Functional Crosstalk between Type I and II Interferon through the Regulated Expression of STAT1

Small "priming" quantities of type I interferon enhance cellular responses to type II interferon by maintaining basal levels of STAT1, explaining the observed crosstalk between these two cytokines.


Introduction
Although type I and type II interferons (IFNs) have distinct roles in immune responses, there is substantial overlap between the genes and cellular responses they regulate. It has been known for some time that many cells secrete small priming quantities of type I IFNs that facilitate more potent responses to subsequent stimuli [1][2][3]. Moreover, cellular responses to CSF-1 or IFNc can be affected by neutralizing type I IFN antibodies or knockout of type I IFN-Receptors (IFNAR) [2,4,5]. Notably, the protective anti-viral effects of IFNc were much less potent in IFNAR1 2/2 than wild-type fibroblasts which appeared to be caused by a lack of type I IFN priming [4,5]. The molecular events that underpin these priming events have not been fully characterized, although it has been proposed that type I and II IFNs shared receptor components [5].
However, as the majority of responses to type I and II IFNs require the expression of the STAT1 transcription factor [6], this is also a possible point of crosstalk between them.
STAT1 is a key mediator of cytokine-induced gene expression as it is activated either as homo-or heterodimer with other STATs by many cytokines including type I and type II IFNs, interleukin (IL)-6 and IL-10. STAT1 activity is of particular importance to the IFN system as STAT1 2/2 mice display many similar phenotypes to mice lacking IFNAR1 or the IFN Receptor (IFNGR)1. In particular, anti-viral, anti-mycobacterial, and anti-tumor responses are compromised [6][7][8][9]. Induction of STAT1 expression is a potential explanation for the priming activity of type I IFN because it is an IFN-stimulated gene (ISG) itself [10][11][12] and its 59 promoter region contains an IRF/gamma activated sequence (GAS) element bound by IFN-stimulated transcription factors [13]. Inducing the expression of STAT1 would increase the pool of this factor available for activation by IFNc. Consistent with such a hypothesis, low expression of STAT1 correlated with IFNresistance in melanoma samples when compared to surrounding normal tissue [14].
In unstimulated cells, STAT1 resides in the cytoplasm as a latent factor that is activated by a series of post-translational modifications initiated when it is recruited to cytokine receptors following receptor ligation [15]. At the receptor, STAT1 is phosphorylated on tyrosine 701, by Janus family kinase (JAK)s, which facilitates its dimerization either with other STAT1 molecules or other STAT proteins depending on the cytokine receptor. In addition, STAT1 proteins are phosphorylated on serine 727 prior to nuclear translocation which is essential for their full transcriptional activity [16]. Conversely, STAT1 activity is negatively regulated by phosphatases, SOCS proteins, and the SUMO ligase Protein Inhibitor of Activated STAT (PIAS)1 [15].
Recently, in the course of our studies on IFNc-activated AP-1 DNA binding, we noticed that IFNc-induced GAS DNA binding was suppressed in c-Jun 2/2 cells compared to wild-type cells [17] and this correlated reduced levels of STAT1 in c-Jun 2/2 cells. The level of STAT1 expression in c-Jun 2/2 murine embryonic fibroblasts (MEFs) were restored to wild-type levels following culture in media conditioned by wild-type fibroblasts suggesting that c-Jun deficiency caused the disruption of an autocrine/ paracrine loop that regulated STAT1 expression. The STAT1inducing component of media conditioned by wild-type fibroblasts was IFNb, because the activity could be blocked by neutralizing antibodies directed against type I IFN and antibodies used were raised against IFNAR and attenuated by targeted knockdown of IFNb by RNA interference (RNAi). While c-Jun has been demonstrated to co-operate with ATF-2, IRF-3, and NFkB for virus-induced production of IFNb [18], to our knowledge our studies are the first to demonstrate that c-Jun is necessary for basal expression of low-level IFNb. Fibroblasts in which this autocrine/ paracrine loop was disrupted by the loss of components of type I IFN receptors also express lower levels of STAT1. As many biological functions of IFN require STAT1 [6,7], this suggested that previous observations of attenuated responses to IFN in IFNAR1 2/2 cells may be related to the reduced STAT1 expression that has been observed [19]. Consistent with this hypothesis, restoring STAT1 expression in IFNAR1 2/2 fibroblasts rescued IFNc-induced gene transcription and anti-viral properties.
In summary, this study provides evidence of an autocrine/ paracrine stimulatory loop that requires the expression of c-Jun, IFNb, and IFNAR to regulate the expression of STAT1. Importantly, this basal IFNb production occurs via a mechanism distinct from the pathogen-stimulated IFNb production mediated by IRF and NFkB pathways [18]. One model to explain crosstalk between type I and II IFNs states that type I and II IFN-R physically interact in a ligand-dependant manner, such that the presence of type I IFNs is essential for a fully competent IFNc response [5]. Herein, we demonstrated that attenuated IFNcmediated gene induction and an associated defective anti-viral response to IFNc that is observed in IFNAR1-deficient cells can be rescued by re-expressing STAT1 and is therefore independent of IFNAR1. We propose that an alternative model to explain the functional synergy between type I and II IFNs is based on the regulated expression of STAT1 via c-Jun-mediated production of basal levels of IFNb.

STAT1 Expression Is Attenuated in c-Jun 2/2 MEFs
In the course of our studies of IFN-induced signaling and gene expression, we performed elecrophoretic mobility shift assays (EMSAs) assessing GAS binding species in nuclear extracts from IFNc-stimulated wild-type and c-Jun 2/2 MEFs. A GAS binding complex was detected in both wild-type and matched c-Jun 2/2 MEFs following 15-30 min of exposure to IFNc, however in the absence of c-Jun, IFNc-induced GAS binding activity was markedly attenuated ( Figure 1A). The decrease in GAS binding activity in c-Jun 2/2 MEFs was a consequence of reduced expression of STAT1. Both STAT1 mRNA and protein were ,10-fold lower in c-Jun 2/2 MEFs compared to wild-type cells ( Figure 1B and C). However, expression of STATs was not globally affected, as expression of STAT3, another GAS-binding transcription factor, remained unchanged ( Figure 1C). Reduced STAT1 expression was not a clone-specific phenomenon as similar results were obtained using an independently derived matched pair of wild-type and c-Jun 2/2 MEFs ( Figure S1).

c-Jun Maintains Levels of STAT1 Expression by Stimulating Autocrine Production of a Soluble Factor
To determine if c-Jun could regulate STAT1 levels by inducing the secretion of a soluble factor that acted in autocrine/paracrine fashion to induce STAT1 expression, conditioned media from wild-type or c-Jun 2/2 MEFs were cultured in (i) fresh media, (ii) media conditioned by c-Jun 2/2 MEFs, or (iii) media conditioned by wild-type MEFs. Cells were harvested after 16 h of culture in conditioned media and STAT1 mRNA and protein expression was assessed. Expression of STAT1 mRNA and protein was unaltered in wild-type MEFs cultured in fresh media or conditioned media from wild-type or c-Jun 2/2 MEFs (Figure 2A and B). In c-Jun 2/2 MEFs, basal expression of STAT1 was much lower than in wildtype cells and was not increased when the cells were cultured in either fresh media or conditioned media from c-Jun 2/2 MEFs (Figure 2A and B). In contrast, when c-Jun 2/2 MEFs were cultured in media conditioned by wild-type MEFs, STAT1 mRNA and protein expression was induced almost to the levels observed in wild-type cells (Figure 2A and B). These data confirmed that fibroblasts secrete a c-Jun-dependent soluble factor that induces STAT1 expression through an autocrine/paracrine feedback loop.

Author Summary
Cells of the immune system release interferons (IFNs) in response to pathogens or tumor cells; these proteins signal to other immune cells to initiate the body's defense mechanisms. The two classes of IFNs-types I and II-have different receptors and distinct effects on the cells; however, there is ''crosstalk'' between them. In particular, small quantities of type I IFN can ''prime'' cells to produce a robust response to type II IFN. In this paper, we provide evidence to explain the molecular basis of this crosstalk. We show that continuous expression of the transcriptional activator c-Jun is responsible for producing basal, priming levels of a type I IFN; this signals to immune cells with the type I IFN receptor (IFNAR1) to maintain expression of STAT1 inside these cells. STAT1 is a key factor for immune cell responses to type II IFN. Thus, signaling by low levels of type I IFN primes the cells with sufficient STAT1 to respond robustly to a subsequent type II IFN signal. This work provides an alternative explanation of the priming phenomenon to a previous proposal that the ligandbound type I receptor, IFNAR1, acts as a component of the type II IFN receptor.

Constitutive Secretion of IFNb Maintains Basal Expression of STAT1
Type I IFN is constitutively secreted from unstimulated fibroblasts and can induce STAT1 expression [10]. To determine if type I IFN was the STAT1-inducing active component of fibroblast conditioned media, c-Jun 2/2 MEFs were cultured in either fresh or conditioned media from wild-type cells in the presence of a type I IFN blocking antibody [20]. STAT1 expression was increased in c-Jun 2/2 MEFs cultured in conditioned media from wild-type cells in the presence of control antibodies ( Figure 3A) and this enhanced expression was entirely blocked by the presence of type I IFN neutralizing antibodies used at concentrations capable of neutralizing ,5 IU/mL IFNb. Additional studies ( Figure 3B and C) revealed that the STAT1-inducing activity of wild-type-conditioned media was almost ablated by a blocking mAb raised against IFNAR1 [21]. Together, these data demonstrate that type I IFN is a component of conditioned media from wild-type cells that is necessary for the rescue of STAT1 expression in c-Jun 2/2 cells.
It has been reported that STAT1 levels are diminished in IFNb 2/2 cells [22] indicating that IFNb could be the key component of the conditioned media from wild-type cells shown to induce expression of STAT1 in c-Jun 2/2 MEFs. Treatment of c-Jun 2/2 MEFs with doses as low as 1 IU/mL IFNb induced STAT1 mRNA and doses between 5 and 10 IU/mL were sufficient to restore STAT1 mRNA and protein expression to levels seen in wild-type cells ( Figure S2A and B). STAT1 mRNA levels were slightly increased in wild-type MEFs treated with IFNb ( Figure S2C), which is consistent with studies demonstrating that STAT1 expression is induced in fibosarcoma cell lines treated with IFNa or b [12] and in splenic leukocytes where STAT1 levels were increased following virus infection in a type I IFN-dependent manner [11]. Comparison of the levels of expression of IFNb mRNA in wild-type and c-Jun 2/2 cells revealed that c-Jun 2/2 MEFs expressed ,50% of the wild-type levels of IFNb mRNA ( Figure 4A). AP-1 sites are known to be important for inducible expression of IFNb [23], but little is known of what regulates EMSAs were performed using radiolabeled oligonucleotides containing a GAS consensus sequence, and nuclear extracts from wild-type or c-Jun 2/2 MEFs treated with 100 IU/mL IFNc for indicated times. (B) RNA was extracted from wild-type or c-Jun 2/2 MEFs, cDNA synthesized, and qRT-PCR performed with primers complementary to murine STAT1. Histograms represent mean and error bars the SEM of four independent experiments and are expressed relative to the levels detected in wild-type cells (* p,0.05). (C) SDS-PAGE and Western blotting with antibodies against STAT1 and STAT3 were performed using whole cell extracts from wild-type or c-Jun 2/2 MEFs. As a control, the expression of a-tubulin was also tested by Western blot. doi:10.1371/journal.pbio.1000361.g001 constitutive production of type I IFN in unstimulated cultured fibroblasts. Chromatin immunoprecipitation (ChIP) assays on unstimulated wild-type and c-Jun 2/2 MEFs demonstrated a .2fold increase in c-Jun bound to the murine IFNb promoter when compared to Ig control samples ( Figure 4B). Together, these data imply that expression of c-Jun and subsequent occupation of the IFNb promoter by c-Jun is required for basal secretion of IFNb.
To determine if IFNb was the type I IFN necessary to maintain STAT1 expression, we used RNAi to knock down IFNb in wild-type MEFs ( Figure 4C) and assessed the ability of conditioned media from these cells to induce the expression of STAT1 mRNA in c-Jun 2/2 MEFs. As expected, STAT1 mRNA levels were greater when c-Jun 2/2 MEFs were cultured in conditioned media from wild-type MEFs or from MEFs expressing a control knockdown vector than if these cells were cultured in fresh media ( Figure 4D). In contrast, the ability of conditioned media from wild-type cells with RNAi-mediated knockdown of IFNb to induce STAT1 expression in c-Jun 2/2 MEFs was significantly reduced ( Figure 4D). These data confirm that IFNb is expressed by unstimulated wild-type fibroblasts and is necessary for the maintenance of STAT1 expression.

IFNAR-Deficient Cells Express Reduced Levels of STAT1
As disruption of autocrine/paracrine stimulation by IFNb affected the level of STAT1 expression in c-Jun 2/2 MEFs, we predicted that cells lacking either chain of the type I IFN receptor would also express less STAT1 than wild-type cells. Primary MEFs ( Figure 5A) and splenocytes ( Figure 5B) from either IFNAR1 2/2 or IFNAR2 2/2 (unpublished data) mice expressed significantly lower levels of STAT1 than wild-type cells. We extended these studies to compare the expression of STAT1 across multiple tissues in wildtype versus IFNAR1 2/2 mice. As shown in Figure 5C, the levels of STAT1 were consistently reduced in all tissues from IFNAR1 2/2 mice compared to their wild-type counterparts, suggesting this defect may have broad physiological importance. Interestingly expression of STAT2 was also reduced in IFNAR11 2/2 MEFs while the levels of STAT3 were unaffected by knockout of the type I IFN receptor ( Figure S3A).
Our model predicted that, unlike c-Jun deficiency that affected production of an autocrine stimulus, IFNAR1 deficiency affects responses to the autocrine stimulus. In support of this model, wildtype-conditioned media was able to rescue the expression of STAT1 in c-Jun 2/2 MEFs, but in IFNAR1 2/2 MEFs STAT1 expression was unaffected by culture in wild-type-conditioned media ( Figure S3B). These data support the existence of an autocrine loop involving IFNb that regulates basal STAT1 expression levels and suggest that defects in any part of this loop are likely to affect the expression of STAT1.

Re-Expression of STAT1 in IFNAR-Deficient Cells Restores IFNc Signaling and Gene Expression
STAT1 is important for not only IFNa/b signaling but also the signaling of several other cytokines, including IFNc [15]. The expression of approximately two-thirds of IFNc-induced genes is dependent upon STAT1 expression, however not all IFNcmediated biological responses are entirely dependent on STAT1 expression [17,24]. It has previously been reported that IFNAR1 2/2 cells are refractory to IFNc treatment due to the proposed interaction between IFNAR1 and IFNGR [5]. To determine if decreased expression of STAT1 may confer the observed decrease of IFNc-mediated responses in IFNAR1 2/2 cells, STAT1 levels were restored in these cells by retroviral transduction ( Figure 6A). GAS binding activity was assessed by EMSA using nuclear extracts from IFNc-treated wild-type MEFs, IFNAR1 2/2 MEFs, and IFNAR1 2/2 MEFs reconstituted with empty vector (IFNAR1 2/2 MSCV) or STAT1 (IFNAR1 2/2 HA-STAT1). Consistent with previous studies [5,25], IFNc induced less GAS binding in IFNAR1 2/2 cells than wild-type cells ( Figure 6B). This low level of GAS binding was also observed in cells transduced with empty vector but was rescued in cells reconstituted with HA-STAT1a. These data demonstrated that the reduced GAS binding observed in IFNAR1 2/2 cells was caused by reduced STAT1 expression rather than being a direct consequence of IFNAR1 deficiency.
Previous studies demonstrated that IFNc-induced gene expression was attenuated in IFNAR1 2/2 cells [5]. We therefore assessed the impact of re-expression of STAT1a in IFNAR1 2/2 cells upon the IFNc-induced expression of genes such as b-2-microglobulin and SOCS3 that require STAT1 expression [26]. Both genes were induced in response to IFNc in wild-type cells, although with differing kinetic profiles, but induction was weak or absent in IFNAR1 2/2 cells. IFNc-induced expression of both b-2-microglobulin and SOCS3 was restored in cells that re-expressed HA-STAT1a, but not in cells transduced with an empty vector ( Figure 6C, D). Similar results were observed when other IFNcresponsive genes were tested ( Figure S4).

Expression of STAT1 in IFNAR-Deficient Cells Restores Their Protective Anti-Viral Response Following Treatment with IFNc
In order to determine whether the reduced levels of STAT1 in IFNAR1 2/2 cells could affect biological responses to IFNc, we investigated whether re-expression of STAT1 in IFNAR1 2/2 cells impacted upon the ability of IFNc to protect them against infection by the cytopathic virus murine encephalomyocarditis virus (EMCV). Wild-type, IFNAR1 2/2 , IFNAR1 2/2 MSCV, and IFNAR1 2/2 HA-STAT1 MEFs were infected with a dose of virus sufficient to induce 100% lysis of wild-type MEFs in the presence or absence of various doses of IFNc, and the cytopathic effects were determined by assessing cell viability after 24 h. As was shown previously [5], the ability of IFNc to protect cells from EMCVmediated lysis was significantly reduced in IFNAR1 2/2 MEFs when compared to wild-type MEFs at most doses of IFNc and the concentration of IFNc (500 IU/ml) required to provide 80% protection from the virus for IFNAR1 2/2 cells was much greater than that required to provide a similar level of protection for wildtype cells (10 IU/ml). The response of IFNAR1 2/2 MEFs transduced with empty vector to IFNc was not significantly different from the untransduced IFNAR1 2/2 MEFs at any dose of IFNc and the concentration of IFNc required to provide 80% protection (450 IU/ml) was of a similar order of magnitude ( Figure 7). In contrast, protection from virus-induced lysis was significantly enhanced in IFNAR1 2/2 HA-STAT1 MEFs at most doses of IFNc. These data provide direct evidence that the attenuated protective anti-viral responses to IFNc observed in IFNAR1 2/2 cells is a consequence of reduced STAT1 expression.

Discussion
Herein we demonstrate that c-Jun is essential for the constitutive production of small quantities of IFNb that initiates autocrine or paracrine feedback loops required to maintain the expression of STAT1 (Figure 8). This system was disrupted either by c-Jun deficiency, which prevents production of IFNb, or by IFNAR deficiency, which affects the ability of cells to respond to the autocrine stimulus. Consistent with our data, others found that cells lacking IFNb also express much lower levels of STAT1 [22] and virus-mediated induction of STAT1 is dependent on type I IFN signaling [11]. As IFNc signaling is attenuated when the autocrine stimulus is blocked (Figure 8) but restored by adding back STAT1, it appears the level of STAT1 expressed by the cell determines the response of the cell to other cytokines. These results suggested the ability of IFNc to induce a protective anti-viral state was due to the type I IFN-mediated maintenance of STAT1 expression rather than the recruitment of IFNAR1 into the IFNcR complex as has been previously proposed [5]. These findings define a novel mechanism through which STAT1-mediated signals can be regulated and highlight the importance of crosstalk between type I and II IFNs for anti-viral immunity.
It has been known for some time that, as well as being produced in large quantities following viral infections, cells can secrete low levels of type I IFN constitutively [1,2,27]. Virusinduced activation of the IFNb enhanceasome is one of the bestcharacterized transcriptional modules [18,23]. Viral activation of the IFNb promoter involves the binding of NFkB, IRF3, and ATF2/c-Jun complexes to a series of DNA elements termed PRD I-IV [23]. In this setting, c-Jun binds to PRD IV of the promoter and facilitates co-operative binding of the other factors. Removing PRD IV from the promoter, or even reversing its orientation, has a major impact on the transcriptional activity of the promoter [23], suggesting the role of c-Jun is critical in the context of viral infection. In contrast, little is known of the molecular mechanisms of constitutive type I IFN production. Our study indicates that PRD IV of the IFNb promoter is occupied by c-Jun even in ''resting'' cultured cells ( Figure 4B). This requirement for c-Jun explains why we found that constitutive IFNb production and hence the expression of STAT1 was attenuated in c-Jun 2/2 cells. In addition to regulating basal expression of IFNb, we have recently demon-  MEFs, IFNAR1 2/2 MEFs, and IFNAR1 2/2 MEFs transduced with empty vector (IFNAR1 2/2 MSCV) or IFNAR1 2/2 MEFs transduced with HA tagged STAT1a (IFNAR1 2/2 STAT1) were subjected to SDS-PAGE and probed with an antibody specific to STAT1,and membranes were stripped and reprobed with antibodies specific to hsp70 as a loading control. (B) EMSAs were performed using radiolabeled oligonucleotides containing a GAS consensus sequence, and nuclear extracts from wild-type MEFs, IFNAR1 2/2 MEFs, IFNAR1 2/2 MEFs transduced with empty vector (IFNAR1 2/2 MSCV), or IFNAR1 2/2 MEFs transduced with HA tagged STAT1a (IFNAR1 2/2 STAT1) treated in the presence or absence of 100 IU/mL IFNc. (C, D) Wild-type MEFs, IFNAR1 2/2 MEFs, and IFNAR1 2/2 MEFs transduced with empty vector (IFNAR1 2/2 MSCV) or IFNAR1 2/2 MEFs transduced with HA-tagged STAT1a (IFNAR1 2/2 STAT1) were treated with 100 IU/mL IFNc for 0, 1, or 6 h. RNA was extracted, cDNA synthesized, and qRT-PCR performed with c-Jun promotes IFNc responses via IFNb and STAT1 strated that c-Jun is activated following IFNc treatment and may also play a direct role in regulating the expression of a subset of IFNc-responsive genes (ISGs) [17]. Indeed we identified ISGs that were dependent on c-Jun for induction by IFNc, others that required STAT1, and others that required both c-Jun and STAT1 for increased expression following treatment with IFNc [17]. These results, coupled with the functional data provided herein, highlight the complex molecular interplay between c-Jun and canonical mediators of type I and II IFN signaling such as STAT1 in regulating a comprehensive response to IFN treatment.
Takaoka and colleagues previously demonstrated the importance of IFNb in the production of an IFNc-mediated anti-viral response [5]. In that paper the authors showed that IFNb 2/2 MEFs were defective in mounting an IFNc-induced antiviral response. These data mirror what we have demonstrated herein where we show that IFNAR1 2/2 MEFs show a similar defect in mounting an IFNc-induced antiviral response. However, we showed that restoring STAT1 expression in IFNAR1 2/2 cells significantly rescued the ability of IFNc to protect cells against EMCV, suggesting that regulating the levels of STAT1 expression through the autocrine loop may play an important role in responses to this challenge. The ability of type I and II IFNs to co-operate, for example, in treatment of melanoma tissue [28] or priming of macrophage cytotoxicity [29] has long been recognized. Interestingly, at a cellular level, IFNAR1 2/2 cells were known to have an anomalously poor response to IFNc with respect to induction of GAS DNA binding, induction of gene expression, and protection against the cytopathic effects of EMCV [4,5,25]. IFNc function is not entirely compromised in IFNAR1 2/2 animals because IFNGR1 2/2 mice have distinct phenotypic differences from IFNAR1 2/2 mice [30]. Inhibiting autocrine priming by type I IFN does not only affect signaling by IFNc. Therefore its is not surprising that IL-6 signaling [31] and CSF-1 signaling are affected by inhibiting priming by type I IFN [2] and that signals induced by IL-10 can be affected by priming with IFNs [32].
It was proposed that the ligand-bound IFNAR1 chain acts as a component of the IFNGR and promotes recruitment of STAT1 to the IFNGR because IFN receptors are clustered within caveolar membrane fractions to facilitate their association [5]. Such a hypothesis is inconsistent with mapping of the docking site of STAT1 to the IFNAR2 chain of the type I IFN-R rather than the IFNAR1 as specified by the shared receptor model [33]. We demonstrated herein that IFNAR1 2/2 cells express lower basal levels of STAT1 relative to wild-type controls ( Figure 5), and as STAT1 is a critical mediator of IFN signaling, this is an alternative reason why these cells may lack sensitivity to IFNc. Our model not primers specific for b-2-microglobulin and SOCS3. mRNA levels are expressed relative to those of wild-type C57/BL6 (B6) splenocytes. Histograms represent the mean and error bars the standard error of four independent experiments. (* p,0.05 for samples that were significantly induced). doi:10.1371/journal.pbio.1000361.g006 only explains the inability of IFNc to prime IFNAR1 2/2 cells for an anti-viral response and the rescue of IFNc function in IFNAR1 2/2 cells by STAT1a expression but also the attenuated responses to other cytokines, such as IL-6 and CSF-1, observed in IFNAR1 2/2 cells [2,31] and predicts they may also be rescued by expression of STAT1. As IFNc function was not entirely recovered following re-expression of STAT1 in IFNAR 2/2 cells, we cannot exclude that the shared receptor mechanism makes a contribution, but there are other reasons why reconstitution of STAT1 may not have fully rescued IFNc function. These include the absence of other as yet unidentified signal transducing proteins from cells of this genotype.
The level of STAT1 expression in cells can have functional consequences with respect to immune responses. In response to viral infection, Ag-specific CD8 + T cells express peak levels of STAT1 for a shorter period of time than CD4 + cells [34]. This decreased sensitivity to IFN-induced growth inhibition allows expansion of Agspecific CD8 + cells while the proliferation of cells with higher STAT1 is inhibited [34]. The relative amounts of different STATs can also affect the biological responses to cytokines. For example STAT1:3 and STAT1:4 ratios have been shown to alter cellular responses, and thus regulating the levels of these transcription factors will affect the outcome of immune responses [11,35].
Our previous studies revealed that loss of IFN signaling abrogated the immune-mediated neo-natal lethality of SOCS1 2/2 deficiency [36], and more recently we discovered that deleting IFNAR1 also rescued this pathology [37] to a level equivalent to SOCS1 2/2 IFN +/2 . Although SOCS1 directly regulated type I IFN signaling, another reason why IFNAR1 deficiency can protect SOCS1 2/2 animals may be the similarities and crosstalk between type I and II IFN signaling pathways. These data highlight the patho-physiological importance and mechanism of crosstalk between type I and II IFN that are important considerations in understanding the contributions of individual cytokines to host defense and in their therapeutic targeting.

ChIP
ChIP assays were performed as described previously [41] using 5 mg of anti c-Jun or rabbit IgG control antibodies. The abundance of specific sequences in ChIP samples was quantitated using the SYBR Green dye detection method (Applied Biosystems, Warrington, UK). Primers used for PCR reactions were mIFN PRDIV (59-ATTCCTCTGAGGCAGAAAGGACCA; 59-GCA-AGATGAGGCAAAGGCTGTCAA) and were designed using Primer Express 2 software. Threshold cycle values (Ct) were measured in the exponential phase, and promoter occupancy was calculated using the formula 2 (Ct Ig 2 Ct c-Jun) .

Statistical Analysis
Statistical significance was tested using one-way ANOVA testing with OriginLab 7.5 software (Northampton, MA, USA) or Prism Software Graphpad (La Jolla, CA, USA). concentrations of IFNc (0-1,000 IU/mL) and cultured for 16 h. As controls, cells were cultured in fresh media alone (100% survival) or with EMCV alone (0% survival). Cells were washed in PBS, formalin fixed (10 min at RT), washed (twice with PBS), and stained in 0.5% Crystal Violet/20% methanol. Stained cells were extensively washed, crystal violet was solubilized in 10% acetic acid, and OD 550 nm was recorded. Viability was calculated by comparison against a standard curve.

Supporting Information
Figure S1 STAT1 expression is decreased in c-Jun knockout cells. SDS-PAGE and Western blotting with antibodies against STAT1 was performed using whole cell extracts from an independently derived set of wild-type or c-Jun 2/2 MEFs. As a control, the expression of a-tubulin was also tested by Western blot.  Figure S4 Reconstitution of STAT1 in IFNAR1 knockout cells restores IFNc-mediated upregulation of IFN response genes. Wild-type MEFs, IFNAR1 2/2 MEFs, and IFNAR1 2/2 MEFs transduced with empty vector (IFNAR1 2/2 MSCV) or IFNAR1 2/2 MEFs transduced with HA-tagged STAT1a (IFNAR1 2/2 STAT1) were treated with 100 IU/mL IFNc for 0, 1, or 6 h. RNA was extracted, cDNA synthesized, and qRT-PCR performed with primers specific for caspase 4 (CASP 4), CISH, CXCL11, and MYD88. mRNA levels are expressed relative to those of wild-type C57/BL6 (B6) splenocytes. Histograms represent the mean and error bars the standard error of four independent experiments (* p,0.05 for samples that were significantly induced). Found at: doi:10.1371/journal.pbio.1000361.s004 (1.56 MB TIF)