Lipopolysaccharide Activates the Expression of ISG15-specific Protease UBP43 via Interferon Regulatory Factor 3*

UBP43 is a protease that specifically removes a ubiquitin-like protein, ISG15, from its targets. Highest levels ofUBP43 expression are detected in macrophages and in cell lines of monocytic lineage. Macrophages are important in host defense against bacterial and viral infections. The lipopolysaccharide (LPS) of the bacterial cell wall can mimic bacteria and activate monocytes/macrophages to provoke inflammatory responses. Here, we report that LPS strongly activates UBP43 expression in macrophages, which is paralleled by changes in UBP43 protein levels. Two interferon regulatory factor (IRF) binding sites in theUBP43 promoter are responsible for the induction ofUBP43 expression by LPS, as well as for basalUBP43 promoter activity. We have identified two members of the IRF family (IRF-2 and IRF-3) that specifically bind to these sites. IRF-3 plays a primary role in the LPS-inducible activation of theUBP43 gene and IRF-2 confers a basal transcriptional activity to the UBP43 promoter. Furthermore, we demonstrate that LPS treatment increases the amount of ISG15-conjugates in macrophages. Coordinated induction of ISG15 andUBP43 suggests that ISG15 conjugation is a dynamic process and that a critical balance of ISG15-modification should be maintained during innate immune response.

Ubiquitin is the most familiar of the proteinaceous protein modifiers, and the enzymology of its activation and transfer to target molecules has been studied extensively. More recently, a sizeable group of ubiquitin-related proteins have come to light; at least a dozen distinct ubiquitin-like proteins (Ubls) 1 similarly ubiquitin form covalent attachments to other macromolecules (1). Ubls mediate an impressive range of cellular functions, including cell-cycle progression, DNA repair, and apoptosis, suggesting that covalent posttranslational modification of proteins is a versatile principle of determining the half-life, intracellular localization, and activity of proteins (reviewed in Ref. 2). Ubiquitin cross-reactive protein, better known as IFN-stimulated gene 15 (ISG15), was the first example of a growing class of ubiquitin-like proteins that includes SUMO-1, Nedd8, and FAT 10 (3). ISG15 is one of the most strongly induced genes after interferon (IFN) treatment (4 -6) and is also significantly induced by influenza B virus (7), lipopolysaccharide (LPS) (8), and genotoxic stress (9). It functions intracellularly as a ubiquitin homolog and can form conjugates with certain cellular proteins, a substantial amount of which are colocalized with intermediate filaments of the cytoskeleton (10). Conjugation of Ubls including ISG15 occurs by a mechanism similar but distinct from ubiquitination (11). It involves a three-step mechanism where specific enzymes (or enzyme complexes) activate and covalently link Ubls to their substrates (12,13). An ISG15-activating enzyme has been recently identified as Ube1L (7). Interestingly, UBE1L was found to be absent in 14 different lung cancer cell lines tested, suggesting a possible link between block of ISG15 conjugation and carcinogenesis (14). Monocytes and lymphocytes can release free ISG15 (15). Cytokine-like properties of ISG15, such as the induction of IFN-␥ production and augmentation of natural killer/lymphokine-activated killer cell proliferation and function (16), suggest an important role of ISG15 in immunomodulation.
Modification of proteins by ubiquitin and Ubls is reversible. Ubiquitin (or Ubl)-substrate deconjugation is performed by members of a diverse group of specialized cysteine proteases called deubiquitinating enzymes or ubiquitin-specific proteases.
UBP43 (USP18), a member of the ubiquitin protease family, has been cloned in our laboratory during the analysis of differential gene expression in hematopoietic tissues of AML1-ETO knock-in mice (17). UBP43 encodes a 43-kDa protein and exhibits homology to catalytic domains of ubiquitin-specific proteases (USPs) that function to release free ubiquitin from ubiquitin-protein conjugates. Recently, we have demonstrated that UBP43 is a major ISG15-specific protease and activity of this enzyme is crucial for maintaining a proper balance of ISG15conjugated proteins in cells (18).
Vertebrates and invertebrates respond to bacterial invasion by activation of a defense mechanism that is part of the innate immune response (19). This response is mainly triggered by the recognition of LPS, which are cell wall components of Gramnegative bacteria (20). In mammals, it is primarily monocytes and macrophages that respond to LPS by releasing cytokines and chemokines to provoke inflammatory responses (21). After exposure to LPS, the macrophages undergo profound changes in protein composition that include alteration of cell surface, secreted, and intracellular products. The changes in LPS-stim-* This work was supported by National Institutes of Health Grant CA79849 and American Cancer Society Grant LBC-99438. The Departmental Molecular Biology Service Laboratory for DNA Sequencing and Oligonucleotide Synthesis was supported in part by the Stein Endowment Fund. This is manuscript 14173-MEM from the Scripps Research Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF388669.
We previously reported that the highest level of murine UBP43 expression was detected in thymus and peritoneal macrophages of normal adult mice. Among various hematopoietic cell lines tested, monocyte/macrophage lines also exhibited the highest level of UBP43 expression (17). Macrophages are primary effector cells in host defense, and high activity of the UBP43 may play an important role in their function as well as have an affect on overall immune response. These facts directed us to identify transacting factors regulating the expression of this gene under normal and stressed conditions. In this report we show that UBP43 is strongly up-regulated by LPS. Two interferon regulatory factor (IRF) binding sites in the UBP43 promoter are important for the basal and LPS-induced levels of UBP43 expression. We show that IRF-3 (a transcription factor known to be involved in regulation of defensive responses) is responsible for LPS-induction of UBP43, whereas IRF-2 mediates the basal level of expression.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture-The murine macrophage-like cell line, RAW 264.7, was generously provided by Dr. M. Ostrowski (Ohio State University, Columbus, OH) and was cultured in RPMI 1640 (Invitrogen) with 5% iron-supplemented bovine calf serum (HyClone, Logan, UT) and 2 mM L-glutamine (Invitrogen) at 37°C with 7% CO 2. RAW 264.7 cells were maintained in cell culture between 1 ϫ 10 5 and 1 ϫ 10 6 cells/ml. IRF-1/IRF-2 double knock-out murine embryonic fibroblasts (MEFs) were generously provided by Dr. Janet Stein (University of Massachusetts Medical School, Worcester, MA) with the permission from Dr. Tadatsugu Taniguchi and were maintained in Dulbecco's modified Eagle's medium (Invitrogen) with 10% fetal bovine serum and 2 mM L-glutamine. LPS (Escherichia coli serotype O127:B8) was purchased from Sigma.
Northern Blot Analysis-Total RNA from mouse brain and thymus tissues was prepared by the guanidine isothiocyanate extraction followed by cesium chloride gradient purification (22). Total RNA from peritoneal macrophages and RAW 264.7 cells was isolated using RNazol B reagent according to the manufacturer's instructions (Tel-Test Inc., Friendswood, TX). Ten g of total RNA from each mouse tissue or time point was separated in an agarose/formaldehyde (0.22 M) gel, blotted on Hybond N ϩ membrane (Amersham Biosciences), and probed with either a full-length UBP43 cDNA (17) or a full-length ISG15 cDNA (Gen-Bank accession no. U58202) that has been amplified by PCR from mouse cytomegalovirus-infected cells.
Immunoblotting-Rabbit polyclonal IgGs against human ISG15 were kindly provided by Dr. E. Borden (Cleveland Clinic Foundation) and were used at final concentration of 0.5 g/ml (23). The production of anti-UBP43 antibodies has been previously described (18). For Western blotting, anti-UBP43 antibodies were used at a final concentration of 0.2 g/ml. Cell lysates were resolved by SDS-PAGE, transferred to nitrocellulose membrane, and immunoblotted using standard procedures (24). Generation of UBP43 knockout (UBP43Ϫ/Ϫ) mouse model in which UBP43 gene has been deleted by homologous recombination will be published elsewhere. 2 Absence of UBP43 expression in UBP43Ϫ/Ϫ mice was confirmed by Western blot analysis using anti-UBP43 antibodies as described above.
RNase Protection Assay-RNase-protection assay was performed with a Riboquant kit (BD PharMingen, San Diego, CA) according to the manufacturer's instructions. The labeled 284-nucleotide riboprobe extended from Ϫ197 to ϩ71 of the murine UBP43 upstream region plus 16 nucleotides transcribed from pBluescript II KS(Ϫ) (Stratagene, La Jolla, CA). Ten g of control yeast tRNA or 10 g of total RNA isolated from mouse thymus or brain were used in each reaction.
Reporter Plasmids-The murine UBP43 promoter was isolated from 129SV murine genomic DNA library (Stratagene). The isolated 3.5-kb promoter fragment including a part of the first exon was cloned into pBluescript II KS(Ϫ) and sequenced. The promoter was then subcloned into the luciferase reporter plasmid pXP2 (25), resulting in p3KUBP43luc. To generate a series of 5Ј-region deletions, p3KUBP43-luc was digested with HindIII/SpeI, HindIII/NheI, or HindIII/HindIII, blunt ended with T4 DNA polymerase, and re-ligated. The resulting plasmids (p1.5KUBP43-luc, p0.7KUBP43-luc, and p0.1KUBP43-luc) were named by indicating the length of 5Ј-flanking region in each construct, respectively. To create p0.2KUBP43-luc, 240 bp of the UBP43 promoter was amplified by PCR using 5Ј-gtgtcctggtctagacgactggactg-3Ј and 5Ј-gcgaagaccgagctccatctgcaaag-3Ј as the upstream and downstream primers, respectively. The PCR product was sequenced and then inserted into pXP2. Mutant constructs p0.7KUBP43 (IRFE1m)-luc, p0.7KUBP43 (IRFE2m)-luc, and p0.7KUBP43 (IRFE1/2m)-luc were created by PCRbased mutagenesis with oligonucleotide pairs of an upstream primer (5Ј-acatctgtaaggatccagcaagcattt-3Ј) and one of three downstream primers (5Ј-gtccaagcttaagttttcc-3Ј (IRFE1m), 5Ј-gtccaagctttcgttttcccctagatccaaagggcagcgagactcaggc-3Ј (IRFE2m), or 5Ј-gtccaagcttaagttttcccctagatccaaagggcagcgagactcaggc-3Ј (IRFE1/2m)) using the wild-type promoter as a template. They were then sequenced and inserted into the BamHI/HindIII sites of p0.1KUBP43-luc. Expression constructs for IRF-1 (pCMVIRF1) and IRF-2 (pCMVIRF2) were generously provided by Transient Transfections-Transfection of RAW 264.7 cells was performed by electroporation (260 V, 975 microfarads) using a Gene Pulser II (Bio-Rad) equipped with a capacitance extender. UBP43-luc constructs (1 pmol/transfection) and an internal control for transfection efficiency, the promoterless Renilla luciferase expression construct, pRL-null (0.03 pmol/transfection), were co-transfected into 2.5 ϫ 10 6 cells in 0.2 ml of complete RPMI in a 0.4-cm cuvette (Bio-Rad). The total amount of DNA was adjusted to 7 g with pBluescript II KS(Ϫ). Cells from three electroporations were pooled together to eliminate differences between individual transfections. The mixture was then equally divided into three wells of a six-well plate (Corning Inc., Corning, NY). They were next allowed to adhere for 2 h before the medium was changed. Cells were harvested 48 h after electroporation and assayed for luciferase activity. For LPS treatment, cells were cultured in fresh medium for 36 h after electroporation. LPS (1 g/ml) was added for 7 h, and cells were then lysed and assayed for firefly and Renilla luciferase activities with the Dual Luciferase assay system (Promega, Madison, WI) using a Monolight 3010 luminometer (BD PharMingen). The firefly luciferase activity was normalized based on Renilla luciferase activity. All data were reported as a mean -fold induction, which was calculated by dividing the normalized reporter activity of each stimulated sample by that of the corresponding unstimulated control sample. From the mean values of three independent experiments, the overall (average) mean and its standard deviation were presented. When the effect of IRF-1 and IRF-2 on UBP43 promoter activity was studied, 0.5 g of the respective expression construct was co-transfected with 2 g of either p0.7KUBP43-luc or p0.7KUBP43 (IRFE1/2m)-luc into IRF-1 Ϫ/Ϫ IRF-2 Ϫ/Ϫ double knock-out MEFs. Transfection of MEFs was performed using Superfect reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. Dose-dependent inhibitory action of dominantnegative IRF-3 mutant was analyzed in RAW264.7 cells by co-transfection of wild-type (p0.7KUBP43-luc) or mutated version (p0.7KUBP43(IRFE1/2m)-luc) of UBP43 promoter (1 pmol/transfection) and dominant-negative mutant of IRF-3 (⌬nIRF-3) (0.3, 0.5, and 1 pmol/transfection, respectively).
Electrophoretic Mobility Shift Assay (EMSA)-Nuclear extracts were prepared by a previously described method (26) with minor modifications. Nuclear proteins were extracted from unstimulated cells and from cells stimulated with 1 g/ml LPS for seven h. Cells were washed in cold phosphate-buffered saline and pelleted. Pellets from 1-5 ϫ 10 7 cells were resuspended in 400 l of Buffer A (10 mM Hepes, pH 7.9, 10 mM KCI, 10 mM NaF, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2 mM benzamidine, 2 g/ml leupeptin, 2 g/ml antipain, 2 g/ml chymostatin). After incubation on ice for 5-15 min, 25 l of 10% Nonidet P-40 was added and lysates were vortexed for 10 min. Pelleted nuclei were resuspended in 150 l of Buffer B (10 mM Hepes, pH 7.9, 400 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 1 mM EGTA, and protease inhibitors as in Buffer A) and incubated for 15 min on ice with occasional shaking. The nuclear lysates were cleared by centrifugation, frozen in aliquots in liquid nitrogen, and stored at Ϫ80°C. The double-stranded oligonucleotides were end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase and purified on polyacrylamide gels. EMSA reactions were performed in 20 l of EMSA buffer (10 mM Hepes, pH 7.9, 100 mM NaCl, 1 mM dithiothreitol, 0.1 mM EDTA, 0.1 mM EGTA, 5% glycerol, and 100 ng/l poly d(I-C). Ten g of nuclear extracts were incubated in EMSA buffer with 1 l of labeled oligonucleotide (5000 -10,000 cpm; 5-10 fmol) for 20 min at room temperature. In competition analysis, 1 l of unlabeled competitors (1 pmol) were added to the reaction mixtures. For supershift analysis, 2 g of respective antibodies (anti-IRF-l (M-20), anti-IRF-2 (C-19), anti-p48/ ISGF3␥ (C-20) (Santa Cruz Biotechnology, Santa Cruz, CA) or anti-IRF-3 (Ref. 27)) were incubated with the extract in EMSA buffer for 10 min at room temperature before the labeled oligonucleotide was added. The EMSA reactions were separated on 7% or 4% polyacrylamide gels in 0.5ϫ TBE (45 mM Tris borate, 1 mM EDTA, pH 8.0) for 2 h at 50 V. The gels were vacuum dried, and the signals were detected by autoradiography.

RESULTS
LPS Activates UBP43 Expression-To investigate the transcriptional regulation of UBP43 in macrophages, we characterized the effect of LPS on UBP43 expression. As shown by Northern blot analysis (Fig. 1A), the level of UBP43 mRNA was significantly increased in thymi and macrophages of LPS-challenged mice. When RAW 264.7 macrophage-like cells were stimulated with LPS, an increase of UBP43 mRNA was detectable 2 h after stimulation, reaching a maximum expression at ϳ10 h (Fig. 1B). As expected, peritoneal macrophages as well as RAW 264.7 cells showed a significant up-regulation of ISG15 mRNA in response to LPS (Fig. 1, A and B). As indicated on the Western blot (Fig. 1C), up-regulation of UBP43 expression on the transcriptional level was paralleled by changes in the level of UBP43 protein in LPS-stimulated RAW 264.7 cells. Unstimulated RAW 264.7 cells showed very low level of UBP43 protein; however, LPS treatment led to a significant increase in the production of UBP43 protein. These results indicate that LPS signaling strongly increases the expression of UBP43.
Effects of UBP43 Expression on the Level of Intracellular ISG15 Conjugates-To identify biochemical changes related to UBP43 expression in control and LPS-stimulated macrophages, we analyzed the overall protein ISG15ylation status in extracts from wild-type (UBP43 ϩ/ϩ ) and UBP43-deficient (UBP43 Ϫ/Ϫ ) cells. Although ISG15 conjugates were undetectable in normal (wild-type macrophages, not treated with LPS) peritoneal exudate macrophages, LPS treatment resulted in a substantial increase of intracellular ISG15 conjugates (Fig. 2, first and third lanes from left). Significantly, the absence of UBP43 in UBP43-deficient macrophages led to an increase in total cell protein ISG15ylation even without LPS treatment (Fig. 2, second lane from left). Upon LPS stimulation, UBP43 Ϫ/Ϫ macrophages showed substantially higher levels of conjugates when compared with that of wild-type cells (Fig. 2, fourth lane from left). These data suggest that UBP43 is essential in the regulation of the level of ISG15 conjugates in control and LPS-treated macrophages.
Identification of the Transcription Initiation Site for the UBP43 Gene-To further study the molecular mechanism of UBP43 activation by LPS, a 3.5-kb fragment containing the 5Ј-flanking region and part of the first exon of UBP43 was isolated by screening a murine genomic library using UBP43 cDNA as a probe. The sequence of this 3.5 kb has been deposited in GenBank under accession number: AF388669. To identify the transcription initiation site of UBP43, RNase protection assays were performed. The riboprobe was 284 nucleotides long and contained 268 nucleotides that correlated to UBP43 DNA sequence from bp Ϫ197 to ϩ71. When this probe was hybridized to total RNA prepared from the thymus of LPS-treated mice and digested with ribonucleases, it generated a major band and three minor bands. Such protected bands were not detectable when brain RNA or yeast tRNA were used as negative controls (Fig. 3). With the calculation based on the molecular weight marker, the results demonstrated that the major UBP43 transcription initiation site is 111 bp upstream of the 3Ј-end of exon 1 (Fig. 4A). No TATA box was identified around the transcription initiation site of the UBP43 gene. However, there are two GC box consensus sequences in the nearest upstream region.
Identification of LPS Response Region in the UBP43 Promoter-To determine whether the isolated 3.5 kb upstream sequence of the UBP43 gene confers inducibility by LPS, we made a luciferase-reporter construct, in which the 3.5-kb upstream sequence of UBP43 was inserted into promoterless luciferase reporter, pXP2, to form p3KUBP43-luc. Promoter activity was readily detectable when p3KUBP43-luc was transfected into RAW 264.7 cells. The luciferase activity of p3KUBP43-luc was ϳ200-fold greater than the promoterless pXP2 construct (data not shown), indicating a strong promoter activity of this fragment. Upon LPS treatment, p3KUBP43-luc showed a 5-fold increase in promoter activity (Fig. 4B). To identify the region in the UBP43 promoter that is critical for LPS response, a series of deletion constructs were created and their activities were examined using transient transfection assays in RAW264.7 cells. Deletion of the UBP43 promoter to Ϫ200 bp (p0.2KUBP43-luc) did not significantly reduce LPS-induced UBP43 promoter activity. A further deletion of the promoter to Ϫ100 bp (p0.1KUBP43-luc) completely abolished LPS induction of the reporter gene (Fig. 4B). These results indicate that the LPS response element of the UBP43 promoter is located between bp Ϫ200 and bp Ϫ100.
Two Sequences Homologous to Interferon Regulatory Factor Binding Element (IRFE) Are Important for Constitutive and LPS-inducible UBP43 Promoter Activity-To identify the regulatory elements in the promoter of UBP43 responsible for LPS induction, we first examined the proximal 200-bp sequence of the UBP43 promoter using the TRANSFAC database (transfac.gbf.de/TRANSFAC/). Two putative IRF binding sites were identified. They were designated as IRFE-1 and IRFE-2 (Fig.  4A). Both sites were located between the bp Ϫ95 to Ϫ130 region of the UBP43 promoter. The sequence of the IRFE-1 site (bp Ϫ95 to Ϫ104) closely resembles the IRFE (28). The second site IRFE-2 (bp Ϫ118 to Ϫ130) is closely related to the interferonstimulated response element (29). To delineate whether the IRFE-1 and IRFE-2 sites were important for regulation of the UBP43 promoter in RAW 264.7 cells, point mutations were introduced at either the IRFE-1 or IRFE-2 site, or at both sites of the UBP43 promoter-luciferase reporter gene construct (p0.7KUBP43-luc) to disrupt the consensus binding sites of the respective elements (Fig. 5A). Transfection analysis of IRFEmutated promoter constructs into RAW 264.7 cells demonstrated that these mutations affected the UBP43 basal promoter activity (Fig. 5B). Mutation at either one of the two IRFEs alone decreased the promoter activity to 50% of the activity in the control. Mutation at both sites together decreased the promoter activity to 12% of the control level, indicating that these two IRFEs are important for the basal UBP43 promoter activity. We also studied the effect of these mutations on LPS-induced UBP43 promoter activity. Loss of IRFE-1 in the UBP43 promoter (p0.7KUBP43(IRFE1m)-luc) caused only a slight decrease of LPS induced promoter activation, whereas mutation of the IRFE-2 site (p0.7KUBP43(IRFE2m)-luc) decreased promoter inducibility to 48% of the control (Fig. 5C). When both sites were mutated simultaneously (p0.7KUBP43 (IRFE1/2m)-luc), it essentially abolished the response to LPS, reducing the inducible promoter activity to the level of the control vector alone (pXP2). These data indicate that an intact IRFE region is required for LPS induction of UBP43 gene expression as well as for its basal promoter activity in RAW 264.7 cells. Furthermore, the IRFE-2 site is more capable of mediating LPS inducibility of the UBP43 promoter. Nevertheless, both cis-acting elements were required to provide optimal responsiveness to LPS.

IRF-2 Is the Major Constitutive Binding Protein on UBP43 IRFEs-
The transfection experiments showed that IRFE-1 and IRFE-2 are functional elements for both constitutive and LPSinduced UBP43 promoter activity in macrophages. To identify the transcription factors interacting with these IRFEs, doublestranded oligonucleotides encompassing IRFE-1 and IRFE-2 sites were synthesized as shown in Fig. 5A. When incubated with nuclear extracts from RAW 264.7 cells treated with LPS, both IRFE probes displayed similar binding patterns in EMSA (Fig. 6A). Nuclear extracts prepared from RAW 264.7 cells without LPS treatment showed identical binding patterns (data not shown). The major shifted band can be competed with unlabeled self-oligonucleotide, an oligonucleotide with another IRFE, and an oligonucleotide containing the consensus IRF binding site (ISG15/IRFE) from the ISG15 gene promoter (30). However, oligonucleotides with IRFE-1 and IRFE-2 mutations or containing a non-IRF related PU.1 transcription factor binding site were not able to efficiently compete for binding. Com-

FIG. 3. Identification of UBP43 transcription initiation site.
Ten g of total RNA prepared from brain or thymus of LPS-treated mice or 10 g of yeast tRNA were hybridized with a 32 P-labeled 284-nucleotide riboprobe. After treatment with ribonucleases A and T1, the protected products were separated on a 6% polyacrylamide sequencing gel. Brackets indicate the positions of the protected products. The major product is marked with an asterisk. An arrow shows the position of undigested riboprobe. Labeled HinfI-digested ⌽174 DNA was used as the molecular size marker. The length assigned to each protected band was estimated to be 8% larger than that predicted from DNA size markers, based on the relative mobility of the undigested riboprobe and a DNA marker of the same size.
plexes designated with asterisks (*) are shifted bands likely generated with either nonspecific complex or degraded proteins. We could not reproducibly observe these bands even in seemingly identical runs with the same nuclear protein preparation. These results (shown in Fig. 6A) suggested that a similar DNA-nuclear protein complex was formed independently of LPS treatment with both IRFE-1 and IRFE-2 oligonucleotides and the protein in the complex was probably an IRF family member. To identify this protein, specific antibodies against either IRF-1 or IRF-2 were used in supershift assays. As shown in Fig. 6B, the majority of IRF complex was abrogated by IRF-2 antibody. The addition of IRF-1 antibody did not affect the complexes formed with IRFE-1 or IRFE-2 probes. These data demonstrated that IRF-2 was the major transcription factor constitutively bound to a critical region of the UBP43 promoter.
IRF-2 Is a Positive Regulator of the UBP43 Promoter-The data presented above demonstrate the ability of IRF-2 to bind both IRFEs of UBP43 promoter in the presence or absence of LPS stimulation in RAW 264.7 cells. This leads to a hypothesis that IRF-2 directs constitutive expression of the UBP43 gene, a role opposite to the transcriptional repression activity associated with this factor (31,32). To directly test this hypothesis, transactivation experiments using IRF-1 and IRF-2 expression plasmids and a UBP43 promoter-luciferase construct were performed. Because RAW 264.7 cells contain a high level of endogenous IRF-2, mouse embryonic fibroblasts with an IRF-1 and IRF-2 double knockout were used in the transactivation assays. As shown in Fig. 7, both IRF-1 and IRF-2 activated the UBP43 promoter specifically via the IRF binding site because the UBP43 promoter with IRFE-1 and IRFE-2 mutations did not show any significant activation. Furthermore, IRF-2 is a stronger activator than IRF-1 (8-fold versus 3-fold activation).

IRF-3 Plays an Important Role in LPS Induction of UBP43
Expression-We noticed an additional slower migrating band in EMSA when samples were electrophoresed for a longer time (Fig. 6B). To analyze this complex, nuclear extracts prepared from RAW 264.7 cell with or without LPS treatment were used in EMSA. The slower migrating band was only visible with nuclear extracts from LPS-treated cells (Fig. 8A) and was more abundant when IRFE-2 oligonucleotide was used as a probe. Furthermore, similar to the IRF-2 complex, it was specifically competed with unlabeled IRFE-1, IRFE-2, and ISG15/IRFE consensus oligonucleotides, but not by mutant oligonucleotides or PU.1 oligonucleotides. These EMSA results indicated that the slower migrating complex is specific and might recruit another IRF family member. Because the IRFE-2-mutant version of the UBP43 promoter (p0.7KUBP43(IRFE2 m)-luc) demonstrated a measurable difference in inducibility by LPS relative to the wild-type promoter (Fig. 5C), we suspected that the transcriptional factors involved in the formation of this complex were likely to mediate LPS induction of UBP43. Recently, a new member of the IRF family, IRF-3, has been identified (33). It has been shown that LPS stimulation is able to induce phosphorylation, nuclear translocation, and subsequent DNA binding of IRF-3 (27). To determine whether IRF-3 participates in the formation of LPS-inducible complex, we performed supershift experiments with antibodies against IRF-3. As shown in Fig. 8B, IRF-3 antibodies caused a specific supershift of the slower migrating complex, whereas no effect was observed on FIG. 4. Sequence and deletion analysis of UBP43 promoter. A, sequence (200 bp) of the UBP43 proximal promoter region and part of the 5Ј-untranslated region of the UBP43 gene is presented. The major transcription initiation site is marked with an arrow and numbered as ϩ1. Potential IRFE sites are shown in bold. GC boxes are underlined. The translation start codon of UBP43 is shown in italic bold. These sequence data are available from GenBank/EMBL/DDBJ under accession no. AF388669. B, LPS induction of UBP43 promoter. Schematic representations of UBP43 promoter-luciferase constructs are shown on the left. These constructs and the promoterless pXP2 vector were transfected into RAW 264.7 cells. The number in the name of each construct indicates the length of the 5Ј-flanking region of UBP43 included in the construct. Half of the transfected cells were treated with LPS for 7 h, and the other half were cultured under normal conditions. The cells were harvested and assayed for luciferase activity as described under "Experimental Procedures." The data are expressed as -fold increase in relative luciferase activity in LPS-stimulated cells over the untreated cells. The data represent the mean -fold of induction of three independent experiments Ϯ S.D. of the mean. the mobility of the IRF-2 complex. Neither p48/ISGF3␥ and Stat1 (data not shown), as the components of ISGF3 complex, nor IRF-1 and IRF-2 ( Fig. 6B) proteins were present in this complex. To test whether IRF-3 is functionally important for the induction of the UBP43 promoter by LPS, the level of UBP43 inducibility was analyzed in the presence of expression constructs of either an empty vector or a dominant negative form of IRF-3 (Fig. 8C). The expression of the dominant negative form of IRF-3 clearly decreased the induction of UBP43 promoter activity by LPS. These results indicate that IRF-3 is critical for the LPS induction of UBP43 expression. DISCUSSION In the present study, we assessed the transcriptional regulation of an ISG15 specific protease, UBP43, in macrophages. We demonstrate that LPS treatment stimulates the expression of UBP43. We also show that LPS increases the level of intracellular ISG15 conjugates. Massive accumulation of ISG15 conjugates observed in LPS-stimulated UBP43 Ϫ/Ϫ macrophages confers the crucial role of UBP43 in maintaining the proper balance of ISG15-conjugated proteins in cells. We describe the isolation and functional characterization of the promoter region of the UBP43 gene and the identification of sequence elements and trans-acting factors involved in the regulation of its expression by LPS.
The proximal UBP43 promoter does not contain classical initiation elements, such as a TATA box, CAAT box, or consensus initiation element (Inr). However, it possesses several GCrich motifs (Fig. 4A), which can functionally substitute a TATA box for directing transcription initiation with multiple transcription start sites (34 -36). Successive 5Ј end deletion of the UBP43 promoter led us to define a 100-bp region necessary and sufficient to promote maximal induction of transcription by LPS. This region contains tandem IRFEs, located 100 bp upstream of the transcription start site. We also found that these sites are indispensable for the basal transcription of the UBP43 gene because mutation of either IRFE decreases the level of UBP43 promoter activity in unstimulated cells.
The supershift results suggest that IRF-2 binds both IRFEs in unstimulated RAW 264.7 cells. Furthermore, IRF-2 has a higher potential to activate the UBP43 promoter compared with IRF-1. IRF-2 was originally regarded as a transcriptional repressor that antagonizes IRF-1 activity by competing for binding to the IRFEs of interferon and IFN-inducible genes (28,31,37). However, recent evidence indicates that IRF-2 is a dual-function transcription factor, as it activates the transcription of EBNA-1 (38), histone H4 (39, 40), VCAM-1 (41), gp91 phox (42), and CIITA (43). The latent transactivation domain located in the central region of IRF-2 possibly accounts for the trans- activating capability of IRF-2 (32). Our findings provide yet another example of IRF-2 being an activator. Binding of IRF-2 to UBP43 IRFEs was not affected by LPS treatment, suggesting therefore that IRF-2 confers a basal transcriptional activity to the UBP43 promoter.
As demonstrated in transfection experiments by mutating the IRFE bases generally known to be indispensable for the binding of IRF family members, an intact IRFE region is re-FIG. 6. Identification of IRF binding to the UBP43 promoter by EMSA. A, the same IRF family members interact with both IRFE-1 and IRFE-2 of the UBP43 promoter. The double-stranded UBP43 promoter bp Ϫ109 to Ϫ89 (IRFE-1) and bp Ϫ135 to Ϫ111 (IRFE-2) were 32 Plabeled and incubated with 1 g of double-stranded poly(dI-dC) in the absence and the presence of nuclear proteins prepared from LPStreated RAW264.7 cells (RAW264.7 NE). Unlabeled oligonucleotides of wild-type IRFE-1 (IRFE wt) and IRFE-2 (IRFE wt), mutant IRFE-1 (IRFE1 mut) and IRFE-2 (IRFE2 mut), IRF binding oligonucleotide from the ISG15 promoter (ISG15/IRFE), and an IRFE-unrelated PU.1 binding site containing oligonucleotide (PU.1) were added at a 100-fold molar excess over the probe oligonucleotide in competition assays. IRF points to the complexes formed between oligonucleotides and full-length IRF proteins. Asterisks (*) mark either the complexes formed between the oligonucleotide probe and degraded IRF protein or nonspecific complexes. B, IRF-2 is the major protein from RAW 264.7 nuclear extracts that binds to IRFE sites of UBP43 promoter. 32 P-Labeled IRFE1 and IRFE2 oligonucleotides were incubated with or without nuclear proteins prepared from LPS-treated RAW 264.7 cells. Two g of antibodies against either IRF-1 (␣-IRF1) or IRF-2 (␣-IRF2) were added to the reaction for supershift assays. The gel was electrophoresed longer than the gel presented in panels A and B. The asterisk (*) marks a newly detected complex that is discussed in Fig. 8.   FIG. 7. IRF-2 is positive regulator of UBP43 expression. Luciferase reporter gene constructs containing either wild-type (p0.7KUBP43-luc) or both IRFE sites mutated (p0.7KUBP43(IRFE1/ 2m)-luc) of UBP43 promoter was co-transfected with IRF-1, IRF-2, or empty vector pcDNA3 expression constructs into IRF-1 Ϫ / Ϫ IRF-2 Ϫ / Ϫ double knockout MEFs. Co-transfected Renilla luciferase construct was used to normalize the transfection efficiency. The data represent the mean -fold induction of three independent experiments. The error bars indicate the S.D. of the mean.
FIG. 8. IRF-3 mediates LPS induction of UBP43 promoter activity. A, the slower migrating complex is specifically induced upon LPS treatment. Nuclear extracts were prepared from RAW 264.7 cells with or without LPS stimulation (ϩLPS and ϪLPS, respectively). EMSA were performed using 32 P-labeled double-stranded oligonucleotides that correspond to bp Ϫ135 to Ϫ111 of IRFE-2 in the UBP43 promoter. The added competitors indicated on the top and other components of the reactions are as described in Fig. 6. The slower migrating band is marked with an asterisk. B, IRF-3 is involved in the formation of slower migrating complex. For supershift assays, nuclear proteins were pre-incubated for 15 min with 2 g of antibodies against either IRF-3 (␣-IRF3) or anti-p48/ISGF3␥ (␣-ISGF3␥) prior to the addition of 32 P-labeled double-stranded oligonucleotide. Arrows with IRF-2 and IRF-3 mark specific complexes formed between UBP43 IRFE-2 and IRF-2 or IRF-3 proteins, respectively. The arrow with SS marks the band supershifted by IRF-3 antibodies. C, dose-dependent inhibition of UBP43 promoter activation by expression of dominant-negative form of IRF-3. RAW 264.7 cells were co-transfected with luciferase reporter gene constructs containing either wild-type (p0.7KUBP43-luc) or mutated IRFE sites (p0.7KUBP43(IRFE1/2m)-luc) of UBP43 promoter and empty pcDNA3 vector or dominant-negative mutant of IRF-3 (⌬nIRF-3). Triangle indicates an increasing amount of ⌬nIRF-3 used in transfection (described under "Experimental Procedures"). LPS was applied 48 h after transfection for 7 h, and then cells were harvested and assayed for luciferase activity. A co-transfected Renilla luciferase construct was used to normalize the transfection efficiency. The data are expressed as -fold increase in relative luciferase activity in LPS-treated cells over the untreated cells. The data represent the mean -fold induction of three independent experiments. The S.D. of the mean is indicated by the error bars. quired for the induction of UBP43 gene expression by LPS (Fig.  5). Our results suggest that IRF-3 mediates the LPS induction of UBP43 in the RAW 264.7 macrophage-like cell line by binding to the IRFEs of the UBP43 promoter.
Among the IRF family members, IRF-3 is of particular interest, because its activation appears to have a direct role in the induction of defensive responses. Recently, viral infection (33, 44 -46), LPS stimulation (27), as well as general genotoxic stress (47) were shown to induce the phosphorylation, nuclear translocation, and subsequent IRFE binding of IRF-3. However, it was suggested that IRF-3 has no intrinsic transactivation capabilities and it may instead require assembly with other co-activators, such as CBP and/or p300, to induce gene expression (44,46,48). LPS activates IRF-3 phosphorylation via a p38 MAPK dependent pathway (27). Our data show a strong increase of UBP43 transcription upon LPS stimulation (Fig. 1), and such induction can be reduced significantly by a p38 MAP kinase inhibitor SB203580 or by co-transfection with a dominant negative p38 MAPK expression construct (data not shown). The anti-IRF-3 antibodies supershifted the IRFE-specific LPS-inducible complex in nuclear extracts from RAW 264.7 cells. Together with the repression of LPS-mediated activation of the UBP43 gene by a dominant-negative IRF-3 mutant (Fig. 8C), these results demonstrate that IRF-3 plays a primary role in the LPS-induced activation of the UBP43 gene.
The family of IRFs is involved in a wide range of host defense mechanisms (reviewed in Refs. 37 and 49)). IRF proteins stimulate the expression of many genes with antiviral, antiproliferative, apoptotic, and immunomodulatory functions. The cloning of UBP43 has recently been reported by three other groups using differential expression analyses (50 -52). In addition to LPS induction as shown in this report, UBP43 expression is also up-regulated by porcine reproductive and respiratory syndrome virus infection (50) and interferon treatment (51,52).
Based on experimental evidence of several laboratories, Taniguchi and co-workers (53) categorized IFN and IFN-inducible genes into four distinct groups in terms of activation. Group one, or "ISGF3 only" group, is totally dependent on the IFN␣/␤-activated transcription factor ISGF3. Second, the "ISGF3/IRF-3" group, can be activated by both virus and virusinduced interferon as well as bacteria and general genotoxic stresses. Groups three and four include IFN genes themselves whose transcription depends on "IRF-3/IRF-7" (IFN␤) or "IRF-7 only" (IFN␣). Based on the data presented here and on work of Kang et al. (52), UBP43 belongs to the second, ISGF3/ IRF-3 group to which ISG15 has also been assigned. These genes have acquired regulatory mechanism, which ensures gene induction even in the absence of IFN␣/␤ signaling, to exert their function in the host defense against extracellular pathogens. Coordinated induction of ISG15 and UBP43 suggests that ISG15 conjugation is a dynamic process and critical balance of ISG15 modification should be maintained at all times. Unlike ubiquitination of proteins, which mostly are destined to degradation, modification by Ubls mediates specific functions depending on the type of Ubls. In this regard, the reversible Ubl modification resembles the phosphorylation and dephosphorylation reaction of proteins, and probably serves the same functions, which are to modulate the structure, activity, or localization of the target proteins.
It is not known whether linkage of ISG15 to its target proteins results in their degradation or rather, as is the case for other ubiquitin-like proteins such as SUMO-1 and Nedd8 (1,54), this linkage modifies the biological activities of the targeted proteins. Because the proteins that are targeted by ISG15 have not been yet identified, the exact function of ISG15 modification remains to be elucidated. The direct identification of UBP43 substrates and the study of cellular response to bacterial infection in the absence of UBP43 expression in the future will provide valuable information regarding the importance of UBP43 and ISG15 modification in innate immunity.