12 h, consistent with the results of western blot (Fig. 8Ae,Be). Levels of CK19 protein, an inflammation marker, were slightly increased by PgLPS (Fig. 8Af,Bf). These results suggest that PgLPSinduced C/EBPb and YY1 and phosphorylation of Smad3 were associated with LPS-induced Amtn gene transcription. Fig. 6. ChIP assay was performed using GE1 cells to investigate C/EBPb, YY1 and Smad3 proteins bind to C/EBP1, C/EBP2 and YY1 elements in the mouse Amtn gene promoter. Binding of C/EBPb to C/EBP1 and C/EBP2 was increased after 24 h stimulation by PgLPS (0.1 lgÁmL À1 ), and C/EBPb binding to YY1 was induced after 12 h stimulation by PgLPS (0.1 lgÁmL À1 ). YY1 and Smad3 binding to C/EBP1, C/EBP2 and YY1 was induced after 12 h stimulation by PgLPS (0.1 lgÁmL À1 ). Fig. 7. Change of C/EBPb, YY1, Smad3 and pSmad3 nuclear proteins from GE1 gingival epithelial cells was examined with western blot analysis. After treatment with PgLPS, protein levels of C/EBPb and YY1 were increased at 12 h of treatment. Timedependent treatment with PgLPS did not change Smad3 protein level; however, pSmad3 protein level was induced by treatment with PgLPS at 12 and 24 h. Expression of HDAC1 was used as control. Densitometric analysis was performed (Table 3).

C/EBPb and YY1 directly interact with Smad3 in GE1 cells
Smad3 is tightly associated with several transcription factors, and can promote or modulate their activities [36]. Hence, we investigated whether the modulation of Smad3 activity by PgLPS involves a stimulusdependent physical interaction. Immunoprecipitation analysis demonstrated that C/EBPb in nuclear extract was co-precipitated with Smad3 in GE1 cells; however, treatment with PgLPS did not influence their interactions. In contrast, for the interaction between YY1 and Smad3, PgLPS increased YY1 binding to Smad3 (Fig. 9). We performed western blot with HDAC1 antibody using flow-through lysates after immunoprecipitation with anti-IgG and anti-Smad3 antibodies as the loading control; these results demonstrated that equal amounts of lysates were used for immunoprecipitation.

Discussion
Amtn gene transcription is tightly regulated and is affected by the activity of inflammatory cytokines and their byproducts. AMTN can lead to pathologies including inflammation [5,6], apoptosis [13,14] and cancer [37]. In this study we have shown that PgLPS induced Amtn gene transcription that was mediated through C/EBPb, YY1 and Smad3 binding to C/EBP1, C/EBP2 and YY1 elements, which was accompanied by physical interaction between C/EBPb, YY1 and Smad3. LPS could induce IL-1b and TNF-a production in THP-1 cells [38], and IL-1b and TNF-a increased Amtn gene transcription in gingival epithelial cells [7,8,31]. Similar to the results of these studies, we demonstrated that PgLPS induced Amtn mRNA levels after 24 h treatment and upregulated mouse Amtn gene transcription via MEK1/2, PI3-K and ALK5 signaling pathways in GE1 cells (Figs 1 and 2). It is known that LPS upregulates ALK5 expression in the induction of TNF-a biosynthesis [39]. Regulatory mechanisms of Amtn gene transcription by PgLPS are consistent with those for IL-1b and TNF-a, indicating that IL-1b and TNF-a induced transcription factor binding to C/EBP1, C/EBP2 and YY1 elements [7,8,31] (Figs 2C, 4 and 6).
We confirmed that C/EBP1, C/EBP2 and YY1 elements are specific for LPS using mutant constructs of À238AMTN, in which basal transcriptional activity of À238AMTN was higher than the basal activity of À116AMTN (statistically significant difference, P < 0.05), whereas LPS-induced À116AMTN and À238AMTN activities were almost the same, in spite of À238AMTN containing YY1 and C/EBP2 sites (Fig. 2A). The results of luciferase mutation assays using À238mC/EBP1, À238mC/EBP/2 and À238mYY1 suggest that these three response elements could be functional (Fig. 3). Further study is necessary to resolve the discrepancy.
Association with Smad2/3 signaling in aggressive periodontitis caused by A. actinomycetemcomitans infection was previously reported [40]. Furthermore, we have shown that TGF-b1-induced Amtn gene transcription was mediated through Smad3 signaling pathways [13]. Hence, we proposed that PgLPS increased Smad3 binding to C/EBPb and YY1, and thereby upregulated Amtn gene transcription. Smad3 is a critical transcription factor for the TGF-b1 signaling pathway in tumor pathogenesis and contributes to cell growth, apoptosis, invasion and metastasis [41]. Physical association of Smad3 and Smad4 with C/EBPb and C/EBPd provided repression of the function of EBPb and C/EBPd in inducing adipogenic differentiation of mesenchymal cells [36]. In this study, the complexes of C/EBPb and Smad3 were clearly detected by immunoprecipitation, whereas the amount of the complexes was not influenced by PgLPS (Fig. 9A). Interestingly, physical interaction between YY1 and Smad3 was dramatically induced by PgLPS (Fig. 9B). Considering to the results of C/EBPb, YY1, Smad3 and pSmad3 protein levels after stimulation by PgLPS ( Fig. 7 and Table 3), LPS-induced C/EBPb, YY1 and pSmad3 directly bind to C/EBP1, 2 and YY1 elements, respectively, mediating their physical interaction with Smad3. In breast cancer cells, overexpression of C/ EBPb increased the protein expression levels of TGF-b1 and pSmad3 and repressed Smad3 expression [42]. Additionally, growth-promoting activity of C/EBPb has been observed in mammary epithelial cells [43], and Smad3 is required for the inhibition of adipogenesis by retinoic acid, mediated through the interference of C/EBPb activity [44]. These reports may represent an interaction of C/EBPb with Smad3, or pSmad3 may be selectively and tightly regulated by cell types and numerous other biological processes. Thus, further research will address the mechanism and phosphorylation of Smad3 involved in the inflammation.
YY1 is a multifunctional zinc-finger transcription factor belonging to the Polycomb protein family. This family is a group of homeobox gene receptors that Fig. 9. Interaction of C/EBPb and Smad3 in the induction of Amtn gene expression by PgLPS. Smad3-precipitated C/EBPb was detected, which indicated physical interaction. C/EBPb and Smad3 complexes were not changed by PgLPS treatment, compared with controls. Immunoprecipitation of IgG was performed as negative controls. Also shown is interaction of YY1 and Smad3 in the induction of Amtn gene expression by PgLPS. Smad3-precipitated YY1 was detected, which indicated physical interaction. YY1 and Smad3 complexes were remarkably induced by PgLPS treatment. HDAC1 expression of flow through after immunoprecipitation did not change as controls. **Indicated non-specific bands.
regulate cell cycle and hematopoiesis, associated with cancer biology [31]. YY1 is tightly associated with cancer; however, a relationship between Amtn and tumors has been reported in a few studies. Amtn expression led to positive staining of eosinophilic material in samples of adenomatoid odontogenic tumor [37], and detected in keratocystic odontogenic tumor cells [45]. However, these studies did not mention a correlation between these tumors and YY1; therefore, further study is required to understand the relationship between YY1 and Amtn in initiation and metastasis of tumors.
In conclusion, this study showed that Complexes of PgLPS-induced C/EBP, YY1 and Smad3 are regulators of Amtn gene transcription via C/EBP1, C/EBP2 and YY1 elements in the mouse Amtn gene promoter. The C/EBPb, YY1 and Smad3 complexes might be crucial player in the defense system of gingival epithelium against periodontal pathogens and inflammation.