The AP-1 transcription factor JunB is required for Th17 cell differentiation

Interleukin (IL)-17-producing T helper (Th17) cells are crucial for host defense against extracellular microbes and pathogenesis of autoimmune diseases. Here we show that the AP-1 transcription factor JunB is required for Th17 cell development. Junb-deficient CD4+ T cells are able to develop in vitro into various helper T subsets except Th17. The RNA-seq transcriptome analysis reveals that JunB is crucial for the Th17-specific gene expression program. Junb-deficient mice are completely resistant to experimental autoimmune encephalomyelitis, a Th17-mediated inflammatory disease, and naive T helper cells from such mice fail to differentiate into Th17 cells. JunB appears to activate Th17 signature genes by forming a heterodimer with BATF, another AP-1 factor essential for Th17 differentiation. The mechanism whereby JunB controls Th17 cell development likely involves activation of the genes for the Th17 lineage-specifying orphan receptors RORγt and RORα and reduced expression of Foxp3, a transcription factor known to antagonize RORγt function.

JunD are viable, c-Jun-deficient embryos die at embryonic day 12.5 (E12.5) with abnormalities in liver and heart, and JunB inactivation causes multiple defects in extra-embryonic tissues, leading to embryonic lethality at E8.5-10.5 19 . Although the three Jun proteins are each capable of effectively binding to the Th17-polarizing AP-1 factor BATF [15][16][17][18] , their role in Th17 development has remained to be elucidated.
In the present study, we show that JunB is required for Th17 cell differentiation. CD4 + T cells deficient in the JunB-encoding gene Junb fail to differentiate into Th17 cells. The critical role of JunB in generation of the Th17-specific gene expression pattern is presented by the RNA-seq transcriptome analysis. Junb-deficient mice are completely resistant to EAE. The dominant role of JunB as a partner of BATF is consistent with the findings that c-Jun is much less abundantly expressed in Th17 cells compared with JunB and JunD, and that JunB but not JunD cooperates with BATF to activate Th17 signature genes. JunB appears to control Th17 cell specification by inducing activation of Rorc and Rora and by reducing expression of Foxp3.

JunB-deficient T helper cells fail to differentiate into Th17 cells.
To know the mechanism underlying Th17 cell differentiation, we immunoprecipitated the Th17-polarizing transcription factor IκBζ and analyzed IκBζ-interacting proteins by liquid chromatography-tandem mass spectrometry (LC-MS/MS) and by various binding assays (Supplementary Figure 1). The analyses led to identification of JunB as a novel IκBζ-binding protein, raising a possibility that JunB also participates in Th17 development. Indeed JunB expression was markedly induced, when naive CD4 + T cells were activated via T cell receptor under Th17 cell-polarizing conditions (IL-6 and TGF-β) (Fig. 1A). To investigate the role of JunB in Th17 cell differentiation, we generated Junb f/f mice (Supplementary Figure 2A-F); the mice were crossed to Meox2 +/Cre mice for deletion of the Junb locus in the embryo proper but not in extraembryonic tissues, because conventional Junb-deficient mice are known to be embryonic lethal due to placental defects 20 .
Consistent with previous reports using similar Junb-deficient mice 20,21 , the present Junb-deficient mice also exhibited myeloproliferative abnormality (Supplementary Figure 3A) and impairment of osteoclast differentiation (Supplementary Figure 3B,C). Ablation of Junb did not affect development of naive CD4 + T cells (Supplementary Figure 3D,E). On the other hand, when Junb-deficient CD4 + T cells cultured under Th17-polarizing conditions, they expressed much less amounts of IL-17A (Fig. 1B) and Il17a mRNA (Fig. 1C) than control CD4 + T cells. Furthermore, expression of other Th17 signature genes encoding IL-17F (Il17f), IL-21 (Il21), and IL-23R (Il23r) was diminished in Junb-deficient T cells (Fig. 1C), indicating an essential role for JunB in Th17 cell differentiation. When OT-II mice-derived CD4 + T cells 22 were cultured with the chicken ovalbumin peptide OVA 323-339 and splenic antigen-presenting cells (APCs) under Th17-polarizing conditions, Junb-deficient OT-II T cells differentiated into IL-17A-producing cells much less efficiently than control cells (Fig. 1D). JunB appears to function in T cells but not in APCs, because differentiation of control naive T cells occurred even in the presence of Junb-deficient APCs (Fig. 1E). Thus, JunB likely plays a crucial role in Th17 cell differentiation.
JunB is crucial for Th17-specific gene expression program. To investigate the effect of JunB deficiency on global gene expression under Th17-polarizing conditions, we performed RNA-seq analyses. Consistent with the results obtained in the qPCR analyses (Fig. 1C), expression of the Th17 signature genes (Il17a,Il17f,Il21,and Il23r) was abrogated in the absence of JunB ( Fig. 2A). Pathway enrichment analyses revealed that genes involved in Th17-related functions were significantly enriched in the JunB-regulated gene set (Supplementary Figure 4). A strong correlation was observed between genes up-regulated during Th17 differentiation and those impaired by JunB deficiency (Fig. 2A, red dots): genes highly activated during Th17 differentiation had a tendency For each gene, log2FC between control and Junb-deficient Th17 cells are plotted against log2FC during Th17 differentiation from control naive or Th0 cells. Red and blue dots indicate genes significantly up-and down-regulated during Th17 differentiation, respectively (FDR < 0.001, |log2FC| > 1). (C) Dendrogram of unsupervised hierarchical clustering of the transcriptomes of control and Junb-deficient naive, Th0, and Th17 cells.
to be effectively impaired by the absence of JunB. We also found that expression of another subset of genes was down-regulated during Th17 differentiation in a JunB-dependent manner (Fig. 2B, blue dots): genes severely down-regulated during Th17 differentiation tended to be activated by JunB deficiency (Fig. 2B). The hierarchical clustering analysis revealed that lack of JunB drastically altered gene expression in CD4 + T cells after culture under Th17-polarizing conditions (Fig. 2C). By contrast, the absence of JunB did not largely affect expression patterns in naive T helper cells and Th0 cells (Fig. 2C). Importantly, the expression profile of Junb-deficient CD4 + T cells under Th17-polarizing conditions was more similar to that of normal Th0 cells than that of Th17 cells (Fig. 2C). Thus JunB is indispensable for generation of a gene expression pattern specific to Th17 cells.
JunB is dispensable for development of Th1, Th2, and Treg cells. In contrast to the essential role of JunB in Th17 differentiation, neither Th1 nor Th2 differentiation was affected by ablation of Junb: interferon-γ (IFN-γ) and IL-4 were normally synthesized under Th1-and Th2-polarizing conditions, respectively ( Fig. 3A-C). In addition, Foxp3 (encoding Foxp3), which specifies differentiation into Treg cells 1,2 , was expressed in Junb-deficient cells as much as in control cells under Treg-polarizing conditions (Fig. 3D). Intriguingly, Foxp3 expression under Th17-polarizing conditions was increased in Junb-deficient cells (Fig. 3D); a similar increase has been also observed by ablation of Irf4 and Batf, each being required for development of Th17 cells 7,8 . These findings indicate that JunB is selectively required for Th17 differentiation.
JunB-deficient mice are completely resistant to EAE induction. To examine the in vivo role of JunB in Th17 cell differentiation, we evaluated the effects of Junb ablation in EAE, because Th17 cells are the major pathogenic population in this disease 3,4 . Junb f/+ mice were immunized with myelin oligodendrocyte glycoprotein peptide 35-55 (MOG  ) and monitored for clinical signs of EAE. As shown in Fig. 4A, all Junb f/+ mice (n = 25) developed severe EAE. In contrast, none of the Junb-deficient mice (n = 22) displayed signs of paralysis during the 48-day period (Fig. 4A). The EAE phenotype of Junb-deficient mice is similar to that of mice lacking Batf 7 or Irf4 8 . The resistance of Junb-deficient mice to EAE induction was confirmed by histopathological analysis of the spinal cords of Junb f/+ and Junb-deficient mice. As shown in Fig. 4B, three to six demyelinated areas were observed in the spinal cord sections of Junb f/+ mice; on the other hand, no demyelinated areas in those of Junb-deficient mice (Fig. 4B). In addition, although both CD3ε + T cells and CD11b + myeloid cells were densely infiltrated into the spinal cord of Junb f/+ mice, no infiltration of immune cells occurred in Junb-deficient mice (Fig. 4C). Thus, clinical and histopathological analyses indicate that Junb-deficient mice are completely resistant to EAE.
To investigate the presence or absence of Th17 cells in MOG-treated Junb-deficient mice, we prepared CD4 + T cells and restimulated them with MOG  in the presence of APCs. As shown in Fig. 4D, IL-17A was abundantly expressed upon restimulation in CD4 + T cells from MOG-immunized Junb f/+ mice. By contrast, CD4 + T cells from immunized Junb-deficient mice failed to produce IL-17A when stimulated with MOG   (Fig. 4D), confirming that Th17 cells are absent from Junb-deficient mice. Taken together with the present findings, we conclude that JunB is required for Th17 differentiation both in vitro and in vivo.
Because epidermis-specific deletion of Junb is known to result in skin inflammation 19 , we studied the effect of systemic Junb deletion in imiquimod-induced dermatitis, a mouse model for psoriasis-like inflammatory disease 23 . Treatment with imiquimod induced ear swelling in Junb-deficient mice to the extent similar to that in control mice (Supplementary Figure 5A). In addition, Junb deletion did not affect the induction of psoriasis-associated genes such as Defb4, Il17f, S100a9, and Il19 in imiquimod-treated skin lesions, although the mRNA level of the two other associated genes IIl23a and Il24 in Junb-deficient mice was slightly higher than that in control mice at day 5 after imiquimod treatment (Supplementary Figure 5B). These findings suggest that JunB plays a marginal, if any, role in imiquimod-induced psoriasis.   Figure 6A,B), an AP-1 protein that is required for Th17 differentiation 7 , and can exist in a complex with BATF on an AP-1 site, as demonstrated by recent analysis using electrophoretic mobility shift assays (EMSAs) [24][25][26]    not appear to play a major role in Th17 development because of its low expression, although c-Jun has an ability to form an AP-1 complex with BATF when overexpressed in HEK293T cells 26 . JunB but not JunD cooperates with BATF to activate Th17 signature genes. To further know the reason why JunB plays the dominant role in Th17 development, we next investigated the role for JunB in BATF-dependent activation of Th17 signature genes and compared it with that for JunD, which is present in Th17 cells at the level comparable to JunB in contrast to the low expression of c-Jun. As described above, JunB as well as JunD is capable of forming a heterodimer with BATF. Although JunB by itself failed to induce transcription via the Il17a promoter, JunB activated the promoter in cooperation with BATF (Fig. 6A,B), indicating that JunB forms a productive dimer with BATF in Il17a activation. On the other hand, JunD did not elicit Il17a transcription even in the presence of BATF. Furthermore, JunB but not JunD cooperated with BATF to activate the Th17 signature genes Il17f and Il23r (Fig. 6C,D). These findings indicate that JunB functions as an indispensable partner of BATF in Th17 development, whereas JunD does not play a major role in regulation of Th17 cells.
We also used Junb-deficient CD4 + T cells to know the role of Jun family proteins in Th17 differentiation. Compared with JunB, retrovirally expressed JunD only marginally restored Th17 development of the Junb-deficient cells (Fig. 6E). The finding supports the idea that JunD is not a major regulator in Th17 differentiation. The activity of JunB likely depends on the N-terminal region, which is not involved in binding to DNA or dimerization with BATF 15 , because truncation of this region resulted in a loss of both activation of Il17a transcription (Fig. 6F) and induction of Th17 differentiation (Fig. 6G). On the other hand, overexpression of c-Jun in the Junb-deficient CD4 + T cells partially restored IL-17A production, but less effective than that of JunB ( Fig. 6E). Although c-Jun thus may have an ability to replace JunB in Th17 differentiation at least partially, JunB but not c-Jun plays the indispensable role, because c-Jun is expressed in Th17 cells to a much lesser extent than JunB (Fig. 5).
The Fosl2-encoded protein Fra2, another member of the AP-1 family, also has been reported to be important for Th17 cell specification 27

JunB is required for expression of Rorc and Rora. The present findings indicate that JunB regulates
Th17 development by forming a heterodimer with BATF. To further know the molecular mechanism whereby JunB functions, we tested its role in expression of the Th17 lineage-specifying factor RORγt (encoded by Rorc) 5,6 . It is known that Rorc expression under Th17-polarizing conditions is impaired in CD4 + T cells deficient in the Th17-polarizing transcription factors BATF 7 , IRF4 8 , and STAT3 9 , whereas the expression is not affected in CD4 + T cells lacking IκBζ (encoded by Nfkbiz), which also participates in Th17 development 13 . As shown in Fig. 7A, expression of Rorc mRNA was prevented in Junb-deficient CD4 + T cells at 24 and 72 h after stimulation. The RORγt-related protein RORα is also known to regulate development of Th17 cells, e.g., RORα and RORγt molecularly cross-compensate for Th17 differentiation 6 . Although expression of Rora was enhanced during Th17 differentiation in Junb f/+ CD4 + T cells, the enhancement did not occur in Junb-deficient CD4 + T cells (Fig. 7A). The impaired expression of Rorc and Rora in Junb-deficient cells was confirmed by RNA-seq analysis ( Fig. 2A). Furthermore, as shown in a chromatin immunoprecipitation analysis (Supplementary Figure 10), at around the transcription start site of Rorc and Rora as well as that of other Th17 signature genes, Junb deficiency resulted in deacetylation of histones H3 and H4, characteristic of transcriptionally inactive chromatin states. It has been reported that Rora expression is prevented also in Batf-deficient CD4 + T cells 7 but not in Nfkbiz-deficient cells 13 . The reduced transcription of Rorc and Rora in Junb-deficient T cells does not seem to be due to perturbation in IRF4 and BATF, because the amounts of these transcription factors were not affected by Junb deficiency in Th17 cells (Fig. 7B). Of note, compensatory elevation in expression of c-Jun and JunD did not occur in the absence of Junb (Fig. 7B). Thus, JunB as well as BATF stimulates the expression of the ROR nuclear receptors RORγt and RORα, which likely contributes to differentiation of Th17 cells.
We next tested whether overexpression of RORγt is able to rescue impaired development of Th17 cells in Junb-deficient cells. Retroviral expression of RORγt restored IL-17A production impaired in Junb-deficient CD4 + T cells under Th17-polarizing conditions, but to a small extent (less than 10%) (Fig. 7C). Similarly, a loss of Th17 differentiation by ablation of Batf or Irf4 is only slightly rescued by overexpression of RORγt 7,8 . The partial restoration indicates that RORγt expression is not sufficient for differentiation of Th17 cells, and thus raises a possibility that RORγt function is also supported by the Th17-polarizing factors JunB, BATF, and IRF4. It has recently been proposed that early binding of BATF and IRF4 to Th17-associated genes governs chromatin accessibility and subsequent recruitment of RORγt 27 , and thus it seems likely that the BATF-partner JunB also participates in initial chromatin accessibility of RORγt for Th17 cell specification. RORγt-dependent expression of IL-17A is not only a feature of Th17 cells but also that of other RORγt + cells including γδT cells 28 and NKp46 − CCR6 + CD4 + group 3 innate lymphoid cells (ILC3) 29 . To know the role of JunB in IL-17A production by these cell types, we analyzed lymphocytes from short intestinal lamina propria of Junb-deficient and control mice. As expected, ablation of Junb led to a severe impairment in IL-17 production by Th17-containing populations such as CD4 + TCRβ + cells (Supplementary Figure 11A) and CD3ε + TCRγδ − cells (Supplementary Figure 11B). By contrast, Junb depletion only modestly reduced IL-17A production by CD4 + TCRβ − cells, containing IL-17-producing ILC3 cells (Supplementary Figure 11A), and that by CD3ε + TCRγδ + cells (γδT cells) (Supplementary Figure 11B). These findings indicate that JunB plays an indispensable role in IL-17A expression by Th17 cells, but not in that by ILC3 and γδT cells.
JunB promotes Th17 differentiation by cooperating with BATF. Because JunB was capable of directly binding to not only BATF but also IκBζ (Supplementary Figure 1), it seemed possible that a functional link may exist between JunB and IκBζ. To test this possibility, we expressed IκBζ retrovirally in CD4 + T cells cultured under Th17-polarizing conditions. As previously reported 13 , overexpression of IκBζ or RORγt in wild-type CD4 + T cells increased a population of IL-17A-producing cells (Supplementary Figure 12A). Since the increase did not occur in Junb-deficient cells (Supplementary Figure 12B), IκBζ may possibly function in cooperation with JunB. However, expression of the Th17-specifying genes Rorc and Rora is dependent on JunB (Fig. 7A) but not on IκBζ, indicating that JunB does not cooperate with IκBζ. Furthermore, BATF effectively replaced IκBζ as a JunB-binding partner (Supplementary Figure 6C), and, unlike IκBζ, BATF regulates expression of both Rorc and Rora 7 . These findings agree with the conclusion that JunB primarily functions by forming a productive dimer with BATF in Th17 development.

Discussion
The present study provides genetic evidence that JunB is required for Th17 cell differentiation: Junb-deficient CD4 + T cells are defective in differentiating into Th17 cells, and Junb-deficient mice are refractory to induction of the Th17 cell-dependent autoimmunity EAE. The conclusion that JunB likely functions via forming an AP-1 complex with BATF, which is also indispensable for Th17 development, may be supported by recent observations in EMSAs using Th17 nuclear extracts: BATF can be complexed with JunB on AP-1-binding sites 7,24-26 . Although a part of such complexes also contains JunD 25 , JunD does not appear to make a major contribution to Th17 development. This is because JunD fails to activate Th17 signature genes in cooperation with BATF (Fig. 6). The role of c-Jun as a productive partner of BATF also seems to be minimal, since c-Jun is not abundantly expressed during differentiation into Th17 cells (Fig. 5) and compensatory elevation in the amount of c-Jun does not occur in the absence of Junb (Fig. 7).
The molecular mechanism whereby JunB controls Th17 differentiation likely involves JunB-dependent expression of the Th17 lineage-specifying factors RORγt and RORα; the expression is impaired by Junb ablation (Fig. 7). Consistent with this, the absence of the JunB-partner BATF also results in an impaired expression of Rorc and Rora 7 . On the other hand, IκBζ, another JunB-binding protein, does not participate in expression of Rorc and Rora 13 . The difference suggests that JunB does not function with IκBζ in Th17 differentiation, which is in agreement with the present observation that JunB interacts with BATF much more strongly with IκBζ (Supplementary Figure 6C). A BATF-containing AP-1 dimer has been shown to interact with IRF4 (or IRF8) on AP-1-IRF composite elements (AICEs), thereby activating a variety of genes that regulate development of immune cells such as Th2, Th17, B, and dendritic cells 15,[24][25][26][27] . BATF in the AP-1 heterodimer and IRF4 are thus considered to cooperatively function in Th17 differentiation at least in part. Like JunB and BATF, IRF4 also regulates Rorc expression 8 . In addition to increased expression of Rorc and Rora, JunB appears to function also by repressing expression of Foxp3, a transcription factor that is known to inhibit Th17 cell differentiation by antagonizing RORγt function 14 . This is because Foxp3 expression is elevated in Junb-deficient cells under Th17-polarizing conditions (Fig. 3), which elevation is also observed in cells lacking the JunB-partner BATF 7 . Furthermore, since early binding of BATF to Th17-associated genes appears to govern chromatin accessibility and subsequent recruitment of RORγt 27 , it seems possible that the BATF-partner JunB also contributes to initial chromatin accessibility of RORγt for Th17 cell specification. This may explain at least partially the reason why forced expression of RORγt only partially rescues Th17 differentiation impaired in Junb-deficient cells (Fig. 7).
The present RNA-seq transcriptome analysis confirms the requirement of JunB for Th17 development: besides induction of Th17 signature genes, JunB turns on the Th17 gene expression program (Fig. 2). On the other hand, JunB deficiency only marginally affects gene expression in naive CD4 + T cells and Th0 cells (Fig. 2), which is consistent with the finding that CD44 and CD62L are normally expressed in Junb-deficient naive CD4 + T cells (Supplementary Figure 3E). The gene expression pattern of Junb-deficient CD4 + T cells under Th17-polarizing conditions falls into the same cluster as that of Th0 cells (Fig. 2C). Thus JunB is not involved in TCR-mediated conversion of naive helper T cells into Th0 cells.
The present study also demonstrates that Junb-deficient CD4 + T cells are able to differentiate into Th1, Th2, and Treg cells under the respective polarizing conditions (Fig. 3). Consistently, the JunB-binding protein BATF is not required for differentiation into these T cell subsets 7 . It has been reported that JunB protein level correlates with the extent of differentiation into Th2 cells [30][31][32] ; accordingly, JunB is considered to facilitate Th2 cell development, although the direct effect of JunB has not been tested using Junb-knockout mice until the present study. Because the facilitation is mainly due to increased production of the autocrine factor IL-4 that is necessary for Th2 lineage commitment [30][31][32] , it seems likely that JunB is dispensable for Th2 differentiation in the presence of a high amount of IL-4, such as under the present Th2 polarizing conditions (Fig. 3). On the other hand, JunD has been reported to negatively regulate differentiation into Th2 cells 33 . Thus JunD appears to play a role distinct from that of JunB in Th2 differentiation as well as in Th17 differentiation.
ScIeNTIFIc RepoRtS | 7: 17402 | DOI:10.1038/s41598-017-17597-3 As shown in the present study, JunB expression is elevated at the mRNA and protein levels during Th17 differentiation (Figs 1 and 5). It is known that the amount of JunB is regulated via various mechanisms. For example, JunB is stabilized at the protein level by CARMA1 (also known as CARD11), a scaffold protein exclusively expressed in lymphoid and myeloid cells 34 . Intriguingly, lack of CARMA1 selectively prevents Th17, but not Th1 or Th2 differentiation, and CARMA1-knockout mice are resistant to EAE 35 . The mechanism for CARMA1 in Th17 differentiation appears to be explained at least in part by the present conclusion that JunB plays a crucial role in Th17 differentiation. It has been also reported that serum glucocorticoid kinase 1 (SGK1) stabilizes JunB by preventing ubiquitination and degradation of this protein, which is mediated via the E3 ligase Nedd4-2 (also known as Itch) and its adaptor protein Ndfip 36 ; and ablation of the SGK1 gene does not affect primary Th17 differentiation but attenuates IL-23R-mediated induction of pathogenic Th17 cells 37 . Thus, in addition to induction of Th17 differentiation, JunB may also contribute to the stabilization of the Th17 cell phenotype and pathogenicity acquisition. Control of JunB protein level at various steps would provide new therapeutic opportunities for human inflammatory autoimmune diseases in which Th17 cells have been implicated, such as psoriasis and multiple sclerosis 38,39 .

Materials and Methods
Mice. Junb f/f mice were generated according to the standard technique 40,41 . The BAC (bacterial artificial chromosome) clones containing mouse Junb gene (C57BL/6J) were obtained from BACPAC resource (Children's Hospital Oakland Research Institute). The targeting vector was constructed in pBluescript (Stratagene) to replace the exon of the Junb gene with the floxed Junb gene containing PGK-neo r (neomycin phosphotransferase gene driven by phosphorglycerate kinase promoter) cassette for positive selection and diphtheria toxin cassette for negative selection. A linearized targeting vector was electroporated into MS12 ES cells derived from blastocysts of C57BL/6J mice 42 . Positive clones were evaluated by Southern blot analysis with probes specific for 5′-and 3′-ends of the recombination site (5′-probe (600 nt) and 3′-probe (493 nt)) and also with an internal probe (475 nt). The confirmed clones were injected into blastocysts. The blastocysts were implanted into the uterus of a pseudo-pregnant mouse to yield chimeric mice. By crossing the chimera mice to C57BL/6 mice, the mice harboring the floxed Junb and PGK-neo r allele were generated. The correct recombination in F2 mice was confirmed by Southern blot analysis using the probes mentioned above and also by DNA sequencing around the recombination site. To remove PGK-neo r cassette, the flippase expression vector (pCAGGD-FLPe, Gene Bridges) was injected into the fertilized eggs obtained from mating wild-type C57BL/6 mice with the mice harboring the floxed Junb and PGKneo r allele 43 . The removal of PGK-neo r cassette by flippase was confirmed by Southern blot analysis 42,44,45 using the internal probe (475 nt) and also by PCR with primers (mJb-F30, 5′-ATGACCCATGTCAGCAACGG-3′; mJb-R22, AAGTGCGTGTTTCTTCTCCACAG). Junb f/+ mice were backcrossed onto wild-type C57BL/6 mice more than 15 times to eliminate the potential mutations caused during the gene manipulation. Junb f/f mice were obtained by mating Junb f/+ mice; the obtained pups had normal Menderian distribution. Junb f/f mice also appeared to develop normally. Junb f/f mice were crossed to Meox2 +/Cre mice 46 (The Jackson laboratory) to obtain Junb f/+ ; Meox2 +/Cre mice. These mice were then crossed with Junb f/+ or Junb f/f mice for generation of Junbdeficient (Junb f/f ; Meox2 +/Cre ) mice. The loss of the Junb gene by Cre-mediated recombination was verified by PCR using the following primers: 5′-CTGACAATTCCAGTTCTTTAAGC-3′, 5′-ATGACCCATGTCAGCAACGG-3′, and 5′-AAGTGCGTGTTTCTTCTCCACAG-3′. All animals were housed and maintained in a specific pathogen-free animal facility at Kyushu University. All experiments were performed in accordance with the guidelines for Proper Conduct of Animal Experiments (Science Council of Japan). The experimental protocol was approved by the Animal Care and Use Committee of Tokyo Metropolitan Institute of Gerontology, Kyushu University (Permit Numbers: A22-005, A24-042, and A26-102), and Toho University School of Medicine (17-43-288). All efforts were made to minimize the number of animals and their suffering.
In vitro differentiation of CD4 + T cells. Naive CD4 + T cells were prepared from the spleen and lymph nodes of 6-9 week old mice by magnetic sorting using Dynabeads ® Mouse CD4 (L3T4, Invitrogen) and DETACHaBEAD ® Mouse CD4 (Invitrogen), as described previously 47

RNA-seq and bioinformatic analysis.
Two and a half micrograms of total RNA was subjected to ribosomal RNA depletion using Ribo-Zero Magnetic Gold Kit (Human/Mouse/Rat) (Epicentre) followed by RNA-seq library preparation using NEBNext Ultra Directional RNA Library Prep Kit for Illumina (NEB) according to the manufacturer's instructions. Each library was quantified using Kapa Library Quantitation Kit (Kappa) and sequenced on Illumina HiSeq. 2500 to generate 101-bp paired-end reads. Raw FASTQ reads were trimmed for adapter sequences and low-quality reads using Trimmomatic version 0.32 49 . Trimmed reads were then aligned to the mouse reference genome mm9 and the transcriptome defined by the mm9 genes.gtf table obtained from the UCSC Genome Browser using TopHat version 2.0.13 50 . Mapped reads were assigned to all exons using feature-Counts 51 . The differential analysis was conducted using the Bioconductor package edgeR 52 , applying trimmed Mean of M-values library normalization. A gene was defined as differentially expressed, if the false discovery rate (FDR) corrected with Benjamini-Hochberg method was less than 0.001 and if the log2 fold change (log2FC) was more than 1 (up-regulated) or less than −1 (down-regulated). Differentially expressed genes were subjected to pathway enrichment analysis using Ingenuity Pathway Analysis (Qiagen). Over-represented pathways were defined as those with Benjamini-Hochberg-corrected P values of Fisher's exact test less than 0.01. The same gene set was also analyzed using ConsensusPathDB 53 , which enables an integrative usage of widely used pathway databases, namely KEGG, Pathway Interaction Database, Reactome, and Wikipathways. To exploit the latest human pathway datasets (Release 32, 2017) of ConsensusPathDB, human orthologues of the differentially expressed genes were identified with HomoloGene (https://www.ncbi.nlm.nih.gov/homologene) and Mouse Genome Informatics (www.informatics.jax.org) and used as a query to find enriched pathways. FPKM values were used for unsupervised hierarchical clustering based on the Spearman correlation distance and the Ward's linkage clustering algorithm with the hclust function implemented in R software.

Mass spectrometric analysis of IκBζ-binding proteins. Modified RAW264.7 cells expressing FLAG-
IκBζ under the control of the zinc-inducible sheep metallothionein Ia promoter 11 were stimulated with 50 nM ZnSO 4 for 3 h and subsequently with LPS (100 ng/ml) for 2 h. Nuclei prepared from the stimulated cells were sonicated in a buffer containing 140 mM KCl, 0.5% NP-40, 5% glycerol, and 20 mM HEPES, pH 8.0; the sonicates were subjected to immunoprecipitation using anti-FLAG (M2) antibody-conjugated agarose (Sigma). After washing with PBS (137 mM NaCl, 2.68 mM KCl, 8.1 mM Na 2 HPO 4 , and 1.47 mM KH 2 PO 4 , pH 7.4) containing 0.1% Triton X-100, the precipitated proteins were separated by SDS-PAGE (10%), followed by staining with Coomassie Brilliant Blue (CBB). Protein identification using liquid chromatography and tandem mass spectrometry (LC-MS/MS) was performed at the Laboratory for Technical Support, Medical Institute of Bioregulation, Kyushu University. Bands separated by SDS-PAGE were excised from the gel, and subjected to reduction with DTT, S-carbamidomethylation with iodoacetamide, and in-gel digestion with trypsin. Fragmented peptides were analyzed by LC-MS/MS as previously described 54 . The analysis reproducibly identified JunB in addition to the NF-κB p50 subunit, a well-known partner of IκBζ 10-12 . An in vitro pull-down binding assay. Maltose-binding protein (MBP)-fusion proteins were expressed in E. coli BL21 strain and purified using Amylose Resin (New England Biolabs). FLAG-tagged proteins were synthesized in vitro using TNT ® T7 Quick for PCR (Promega). The MBP-fusion protein bound to the resin was mixed with the FLAG-tagged protein in a binding buffer (150 mM NaCl, 5% glycerol, 1 mM DTT, 1 mM EDTA and 25 mM Tris-Cl, pH 8.0) and incubated for 1 h at 4 °C. After washing with the binding buffer containing 0.05% Triton X-100, proteins were eluted with 20 mM maltose. The eluate was subjected to SDS-PAGE, followed by staining with CBB or by immunoblot analysis with the anti-FLAG antibody (M2, Sigma).

Imiquimod-induced model of psoriasis-like skin inflammation. Psoriasis-like dermatitis was
induced by treatment of mice with imiquimod, according to the method by Yoshiki et al. 60 . Mice were topically treated with a daily dose of 50 mg of imiquimod cream (5%) (Mochida Pharmaceutical) on one ear for 5 consecutive days. Severity of the dermatitis was quantified by the extent of ear swelling and by the induction level of psoriasis-associated genes in samples obtained by ear biopsy. Ear thickness was measured daily with a Mitutoyo digmatic micrometer (Mitsutoyo). For estimation of psoriasis-associated gene expression, total RNA was extracted with TRIsure (BIOLINE) using Micro Smash MS-100 (TOMY), followed by RT-qPCR analysis. Expression level of mRNA was normalized to that of Hrpt.