Type I IFN Signaling Infection by Up-Regulating Typhimurium

Enhanced Antibacterial Potential in UBP43-Deficient Mice against Salmonella typhimurium Infection by Up-Regulating Type I IFN Signaling This information is current as of September 24, 2021. Keun Il Kim, Oxana A. Malakhova, Kasper Hoebe, Ming Yan, Bruce Beutler and Dong-Er Zhang J Immunol 2005; 175:847-854; ; doi: 10.4049/jimmunol.175.2.847 http://www.jimmunol.org/content/175/2/847 Downloaded from

References This article cites 45 articles, 21 of which you can access for free at:

http://www.jimmunol.org/content/175/2/847.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 24, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Enhanced Antibacterial Potential in UBP43-Deﬁcient Mice against Salmonella typhimurium Infection by Up-Regulating Type I IFN Signaling1

Keun Il Kim,2* Oxana A. Malakhova,* Kasper Hoebe,† Ming Yan,* Bruce Beutler,† and Dong-Er Zhang3*

ISG15 is an IFN-inducible ubiquitin-like protein and its expression and conjugation to target proteins are dramatically induced upon viral or bacterial infection. We have generated a UBP43 knockout mouse model that is lacking an ISG15-specific isopeptidase to study the biological role of the protein ISGylation system. We report that UBP43-deficient mice are hypersensitive to LPS- induced lethality and that TIR domain-containing adapter inducing IFN-␤ 3 IFN regulatory factor 3 3 type I IFN is the major ؊/؊ axis to induce protein ISGylation and UBP43 expression in macrophages upon LPS treatment. In ubp43 macrophages, upon Downloaded from LPS treatment we detected increased expression of IFN-stimulated genes, including genes for several cytokines and chemokines involved in the innate immune response. The ubp43؊/؊ mice were able to restrict the growth of Salmonella typhimurium more efficiently than wild-type mice. These results clearly demonstrate two aspects of IFN-signaling, a beneficial effect against pathogens but a detriment to the body without strict control. The Journal of Immunology, 2005, 175: 847–854.

nterferons are well-known cytokines that play pivotal roles in ISG15 conjugation process ISGylation follows an analogous http://www.jimmunol.org/ antiviral, antibacterial, cell growth, differentiation, and anti- mechanism to other ubiquitin-like protein modifications (8–10). I tumor responses (1, 2). IFNs are also considered as signals However, the ISGylation system is unique among ubiquitin-like for modulating innate and adaptive immunity (3). Type I IFNs protein systems. The expression of ISG15 and currently known induce several hundred IFN-stimulated genes (ISGs),4 including enzymes for ISGylation and deISGylation are all highly induced cytokine genes, through the Jak/Stat signaling pathway (4). It has by type I IFN. Furthermore, the ISGylation system has been de- been known that Jak/Stat signaling is tightly regulated by various tected only in vertebrate genomes, whereas most ubiquitin-like means and dysregulation of this signaling is associated with a va- protein systems are highly conserved throughout all eukaryotes riety of immune disorders (5). There are three groups of negative from yeast to human. Thus, the ISGylation system may have spe- regulators for type I IFN and other such cytokine signaling: several cialized functions related to host defense against pathogens in by guest on September 24, 2021 protein tyrosine phosphatases (Src homology proteins 1 and 2, higher eukaryotes. To date, three enzymes, UBE1L (E1), UBC8 CD45, protein tyrosine phosphatase 1B, and T cell protein tyrosine (E2), and UBP43 (USP18, isopeptidase) have been identified for phosphatase), suppressor of cytokine signaling (SOCS) proteins, the ISGylation and deISGylation processes (11–14). Genetic and protein inhibitor of activated STAT family members (5). knockout studies demonstrated that UBP43 activity is not essential Recently, we have demonstrated that UBP43, a deconjugating for the precursor processing of ISG15 in vivo because UBP43- protease of ISG15, is a new negative regulator of the type I IFN deficient cells can generate ISGylated proteins upon IFN treatment signaling pathway (6). ISG15 is an IFN-inducible ubiquitin-like (6). The most pronounced phenotype of UBP43 deficiency is hy- protein and its expression and conjugation to target proteins are persensitivity to type I IFN, with enhanced and prolonged activa- highly induced upon viral or bacterial infection (7, 8). The protein tion of Jak/Stat signaling (6). Thus, UBP43 is another member of the negative regulators of type I IFN and the Jak/Stat signaling pathway. Departments of *Molecular and Experimental Medicine and †Immunology, The Type I IFNs are produced in macrophages upon activation of Scripps Research Institute, La Jolla, CA 92037 TLR4 signaling by bacterial LPS, a pathogenic agent in endotoxin Received for publication March 21, 2005. Accepted for publication April 28, 2005. shock (15). Although TLRs have been shown to act as primary The costs of publication of this article were defrayed in part by the payment of page immune sensors for invading pathogens, over-activation of inflam- charges. This article must therefore be hereby marked advertisement in accordance matory signaling often causes a lethal shock to the host accompa- with 18 U.S.C. Section 1734 solely to indicate this fact. nied by multiple tissue damage and organ failure (16). TNF-␣ has 1 This work is supported by National Institutes of Health Grant CA079849 (to D.E.Z.) and GM060031 (to B.B.). The Stein Endowment Fund has partially supported the been known to be a major contributor to LPS-induced septic shock Molecular and Experimental Medicine Departmental Molecular Biology Service Lab- (17–19). In addition to TNF-␣, a recent report demonstrated that oratory for DNA Sequencing and Oligonucleotide Synthesis. This is manuscript no. type I IFNs, especially IFN-␤, are central effecters in LPS-induced 17087-MEM from The Scripps Research Institute. lethality (20). Several microarray studies showed that ISG15 is a 2 Current address: Department of Biological Science, Sookmyung Women’s Univer- sity, 53-12 Chungpa-dong 2 Ka, Yongsan-gu, Seoul 140-742, Korea. highly up-regulated gene in human macrophages during various 3 Address correspondence and reprint requests to Dr. Dong-Er Zhang, MEM-L51, bacterial infections (21, 22) and both ISG15 and UBP43 were iden- The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. tified as LPS-induced genes in mouse macrophages (23). We ques- E-mail address: [email protected] tioned whether ISG15 and/or UBP43 are involved in antibacterial 4 Abbreviations used in this paper: ISG, IFN-stimulated gene; BMM, bone marrow- responses mediated by TLR4 signaling and addressed this question derived macrophage; ISGylation, ISG15 protein conjugation; IRF, IFN regulatory factor; SOCS, suppressor of cytokine signaling; m, murine; IP-10, IFN-␥-inducible by using the UBP43 knockout mouse model, which is lacking an protein-10, Trif, TIR domain-containing adapter inducing IFN-␤. ISG15-specific deconjugating enzyme.

We report that LPS-induced UBP43 expression and protein IS- with the indicated magnification. The number of particles in the cells was Gylation is mainly a TIR domain-containing adapter inducing also counted from a total of 200 cells to get the average number of particles IFN-␤ (Trif) and type I IFN-dependent event in macrophages. We from each genotype of macrophages. also show that UBP43-deficient mice have a higher sensitivity to Northern blotting endotoxin-induced septic shock than wild-type mice. However, ␣ Total RNA from BMM was isolated using RNA Bee reagent according to TNF- production in UBP43-deficient macrophages was compa- the manufacturer’s instructions (Tel-Test). Five micrograms of total RNA rable to that of wild-type macrophages upon LPS treatment. In- from each time point was separated in an agarose/formaldehyde gel (0.22 stead, ubp43Ϫ/Ϫ macrophages exhibited enhanced and prolonged M), blotted on Hybond Nϩ membrane (Amersham Biosciences), and 32 activation of Jak/Stat signaling resulting in elevated expression of probed with P-labeled cDNAs. As an equal loading control, 28 S rRNA was detected using 32P-labeled probe (5Ј-AACGATGAGAGTAGTGG-3Ј) ISGs, as well as IFN-inducible cytokine and chemokine genes, by hybridizing at 43°C and washing with 6ϫ SSC containing 0.1% SDS which indicates a negative regulatory role of UBP43 in type I IFN at 43°C. signaling. ubp43Ϫ/Ϫ macrophages also showed enhanced phagocytic activity to IgG opsonized yeast particles. With these characteristics, Quantitative real-time PCR Ϫ Ϫ ubp43 / mice were more capable of restricting the growth of Sal- For quantitative RT-PCR analysis, cDNA was generated from total RNA using monella typhimurium than wild-type mice. These results clearly dem- TaqMan Reverse Transcription Reagents kit (Applied Biosystems) according onstrate two aspects of IFN-signaling, a beneficial effect against to manufacturer’s protocol. Real-time PCR was performed with the following primers: murine (m) TNF-␣ (forward) 5Ј-ACAAGGCTGCCCCGACTAC-3Ј, pathogens but detrimental to the body without strict control. (reverse) 5Ј-ACTTTCTCCTGGTATGAGATAGCAAAT-3Ј, (probes) 5Ј- TGCTCCTCACCCACACCGTCAGC-3Ј; mIL-1␤ (forward) 5Ј-AACCT Materials and Methods GCTGGTGTGTGACGTT-3Ј, (reverse) 5Ј-CAGCACGAGGCTTTTTTG Downloaded from Ј Ј Ј Mice TTG-3 , (probes) 5 -CCATTAGACAACTGCACTACAGGCTCCGA-3 ; mIL-6 (forward) 5Ј-GTCAATTCCAGAAACCGCTATGA-3Ј, (reverse) 5Ј- The generation of ubp43Ϫ/Ϫ, ifnar1Ϫ/Ϫ, trifLps2/Lps2, myd88Ϫ/Ϫ, and CACCAGCATCAGTCCCAAGA-3Ј, (probes) 5Ј-CTCTCTGCAAGAGA myd88Ϫ/Ϫ mice harboring trif Lps2 mutation has been described (24–27). CTTCCATCCAGT-3Ј; mMCP-1 (forward) 5Ј-AACTCTCACTGAAGC ubp43Ϫ/Ϫ mice were backcrossed to C57BL/6 mice to five generations CAGCTCTCT-3Ј, (reverse) 5Ј-TGAACAGCAGGCCCAGAAG-3Ј, (probes) because further backcross to C57BL/6 caused embryonic lethality. Other 5Ј-TCCTCCACCACCATGCAGGTC-3Ј; murine IFN-␥-inducible protein-10 mice were in a C57BL/6 background. Mice were housed in a pathogen-free (mIP-10) (forward) 5Ј-ATGAGGGCCATAGGGAAGCTT-3Ј, (reverse) 5Ј- http://www.jimmunol.org/ facility and procedures were approved by the Institutional Animal Care and ATCGTGGCAATGATCTCAACAC-3Ј, (probes) 5Ј-AATCATCCCTGC Use Committee of The Scripps Research Institute. GAGCCTATCCT-3Ј; and mMIP-1␣ (forward) 5Ј-GCGAGTACCAGTC CCTTTTCTG-3Ј, (reverse) 5Ј-GGTGGAGACCTTCATGATGTTGA-3Ј, Preparation of macrophages and treatment with TLR ligands (probes) 5Ј-CTGCTGACAAGCTCACCCTCTGTCACC-3Ј. TaqMan probes were labeled with FAM as a reporter and Black Hole Quencher as a quencher. To generate bone marrow-derived macrophages (BMM), bone marrow Thermal cycling and detection of the real-time amplification were performed cells from wild-type and knockout mice were cultured in DMEM supple- using the ABI PRISM 7900HT Sequence Detection System (Applied Biosys- mented with 10% FBS and 20% M-CSF conditional medium from L929 tems). The quantity of target gene is given relative to the GADPH control cells. The nonadherent cells were collected the next day and continually based on cycle threshold Ct values determined during the exponential phase cultured for 6 days for the development of BMM. For the preparation of of PCR. peritoneal macrophages, mice were injected with 3 ml of 4% thioglycolate by guest on September 24, 2021 i.p. Peritoneal exudates were collected 72 h postinjection by washing with Western blot analysis 10 ml of ice-cold PBS. Macrophages were plated in cell culture plates and incubated for 12 h prior to experiments. The production of anti-UBP43 Abs has been previously described (12). Macrophages were treated with various TLR ligands with concentra- Rabbit anti-mouse ISG15 polyclonal Abs were generated using full-length tions as indicated in the figure legends. Peptidoglycan (Staphylococcus mouse ISG15 and immunoaffinity purified as described in detail (28). Phos- aureus) was purchased from Fluka, poly(I:C) from Sigma-Aldrich, and phospecific Abs for STAT1, ERK, JNK, and p38 were purchased from Cell resiquimod from Novartis Pharmaceuticals. Phosphorothioate-stabilized Signaling Technology, and Abs for STAT1, ERK, JNK, p38, and I␬B␣ CpG oligodeoxynucleotide (5Ј-TCCATGACGTTCCTGATGCT-3Ј) was were from Santa Cruz Biotechnology. Western blotting was performed as obtained from Integrated DNA Technologies. previously described (12). LPS injection and cell counting Bacterial infection, bacterial load counting, and checking survival LPS (Escherichia coli serotype O127:B8; Sigma-Aldrich) was dissolved in PBS and i.p. injected to mice at 15 mg per kg body weight. Blood and bone S. typhimurium 14028 was obtained from American Type Culture Collec- marrow nucleated cells were counted in Tu¨rk solution (3% acetic acid and tion. Overnight cultures of bacteria were freshly inoculated in Nutrient 0.01% crystal violet in H2O). broth from BD Biosciences. The bacteria were collected in the log phase of growth, chilled on ice, and the concentration estimated from the absor- TNF-␣ detection bance at 600 nm using a predetermined calibration curve. For injections, bacteria were washed three times in PBS and serial dilutions were prepared Thioglycolate elicited peritoneal macrophages (5 ϫ 105 cells/well) were in the same buffer. The number of bacteria injected was subsequently con- cultured on 24-well plates for 12 h and then treated with 1 ␮g/ml LPS. firmed by viable bacterial counts on agar plates. Age-matched wild-type Culture supernatant was collected at indicated time points. The amount of Ϫ Ϫ and ubp43 / mice were injected i.p. with bacterial suspensions prepared TNF-␣ was measured by ELISA as specified by the manufacturer (BD in 500 ␮l of PBS. Mice were sacrificed 72 h postinfection, livers and Pharmingen). spleens were removed aseptically, and tissues were then homogenized in Phagocytosis assay sterile PBS. Bacterial counts were determined by plating 10-fold serial dilutions of liver or spleen homogenates on Nutrient Agar plates (BD Bio- The phagocytic ability of macrophages was assayed by measuring uptake sciences). For survival curves, mice were infected with S. typhimurium as of Texas Red zymosan particles (Molecular Probes). Thioglycolate-elicited described and monitored until terminated. peritoneal macrophages (1 ϫ 104 cells/well) from wild-type and ubp43Ϫ/Ϫ mice were plated on eight-well glass slide chambers (Nalge Nunc Inter- Results national) and allowed to adhere for1hat37°C. Texas Red zymosan par- Trif-dependent induction of protein ISGylation by TLR3 and ticles were prepared in PBS containing 20 mg/ml BSA, 2 mM sodium azide, and opsonized by incubating with 10% mouse IgG for 15 min at TLR4 ligands in BMM 5 37°C. Opsonized yeast particles (5 ϫ 10 , 50 times excess) were added to ISG15 has been reported as one of the most up-regulated genes in the cells and incubated for1hat37°C. Cells were then washed three times with ice-cold PBS and the fluorescence from uningested particles was human macrophages during various bacterial infections (21, 22). quenched with 0.4% trypan blue. The fluorescence of the particles in the Both ISG15 and UBP43 were identified as LPS-induced genes in cell was visualized using a fluorescence microscope (Leica Microsystems) mouse macrophages (23). To understand the signaling pathway The Journal of Immunology 849 that induces ISG15 expression and its conjugation upon bacterial major receptors that can induce protein ISGylation in infection, we first investigated which TLR signaling pathway is macrophages. responsible for the induction of the ISGylation system in macro- TLR3 and TLR4 are known to share overlapping adaptor pro- phages. BMM from C57BL/6 mice were treated with various TLR teins to transduce signals from the receptor to the nucleus (29). To ligands and ISG15 conjugates were detected by Western blot anal- identify which downstream signaling pathway is involved in the ysis. IFN-␤ treatment clearly induced protein ISGylation in mac- activation of the protein ISGylation system, we compared ISG15 rophages (Fig. 1A). The TLR3 ligand (poly(I:C)) and TLR4 ligand conjugation in BMM from wild-type, myd88Ϫ/Ϫ, trifLps2/Lps2 mice (LPS) also induced ISG15 conjugates to a similar extent as IFN-␤. that contain homozygous Lps2 alleles of Trif, and myd88Ϫ/Ϫ/ Other TLR ligands including peptidoglycan (TLR2), resiquimod trifLps2/Lps2 mice upon LPS treatment (Fig. 1B). We also checked (TLR7), and unmethylated CpG DNA (TLR9) did not induce pro- individual gene expressions for ISG15, UBE1L, and UBP43 in a tein ISGylation, although a basal level of ISGylated proteins were similar set of experiments (Fig. 1C). MyD88-deficient macro- still detectable in those cells. Thus, TLR3 and TLR4 seem to be the phages showed the similar levels of ISG15 conjugates and expression of the three listed genes compared with wild-type cells. Con- versely, trifLps2/Lps2 macrophages expressed significantly decreased amounts of ISG15, UBE1L, and UBP43 compared with wild-type or myd88Ϫ/Ϫ macrophages, and as a result, showed only small amounts of ISG15 conjugate upon LPS treatment. Further- more, myd88Ϫ/Ϫ/trifLps2/Lps2 macrophages did not show any LPS-

induced protein ISGylation. Taken together, these results suggest Downloaded from that LPS induces the ISGylation system mainly through a Trif- dependent pathway. However, the MyD88 pathway still contributes to the maximum activation of ISGylation, probably by inducing as yet unknown components of ISGylation.

Type I IFN signaling-mediated activation of protein ISGylation http://www.jimmunol.org/ in LPS-treated BMM Because Trif is an adaptor protein involved in IFN regulatory factor (IRF) 3 activation upon LPS response and thus induces type I IFN production in macrophages (27, 30), we analyzed the contri- bution of type I IFN signaling to protein ISGylation upon LPS treatment in BMM. We compared protein ISGylation and the expression of ISG15, UBP43, and UBE1L in wild-type and ifnar1Ϫ/Ϫ BMM upon LPS treatment. Wild-type BMM showed clear induction of protein ISGylation with LPS treatment, whereas ifnar1Ϫ/Ϫ by guest on September 24, 2021 BMM produced only very minor amounts of ISGylated proteins (Fig. 2A). LPS induced UBP43 expression only in wild-type BMM

FIGURE 1. Trif-dependent activation of the ISGylation system in macrophages. A, BMMs prepared from C57BL/6 mice were treated with IFN-␤ (500 U/ml) and various TLR ligands; 10 ␮g/ml peptidoglycan (PGN) for TLR2, 0.1 ␮g/ml poly(I:C) for TLR3, 1 ␮g/ml LPS for TLR4, 1 ␮g/ml resiquimod for TLR7, and 100 ␮M unmethylated CpG DNA for TLR9. Macrophages were harvested and subjected to SDS-PAGE followed by immunoblotting against an anti-ISG15 Ab. B, BMMs from wild-type FIGURE 2. Type I IFN-mediated activation of protein ISGylation in (WT), myd88Ϫ/Ϫ, trifLps2/Lps2 (trifLps2), and myd88Ϫ/Ϫ/trifLps2/Lps2 double LPS treated macrophages. A, BMMs from wild-type and ifnar1Ϫ/Ϫ mice knockout (DKO) mice were treated with 1 ␮g/ml LPS for 24 h and ISGy- were treated with 1 ␮g/ml LPS and protein ISGylation and UBP43 ex- lated proteins were detected by immunoblotting. Ponceau staining was pression was detected by immunoblotting. B, Total RNA was isolated from used as a loading control for A and B. C, Gene expression for the compo- the macrophages with the same treatment and subjected to Northern blot nents of ISGylation and deISGylation were detected by Northern blotting analysis using cDNA probes for IL-1␤, ISG15, UBE1L, UBCm8, and in BMMs treated with 1 ␮g/ml LPS. A 28 S rRNA signal on the same UBP43. Equal loading was evaluated by visualizing 28 S rRNA on the membrane was used as a loading control. same membrane. 850 ENHANCED ANTIBACTERIAL EFFICIENCY IN ubp43Ϫ/Ϫ MICE but not in ifnar1Ϫ/Ϫ (Fig. 2A). All of the components for the protein ISGylation system tested in this experiment, namely ISG15, UBE1L, and UBP43, showed LPS-induced expression patterns by Northern blotting (Fig. 2B). However, individual genes appeared to be regulated in different ways. ISG15 reached maximum expression levels 4–12 h after LPS treatment in wild-type BMM (Fig. 2B). In contrast, in ifnar1Ϫ/Ϫ BMM, ISG15 expression increased only for a very short time (2–4 h) and to a much lesser extent compared with wild-type BMM (Fig. 2B). This result sug- gests that LPS induces ISG15 gene expression through two independent mechanisms in BMM and type I IFN signaling is responsible for expression of ISG15 at later time points. The UBE1L and UBP43 genes were also induced by LPS in an IFN-dependent manner (Fig. 2B). However, UBE1L showed a basal level expression in ifnar1Ϫ/Ϫ BMM whereas UBCm8 and UBP43 expression was not detectible even with very long exposure (Fig. 2B and data not shown). Taken together, the components of ISGylation system are induced simultaneously by LPS mainly through type I IFN- signaling pathway in BMM but individual genes are regulated in Downloaded from different ways. One interesting finding was that the protein level of UBP43 FIGURE 3. LPS hypersensitivity of ubp43Ϫ/Ϫ mice. A, Five- to 6-wk- reached a maximum level around 4–12 h after LPS treatment and Ϫ Ϫ old wild-type and ubp43 / mice (n ϭ 7 for each population) were injected decreased significantly at 24 h, whereas ISG15 conjugates were i.p. with 15 mg per kg body weight of LPS. LPS-injected mice were then accumulated over 24 h (Fig. 2A), suggesting that UBP43 is an monitored for 5 days and the percentage of survival was calculated for each unstable protein and was degraded shortly after its expression by a genotype. B, The number of total white blood cells in peripheral blood was http://www.jimmunol.org/ proteolytic system. Indeed, UBP43 binds to Skp2 of SCF ubiquitin counted 72 h after the LPS injection. C, Thioglycolate elicited peritoneal E3 complex and can be degraded via the proteasome pathway upon macrophages from wild-type and ubp43Ϫ/Ϫ mice were stimulated with 1 ubiquitination (31). These results suggest that the cellular amount ␮g/ml LPS for the indicated time. The amount of secreted TNF-␣ was of UBP43 is tightly controlled by both transcriptional and protein measured by ELISA and it is a representative result from three independent stability levels. experiments. Hypersensitivity of UBP43-deficient mice to LPS-induced septic shock are mainly dependent on type I IFN signaling in macrophages, we We previously found that LPS strongly activated the expression of hypothesized that hypersensitivity of ubp43Ϫ/Ϫ mice to LPS might by guest on September 24, 2021 UBP43 and ISG15 and also induced the conjugation of ISG15 to be caused by altered sensitivity to type I IFN. We first checked cellular targets in macrophages (32). To test the possible involve- type I IFN signaling in wild-type and UBP43-deficient BMM in ment of UBP43 and the protein ISGylation system in endotoxin response to LPS treatment. Phosphorylation of STAT1 on tyrosine responses, we injected wild-type and ubp43Ϫ/Ϫ mice with suble- occurred 2–4 h after LPS treatment in wild-type BMM (Fig. 4A). thal doses of LPS (15 mg per kg body weight) i.p. and monitored However, in UBP43-deficient BMM, weak STAT1 phosphoryla- mouse survival for 5 days. All wild-type mice survived and re- tion was detected even before LPS treatment, and became around covered from the shock within 3 days (Fig. 3A). In contrast, over 4-fold higher compared with wild-type BMM upon LPS treatment. 50% of ubp43Ϫ/Ϫ mice died by the same dosage of LPS injection STAT phosphorylation was detectable for over 24 h after LPS (Fig. 3A), accompanied by damage to multiple organs including treatment. At the same time, substantially stronger protein ISGy- the spleen, thymus, and kidneys (data not shown). We also de- lation signal was detected in ubp43Ϫ/Ϫ BMM compared with wild- tected a significant decrease in total white blood cells in UBP43- type BMM before and after LPS treatment. IFN-inducible genes deficient mice upon LPS treatment (Fig. 3B). TNF-␣ has been including ISG15, UBE1L, IRF7, and inducible NO synthase genes known to be a major factor involved in lethality by LPS-induced were also highly expressed in UBP43-deficient compared with sepsis (17–19). We checked TNF-␣ production in thioglycolate wild-type BMM under the same conditions (Fig. 4B). These results elicited peritoneal macrophages from wild-type and ubp43Ϫ/Ϫ indicate that UBP43-deficient BMMs are hyperresponsive to type mice. We did not detect significant differences in the production of I IFN signaling upon LPS treatment and this might be a cause of TNF-␣ between wild-type and UBP43-deficient macrophages after hypersensitivity of ubp43Ϫ/Ϫ mice to LPS. LPS administration (Fig. 3C). These results indicate that UBP43 LPS also activates other transcription factors such as AP-1 and deficiency causes LPS-induced hypersensitivity in mice indepen- NF-␬B through MAPKs and I␬B kinases, respectively, in a dent of TNF-␣ production. MyD88-dependent manner (25). We analyzed the activation of signaling cascades in wild-type and ubp43Ϫ/Ϫ BMM upon LPS chal- Hyperactivation of Jak/Stat signaling but normal activation of lenge by detecting the phosphorylation of kinases (ERK1/2, JNK, other MyD88-dependent kinase signaling in UBP43-deficient and p38) or the degradation of I␬B␣ protein for NF-␬B signaling. BMM upon LPS treatment LPS treatment induced the phosphorylation of ERK1/2, JNK, and Recently, type I IFNs were proposed to be other central factors p38 without any differences between wild-type and ubp43Ϫ/Ϫ involved in developing LPS-induced endotoxin shock (20). We BMM in the intensity and the time course of phosphorylation (Fig. previously revealed that UBP43-deficient cells exhibited an ele- 4C). The LPS-induced degradation of I␬B␣ also showed no dif- vated response to type I IFN accompanied by hyperphosphoryla- ference between wild-type and ubp43Ϫ/Ϫ BMM as judged by im- tion of Stat1 and increased expression of ISGs (6). Because we munoblotting. Taken together, these results suggest that UBP43 found that LPS-induced UBP43 expression and protein ISGylation deficiency causes hyperactivation of type I IFN-dependent gene The Journal of Immunology 851

obtained the same extent of TNF-␣ expression in wild-type and ubp43Ϫ/Ϫ macrophages from a short time point (1 h) upon LPS treatment in an independent experiment (data not shown). Inter- estingly, the IL-1␤ message level in UBP43-deficient BMM was decreased ϳ2-fold over tested time periods of LPS treatment (Fig. 5A). In contrast, we detected a 2-fold increase in IL-1␤ expression in LPS-treated ifnar1Ϫ/Ϫ BMM compared with wild-type from Northern blot analysis (Fig. 3C) and also from quantitative RT- PCR (Fig. 5B). These results indicate that hypersensitivity to type I IFN in UBP43-deficient macrophages causes down-regulation of IL-1␤ expression during the LPS response. The mRNA levels for cytokine IL-6 and chemokines including MCP-1, IP-10, and MIP-1␣ in wild-type BMM reached a maximum at ϳ4 h after treatment with LPS and rapidly decreased with time to basal levels (Fig. 5A). Initial expression of these genes in UBP43-deficient BMM was comparable to or slightly higher (for MCP-1) than in wild-type BMM at 4 h post LPS challenge. How- ever, a substantial amount of mRNA was still detected even after

24 h of LPS treatment (Fig. 5A). The expression of these genes was Downloaded from impaired in ifnar1Ϫ/Ϫ BMM treated with LPS (Fig. 5B). Together, these results suggest that type I IFN plays an important role in regulating the production of subsets of cytokines and chemokines, including IL-6, MCP-1, IP-10, and MIP-1␣ upon LPS treatment. In ubp43Ϫ/Ϫ macrophages, the production of type I IFN-inducible

cytokines and chemokines is prolonged and/or enhanced in accor- http://www.jimmunol.org/ dance with prolonged activation of Jak/Stat signaling. Enhanced phagocytic activity in UBP43-deﬁcient macrophage One of the most important roles of macrophages upon bacterial infection along with cytokine and chemokine production is to engulf and degrade invading bacteria to clear the infected host body (33). We analyzed the phagocytic activity of macrophages from wild-type and ubp43Ϫ/Ϫ mice. Thioglycolate elicited peritoneal

FIGURE 4. Hyperactivation of Jak/Stat signaling but normal activation by guest on September 24, 2021 of MyD88-dependent signaling in UBP43-deficient macrophages by LPS. macrophages were incubated with a 50-fold excess amount of IgG A, BMMs from wild-type (WT) and ubp43Ϫ/Ϫ mice were treated with 1 opsonized zymosan particles, and particle uptake was visualized ␮g/ml LPS and the tyrosine phosphorylation of STAT1, STAT1 expres- on the fluorescence microscope. As shown in Fig. 6, certain pop- Ϫ Ϫ sion, protein ISGylation and UBP43 expression at indicated time points ulations of ubp43 / macrophages could take up a large number were detected by immunoblotting. Ponceau staining was used as a loading of fluorescent particles (up to 15–18 particles), whereas wild-type control. B, The expression of ISGs was detected in wild-type and ubp43Ϫ/Ϫ macrophages only contained a small number of particles. Average BMM upon 1 ␮g/ml LPS treatment by Northern blotting. The amount of 28 particle numbers for wild-type and ubp43Ϫ/Ϫ macrophages S rRNA was used as a loading control. C, The activation of the MyD88- counted from a total of 200 cells were 3–4 and 7–9, respectively. dependent signaling pathway was examined in LPS-treated BMM derived This result indicates that a subpopulation of ubp43Ϫ/Ϫ macro- from wild-type or ubp43Ϫ/Ϫ mice. The phosphorylated form of each kinase phages were preactivated and had an enhanced ability to engulf was examined as an indication of the activation of Erk, JNK, and p38 and the unphosphorylated forms were used as equal loading controls. For the particles compared with wild-type cells. ␬ ␬ ␣ activation of NF- B signaling, the degradation of I B was analyzed by Enhanced antibacterial potential in UBP43-deficient mice immunoblotting. against S. typhimurium infection Because we detected hyperactivation of type I IFN-related re- expression but does not affect MyD88-dependent signaling path- sponses in UBP43 knockout macrophages, we questioned the bi- ways upon TLR4 activation by LPS. ological effect of UBP43 deficiency in mice against live pathogens, especially against Gram-negative bacteria. To address this matter, Altered expression of cytokine and chemokine genes in wild-type and ubp43Ϫ/Ϫ mice were infected with S. typhimurium UBP43-deficient BMM upon LPS treatment (ϳ40 CFU) i.p. and the bacterial load was determined 72 h postin- One characteristic event in macrophages in response to stimulatory fection from the liver and spleen, which are the major target organs molecules such as LPS is the production and release of cytokines for Salmonella proliferation in the host (34). At the same time, and chemokines to initiate inflammatory responses. Because we mouse survival was also monitored with the same dose of bacterial detected that hyperactivation of type I IFN response resulted in an infection. At day 3 after Salmonella infection, we detected numer- elevated expression of ISGs upon LPS treatment, we examined the ous gray-white lesions on the livers from wild-type mice whereas expression of cytokine and chemokine genes in LPS-treated BMM livers from ubp43Ϫ/Ϫ mice were relatively free of such lesions using quantitative RT-PCR. The LPS-induced expression of (Fig. 7A). Spleens from wild-type mice also showed similar lesions TNF-␣ showed a slight but not significant increase in ubp43Ϫ/Ϫ but it was less severe than lesions found in the liver (data not cells (Fig. 5A), consistent with our previous data from LPS-treated shown). H&E staining of liver sections also showed a significant peritoneal macrophages, which showed a similar amount of TNF-␣ difference between wild-type and ubp43Ϫ/Ϫ mice. The morphol- production in wild-type and ubp43Ϫ/Ϫ macrophages (Fig. 3C). We ogy of the liver from ubp43Ϫ/Ϫ mice was quite similar to that of 852 ENHANCED ANTIBACTERIAL EFFICIENCY IN ubp43Ϫ/Ϫ MICE

FIGURE 5. Altered expression of cytokine and chemokine genes in UBP43-deficient macrophages upon LPS treatment. A, BMMs from wild- type and ubp43Ϫ/Ϫ mice were stimulated with 1 ␮g/ml LPS for the indicated time periods. The expression of genes was analyzed by quantitative RT-PCR. The fold of induction was calculated by calibrating to 0 h LPS stimulation in wild-type cells. Downloaded from GAPDH gene was used as an internal control for the reaction. B, Gene expression was measured in wild-type and ifnar1Ϫ/Ϫ BMM treated with 1 ␮g/ml LPS for4hasdescribed. http://www.jimmunol.org/ by guest on September 24, 2021 wild-type mice without Salmonella infection except for the pres- from ubp43Ϫ/Ϫ mice both in the liver and spleen were at least ence of many infiltrating immune cells (Fig. 7B). However, as we 10-fold lower than the numbers found in wild-type mice (Fig. 7C). found lesions in liver of wild-type mice, H&E-stained liver sec- These results suggest that ubp43Ϫ/Ϫ mice can restrict bacterial tions also showed large areas of necrotic cell death (Fig. 7B). Bac- growth much more efficiently than wild-type mice. terial load in the liver and spleen showed significant differences Virulent strains of Salmonella grow in the mouse and finally between wild-type and ubp43Ϫ/Ϫ mice. Bacterial numbers counted causes host death even with few numbers of initial infections (34). Surprisingly, the survival rate of ubp43Ϫ/Ϫ mice with a low dosage of Salmonella infection (ϳ40 PFU) was not significantly increased compared with wild-type mice. All wild-type mice died within 8 days postinfection and ubp43Ϫ/Ϫ mice survived slightly longer than wild-type hosts but all died at 11 days postinfection (Fig. 7D). This result indicates that although UBP43 deficiency provided an advantage in protecting the host from bacterial growth, it also caused detrimental effects to the host because of probable hypersensitivity to LPS. Discussion The innate immune system is the first line of the host defense against invading pathogens (35, 36). TLRs are now generally accepted as receptors, which recognize pathogen-associated molecular patterns. Among them, TLR4 has been known as a specific receptor of LPS, the cell wall component of Gram-negative bac- FIGURE 6. Enhanced phagocytic activity of UBP43-deficient macro- teria (37). Binding of LPS to TLR4 initiates a signaling cascade to 4 phages. Thioglycolate elicited peritoneal macrophages (1 ϫ 10 cells/well) activate transcription factors including NF-␬B and IRF3, resulting Ϫ/Ϫ from wild-type and ubp43 mice were plated on eight-well glass slide in the production of immunoregulatory cytokines and chemokines chambers and allowed to adhere for1hat37°C. Texas Red labeled yeast (15). We previously found that LPS induces the expression of particles opsonized with mouse IgG (5 ϫ 105 particles, 50 times excess of cell number) were added to the cells and incubated for1hat37°C. Cells ISG15 and UBP43 as well as the conjugation of ISG15 to numer- were then washed three times with ice-cold PBS and the fluorescence from ous target proteins in macrophages (32). Thus we questioned uningested particles was quenched with 0.4% trypan blue. The fluores- whether the protein ISGylation system and UBP43 are involved in cence of the particles in cells was visualized using a fluorescence micro- the TLR4-mediated host defense mechanism. First we examined scope with ϫ100 or ϫ400 magnification. which TLR signaling other than TLR4 can activate protein The Journal of Immunology 853

mokines (MCP-1, IP-10, and MIP-1␣) were also highly expressed in ubp43Ϫ/Ϫ macrophages compared with wild-type cells by LPS treatment. Moreover, the expression of these genes were impaired or significantly reduced in ifnar1-deficient macrophages. Cytokine and chemokine production during the inflammatory response is regulated by complicated mechanisms. For example, IL-6 expression by LPS has been known to require at least two transcription factors: NF-␬B and the nuclear protein I␬B␨ (25, 39). Now, our results clearly demonstrated that the type I IFN pathway plays a pivotal role in producing the maximum amount of inflammatory cytokines and chemokines synergistically with other signaling pathways and UBP43 is a crucial component for finishing type I IFN signaling upon LPS challenge. A defect in any one of these signaling molecules causes a significant reduction or overproduc- tion of a subset of cytokines and chemokines. Recently, type I IFN has been identified as a major factor for LPS-induced mouse death (20). UBP43-deficient mice are not able to terminate inflammatory response efficiently due to its intrinsic hypersensitivity to type I

IFN. Most probably, this is why UBP43 mice are more susceptible Downloaded from to LPS-induced septic shock compared with wild-type mice. FIGURE 7. Enhanced antibacterial potential in UBP43-deficient mice. In contrast, the hypersensitivity to type I IFN helps mice to Age-matched wild-type and ubp43Ϫ/Ϫ mice (n ϭ 7 for each genotype) restrict bacterial growth in the body. Salmonella growth in were infected i.p. with ϳ40 CFU of S. typhimurium. Livers and spleens ubp43Ϫ/Ϫ mice was substantially reduced compared with growth were harvested 72 h postinfection. Representative pictures of livers for found in wild-type mice. UBP43-deficient macrophages exhibited both genotypes are presented in A. Portions of the livers from infected elevated phagocytic activity, enhanced expression of antibacterial http://www.jimmunol.org/ Ϫ/Ϫ wild-type and ubp43 mice were subjected to paraffin section and stained inducible NO synthase as well as cytokine and chemokine genes. with H&E (B). Total viable bacterial counts in the liver and spleen were All together, hypersensitivity to type I IFN was beneficial to the extrapolated from the bacterial counts in the respective tissue samples (C). Ϫ/Ϫ Ϫ/Ϫ ϭ antibacterial response in ubp43 mice. However, survival of Survival of wild-type and ubp43 mice (n 6 for each genotype) were Ϫ/Ϫ monitored after infecting them i.p. with ϳ40 CFU of S. typhimurium. The ubp43 mice was only slightly extended after peritoneal infec- result is expressed as a percentage of survival at times after infection (D). tion of S. typhimurium compared with the wild-type mice. Al- though ubp43Ϫ/Ϫ mice can delay bacterial growth, it cannot erad- icate the infecting bacteria. Once the bacterial number has reached ISGylation and how exactly TLR4 signaling induces protein a certain point, it might cause systemic inflammation, resulting in ISGylation in macrophages. Our experiments revealed that TLR3 rapid death of ubp43Ϫ/Ϫ mice with their intrinsic hypersensitivity by guest on September 24, 2021 and TLR4 are major receptors inducing protein ISGylation in mac- to type I IFN, as was the case seen in the LPS-injection experi- rophages and that Trif 3 IRF3 3 type I IFN is the main axis for ment. Thus, type I IFN signaling is beneficial to setup an antibac- inducing gene expression for the protein ISGylation system and for terial state against the Gram-negative bacteria S. typhimurium, UBP43 in macrophages upon LPS induction. Interestingly, the though at the same time, it is detrimental to the body without ISG15 gene expression pattern showed a biphasic response upon proper termination of the signaling. LPS challenge. The expression at the late time period (8–12 h after The mode of action of UBP43 as a negative regulator in type I LPS treatment) comes from the activation of Jak/Stat signaling by IFN signaling is similar to that of SOCS-1 in IFN-␥ signaling in type I IFN because ifnar1Ϫ/Ϫ macrophages have a defect in ex- that both proteins are induced by the signaling pathway and, in pressing ISG15 gene at this point. The early expression of ISG15 turn, function as negative regulators to terminate the signaling pro- (2–4 h) is probably induced by IRF3 because it is well known that cess (40). SOCS-1 has been identified as a negative regulator of IRF3 can directly activate ISG15 gene expression upon viral infection LPS-induced TLR4 signaling as well (41–43). The socs-1Ϫ/Ϫ (38). In contrast, LPS induction of UBE1L, UBCm8, and UBP43 was mouse itself, or with an IFN-␥-deficient background, is hypersen- mainly dependent on type I IFN signaling. Thus, type I IFN signaling sitive to LPS-induced lethal effects as are the ubp43Ϫ/Ϫ mice. plays a pivotal role in inducing protein ISGylation in macrophages However, the mechanism whereby SOCS-1 deficiency causes LPS upon LPS treatment. As reported previously, IFN production upon hypersensitivity is controversial. Two groups claimed that SOCS-1 LPS treatment is due to the phosphorylation of IRF3 via a Trif path- directly suppresses LPS-induced NF-␬B activation (41, 42) but way and the induction of NF-␬B via a MyD88 pathway (15). recently another group has found that type I IFN signaling but not We previously generated UBP43 knockout mice to study the NF-␬B signaling is a target of SOCS-1 upon LPS treatment (43). biological function of the protein ISGylation system (6, 26). We It is clear that UBP43 is a specific negative regulator for type I IFN used this mouse model to investigate the possible involvement of signaling because we did not detect any changes in the activation ISGylation or UBP43 in response to LPS-mediated TLR4 signal- of MyD88-dependent TLR4 signaling by LPS. More importantly, ing. In our experiment, UBP43-deficient mice showed a hypersen- the proper termination of type I IFN-signaling by UBP43 is a crit- sitivity to LPS injection, resulting in over 50% death of ubp43Ϫ/Ϫ ical requirement to prevent over-activation of TLR4 signaling by mice with a sublethal (for wild-type mice) dose of LPS injection. LPS. In the absence of UBP43, cells accumulate much higher lev- TNF-␣ production was almost the same in wild-type and UBP43- els of ISGylated proteins. To date, only five cellular targets for deficient macrophages. Instead, we detected hyperactivation of the protein ISGylation have been reported, including Serpin 2a, Jak1, Jak/Stat signaling pathway in UBP43-deficient macrophages upon Stat1, phospholipase C␥1, and Erk1/2 (28, 44), among possibly LPS treatment characterized by enhanced and prolonged phos- hundreds of ISGylated proteins. We recently report that UBP43- phorylation of STAT1 and elevated expression of ISGs. One in- deficient mice are resistant to certain virus infections (45). Cur- teresting finding is that several genes for cytokines (IL-6) and che- rently, it is not clear how protein ISGylation contributes to innate 854 ENHANCED ANTIBACTERIAL EFFICIENCY IN ubp43Ϫ/Ϫ MICE immunity. Further identification of ISG15 targets and intensive 19. Rothe, J., W. Lesslauer, H. Lotscher, Y. Lang, P. Koebel, F. Kontgen, A. Althage, studies for the consequence of ISGylation of these target proteins R. Zinkernagel, M. Steinmetz, and H. Bluethmann. 1993. Mice lacking the tu- mour necrosis factor receptor 1 are resistant to TNF-mediated toxicity but highly will eventually provide molecular explanation of how protein IS- susceptible to infection by Listeria monocytogenes. Nature 364: 798–802. Gylation is involved in innate immune responses against bacterial 20. Karaghiosoff, M., R. Steinborn, P. Kovarik, G. Kriegshauser, M. Baccarini, B. Donabauer, U. Reichart, T. Kolbe, C. Bogdan, T. Leanderson, et al. 2003. and viral infection. Central role for type I interferons and Tyk2 in lipopolysaccharide-induced en- In conclusion, UBP43, a deconjugating enzyme for the protein dotoxin shock. Nat. Immunol. 4: 471–477. ISGylation system, is a crucial component in endotoxin-induced 21. Nau, G. J., J. F. Richmond, A. Schlesinger, E. G. Jennings, E. S. Lander, and R. A. Young. 2002. Human macrophage activation programs induced by bacterial inflammatory responses by terminating type I IFN signaling. De- pathogens. Proc. Natl. Acad. Sci. USA 99: 1503–1508. ficiency of UBP43 causes hypersensitivity to LPS-induced sepsis. 22. Nau, G. J., A. Schlesinger, J. F. Richmond, and R. A. Young. 2003. Cumulative ubp43Ϫ/Ϫ mice could restrict Salmonella growth more efficiently Toll-like receptor activation in human macrophages treated with whole bacteria. Ϫ/Ϫ J. Immunol. 170: 5203–5209. than wild-type mice, though survival rates of ubp43 mice were 23. Gao, J. J., V. Diesl, T. Wittmann, D. C. Morrison, J. L. Ryan, S. N. Vogel, and only slightly extended after Salmonella infection as compared with M. T. Follettie. 2002. Regulation of gene expression in mouse macrophages the wild-type mice, the most likely reason being the intrinsic hy- stimulated with bacterial CpG-DNA and lipopolysaccharide. J. Leukocyte Biol. Ϫ/Ϫ 72: 1234–1245. persensitivity to LPS in the ubp43 mice. These results clearly 24. Muller, U., U. Steinhoff, L. F. Reis, S. Hemmi, J. Pavlovic, R. M. Zinkernagel, demonstrate a good model that the tight regulation of cytokine and M. Aguet. 1994. Functional role of type I and type II interferons in antiviral signaling is critical to protect the host from harmful overresponse defense. Science 264: 1918–1921. 25. Kawai, T., O. Adachi, T. Ogawa, K. Takeda, and S. Akira. 1999. Unresponsive- to inflammatory signaling, which may have lethal effects. ness of MyD88-deficient mice to endotoxin. Immunity 11: 115–122. 26. Ritchie, K. J., M. P. Malakhov, C. J. Hetherington, L. Zhou, M. T. Little,

Acknowledgments O. A. Malakhova, J. C. Sipe, S. H. Orkin, and D. E. Zhang. 2002. Dysregulation Downloaded from We thank Drs. Michel Aguet (Swiss Institute for Experimental Cancer of protein modification by ISG15 results in brain cell injury. Genes Dev. 16: 2207–2212. Research, Lausanne, Switzerland) and Shizuo Akira (Department of Host 27. Hoebe, K., X. Du, P. Georgel, E. Janssen, K. Tabeta, S. O. Kim, J. Goode, P. Lin, Defense, Research Institute for Microbial Diseases, Osaka University, N. Mann, S. Mudd, et al. 2003. Identification of Lps2 as a key transducer of Osaka, Japan) for ifnar1Ϫ/Ϫ and myd88Ϫ/Ϫ mice, respectively, Drs. Li Li MyD88-independent TIR signalling. Nature 424: 743–748. and Chang Hahn (Aventis Pharmaceuticals, Bridgewater, NJ) for technical 28. Malakhov, M. P., K. I. Kim, O. A. Malakhova, B. S. Jacobs, E. C. Borden, and assistance, and Dr. Herbert Virgin (Washington University School of Med- D. E. Zhang. 2003. High-throughput immunoblotting: ubiquitiin-like protein ISG15 modifies key regulators of signal transduction. J. Biol. Chem. 278: http://www.jimmunol.org/ icine, St. Louis, MO) and members of the Zhang laboratory for valuable 16608–16613. discussions and critical reading of the manuscript. 29. Yamamoto, M., K. Takeda, and S. Akira. 2004. TIR domain-containing adaptors define the specificity of TLR signaling. Mol. Immunol. 40: 861–868. Disclosures 30. Yamamoto, M., S. Sato, H. Hemmi, K. Hoshino, T. Kaisho, H. Sanjo, O. Takeuchi, M. Sugiyama, M. Okabe, K. Takeda, and S. Akira. 2003. Role of The authors have no financial conflict of interest. adaptor TRIF in the MyD88-independent Toll-like receptor signaling pathway. Science 301: 640–643. References 31. Tokarz, S., C. Berset, J. La Rue, K. Friedman, K. Nakayama, K. Nakayama, 1. Stark, G. R., I. M. Kerr, B. R. Williams, R. H. Silverman, and R. D. Schreiber. D. E. Zhang, and S. Lanker. 2004. The ISG15 isopeptidase UBP43 is regulated 1998. How cells respond to interferons. Annu. Rev. Biochem. 67: 227–264. by proteolysis via the SCFSkp2 ubiquitin ligase. J. Biol. Chem. 279: 2. Samuel, C. E. 2001. Antiviral actions of interferons. Clin. Microbiol. Rev. 14: 46424–46430. 778–809. 32. Malakhova, O., M. Malakhov, C. Hetherington, and D. E. Zhang. 2002. Lipo- by guest on September 24, 2021 3. Pestka, S., J. A. Langer, K. C. Zoon, and C. E. Samuel. 1987. Interferons and their polysaccharide activates the expression of ISG15-specific protease UBP43 via actions. Annu. Rev. Biochem. 56: 727–777. interferon regulatory factor 3. J. Biol. Chem. 277: 14703–14711. 4. Der, S. D., A. Zhou, B. R. Williams, and R. H. Silverman. 1998. Identification of 33. Sansonetti, P. 2001. Phagocytosis of bacterial pathogens: implications in the host genes differentially regulated by interferon ␣, ␤,or␥ using oligonucleotide ar- response. Semin. Immunol. 13: 381–390. rays. Proc. Natl. Acad. Sci. USA 95: 15623–15628. 34. Richter-Dahlfors, A., A. M. Buchan, and B. B. Finlay. 1997. Murine salmonel- 5. Shuai, K., and B. Liu. 2003. Regulation of JAK-STAT signalling in the immune losis studied by confocal microscopy: Salmonella typhimurium resides intracel- system. Nat. Rev. Immunol. 3: 900–911. lularly inside macrophages and exerts a cytotoxic effect on phagocytes in vivo. 6. Malakhova, O. A., M. Yan, M. P. Malakhov, Y. Yuan, K. J. Ritchie, K. I. Kim, J. Exp. Med. 186: 569–580. L. F. Peterson, K. Shuai, and D. E. Zhang. 2003. Protein ISGylation modulates 35. Beutler, B. 2004. Innate immunity: an overview. Mol. Immunol. 40: 845–859. the JAK-STAT signaling pathway. Genes Dev. 17: 455–460. 36. Finlay, B. B., and R. E. Hancock. 2004. Can innate immunity be enhanced to treat 7. Loeb, K. R., and A. L. Haas. 1992. The interferon-inducible 15-kDa ubiquitin microbial infections? Nat. Rev. Microbiol. 2: 497–504. homolog conjugates to intracellular proteins. J. Biol. Chem. 267: 7806–7813. 37. Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. Van Huffel, X. Du, D. Birdwell, 8. Kim, K. I., and D. E. Zhang. 2003. ISG15, not just another ubiquitin-like protein. E. Alejos, M. Silva, C. Galanos, et al. 1998. Defective LPS signaling in C3H/HeJ Biochem. Biophys. Res. Commun. 307: 431–434. and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282: 2085–2088. 9. Hochstrasser, M. 2000. Evolution and function of ubiquitin-like protein-conju- 38. Nakaya, T., M. Sato, N. Hata, M. Asagiri, H. Suemori, S. Noguchi, N. Tanaka, gation systems. Nat. Cell Biol. 2: E153–E157. and T. Taniguchi. 2001. Gene induction pathways mediated by distinct IRFs 10. Pickart, C. M. 2001. Mechanisms underlying ubiquitination. Annu. Rev. Biochem. during viral infection. Biochem. Biophys. Res. Commun. 283: 1150–1156. 70: 503–533. 39. Yamamoto, M., S. Yamazaki, S. Uematsu, S. Sato, H. Hemmi, K. Hoshino, 11. Yuan, W., and R. M. Krug. 2001. Influenza B virus NS1 protein inhibits conju- T. Kaisho, H. Kuwata, O. Takeuchi, K. Takeshige, et al. 2004. Regulation of gation of the interferon (IFN)-induced ubiquitin-like ISG15 protein. EMBO J. 20: Toll/IL-1-receptor-mediated gene expression by the inducible nuclear protein 362–371. I␬B␨. Nature 430: 218–222. 12. Malakhov, M. P., O. A. Malakhova, K. I. Kim, K. J. Ritchie, and D. E. Zhang. 2002. UBP43 (USP18) specifically removes ISG15 from conjugated proteins. 40. Alexander, W. S., and D. J. Hilton. 2004. The role of suppressors of cytokine Annu. Rev. J. Biol. Chem. 277: 9976–9981. signaling (SOCS) proteins in regulation of the immune response. Immunol. 13. Zhao, C., S. L. Beaudenon, M. L. Kelley, M. B. Waddell, W. Yuan, 22: 503–529. B. A. Schulman, J. M. Huibregtse, and R. M. Krug. 2004. The UbcH8 ubiquitin 41. Kinjyo, I., T. Hanada, K. Inagaki-Ohara, H. Mori, D. Aki, M. Ohishi, H. Yoshida, E2 enzyme is also the E2 enzyme for ISG15, an IFN-␣/␤-induced ubiquitin-like M. Kubo, and A. Yoshimura. 2002. SOCS1/JAB is a negative regulator of LPS- protein. Proc. Natl. Acad. Sci. USA 101: 7578–7582. induced macrophage activation. Immunity 17: 583–591. 14. Kim, K. I., N. V. Giannakopoulos, H. W. Virgin, and D. E. Zhang. 2004. Inter- 42. Nakagawa, R., T. Naka, H. Tsutsui, M. Fujimoto, A. Kimura, T. Abe, E. Seki, feron-inducible ubiquitin E2, Ubc8, is a conjugating enzyme for protein ISGy- S. Sato, O. Takeuchi, K. Takeda, et al. 2002. SOCS-1 participates in negative lation. Mol. Cell. Biol. 24: 9592–9600. regulation of LPS responses. Immunity 17: 677–687. 15. Beutler, B., and E. T. Rietschel. 2003. Innate immune sensing and its roots: the 43. Gingras, S., E. Parganas, A. de Pauw, J. N. Ihle, and P. J. Murray. 2004. Re- story of endotoxin. Nat. Rev. Immunol. 3: 169–176. examination of the role of SOCS1 in the regulation of Toll-like receptor signal- 16. Cohen, J. 2002. The immunopathogenesis of sepsis. Nature 420: 885–891. ing. J. Biol. Chem. 279: 54702–54707. 17. Beutler, B., I. W. Milsark, and A. C. Cerami. 1985. Passive immunization against 44. Hamerman, J. A., F. Hayashi, L. A. Schroeder, S. P. Gygi, A. L. Haas, cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. L. Hampson, P. Coughlin, R. Aebersold, and A. Aderem. 2002. Serpin 2a is Science 229: 869–871. induced in activated macrophages and conjugates to a ubiquitin homolog. J. Im- 18. Pfeffer, K., T. Matsuyama, T. M. Kundig, A. Wakeham, K. Kishihara, munol. 168: 2415–2423. A. Shahinian, K. Wiegmann, P. S. Ohashi, M. Kronke, and T. W. Mak. 1993. 45. Ritchie, K. J., C. S. Hahn, K. I. Kim, M. Yan, D. Rosario, L. Li, J. C. de la Torre, Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to en- and D. E. Zhang. 2004. Role of ISG15 protease UBP43 (USP18) in innate im- dotoxic shock, yet succumb to L. monocytogenes infection. Cell 73: 457–467. munity to viral infection. Nat. Med. 10: 1374–1378.