The Transcription Factor C/EBP-β Mediates Constitutive and LPS-Inducible Transcription of Murine SerpinB2

SerpinB2 or plasminogen activator inhibitor type 2 (PAI-2) is highly induced in macrophages in response to inflammatory stimuli and is linked to the modulation of innate immunity, macrophage survival, and inhibition of plasminogen activators. Lipopolysaccharide (LPS), a potent bacterial endotoxin, can induce SerpinB2 expression via the toll-like receptor 4 (TLR4) by ∼1000-fold over a period of 24 hrs in murine macrophages. To map the LPS-regulated SerpinB2 promoter regions, we transfected reporter constructs driven by the ∼5 kb 5'-flanking region of the murine SerpinB2 gene and several deletion mutants into murine macrophages. In addition, we compared the DNA sequence of the murine 5′ flanking sequence with the sequence of the human gene for homologous functional regulatory elements and identified several regulatory cis-acting elements in the human SERPINB2 promoter conserved in the mouse. Mutation analyses revealed that a CCAAT enhancer binding (C/EBP) element, a cyclic AMP response element (CRE) and two activator protein 1 (AP-1) response elements in the murine SerpinB2 proximal promoter are essential for optimal LPS-inducibility. Electrophoretic mobility shift (EMSA) and chromatin immunoprecipitation (ChIP) assays demonstrated that LPS induces the formation of C/EBP-β containing complexes with the SerpinB2 promoter. Importantly, both constitutive and LPS-induced SerpinB2 expression was severely abrogated in C/EBP-β-null mouse embryonic fibroblasts (MEFs) and primary C/EBP-β-deficient peritoneal macrophages. Together, these data provide new insight into C/EBP-β-dependent regulation of inflammation-associated SerpinB2 expression.


Introduction
The inflammatory response is a double-edged sword. Properly orchestrated, it results in the clearing of foreign molecules and invading pathogens from the body. Uncontrolled, it may lead to organ damage, sepsis, and even cancer [1][2][3]. Many of the pathological manifestations of the inflammatory response are mediated by cytokines and other inducible gene products expressed by macrophages upon exposure to the gram-negative bacterial cell wall component LPS. As macrophages are key effectors of pathogen-induced innate immune responses, their survival is critical for initial pathogen neutralization and subsequent development of adaptive immune responses. One of the most LPS-inducible macrophage gene products known is the ovalbumin-like serine protease inhibitor (ov-serpin) SerpinB2, a widely recognized macrophage survival factor [4;5]. SerpinB2 was first identified as an inhibitor of urokinase-type plasminogen activator (uPA) [6][7][8], a serine protease involved in the degradation and turnover of the extracellular matrix through the activation of plasminogen [7;9]. Such function requires SerpinB2 to be secreted from the cell yet SerpinB2 exists primarily as a nonglycosylated intracellular protein [10]. Over the past decade, intracellular roles for SerpinB2 in cell survival [11][12][13][14][15][16][17], proliferation and differentiation [18][19][20][21], signal transduction [15;22;23] and innate immunity [24][25][26][27][28], have been described.
The SerpinB2 gene is highly regulated in a cell type specific manner analogous to that of cytokines and oncogenes [29;30]. It is one of the most responsive genes known [31], and can be induced over 1000-fold by LPS [31][32], and is up-regulated by a range of inflammatory mediators [9]. LPS activates immune responses through multiple signalling pathways. The toll-like receptor 4 (TLR4) is responsible for the recognition of LPS and other microbial products and plays a central role in the initiation of innate immune responses, including cytokine release. The binding of LPS to TLR4 on the surface of macrophages leads to the recruitment of adaptor molecules and the activation of protein kinases, generating signals to the nuclear factor-kB (NF-kB), mitogen-activated protein kinase (MAPK) and/or phosphoinositide 3(PI3)-kinase pathways [33].
In studies aimed at identifying LPS-inducible pro-survival factors downstream of p38 MAPK, SerpinB2 was identified as a factor whose expression was upregulated by cooperation of the IKKb/NF-kB and p38 MAPK/CREB pathways [16]. Our previously published data indicated that SerpinB2 is distinctly regulated from other LPS-inducible genes in terms of kinetics, LPS dose response and sensitivity to IFN-c co-stimulation [4]; however, the cis-acting elements in the SerpinB2 promoter responsible for LPS-dependent transcription in macrophages and the specific LPS-responsive transcription factors that bind the SerpinB2 promoter were not defined. Here we show that LPS induction of SerpinB2 is dependent upon cis-acting regulatory sequences in the region between nucleotides 2189 and 2539 of the murine SerpinB2 promoter, and is critically dependent upon a C/EBP binding site at 2203/2195. C/EBP-b directly bound to this site in vivo and its deficiency abrogated constitutive SerpinB2 expression and SerpinB2 induction by LPS. Importantly, a C/EBP-b phosphoacceptor site was found to negatively regulate LPS-induced SerpinB2 promoter activity. Together, these findings provide new insight into the transcriptional regulation of the SerpinB2 gene.

DNA Sequence Analysis
The DNA sequence of the murine SerpinB2 promoter was determined by sequencing the pUC-based plasmids pDB9406, pDB9402-41 and pDB9402-42, in addition to plasmids prepared from pDB9402-41 and pDB9402-42 containing deletions introduced by exonuclease III digestion. Plasmids pDB9406, pDB9402-41 and pDB9402-42 containing genomic DNA isolated from a lFIXII (Stratagene) genomic library prepared from a 129 mouse strain, were kindly provided by Dr. Dominique Belin, University of Geneva. pDB9406 contains a 4.4 kb EcoRI/SpeI genomic fragment spanning the transcription initiation site; pDB9402-41 and pDB9402-42 contain a 1.2 kb EcoRI genomic fragment, located immediately upstream of the pDB9406 EcoRI fragment, cloned in opposite orientations. Subcloned inserts were verified by restriction enzyme digestion and DNA sequencing. The nucleotide sequence of the 4480 bp murine SerpinB2 gene 59 flanking region was determined using the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin-Elmer) and a PE 373A sequencer (Perkin-Elmer). This sequence was deposited in GenBank/EMBL/DDBJ Data Bank with Accession No. AF339731.

Construction of SerpinB2 Reporter Gene Plasmids
The PCR primers DW5'LUC, containing a Kpn I restriction site, and DW3'LUC, containing an Xho I restriction site, were used to PCR amplify and clone the SerpinB2 promoter (23261 to +92) from pDB9406 into the Kpn I/Xho I polylinker restriction sites of pGL3 Basic (Promega) to produce pGLmP-3261. The EcoR I insert of pDB9402-42 was sub-cloned immediately upstream of the SerpinB2 promoter EcoR I site (23261) of pGLmP-3261 to produce pGLmP-4480. Additional murine SerpinB2 luciferase reporter constructs (pGLmP-2751, pGLmP-2614, pGLmP-1686, pGLmP-1341, pGLmP-694, pGLmP-539 and pGLmP-189) were generated by digesting pGLmP-3261 with EcoR I and a second restriction enzyme (Sfi I, Apa I, BstX I, Bsu36 I, Pac I, Hae III and Apo I respectively), blunt ending the resultant 59 or 39 overhangs with T4 DNA polymerase (NEB) and re-ligating the vector ends with T4 DNA ligase. The PCR primers BSPCR2 and DW3'LUC were used to subclone the murine SerpinB2 promoter regions 287 to +92 into the Kpn I/Xho I polylinker restriction sites of pGL3 Basic to produce pGLmP-87. The control empty vector was pGL3 Basic (Promega).
Transient Transfection and Luciferase Assays (RAW264.7 Macrophages) RAW 264.7 cells (2610 7 ) growing in log phase were transfected with the indicated luciferase reporter plasmid (20 mg) along with the pRL-thymidine kinase (TK) (Promega) internal control reporter plasmid (2 mg) by electroporation using a Bio-Rad Gene Pulser with a Capacitance Extender (0.25 kV, 960 mFd). pGL3 control plasmid which encodes the SV40 promoter and enhancer was included as a positive control for transfection efficiency, and as an internal standard for promoter and enhancer activities. Transfected cells were transferred to 10 ml of pre-warmed media in 6-well tissue culture plates, divided into two identical cell pools and incubated 16 hrs in a 5% CO 2 and 95% humidified air atmosphere at 37uC either in the presence or absence of 100 ng/ ml LPS. Luciferase activity was measured using a Dual-Luciferase Reporter Assay System kit (Promega). Measurements represent the results of at least three independent experiments. Promoter activity is expressed as the number of firefly luciferase light units normalized either to pRL-TK renilla luciferase light units or to cellular protein concentration (where co-transfected C/EBP-b or LPS affected pRL-TK activity). Protein concentration was determined using the Bio-Rad protein microassay reagent.

Lentiviral shRNAs, Packaging and Transduction
pLKO.1-puro lentiviral vectors carrying short hairpin RNAs (shRNA) specific for human and mouse cebpb were used in these studies. Because the human and mouse cebpb 39 untranslated regions are not identical, these species-specific shRNAs cannot knockdown expression of endogenous C/EBP-b when used on cells of the other species; therefore we used the human CEBPB shRNA as a control in these experiments as in [38]. To produce lentiviral particles, HEK-293T cells were transfected with a mixture of plasmids: each shRNA expression plasmid (1 mg), pCMV-DR8.2dvpr packaging plasmid (0.75 mg), and pCMV-VSV-G envelope plasmid (0.25 mg) using Lipofectamine 2000 reagent (Invitrogen). The lentiviral supernatant was collected 48 hrs after transfection, cleared by centrifugation at 2,000 g for 10 mins and passed through a 0.45 mm filter. The target cells were treated with the lentiviral supernatant and 8 mg/ml Polybrene (American Bioanalytical) for 24 hrs. The lentiviral supernatant was replaced with fresh growth media and incubated further for 72 hrs to allow for effective gene knockdown. C/EBP-b knockdown was confirmed by western blot analysis.

Transient Transfection and Luciferase Assays (MEF Cells)
Cebpb 2/2 MEFs (1610 5 ) [34] were transfected with the indicated luciferase reporter plasmid (400 ng) along with a bactin-b-galactosidase reporter plasmid (200 ng) by electroporation using the Invitrogen Neon TM system (1 pulse, 1350 V, 30 msec). Transfected cells were transferred to pre-warmed media in 24 well plates and incubated for 48 hrs prior to incubation with LPS (100 ng/ml) for 4 hrs where indicated. In some experiments, plasmids encoding C/EBP-b or C/EBP-b phospho-acceptor mutants (T 188 A, T 217 A, S 64 A) [38][39][40] or control vector were co-transfected (0.6-1.0 mg total DNA). Luciferase activity was determined and normalized to that of b-galactosidase [38] using the Luciferase Assay System and b-Galactosidase Enzyme Assay System Kits, respectively (Promega). Each experiment was repeated at least three times, and triplicate samples were employed for each sample. Expression of the C/EBP-b phospho-acceptor mutant proteins was checked for equal expression by western blot.

Chromatin Immunoprecipitation (ChIP) Assay
ChIP assays were performed using a commercially available Magna-ChIP TM kit (Millipore), as recommended by the manufacturer, with minor modifications. Briefly, after crosslinking the chromatin with 1% formaldehyde at room temperature for 10 min and neutralizing with glycine for 5 min at room temperature, cells were washed with cold PBS, scraped and collected on ice. Cells extracts were prepared using a commercially available kit (Millipore). Nuclear lysates were sonicated 5 times for 15 sec with 1 min intervals on ice using a Sonic Dismembrator (Fisher). An equal amount of chromatin was immunoprecipitated at 4uC overnight with at least 1 mg of the following antibodies: C/EBPb (sc-150X), p-C/EBP-b (T217) (sc-16993X), normal rabbit IgG (sc-2027)(Santa Cruz Biotechnologies) and RNA polymerase II (Clone CTD4H8)(Millipore). Immunoprecipitated products were collected after incubation with Protein G coated magnetic beads (Millipore). The beads were washed, the bound chromatin was eluted in ChIP Elution Buffer (Millipore) and the proteins were digested with Proteinase K for 2 hrs at 62uC. The DNA was then purified using the QIAquick PCR Purification Kit (Qiagen). DNA was amplified by semi-quantitative PCR or by qPCR using the SYBR green method and primers specific for the SerpinB2 proximal promoter: forward (2338/2315) 59

Statistical Analysis
Data are presented as mean 6 SEM per group. Results were analyzed using the analysis of variance (ANOVA) or Student's t test where relevant. P-values ,0.05 were considered significant.

Results
The SerpinB2 Gene is Highly Responsive to LPS When RAW264.7 macrophages were exposed to LPS, SerpinB2 mRNA was detectable as early as 30 min following LPS challenge, reaching maximal levels at 24 hrs (Fig. 1A). A similar strong induction of LPS-inducible SerpinB2 mRNA expression has been reported previously in murine peritoneal macrophages and human peripheral blood mononuclear cells [4;32]. LPS-induced SerpinB2 expression involves both an increase in gene transcription and stabilization of the mRNA [5;29;42-44]. SerpinB2 protein expression was also induced as has been reported in other cell types [31], and detectable after 8 hrs of LPS treatment (Fig. 1).

The SerpinB2 Proximal Promoter Confers LPS Responsiveness
To investigate cis-acting regulatory elements responsive to LPS in the 59 flanking region of the murine SerpinB2 gene, nucleotides 24480 to +92 and a series of generated deletion mutants of the 59 flanking region were cloned upstream of a promoter-less firefly luciferase reporter gene (pGL3 Basic) ( Fig. 2A). The reporter constructs were then transiently transfected into sub-confluent RAW264.7 macrophages and assayed for luciferase activity in the presence and absence of LPS or PMA, for comparison. PMAinduced SerpinB2 gene regulation has been extensively studied in human macrophage cell lines [45][46][47], and has been shown to occur through several proximal and distal AP-1 responsive elements [30;37;47-51]. As shown in Fig. 2B, the SerpinB2 59 flanking region from 24480 to +92 directs both PMA-and LPSinducible transcription, approximately 2-fold and 7-fold, respectively. Deletion of the murine SerpinB2 promoter from 21686 to 21341 increased LPS-inducibility to approximately 16-fold, indicating the presence of a silencer element in this region. Further deletion beyond 2539 abolished the LPS-response of the promoter, indicating the presence of an essential LPS response element in the region between 2539 and 2189; however, the response of the murine SerpinB2 promoter to PMA is less affected by this deletion. While deletion of the SerpinB2 promoter from 2189 to 287 eliminated the LPS response and marginally reduced the PMA response, the 287 murine SerpinB2 promoter construct was still partially responsive to PMA, indicating that cisacting elements mediating the response of the murine SerpinB2 promoter to PMA also lie downstream of nucleotide 287.

Sequence Conservation within the Human and Murine SerpinB2 Proximal Promoters
To look for potential cis-acting elements that might mediate transcription of the murine SerpinB2 gene and the response to LPS, we aligned the murine and human SerpinB2 59 flanking regions. We reasoned that the presence of evolutionarily conserved, potential transcription factor binding sites in this region might play a role in the regulation of SerpinB2 gene expression [52]. The presence of several repetitive sequence elements delineated five broadly homologous regions (A-E) between the human and murine promoters (Fig. 2C). The proximal promoter (Region E), which contains the essential LPS response element, exhibited the greatest homology. Further analysis of Region E revealed that several of the cis-acting regulatory elements defined in the human SerpinB2 proximal promoter are conserved in the murine SerpinB2 promoter (Fig. 3). Specifically, a TATA consensus sequence is  To investigate regulatory elements between 2539 and 2189 essential for LPS-inducible transcription, several candidate binding sites for transcription factors previously reported to mediate LPS-inducible transcription in other genes [57] were targeted by nucleotide substitution designed to disrupt transcription factor binding to the pGLmP-539 murine SerpinB2 promoter-luciferase reporter gene construct (Fig. 4A). As shown in Fig. 4B, mutation of the consensus E box (2538/2533), PU.1 (2412/2407), or the variant Oct-1 (2296/2288) site did not decrease LPS-induced promoter activity. In contrast, mutation of the C/EBP site (2203/ 2192) completely eliminated promoter activity, similar to the levels observed for the 2189 SerpinB2 promoter deletion construct.  Others have also recently implicated this C/EBP site in LPSinduced activation of the SerpinB2 gene [58].
Considering the conserved sequence and position of the CRE and two AP-1 sites in the murine SerpinB2 promoter and their demonstrated involvement in the PMA-responsiveness of the human SERPINB2 promoter [37], these sites were also mutated by nucleotide substitution to investigate whether they played a role in the LPS response of the murine SerpinB2 promoter. As shown in Fig. 4B, mutation of the CRE at 2177/2172 or either of the two AP-1 sites at 2106/2100 and 294/288, completely or significantly reduced LPS-inducible luciferase activity from the murine SerpinB2 promoter. These data show that the C/EBP element, as well as the CRE and both AP-1 cis-acting elements are critical for LPS-inducible transcription from the SerpinB2 proximal promoter.

The C/EBP Element is Bound by an LPS-induced Nuclear Factor from RAW 264.7 Macrophages
Nuclear factor binding to the putative C/EBP element (2203/ 2192) was investigated by electrophoretic mobility shift assay (EMSA) using nuclear extracts from untreated and LPS-treated RAW 264.7 macrophages. Three different double stranded oligonucleotide probes were used for EMSA, representing (1) the putative SerpinB2 C/EBP element (2203/2192), (2) a mutant SerpinB2 element containing the same mutation as in pGLmP-539mC/EBP and (3) the rat albumin promoter distal element 1 (DEI) region containing a high affinity C/EBP binding site. Four bands of different mobilities, representing DNA-nuclear protein complexes were detected (Fig. 5A). Three of these bands (I, II, III) represent complexes with single stranded DNA (Fig. 5A) while the uppermost (slowest migrating) complex was induced by LPS. The LPS-inducible complex was not detected using the mutant C/EBP oligonucleotide probe and could be abolished by an excess of unlabeled double-stranded oligonucleotide, carrying either the same sequence (Fig. 5B, lanes 4 and 5) or the sequence of a known C/EBP binding site from the rat albumin promoter (Fig. 5B, lanes  8 and 9), but not by the mutated oligonucleotide (Fig. 5B, lanes 6  and 7). Together these data indicate that the putative SerpinB2 C/ EBP site at 2203/2192 binds a LPS-inducible complex that is likely to contain a member of the C/EBP family of transcription factors.

C/EBP-b is a LPS-induced Nuclear Factor that Binds to the C/EBP Element of the SerpinB2 Proximal Promoter
The C/EBP family of basic leucine zipper transcription factors are known for their roles in cellular differentiation and inflammation  [59]. Consisting of six members, the C/EBP transcription factors can homo2/heterodimerize and display similar DNA binding specificities [60]. Four family members, C/EBP-a, C/EBP-b, C/ EBP-d and C/EBP-e, are present in myeloid cells and play different roles in differentiating myeloid cells depending on the extracellular environment [61]. To determine which C/EBP proteins were involved in the formation of the different nucleo-protein complexes, and particularly of the LPS-inducible complex, EMSA was performed after incubating the nucleo-protein complexes with antibodies specific for C/EBP-a, C/EBP-b, C/EBP-d and C/EBPe. Each antibody detects the carboxy-terminal DNA-binding region of the respective protein, so that pre-incubation of antibody with nuclear extract is expected to abolish DNA binding by EMSA [62].
Only antibodies against C/EBP-b abolished the formation of the LPS-inducible complex (Fig. 5C). Taken together these data show that the LPS-induced complex with the C/EBP element (2203/ 2192) contains C/EBP-b.

C/EBP-b Mediates both Constitutive and LPS-induced SerpinB2 mRNA Expression in MEFs and Inflammatory Primary Macrophages
To investigate the importance of C/EBP-b to endogenous SerpinB2 mRNA expression in response to LPS, we utilized C/ EBP-b-null (Cebpb 2/2 ) and wild-type MEFs (Cebpb +/+ ), since RAW264.7 cells constitutively express endogenous C/EBP-b. Wild-type MEFs express low levels of endogenous SerpinB2 and the absence of C/EBP-b attenuated endogenous SerpinB2 mRNA expression. LPS stimulated an increase in SerpinB2 mRNA expression in wild-type MEFs (Fig. 6A), whereas LPS-stimulated SerpinB2 mRNA expression was significantly dampened in Cebpb 2/ 2 MEFs as compared to wild-type. We next tested if similar effects could be seen in thioglycollate-elicited inflammatory macrophages in which C/EBP-b expression was knocked down using speciesspecific lentiviral shRNAs (Fig. 6B, left). As was observed in MEFs, both constitutive and LPS-induced SerpinB2 mRNA expression was significantly decreased in C/EBP-b-deficient inflammatory macrophages (Fig. 6B, right). These data show that C/EBP-b is critical for mediating constitutive and LPS-inducible transcription of endogenous SerpinB2 mRNA.

C/EBP-b Binds the Murine SerpinB2 Proximal Promoter in vivo in an LPS-inducible Manner
We investigated the temporal dynamics of C/EBP-b recruitment to the murine SerpinB2 promoter in response to LPS in vivo by chromatin immunoprecipitation (ChIP). RAW 264.7 cells were stimulated with LPS for up to 8 hrs, soluble chromatin was immunoprecipitated with antibodies against DNA binding proteins, and the enriched DNA amplified by both semiquantitative and qPCR using SerpinB2 proximal promoter specific primers (illustrated in fig. 4A). The results showed that C/EBP-b is constitutively present at the SerpinB2 promoter as demonstrated by its association with the promoter in unstimulated macrophages and increased temporally in response to LPS reaching as much as 10-fold over unstimulated cells after 8 hrs (Fig. 7). Since changes in C/EBP-b phosphorylation states can affect C/EBP-b's ability to transactivate target genes [38;63;64], we investigated recruitment to the SerpinB2 proximal promoter of the T 217 phosphorylated C/EBP-b isoform (p-C/EBP-b T217 ), which has been associated with cell survival [65]. p-C/EBP- b T217 was constitutively bound to the SerpinB2 proximal promoter and also present after 1 hr of LPS stimulation (Fig. 7). In contrast to total C/EBP-b, the binding affinity of p-C/EBP-b T217 for the SerpinB2 proximal promoter diminished with LPS stimulation at later timepoints (4 and 8 hrs) (Fig. 7). These data show an inverse relationship between C/EBP-b and p-C/EBP-b T217 recruitment, and indicate that T 217 -phosphorylated C/EBP-b may not be responsible for increased transcription from the SerpinB2 promoter in response to LPS.

C/EBP-b Promotes LPS-inducible Murine SerpinB2 Proximal Promoter Activity
Since C/EBP-b binds to the SerpinB2 proximal promoter in an LPS-inducible manner both in vitro and in vivo, we wanted to address the question of whether C/EBP-b was an essential factor for driving transcription from the SerpinB2 promoter in cells in response to LPS. The ability of endogenous C/EBP-b to direct transcription from the SerpinB2 proximal promoter was examined by transfection of the pGLmP-539 murine SerpinB2 luciferase reporter construct into Cebpb +/+ and Cebpb 2/2 MEFs. We found that LPS-stimulated SerpinB2 promoter activity was significantly increased in Cebpb +/+ MEFs and abrogated in Cebpb 2/2 MEFs (Fig. 8A), indicating that endogenous C/EBP-b is required for LPS-induced SerpinB2 proximal promoter activity.

Phosphorylation of C/EBP-b at Serine 64 negatively Regulates LPS-stimulated SerpinB2 Promoter Activity
Phosphorylation of C/EBP-b is well recognized to modulate its transactivation potential [63]. To investigate C/EBP-b phosphorylation sites that may be important for LPS-stimulated SerpinB2 proximal promoter activity (Fig. 8B), we co-expressed several C/ EBP-b phospho-acceptor mutants in which the critical threonine or serine residue was mutated to an alanine, along with the pGLmP-539 murine SerpinB2 luciferase reporter in Cebpb 2/2 Figure 6. C/EBP-b is essential for constitutive and LPS-induced SerpinB2 mRNA expression. (A) Endogenous SerpinB2 mRNA expression is abrogated in Cebpb 2/2 MEFs compared to Cebpb +/+ MEFs in the absence and presence of LPS. qPCR analysis of murine SerpinB2 mRNA expression in untreated Cebpb +/+ and Cebpb 2/2 MEFs, and after simulation with LPS (100 ng/ml) for the indicated times. (B) Endogenous SerpinB2 expression is abrogated in C/EBP-b-deficient inflammatory macrophages. Thioglycollate-elicited peritoneal macrophages (TG macs) were infected with human and murine specific lentiviral shRNAs. Human CEBPB shRNA serves as the non-silencing control since it does not target the murine Cebpb sequence [38]. Lentiviral transduced macrophages were stimulated with LPS (100 ng/ml) for 4 hrs. Left: Western blot analysis shows effective knockdown of endogenous C/EBP-b following infection with murine Cebpb shRNA and not human CEBPB shRNA. Right: qPCR analysis of murine SerpinB2 mRNA expression in the lentiviral transduced peritoneal macrophages. The results represent the mean and SEM of two independent experiments performed in duplicate or triplicate. (*, p,0.05, two-way ANOVA). doi:10.1371/journal.pone.0057855.g006 MEFs. Re-expression of wild-type C/EBP-b in Cebpb 2/2 MEFs significantly stimulated SerpinB2 luciferase reporter gene expression in the presence of LPS by ,3 fold (Fig. 8C, left), confirming the importance of C/EBP-b to LPS-induced SerpinB2 gene transcription. Expression of the C/EBP-b phospho-acceptor mutant, C/ EBPb T217A , in Cebpb 2/2 MEFs did not significantly increase SerpinB2 promoter activity above that of wild-type C/EBPb (Fig. 8C, right); confirming that phosphorylation of C/EBPb at T 217 is not a major factor in the regulation of SerpinB2 promoter activity in response to LPS.
C/EBP-b contains additional phosphorylation sites, C/EBPb T188 and C/EBP-b S64 (Fig. 8B), which may be involved in modulating C/EBP-b-dependent SerpinB2 gene transcription. C/ EBP-b T188 is implicated in regulating DAPK1, an IFNc-inducible gene involved in the regulation of cell cycle and apoptosis [38], processes with which SerpinB2 has also been associated [17]. C/ EBP-b S64 is important for LPS-induced transcription of the cytokines IL-6 and MCP-1 [64]. Similarly, SerpinB2 is induced by LPS and regulated in a manner similar to cytokines [29]. Given the similarities in the functional significance of these phosphospecific isoforms of C/EBP-b and SerpinB2, we investigated whether these phospho-acceptor sites may play a role in SerpinB2 gene expression. C/EBP-b T188A -transfected MEFs exhibited SerpinB2 promoter activity similar to that of wild-type C/EBP-btransfected MEFs, whereas the expression of C/EBP-b S64A potentiated SerpinB2 promoter activity in response to LPS (Fig. 8C, right). These data suggest that phosphorylation of C/ EBP-b at S 64 acts to negatively regulate SerpinB2 proximal promoter activity.

Discussion
Macrophages are key mediators of the innate immune response, and consequently provide the first line of defense against pathogens. Pro-inflammatory stimuli, such as the bacterial endotoxin LPS, stimulate macrophages to mount an anti-  pathogenic response which involves massive induction of the prosurvival factor SerpinB2. SerpinB2 is transcriptionally induced by cross talk between the IKKb/NF-kB and p38MAPK signaling modules in response to LPS [16]. Here we report that SerpinB2 gene transcription in response to LPS is conferred by the SerpinB2 proximal promoter and is greatly dependent upon C/EBP-b. LPSinduced C/EBP-b was shown to specifically bind the C/EBP response element in the SerpinB2 proximal promoter in vitro and in vivo, and loss of C/EBP-b abrogates constitutive SerpinB2 gene transcription and the response to LPS.
The murine SerpinB2 proximal promoter region between nucleotides -539 and +92 mediated both PMA-and LPS-inducible gene transcription, with induction by PMA being less intense and more transient than that by LPS. Inspection of the murine SerpinB2 proximal promoter sequence shows that a CRE and two AP-1-like elements, demonstrated to mediate PMA-stimulated transcription of the human SERPINB2 gene [37], also are present in the murine SerpinB2 proximal promoter between nucleotides 2189 and 287. These sites may therefore also play a role in mediating PMAinducible transcription of the murine SerpinB2 gene. In contrast to the pattern of incremental increases in PMA-induced transcriptional activity conferred by regions of the murine SerpinB2 promoter containing these sites, most of the LPS-inducible response is dependent upon cis-acting regulatory sequences in the region between nucleotides 2189 and 2539. LPS responsiveness absolutely required the C/EBP binding site located in the region between nucleotides 2189 and 2539, with the downstream CRE and AP-1-like elements also being critical. Of note, there are previous reports of combinatorial interactions between C/EBPb and CRE binding proteins (CREB) and AP-1 [63], and C/EBPb has been reported to physically interact with AP-1, and NF-kB to promote gene expression of inflammatory mediators [63;66]. Additionally, CREB has been shown to control transcription of the C/EBP-b gene [67].
In this study, C/EBP-b was found to be a major requirement for both constitutive and LPS-induced SerpinB2 gene transcription. In a previous microarray expression profiling study, SerpinB2 was identified as gene whose induction in C/EBP-b-deficient peritoneal macrophages by LPS and IFNc was severely impaired compared to wild-type macrophages [68]. Our qPCR results validate this finding, as we found that both constitutive and LPSinducible SerpinB2 mRNA expression is significantly abrogated in C/EBP-b-shRNA transduced peritoneal macrophages, emphasizing the link between C/EBP-b and SerpinB2 gene transcription. While C/EBP proteins can act as either homodimers or heterodimers [63], we identified C/EBP-b as the only LPSinducible C/EBP isoform to bind the SerpinB2 C/EBP response element, suggesting a predominant role for C/EBP-b in LPSinduced SerpinB2 gene expression. C/EBP-b, like SerpinB2, plays an important role in inflammation, as it is upregulated by LPS and a host of other inflammatory cytokines [59;62]. Furthermore, C/ EBP-b-null mice are susceptible to bacterial infection [69;70] and SerpinB2 has been demonstrated to protect from bacterial and viral-induced cell death [14-16;25]. Thus the regulation of SerpinB2 gene expression by C/EBP-b is consistent with its functional role in inflammation.
Phosphorylation of C/EBP-b at several different amino acid residues has been shown to modulate transactivation of its target genes [63;71;72]. C/EBP-b phosphorylated on T 217 has been reported to rescue macrophages from apoptosis induced by Bacillus anthracis lethal toxin (LT) [65], an activity that has also been attributed to SerpinB2 [16]. However, expression of a C/EBPb phospho-acceptor site mutant, C/EBPb T217A , in Cebpb 2/2 MEFs did not increase SerpinB2 proximal promoter activity over that of wild-type C/EBP-b, even though recruitment of C/EBPb T217 to the SerpinB2 promoter was observed to decrease following LPS stimulation of RAW264.7 cells. These data indicate that the T 217 phospho-acceptor site is not important for the regulation of SerpinB2 gene expression. Similarly, expression of the C/EBPb phospho-acceptor site mutant, C/EBP-b T188A , did not affect SerpinB2 promoter activity differently from that of wild-type. In contrast, expression of C/EBP-b S64A significantly enhanced LPSinduced SerpinB2 promoter activity, indicating that phosphorylation of C/EBP-b at S 64 negatively regulates LPS-induced SerpinB2 promoter activity. Since C/EBP-b S 64 is constitutively phosphorylated in both RAW264.7 cells and MEFs [73], it is likely that dephosphorylation at this site may be a critical event during LPSinduced transcription of SerpinB2 to increase its promoter activity. Roy and colleagues demonstrated that Mixed lineage kinase-3 (MLK3)-driven dephosphorylation of C/EBP-b S 64 was important for IFNc-regulated signaling pathways [73]. Our data suggest that MLK3-driven dephosphorylation of S 64 may also be involved in LPS-signaling pathways.
In recent years it has become apparent that persistent infection is integrally linked to chronic inflammation and cancer, and immune cells such as macrophages can either promote or attenuate cancer progression [2;74-76]. SerpinB2 expression has been associated with both inflammation and cancer, and is a favorable or unfavorable prognostic indicator depending on cancer type [77]. The presence of SerpinB2 has been shown to modulate cytokine profiles which can affect immune cell polarization [27;28;77;78]. Interestingly C/EBP-b has also been shown to modulate cytokine secretion from immune cells, thereby modifying their phenotype [27;79-81]. Our study has demonstrated that C/EBP-b plays an important role in mediating both constitutive and LPS-induced transcription of the SerpinB2 gene, which may have implications for the inflammatory phenotype of infiltrating immune cells in the tumor microenvironment.
In summary, our studies show that the C/EBP site (2203/ 2195) in the murine SerpinB2 proximal promoter is necessary to support both constitutive and LPS-induced SerpinB2 gene expression. Importantly, we were able to uncover a previously unknown role for C/EBP-b S64 in negatively regulating SerpinB2 promoter activity. Taken together these data provide new insight into the regulation of inflammation-associated SerpinB2 gene expression.