The Journal of Immunology

Ndfip1 Regulates Itch Ligase Activity and Airway Inflammation via UbcH7

Mahesh Kathania,* Minghui Zeng,* Viveka Nand Yadav,† Seyed Javad Moghaddam,‡ Baoli Yang,x and K Venuprasad*

The ubiquitin-ligating enzyme (E3) Itch plays a crucial role in the regulation of inflammation, and Itch deficiency leads to severe airway inflammation. However, the molecular mechanisms by which Itch function is regulated remain elusive. In this study, we found that nontypeable Haemophilus influenzae induces the association of Itch with Ndfip1. Both Itch2/2 and Ndfip12/2 mice exhibited severe airway inflammation in response to nontypeable Haemophilus influenza, which was associated with elevated expression of proinflammatory cytokines. Ndfip1 enhanced Itch ligase activity and facilitated Itch-mediated Tak1 ubiquitination. Mechanistically, Ndfip1 facilitated recruitment of ubiquitin-conjugating enzyme (E2) UbcH7 to Itch. The N-terminal region of Ndfip1 binds to UbcH7, whereas the PY motif binds to Itch. Hence, Ndfip1 acts as an adaptor for UbcH7 and Itch. Reconstitution of full-length Ndfip1 but not the mutants that fail to interact with either UbcH7 or Itch, restored the defect in Tak1 ubiquitination and inhibited elevated proinflammatory cytokine expression by Ndfip12/2 cells. These results provide new mechanistic insights into how Itch function is regulated during inflammatory signaling, which could be exploited therapeutically in inflammatory diseases. The Journal of Immunology, 2015, 194: 2160–2167.

nflammation is essential for host defense against invading valent attachment of Ub to a conserved cysteine residue in the pathogens, such as viruses and bacteria (1). The inflammatory HECT domain, and this Ub is transferred to the target protein (5). I response must be resolved after the pathogens are cleared, Itch belongs to the HECT family of E3 ligases, which contains because unchecked inflammation can cause host tissue damage, a PKC-related C2 domain, four WW domains, and the HECT ligase leading to pathological conditions, such as autoimmunity and ma- domain. Deficiency of Itch in mice results in severe multiorgan lignancy (2). However, the mechanism that controls resolution of inflammatory disease, including chronic pulmonary interstitial in- the inflammatory response is incompletely understood. flammation and alveolar proteinosis (6, 7). Further, a truncated Ubiquitin (Ub)-mediated degradation of signaling intermediates human Itch mutation results in severe lung inflammation in a group is an important means of termination of the inflammatory response of pediatric patients (8). (3). Ubiquitination is achieved via an enzyme cascade consisting of Originally, Ndfip1 was identified as a Nedd4-interacting protein (9). ubiquitin-activating, ubiquitin-conjugating (E2), and ubiquitin-li- It is composed of three highly conserved transmembrane domains and gating (E3) enzymes (4, 5). Ub is activated by the Ub-activating cytoplasmic PY motifs, and it is localized to the Golgi, endosomes, enzymes and is transferred to E2s. E2s, in turn, interact with E3s, and multivesicular bodies (10). Ndfip1 also was shown to interact which mediate conjugation of Ub to target proteins. There are two with Itch. Mice that are deficient in Ndfip1 develop inflammation in classes of E3 ligases: the RING domain class and the HECT domain the skin and lungs similar to Itch2/2 mice and die prematurely (11). class. RING domain E3s facilitate the direct transfer of Ub from E2 Nontypeable Haemophilus influenzae (NTHi) is a major cause to a substrate without the formation of a transient covalent E3-Ub of respiratory infections and is the most common colonizer of thioester intermediate. The HECT domain-containing E3 enzymes airways in patients with chronic obstructive pulmonary disease, act as intermediate acceptors of Ub. The catalytic activity of the leading to its exacerbation (12). NTHi was shown to induce an HECT domain ubiquitinates the target protein and catalyzes for- NF-kB–mediated inflammatory response through activation of mation of polyubiquitin chains. The E2s catalyze a thioester co- the kinase Tak1 (13). In this article, we demonstrate that Ndfip1 regulates Itch ligase activity toward Tak1 by promoting the binding *Baylor Institute for Immunology Research, Baylor Research Institute, Dallas, TX of the E2 UbcH7 and inhibits NTHi-induced airway inflammation. 75204; †Department of Pathology, University of Michigan, Ann Arbor, MI 48109; ‡Pulmonary Medicine, The University of Texas MD Anderson Cancer Center, Hous- ton, TX 77030; and xDepartment of Obstetrics and Gynecology, Carver College of Materials and Methods Medicine, University of Iowa, Iowa City, IA 52242 Mice

Received for publication October 29, 2014. Accepted for publication December 29, 2/2 2/2 2014. Itch and Ndfip1 mice were described previously (11, 14). C57BL/6 mice were purchased from Charles River Laboratory. All mice were This work was supported by an American Cancer Society Research Scholar grant housed in microisolator cages in the barrier facility of the Baylor Institute (122713-RSG-12-260-01-LIB) and a Baylor Charles A. Sammons pilot project (to KV.). for Immunology Research. All experiments were performed in accordance Address correspondence and reprint requests to Dr. K Venuprasad, Baylor Institute with the guidelines of the Institutional Animal Care and Use Committee for Immunology Research, Baylor Research Institute, 3410 Worth Street, Suite 600, of Baylor Research Institute. Dallas, TX 75204. E-mail address: [email protected] The online version of this article contains supplemental material. Plasmid construction and cell transfection Abbreviations used in this article: BMDM, bone marrow–derived macrophage; E2, ubiquitin-conjugating enzyme; E3, ubiquitin-ligating enzyme; HA, hemagglutinin; Flag-Itch, Flag-Tak1, hemagglutinin (HA)-Ub (wild-type [WT]), HX-Itch NTHi, nontypeable Haemophilus influenzae; Ub, ubiquitin; WT, wild-type. (C830A), and Myc-Ndfip1 were described previously (14, 15). HA-Ube2L3 (UbcH7) was obtained from Addgene (cat. no. 27561). Myc–Ndfip1-DN, Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 Myc–Ndfip1-DM, and Myc–Ndfip1-DC were created from Myc-Ndfip1 by www.jimmunol.org/cgi/doi/10.4049/jimmunol.1402742 The Journal of Immunology 2161 subcloning into pcDNA3.1/myc-His (-) A (Invitrogen) between the XhoI and Lentiviral transduction HindIII restriction sites. Flag–Ndfip1-D1–41 and Flag–Ndfip1-DPY were created from Myc-Ndfip1 by subcloning into pCMVtag2B between HindIII Lentivirus transfection was performed as described earlier (14). Expression and XhoI sites. All clone sequences were confirmed. Transient transfection of clones were created by Gateway cloning technology, according to the manu- 293T cells was performed using Lipofectamine 2000 (Invitrogen), according facturer’s instructions (Invitrogen). The full-length Ndfip1 expression clones to the manufacturer’s instructions (14). were constructed first by PCR subcloning of Ndfip1 into the pENTR-3C entry vector (Invitrogen) between EcoRI and XhoI, followed by recombination by Protein identification by liquid chromatography-tandem mass Gateway LR cloning into pLenti6.2/N-Lumio/V5-DEST (Invitrogen). Ex- D D spectrometry pression clones for Ndfip1- 1–41 and Ndfip1- PY deletion mutants were similarly prepared first by PCR subcloning into pENTR-3C entry vector Immunoprecipitated proteins were separated by SDS-PAGE. In-gel di- (Invitrogen), followed by recombination by LR into pLenti6.2/N-Lumio/V5- gestion with trypsin, followed by protein identification using liquid DEST (Invitrogen). The ViraPower lentiviral expression system (Invitrogen) chromatography-tandem mass spectrometry, was performed as described was used for lentiviral transduction of bone marrow–derived macrophages elsewhere (16). Briefly, tryptic peptides were resolved on a nano-LC (BMDMs). First, 293FT cells were used for packaging and production of column (Magic AQ C18; Michrom Bioresources) and introduced into an viruses. BMDMs were incubated with the virus-containing supernatants Orbitrap mass spectrometer (Thermo Scientific). The Orbitrap was set to overnight at 37˚C in a 5% CO2 incubator. collect a high-resolution MS1 (FWHM 30,000@400 m/z), followed by NTHi lysate preparation and exposure data-dependent collision-induced dissociation spectra on the “top 9” ions in the linear ion trap. Spectra were searched against a human protein da- NTHi lysates were prepared as described earlier (18). In brief, bacteria were tabase (UniProt release 2011_05) using the X!Tandem/TPP software suite grown on chocolate agar plates at 37˚C in the presence of 5% CO2 and (17). Proteins identified with a ProteinProphet probability $ 0.9 false inoculated in brain heart infusion broth supplemented with 3.5 mg/ml NAD discovery rate , 2% were considered for further analysis. and hemin. The bacterial lysate was prepared by sonication. The final protein concentration was adjusted to 2.5 mg/ml in PBS, and aliquots were Abs and reagents frozen at 280˚C. Mouse inhalation of NTHi lysate was performed with an The following Abs were used in this study: anti–c-Myc (sc-40; Santa Cruz), anti- Aeromist CA-209 nebulizer (CIS-US, Bedford, MA). In brief, a thawed Flag (F1804, F7425; Sigma), anti-Itch (611198; BD), anti–b-actin (A5441; NTHi lysate was placed in the nebulizer driven by 10 l/min of room air Sigma), anti-Ub (sc-8017; Santa Cruz), anti-Xpress (R910-25; Invitrogen), anti- supplemented with 5% CO2 for 20 min. HA (sc-805; Santa Cruz), anti-UBE2G2 (sc-100613; Santa Cruz), anti-V5 Ab Lung inflammation was scored by examining H&E-stained lung sections (R960-25; Invitrogen), anti-Tak1 (Sc-7976; Santa Cruz), and Lys48-specific using a previously described five-point lung injury scoring system (peri- anti-Ub (05-1307; Millipore). The Tak1 inhibitor [(5Z)-7-Oxozeaenol, 3064] vascular and peribronchial inflammation, hyaline membranes, alveolar and was purchased from Tocris Bioscience. Ready-SET-Go! ELISA kits for mouse interstitial infiltrates, and alveolar hemorrhage) (19). IL-6, TNF, and IL-1b were purchased from eBioscience. TLR2/TLR4 inhibitor (OxPAPC, tlrl-oxpl) was purchased from Invitrogen. Real-time PCR analysis Ubiquitination assay Total RNAwas prepared using an RNeasy Mini Kit (QIAGEN), followed by cDNA synthesis using a Verso cDNA Kit (Thermo Scientific). Quantitative 293T cells were transfected with Flag-Tak1 and various constructs, as real-time PCR was performed on a Mastercycler ep realplex (Eppendorf). indicated. MG132 was added 4 h before cell lysis. Cells were washed three LightCycler 480 SYBR Green I Master reaction mix (Roche) was used in times with PBS and lysed in Nonidet P-40 lysis buffer. Immunoprecipitation a 20-ml reaction volume. The expression of individual genes was normalized was performed using anti-Flag Ab. Tak1-associated Ub was analyzed by to the expression of actin. Cycling conditions were 95˚C for 2 min, followed immunoblot using Ab against HA, as we described previously (14). by 50 cycles of 95˚C for 15 s, 55˚C for 15 s, and 72˚C for 20 s.

FIGURE 1. NTHi induces severe lung in- flammation in Itch2/2 and Ndfip12/2 mice. (A) H&E staining of WT, Itch2/2, and Ndfip12/2 mice lungs with and without multiple exposures to NTHi lysate. Scale bar, 25 mm. (B) Inflam- mation score. (C) Relative mRNA levels of proinflammatory cytokines IL-6, IL-1b, and TNF from the lung tissue of WT, Itch2/2, and Ndfip12/2 mice. Data are representative of three or more independent experiments. 2162 Ndfip1 REGULATES ITCH LIGASE ACTIVITY VIA UbcH7

Statistical analysis Tak1 for ubiquitination (14), we sought to get molecular insights into Data were analyzed with GraphPad Prism 4 software to determine statistical Tak1 ubiquitination by a proteomics approach using mass spec- significance using one-way ANOVA. Data are expressed as mean 6 SD. trometry. We transiently transfected 293T cells with Myc-Itch, HA- A p value , 0.05 was considered significant. Ub, and Flag-Tak1 and immunoprecipitated Tak1 using anti-Flag Ab. The immunoprecipitates were subjected to mass spectrometry analysis following in-gel trypsin digestion, and Ndfip1 was identified Results 2/2 2/2 in the Tak1 complex (Supplemental Fig. 1). Itch and Ndfip1 mice are hyperresponsive to NTHi- Because Ndfip1 was shown to interact with Itch (11), and defi- induced lung inflammation ciency of Ndfip1 resulted in chronic inflammation (11), we hy- It was demonstrated that NTHi activated the Tak1–NF-kB pathway pothesized that Ndfip1 may cooperate with Itch to limit lung and initiated an inflammatory response in the lung (13). Tak1 acti- inflammation. To test this hypothesis, we adopted a mouse model of vation was shown to be regulated by ubiquitination of lysine residues NTHi-induced airway inflammation. WT, Itch2/2,andNdfip12/2 within Tak1 (20–24). Because we reported earlier that Itch targets mice were exposed to aerosolized UV-treated NTHi lysate weekly

FIGURE 2. Elevated proinflammatory cytokine production by Itch2/2 and Ndfip12/2 BMDMs in response to NTHi-P6. (A) ELISA of IL-6, IL-1b, and TNF from the BMDMs of WT, Itch2/2, and Ndfip12/2 mice treated with NTHi outer membrane protein P6 for 24 h. (B) Relative mRNA levels of IL-6, IL- 1b, and TNF from WT, Itch2/2,andNdfip12/2 BMDMs treated with P6 in the presence or absence of 5Z-7-Oxozeaenol (50 nM). Data represent three inde- pendent experiments. (C) ELISA of IL-6, IL-1b, and TNF from culture supernatants of WT, Itch2/2,andNdfip12/2 BMDMs treated with P6 in the presence or absence of 5Z-7-Oxozeaenol (50 nM). Data represent three independent experiments. (D) ELISA of IL-6, IL-1b, and TNF from culture supernatants of WT, Itch2/2,andNdfip12/2 BMDMs treated with P6 in the presence or absence of OxPAPC (30 mg/ml). Data represent three independent experiments. The Journal of Immunology 2163 for 4 wk, as previously described (18). Histological analysis of TLR2 pathway during NTHi infection. We incubated WT, Itch2/2, the lung tissue showed aggravated inflammation in Itch2/2 and and Ndfip12/2 BMDMs with purified P6 for 24 h. The culture 2 2 Ndfip1 / mice compared with WT mice (Fig. 1A, 1B). Next, we supernatants were assayed for IL-6, TNF, and IL-1b by ELISA. As analyzed the expression of proinflammatory cytokines in the lung shown in Fig. 2A, both Itch2/2 and Ndfip12/2 macrophages secreted tissues by real-time PCR. As shown in Fig. 1C, increased levels of elevated levels of proinflammatory cytokines compared with WT 2 2 2 2 TNF, IL-6, and IL-1b were observed in Itch / and Ndfip1 / lung macrophages. To confirm that the elevated proinflammatory cytokine tissues exposed to NTHi, corroborating the histopathological phe- expression was through the P6-TLR2 pathway (13), we used a TLR2 notype of enhanced pulmonary inflammation. inhibitor, OxPAPC (29, 30). We found that OxPAPC significantly Because myeloid-derived proinflammatory cytokines play crucial inhibited P6-induced proinflammatory cytokines in Itch2/2 and roles in NTHi-induced lung inflammation and chronic obstructive Ndfip12/2 cells (Fig. 2B). To investigate whether a defect in the + pulmonary disease (18, 25–28), we used FACS to sort CD11b cells regulation of Tak1 led to enhanced production of proinflam- 2/2 2/2 from the lung tissues of WT, Itch ,andNdfip1 mice. Ex- matory cytokines, we preincubated the BMDMs with (5Z)-7- pression of proinflammatory cytokines was analyzed by real-time Oxozeaenol, a Tak1 selective inhibitor (31). We found that (5Z)- PCR. As shown in Supplemental Fig. 2A, we found significantly 7-Oxozeaenol treatment markedly inhibited proinflammatory 2/2 higher expression levels of IL-6, IL-1b, and TNF in Itch and cytokine expression in Itch2/2 and Ndfip12/2 cells in response 2/2 Ndfip1 cells than in WT cells. These results suggested that Itch to P6 (Fig. 2C, 2D). Further, (5Z)-7-Oxozeaenol treatment and Ndfip1 negatively regulated the production of proinflammatory inhibited proinflammatory cytokine expression by NTHi-infected cytokines during NTHi infection. macrophages (Supplemental Fig. 2B). Next, we investigated whether Itch and Ndfip1 associate with NTHi-P6 promotes Itch and Ndfip1 association Tak1. We transiently transfected 293T cells with Flag-Tak1, Flag- TheoutermembranelipoproteinP6wasshowntobindtoTLR2and Itch, and Myc-Ndfip1 and immunoprecipitated the cell lysates with activate NF-kB via Tak1 (13). Therefore, we investigated whether anti-Flag Ab. The immunoprecipitates were blotted with anti-Myc Itch and Ndfip1 regulated signaling pathways downstream of the P6- Ab and then reprobed with anti-Flag Ab. As shown in Fig. 3A, we

FIGURE 3. Ndfip1 forms a complex with Tak1 and enhances Itch ligase activity. (A) 293T cells were transfected with Flag-Itch, Myc-Ndfip1, and Flag- Tak1 plasmids. Cell lysates were immunoprecipitated with anti-Flag Ab and immunoblotted using anti-Myc and anti-Flag Abs. (B) Immunoassay of lysates from P6-treated WT BMDMs, followed by immunoprecipitation with anti-Tak1 Ab and immunoblot analysis with anti-Itch, anti-Ndfip1, and anti-Tak1 Abs. (C) Immunoassay of lysates from NTHi-treated WT BMDMs, followed by immunoprecipitation with anti-Tak1 Ab and immunoblot analysis with anti- Itch, anti-Ndfip1, and anti-Tak1 Abs. (D) Immunoassay of BMDMs isolated from WT and Ndfip12/2 mice treated with NTHi outer membrane protein P6 for 30 min, followed by immunoprecipitation of lysates with anti-Tak1 Ab and immunoblot analysis with Apu2 Ab (anti-Ub Lys 48 specific). (E) Im- munoassay of 293T cells transfected with various combinations of Flag-Tak1, Myc-Itch, Myc-Ndfip1, HX-Itch (C830A), and HA-UbK48 expression vectors, followed by immunoprecipitation of lysates with anti-Flag Ab and immunoblot with anti-HA Ab. 2164 Ndfip1 REGULATES ITCH LIGASE ACTIVITY VIA UbcH7 found that Tak1, Itch, and Ndfip1 were coprecipitated. Next, we in- Ndfip1 acts as an adaptor for UbcH7 and Itch vestigated whether P6 induced the formation of the Itch-Ndfip-Tak1 The function of Itch is believed to be regulated at the point of substrate complex. We stimulated BMDMs with P6 for 30 min and performed recognition through the WW domains (32). Itch ligase activity was coimmunoprecipitation experiments using anti-Itch, anti-Tak1, and shown to be enhanced by JNK1-mediated phosphorylation (32, 33). anti-Ndfip1 Abs. As shown in Fig. 3B, Itch, Tak1, and Ndfip1 Because Itch is a HECT-type E3 ligase, which requires the formation coprecipitated in P6-treated cells, suggesting a trimeric association. of a thioester covalent attachment of Ub to a highly conserved Ndfip1 enhances ubiquitination of Tak1 cysteine residue in its HECT domain by the E2 enzyme (5), we To investigate whether Ndfip1 regulates Itch-mediated Tak1 ubiq- hypothesized that Ndfip1 may regulate Itch ligase activity by re- uitination, we precipitated Tak1 from WT and Ndfip12/2 BMDMs cruiting Ub-loaded E2. To test this hypothesis, we checked whether following P6 stimulation. We immunoblotted the membranes with Ndfip1 interacts with UbcH7, the E2 that forms the thioester linkage Apu2 Ab, which specifically detects K48-linked Ub chains. We for Itch (34–36). Interestingly, we found that UbcH7 coprecipitated observed Tak1 ubiquitination in WT cells upon P6 stimulation but with Ndfip1 in P6-stimulated BMDMs (Fig. 4A). To confirm the not in Ndfip12/2 cells (Fig. 3C), suggesting an essential role for specificity of UbcH7 as an E2 that is recruited by Ndfip1, we per- Ndfip1 in Tak1 ubiquitination. formed coimmunoprecipitation using another E2, UBE2G2. As Because it was demonstrated that Ndfip1 enhances Itch ligase showninFig.4B,Ndfip1failedtointeractwithUBE2G2. activity and regulates JunB ubiquitination during Th2 cell differ- To further confirm that Ndfip1 recruits UbcH7 to Itch and 2/2 entiation (11, 15), we tested whether Ndfip1 also enhances Tak1 facilitates Tak1 ubiquitination, we stimulated WT and Ndfip1 ubiquitination. We transfected 293T cells with different combina- BMDMs with P6. The cell lysates were immunoprecipitated with tions of Flag-Tak1, Myc-Itch, Myc-Ndfip1, HA-UbK48 (in which anti-Itch Ab, and the membranes were immunoblotted with anti- all of the lysine residues except K48 were mutated), and HX-Itch- UbcH7 and anti-Ndfip1 Abs. UbcH7 immunoprecipitated with 2/2 C830A (ligase mutant). We immunoprecipitated the cell lysates Itch in WT cells but not in Ndfip1 cells (Fig. 4C). Similarly, with anti-Flag Ab and blotted the membranes with anti-HA Ab. As UbcH7 coimmunoprecipitated with Itch in the WT lung tissue ex- 2/2 shown in Fig. 3D, increased levels of polyubiquitinated Tak1 were posed to NTHi lysate but not in the tissue from Ndfip1 mice, observed when Itch, Ndfip1, and Tak1 were coexpressed. This which provided further in vivo evidence that Ndfip1 plays a crucial suggests that Ndfip1 promotes Itch ligase activity and enhances role in the recruitment of UbcH7 to Itch (Fig. 4D). We also checked K48-linked ubiquitination of Tak1. We also tested whether Ndfip1 whether a TLR4-induced signal could help Ndfip1 recruit UbcH7. promoted K63-linked ubiquitination by transfecting Itch in the Interestingly, we found that UbcH7 coprecipitated with Ndfip1 in presence of Ndfip1 and an Ub mutant in which all of the lysine LPS-stimulated BMDMs (Supplemental Fig. 3B). residues except K63 were mutated (HA-UbK63). As shown in To investigate whether Ndfip1 acts as an adaptor for Itch and Supplemental Fig. 3A, we did not detect K63-linked ubiquitination. UbcH7, we generated Ndfip1-deletion mutants lacking the first 76 Together, these data suggest that Ndfip1 promotes K48-linked, but N-terminal amino acids (Ndfip1-DN), lacking aa 77–143 in the middle not K63-linked, ubiquitination of Tak1. (Ndfip1-DM), and lacking aa 144–221 in the C terminus (Ndfip1-DC)

FIGURE 4. Ndfip1 acts as an adaptor for UbcH7 and Itch. (A) Immunoassay of lysates from P6-treated WT BMDMs, followed by immunoprecipitation with anti-Ndfip1 Ab and mouse IgG and im- munoblot analysis with anti-UbcH7 or anti-Ndfip1 Abs. (B) Immunoassay of lysates from P6-treated WT BMDMs, followed by immunoprecipitation with anti-Ndfip1 Ab and immunoblot analy- sis with anti-UbcH7, anti-UBEG2, and anti-Ndfip1 Abs. (C) Immunoassay of lysates from P6-treated WT and Ndfip12/2 BMDMs, followed by immunoprecipita- tion with anti-Itch Ab and immunoblot analysis with anti-UbcH7, anti-Ndfip1, or anti-Itch Abs. (D) Lysates were prepared from WT and Ndfip12/2 lungs post-NTHi exposure and immunoprecipitated with anti-Itch Ab, followed by immunoblot analysis with anti-UbcH7, anti-Ndfip1, or anti-Itch Ab. The Journal of Immunology 2165

FIGURE 5. N-terminal end of Ndfip1 interacts with UbcH7. (A) 293T cells were transfected with Myc-Ndfip1, Myc–Ndfip1-DN, Myc–Ndfip1-DM, Myc– Ndfip1-DC, and HA-UbcH7 plasmids. Cells lysates were immunoprecipitated with anti-Myc Ab and immunoblotted using anti-HA and anti-Myc Abs. (B) 293T cells were cotransfected with Myc-Ndfip1, Myc–Ndfip1-DN, and Flag-Itch plasmids. Cells lysates were immunoprecipitated with anti-Flag Ab and immunoblotted using anti-Flag and anti-Myc Abs. (C) 293T cells were transfected with Myc-Ndfip1, Myc–Ndfip1-D1–41, Myc–Ndfip1-DPY, Flag-Itch, and HA- UbcH7 plasmids. Cell lysates were immunoprecipitated with anti-Myc Ab and immunoblotted using anti-HA, anti-Flag, and anti-Myc Abs. (D) Ndfip2/2 BMDMs were reconstituted with full-length Ndfip1, Ndfip1-D1–41, or Ndfip1-DPY via lentiviral transduction. Cells lysates were immunoprecipitated with anti-V5 Ab and immunoblotted using anti-Itch, anti-UbcH7, and anti-V5 Abs. (E) Ndfip12/2 BMDMs were reconstituted with full-length Ndfip1, Ndfip1-D1–41, or Ndfip1-DPY via lentiviral transduction. Cell lysates were immunoprecipitated with anti-Tak1 Ab and immunoblotted using Apu2 (anti-Ub Lys 48 specific) and anti-Tak1 Abs. (F) Relative mRNA levels of proinflammatory cytokines IL-6, IL-1b, and TNF from BMDMs isolated from Ndfip12/2 mice reconstituted with full-length Ndfip1, Ndfip1-D1–41, or Ndfip1-DPY via lentiviral transduction and stimulated with P6 for 4 h. (G) ELISA of proinflammatory cytokines IL-6, IL-1b, and TNF from the culture supernatants of BMDMs isolated from Ndfip12/2 mice reconstituted with full-length Ndfip1, Ndfip1-D1–41, or Ndfip1-DPY via lentiviral transduction and stimulated with P6 for 24 h. (H) Model depicting the adaptor function of Ndfip1 for Itch and UbcH7 during Tak1 ubiquitination. n.s., not significant. 2166 Ndfip1 REGULATES ITCH LIGASE ACTIVITY VIA UbcH7

(Supplemental Fig. 4A). To investigate which region of Ndfip1 and IL-1b, which is mediated by MAPK and NF-kB through binds to UbcH7, we transiently transfected HA-UbcH7 along with TLRs (40). It was demonstrated that the NTHi membrane lipo- either the full-length Myc-Ndfip1 or its deletion mutants. The cell protein P6 activates Tak1, the upstream kinase that is responsible lysate was immunoprecipitated with anti-Myc Ab and blotted with for NF-kB activation (13, 41). We demonstrated earlier that Itch anti-HA Ab. As shown in Fig. 5A, full-length Ndfip1, Ndfip1-DM, forms an Ub-editing complex with the deubiquitinase Cyld. This and Ndfip1-DC coprecipitated with UbcH7. However, Ndfip1-DN complex sequentially cleaves the K63-linked Ub chains on Tak1 failed to bind to UbcH7, suggesting that UbcH7 binds at the and catalyzes K48-linked ubiquitination to promote Tak1 degra- N terminus of Ndfip1. The N terminus of Ndfip1 also contained dation (14). In this article, we demonstrate that Ndfip1 promotes the potential Itch-interacting PY motifs (between aa 42 and 76; Itch-mediated K48-linked ubiquitination by acting as an adaptor Supplemental Fig. 4A). To test whether Ndfip1 interacted with for UbcH7 and Itch (Fig. 5G). It is not known whether P6 induces Itch through these PY motifs, we coexpressed Itch with either full- K63-linked ubiquitination to activate Tak1 during the early phase length Ndfip1 or the Ndfip1-DN mutants. Itch coprecipitated with of infection. Such a phenomenon was reported in the case of Hel- full-length Ndfip1 but not Ndfip1-DN, suggesting that Itch binds to icobacter pylori infection, in which TRAF6 catalyzes K63-linked the PY motifs on Ndfip1 (Fig. 5B). To further narrow down the ubiquitination of Tak1 (20). It is possible that the Itch-Cyld complex UbcH7-interacting region in the N terminus of Ndfip1, we generated switches K63-linked ubiquitination to K48-linked ubiquitination on Ndfip1 deletion mutants lacking aa 1–41 (Ndfip1-D1–41) and lacking Tak1 to inhibit NTHi-induced airway inflammation, and Ndfip1 the PY motif (aa 42–76; Ndfip1-DPY). We transfected 293T cells may facilitate Itch function by recruiting Ub-loaded UbcH7. with UbcH7 and Itch along with full-length Ndfip1, Ndfip1-D1–41, Itch function was proposed to be regulated at the point of or Ndfip1-DPY. As shown in Fig. 5C, Itch and UbcH7 coprecipitated substrate recognition through its WW domains and via intra- in the presence of full-length Ndfip1 but not the Ndfip1-D1–41 mu- molecular autoinhibition (32). The autoinhibitory interaction be- tant. This suggested that UbcH7 binds to the first 41 N-terminal tween the WW domains and the HECT domain could prevent E2 amino acids, whereas Itch binds to the PY motifs (Fig. 5C). binding or inhibit the transthiolation reaction between E2 and E3 (42). Upon activation of the cells, JNK1-mediated phosphoryla- Reconstitution of full-length Ndfip1, but not Ndfip1 mutants, tion was shown to cause a conformational change that results in rescues elevated proinflammatory cytokine expression by the destabilization of the inhibitory intramolecular interactions 2 2 Ndfip1 / cells and promotes Itch ligase activity (33). Additionally, Ndfip1 was To further investigate the role of Ndfip1 as an adaptor for UbcH7 shown to enhance the Itch ligase activity (11). Our results suggest and Itch during NTHi-induced inflammatory responses, we recon- that Ndfip1 does it by recruiting UbcH7, suggesting that there is stituted Ndfip12/2 BMDMs with full-length Ndfip1, Ndfip1-D1–41, additional regulation of Itch function at the level of E2 recruit- 2/2 or Ndfip1-DPY via lentiviral transduction. We stimulated the cells ment. Our results from reconstitution of Ndfip1 cells with with P6 and analyzed Itch, Ndfip1, and UbcH7 interactions by Ndfip1 truncation mutants suggest that interaction of Ndfip1 with coimmunoprecipitation experiments. As shown in Fig. 5D, full- Itch, as well as UbcH7, is essential for effective termination of the length Ndfip1 interacted with Itch and UbcH7 in P6-treated cells. inflammatory response. However, it remains unclear whether the Ndfip1-DPY interacted with UbcH7 but not with Itch. Conversely, Itch–Ndfip1 interaction or the UbcH7–Ndfip1 interaction plays Ndfip1-D1–41 interacted with Itch but not with UbcH7. We further a predominant role. tested the effect of reconstituting Ndfip12/2 BMDMs on Tak1 Regulation of Itch ligase activity by Ndfip1 may not be re- ubiquitination. The transduced Ndfip12/2 cells were stimulated with stricted to the TLR2 pathway. Our results suggest that LPS also P6 protein, and the endogenous Tak1 was immunoprecipitated using promoted an Ndfip1–UbcH7 interaction in BMDMs (Sup- anti-Tak1 Ab, followed by immunoblotting with Apu2 Ab. As plemental Fig. 2B). Because NTHi is a Gram-negative bacte- shown in Fig. 5E, reconstitution of full-length Ndfip1, but not the rium with LPS as a component of its cell wall, it is likely that deletion mutants, restored the defect in Tak1 ubiquitination in Itch and Ndfip1 also regulate Tak1 ubiquitination to down- Ndfip12/2 cells. We also analyzed the level of proinflammatory modulate TLR4-induced Tak1 activation in NTHi-infected macro- cytokine expression in Ndfip12/2 BMDMs. As shown in Fig. 5F, phages. Whether Ndfip1 regulates Tak1 in other signaling pathways 5G, and Supplemental Fig. 4B, reconstitution of Ndfip12/2 cells remains unknown. Furthermore, it remains unclear whether the with full-length Ndfip1, but not with Ndfip1-D1–41 or Ndfip1-DPY interaction among Itch, Ndfip1, UbcH7, and Tak1 is direct. Further mutants, rescued elevated proinflammatory cytokine expression detailed molecular and genetic analyses will be required to deter- by Ndfip12/2 BMDMs. These results suggest that Ndfip1 plays a mine the exact mechanisms of interaction among these proteins. crucial role as an adaptor molecule bridging Itch and UbcH7 to A better understanding of Ndfip1-mediated regulation of Itch ligase activity is essential for inhibiting or enhancing Itch function in in- facilitate termination of inflammation signaling (Fig. 5H). flammatory pathways.

Discussion Acknowledgments By virtue of its target substrates, Itch plays a crucial role in in- We thank Drs. Sangkon Oh and Carson Harrod for critical reading of the flammatory signaling pathways (32, 37), but how its function is manuscript, Dr. Andrea Cooper for helpful discussions, Dr. Venkatesha controlled remains unclear. In this article, we demonstrate that Basrur for MS analysis, and Dr. Balamayooran Theivanthiran for histolog- Itch ligase activity is regulated by Ndfip1 by facilitating the re- ical analysis of lung tissues. cruitment of the E2 enzyme UbcH7. Itch or Ndfip1 deficiency led to elevated airway inflammation in response to NTHi as a result of defects in Itch-mediated Tak1 ubiquitination. Disclosures NTHi is associated with chronic upper and lower respiratory The authors have no financial conflicts of interest. disease and is the dominant species isolated from the lower airways of children and adults with chronic respiratory symptoms (38, 39). References The inflammatory response after NTHi infection is characterized 1. Akira, S., S. Uematsu, and O. Takeuchi. 2006. Pathogen recognition and innate by the upregulation of proinflammatory cytokines, such as TNF immunity. Cell 124: 783–801. The Journal of Immunology 2167

2. Grivennikov, S. I., F. R. Greten, and M. Karin. 2010. Immunity, inflammation, lysine 158 is required for tumor necrosis factor alpha- and interleukin-1beta- and cancer. Cell 140: 883–899. induced IKK/NF-kappaB and JNK/AP-1 activation. J. Biol. Chem. 285: 5347– 3. Bhoj, V. G., and Z. J. Chen. 2009. Ubiquitylation in innate and adaptive im- 5360. munity. Nature 458: 430–437. 25. Berenson, C. S., C. T. Wrona, L. J. Grove, J. Maloney, M. A. Garlipp, 4. Venuprasad, K., C. Yang, Y.Shao, D. Demydenko, Y.Harada, M. S. Jeon, and Y.C. Liu. P. K. Wallace, C. C. Stewart, and S. Sethi. 2006. Impaired alveolar macrophage 2006. Immune regulation by ubiquitin conjugation. Adv. Exp. Med. Biol. 584: 207–217. response to Haemophilus antigens in chronic obstructive lung disease. Am. J. 5. Pickart, C. M. 2001. Mechanisms underlying ubiquitination. Annu. Rev. Bio- Respir. Crit. Care Med. 174: 31–40. chem. 70: 503–533. 26. Caramori, G., I. M. Adcock, A. Di Stefano, and K. F. Chung. 2014. Cytokine 6. Perry, W. L., C. M. Hustad, D. A. Swing, T. N. O’Sullivan, N. A. Jenkins, and inhibition in the treatment of COPD. Int. J. Chron. Obstruct. Pulmon. Dis. 9: N. G. Copeland. 1998. The itchy locus encodes a novel ubiquitin protein ligase 397–412. that is disrupted in a18H mice. Nat. Genet. 18: 143–146. 27. Rotta Detto Loria, J., K. Rohmann, D. Droemann, P. Kujath, J. Rupp, 7. Fang, D., C. Elly, B. Gao, N. Fang, Y. Altman, C. Joazeiro, T. Hunter, N. Copeland, T. Goldmann, and K. Dalhoff. 2013. Nontypeable Haemophilus Influenzae In- N. Jenkins, and Y. C. Liu. 2002. Dysregulation of T lymphocyte function in itchy fection Upregulates the NLRP3 Inflammasome and Leads to Caspase-1- mice: a role for Itch in TH2 differentiation. Nat. Immunol. 3: 281–287. Dependent Secretion of Interleukin-1b - A Possible Pathway of Exacerbations 8. Lohr, N. J., J. P. Molleston, K. A. Strauss, W. Torres-Martinez, E. A. Sherman, in COPD. PLoS ONE 8: e66818. R. H. Squires, N. L. Rider, K. R. Chikwava, O. W. Cummings, D. H. Morton, and 28. Gaschler, G. J., M. Skrtic, C. C. Zavitz, M. Lindahl, P. O. Onnervik, E. G. Puffenberger. 2010. Human ITCH E3 ubiquitin ligase deficiency causes T. F. Murphy, S. Sethi, and M. R. Sta¨mpfli. 2009. Bacteria challenge in smoke- syndromic multisystem autoimmune disease. Am. J. Hum. Genet. 86: 447–453. exposed mice exacerbates inflammation and skews the inflammatory profile. Am. 9. Jolliffe, C. N., K. F. Harvey, B. P. Haines, G. Parasivam, and S. Kumar. 2000. J. Respir. Crit. Care Med. 179: 666–675. Identification of multiple proteins expressed in murine embryos as binding partners for 29. Erridge, C., S. Kennedy, C. M. Spickett, and D. J. Webb. 2008. Oxidized the WW domains of the ubiquitin-protein ligase Nedd4. Biochem. J. 351: 557–565. phospholipid inhibition of toll-like receptor (TLR) signaling is restricted to 10. Harvey, K. F., L. M. Shearwin-Whyatt, A. Fotia, R. G. Parton, and S. Kumar. TLR2 and TLR4: roles for CD14, LPS-binding protein, and MD2 as targets for 2002. N4WBP5, a potential target for ubiquitination by the Nedd4 family of specificity of inhibition. J. Biol. Chem. 283: 24748–24759. proteins, is a novel Golgi-associated protein. J. Biol. Chem. 277: 9307–9317. 30. Li, D., H. Lei, Z. Li, H. Li, Y. Wang, and Y. Lai. 2013. A novel lipopeptide from 11. Oliver, P. M., X. Cao, G. S. Worthen, P. Shi, N. Briones, M. MacLeod, J. White, skin commensal activates TLR2/CD36-p38 MAPK signaling to increase anti- P. Kirby, J. Kappler, P. Marrack, and B. Yang. 2006. Ndfip1 protein promotes the bacterial defense against bacterial infection. PLoS ONE 8: e58288. function of itch ubiquitin ligase to prevent T cell activation and T helper 2 cell- 31. Ninomiya-Tsuji, J., T. Kajino, K. Ono, T. Ohtomo, M. Matsumoto, M. Shiina, mediated inflammation. Immunity 25: 929–940. M. Mihara, M. Tsuchiya, and K. Matsumoto. 2003. A resorcylic acid lactone, 12. Sethi, S., and T. F. Murphy. 2008. Infection in the pathogenesis and course of 5Z-7-oxozeaenol, prevents inflammation by inhibiting the catalytic activity of chronic obstructive pulmonary disease. N. Engl. J. Med. 359: 2355–2365. TAK1 MAPK kinase kinase. J. Biol. Chem. 278: 18485–18490. 13. Shuto, T., H. Xu, B. Wang, J. Han, H. Kai, X. X. Gu, T. F. Murphy, D. J. Lim, 32. Melino, G., E. Gallagher, R. I. Aqeilan, R. Knight, A. Peschiaroli, M. Rossi, and J. D. Li. 2001. Activation of NF-kappa B by nontypeable Hemophilus F. Scialpi, M. Malatesta, L. Zocchi, G. Browne, et al. 2008. Itch: a HECT-type influenzae is mediated by toll-like receptor 2-TAK1-dependent NIK-IKK alpha E3 ligase regulating immunity, skin and cancer. Cell Death Differ. 15: 1103– /beta-I kappa B alpha and MKK3/6-p38 MAP kinase signaling pathways in 1112. epithelial cells. Proc. Natl. Acad. Sci. USA 98: 8774–8779. 33. Gallagher, E., M. Gao, Y. C. Liu, and M. Karin. 2006. Activation of the E3 14.Ahmed,N.,M.Zeng,I.Sinha,L.Polin,W.Z.Wei,C.Rathinam,R.Flavell, ubiquitin ligase Itch through a phosphorylation-induced conformational change. R. Massoumi, and K. Venuprasad. 2011. The E3 ligase Itch and deubiquitinase Cyld Proc. Natl. Acad. Sci. USA 103: 1717–1722. act together to regulate Tak1 and inflammation. Nat. Immunol. 12: 1176–1183. 34. Raimondo, D., A. Giorgetti, F. Bernassola, G. Melino, and A. Tramontano. 2008. 15. Mund, T., and H. R. Pelham. 2009. Control of the activity of WW-HECT domain E3 ubiquitin ligases by NDFIP proteins. EMBO Rep. 10: 501–507. Modelling and molecular dynamics of the interaction between the E3 ubiquitin 16. Maine, G. N., H. Li, I. W. Zaidi, V. Basrur, K. S. Elenitoba-Johnson, and ligase Itch and the E2 UbcH7. Biochem. Pharmacol. 76: 1620–1627. E. Burstein. 2010. A bimolecular affinity purification method under denaturing 35. Zhang, P., C. Wang, K. Gao, D. Wang, J. Mao, J. An, C. Xu, D. Wu, H. Yu, conditions for rapid isolation of a ubiquitinated protein for mass spectrometry J. O. Liu, and L. Yu. 2010. The ubiquitin ligase itch regulates apoptosis by analysis. Nat. Protoc. 5: 1447–1459. targeting thioredoxin-interacting protein for ubiquitin-dependent degradation. J. 17. Nesvizhskii, A. I., A. Keller, E. Kolker, and R. Aebersold. 2003. A statistical Biol. Chem. 285: 8869–8879. model for identifying proteins by tandem mass spectrometry. Anal. Chem. 75: 36. Rossi, M., B. Rotblat, K. Ansell, I. Amelio, M. Caraglia, G. Misso, 4646–4658. F. Bernassola, C. N. Cavasotto, R. A. Knight, A. Ciechanover, and G. Melino. 18. Moghaddam, S. J., C. G. Clement, M. M. De la Garza, X. Zou, E. L. Travis, 2014. High throughput screening for inhibitors of the HECT ubiquitin E3 ligase H. W. Young, C. M. Evans, M. J. Tuvim, and B. F. Dickey. 2008. Haemophilus ITCH identifies antidepressant drugs as regulators of autophagy. Cell Death Dis. influenzae lysate induces aspects of the chronic obstructive pulmonary disease 5: e1203. phenotype. Am. J. Respir. Cell Mol. Biol. 38: 629–638. 37. Venuprasad, K. 2010. Cbl-b and itch: key regulators of peripheral T-cell toler- 19. Bellmeyer, A., J. M. Martino, N. S. Chandel, G. R. Scott Budinger, D. A. Dean, ance. Cancer Res. 70: 3009–3012. and G. M. Mutlu. 2007. Leptin resistance protects mice from hyperoxia-induced 38. Kapur, N., K. Grimwood, I. B. Masters, P. S. Morris, and A. B. Chang. 2012. acute lung injury. Am. J. Respir. Crit. Care Med. 175: 587–594. Lower airway microbiology and cellularity in children with newly diagnosed 20. Lamb, A., X. D. Yang, Y. H. Tsang, J. D. Li, H. Higashi, M. Hatakeyama, non-CF bronchiectasis. Pediatr. Pulmonol. 47: 300–307. R. M. Peek, S. R. Blanke, and L. F. Chen. 2009. Helicobacter pylori CagA 39. King, P. T., S. R. Holdsworth, N. J. Freezer, E. Villanueva, and P. W. Holmes. activates NF-kappaB by targeting TAK1 for TRAF6-mediated Lys 63 ubiq- 2007. Microbiologic follow-up study in adult bronchiectasis. Respir. Med. 101: uitination. EMBO Rep. 10: 1242–1249. 1633–1638. 21. Liang, L., Y. Fan, J. Cheng, D. Cheng, Y. Zhao, B. Cao, L. Ma, L. An, W. Jia, 40. Moghaddam, S. J., C. E. Ochoa, S. Sethi, and B. F. Dickey. 2011. Nontypeable X. Su, et al. 2013. TAK1 ubiquitination regulates doxorubicin-induced NF-kB Haemophilus influenzae in chronic obstructive pulmonary disease and lung activation. Cell. Signal. 25: 247–254. cancer. Int. J. Chron. Obstruct. Pulmon. Dis. 6: 113–123. 22. Sakurai, H. 2012. Targeting of TAK1 in inflammatory disorders and cancer. 41. Chen, R., J. H. Lim, H. Jono, X. X. Gu, Y. S. Kim, C. B. Basbaum, T. F. Murphy, Trends Pharmacol. Sci. 33: 522–530. and J. D. Li. 2004. Nontypeable Haemophilus influenzae lipoprotein P6 induces 23. Fan, Y., Y. Yu, R. Mao, H. Zhang, and J. Yang. 2011. TAK1 Lys-158 but not Lys- MUC5AC mucin transcription via TLR2-TAK1-dependent p38 MAPK-AP1 and 209 is required for IL-1b-induced Lys63-linked TAK1 polyubiquitination and IKKbeta-IkappaBalpha-NF-kappaB signaling pathways. Biochem. Biophys. Res. IKK/NF-kB activation. Cell. Signal. 23: 660–665. Commun. 324: 1087–1094. 24. Fan, Y., Y. Yu, Y. Shi, W. Sun, M. Xie, N. Ge, R. Mao, A. Chang, G. Xu, 42. Kee, Y., and J. M. Huibregtse. 2007. Regulation of catalytic activities of HECT M. D. Schneider, et al. 2010. Lysine 63-linked polyubiquitination of TAK1 at ubiquitin ligases. Biochem. Biophys. Res. Commun. 354: 329–333.