Meet the terminator: the phosphatase PP2A puts brakes on IRF-3 activation

Abstract

Cellular interferon response to microbial infection is transient; In a recently paper in Immunity, Long et al identifies PP2A phosphatase as a deactivator of phospho-IRF-3, the key transcription factor for interferon synthesis, thus providing one basis for the observed transiency.

Innate immunity provides the first line of cellular defense against microbial infections. Various cellular sensors recognize specific microbial components and trigger cytoplasmic signaling cascades that culminate in activating specific transcription factors to induce transcription of their target anti-microbial genes. The proteins encoded by these genes, such as interferon (IFN), inhibit microbe replication and protect the host cell. However, continued expression of many of these proteins is detrimental to cells and organisms and hence is the need to temporally down-regulate the action of the active transcription factors (Gonzalez-Navajas et al., 2012). A critical transcription factor, in this context, is the interferon regulatory factor 3 (IRF-3). Although much is known about how IRF-3 is activated by its phosphorylation by different signaling pathways, relatively less is known about the mechanisms of its deactivation (Hiscott, 2007). In a recent paper in Immunity, Long et al report how the phosphatase, PP2A, dephosphorylates active IRF-3, thus turning off the transcriptional signal and replenishing the cytoplasmic IRF-3 pool (Long et al., 2014).

Infection is detected by cellular pattern recognition receptors (PRR) which recognize microbial nucleic acids, proteins, lipids or sugars; in some cases, they can also recognize proteins, DNA or RNA produced from dead or stressed cells. PRRs include the membrane-bound Toll-like receptors and the cytoplasmic nucleic acid-detecting receptors (Kawai and Akira, 2011). Binding of the cognate ligands activates the receptors and triggers interactions with the adaptor proteins, such as MyD88, TRIF or MAVS, which leads to the recruitment of multiple proteins comprising the signaling complexes, including protein kinases and ubiquitin E3 ligases. The resultant signaling pathways branch out and activate specific transcription factors, most notably NF-κB, IRF-3 or IRF-7 and AP1. The activated transcription factors induce their target genes, in the nucleus, by binding, either singly or in combinations, to specific cis-acting sequences present in the regulatory regions of these genes. All nine members of the IRF family recognize the same sequence, ISRE, but some induce transcription while others repress it (Ikushima et al., 2013). IRF-3, the most prominent inducer, is activated through signaling pathways triggered by many PRRs, such as TLR3, TLR4, RLR, cGAS/STING, all converging to activate the protein kinase TBK1 or IKKε, which phosphorylates multiple serine residues near the C-terminus of IRF-3. Specific phosphorylation of IRF-3 changes its conformation and allows it to dimerize, translocate to the nucleus, bind to the ISRE of the target genes, interact with the obligatory co-activator, deacetylated β-catenin, and promote transcription (Chattopadhyay et al., 2013).

Activation of IRF-3 is beneficial to the hosts for protecting them from microbial (especially viral) infections; however, prolonged gene induction by activated IRF-3 cannot be tolerated by cells, because many IRF-3-induced proteins negatively affect cell growth and survival. Moreover, at the organismal level, over-production of induced cytokines causes hyper-inflammation and autoimmune diseases. For these reasons, all PRR signals are transient and negative feed-back loops are built into the systems. Long et al provide an important missing element of these negative regulatory loops, by describing how activated IRF-3 is recognized and dephosphorylated by PP2A, a Ser/Thr phosphatase (Long et al., 2014). They show that PP2A binds to IRF-3 using its adaptor protein RACK1 (Figure 1). They mapped the IRF-3/RACK1 mutual recognition domains, although the nature of RACK1/PP2A interaction is not clear. IRF-3, activated by TLR3, TLR4 or RLR, is deactivated, but not degraded, by the PP2A-RACK1 complex. The biological significance of this negative regulation was verified using macrophage-specific PP2A knock-out mice and cells derived from them. As expected, in the absence of PP2A, more IFN was induced, suppressing the replication of vesicular stomatitis virus and inhibiting the resultant mortality of infected mice. Other known negative regulations of IRF-3 actions are achieved by its sumoylation or ubiquitination and proteasomal degradation. Some viruses, such as Sendai virus, trigger degradation of IRF-3, thus allowing them to establish persistent infection. Long et al demonstrate that SeV also activates the PP2A pathway of IRF-3 deactivation early after infection. The new pathway appears to be more nimble and reversible in nature and hence it is likely to be used in many physiological situations requiring IRF-3's functional shut-off. IRF-3 is not only a transcription factor but also a pro-apoptotic factor and the two functions are not interdependent (Chattopadhyay et al., 2010). The apoptotic action of IRF-3 is critical for protecting the host from many RNA viruses, HTLV1 and alcoholic liver diseases. IRF-3 activation as an apoptotic protein also requires its phosphorylation by TBK1, but of serine residues different from those required for its activation as a transcription factor (Figure 1). It will be interesting to examine whether PP2A regulates the apoptotic activation of IRF-3 as well.

Figure 1. RACK1/PP2A deactivates transcriptional function of IRF-3.

Inactive IRF-3 remains in a closed conformation in the cytoplasm. Microbial infection-induced signaling pathways activate IRF-3 by phosphorylation of its specific Ser/Thr residues by TBK1, causing conformational changes, dimerization and nuclear translocation of IRF-3. In the nucleus, IRF-3 binds to ISRE of target genes, recruits the essential co-activators and induces gene transcription. Phosphorylated nuclear IRF-3 is exported to the cytoplasm and is recruited to a Ser/Thr phosphatase, PP2A by the adaptor protein, RACK1. PP2A dephosphorylates IRF-3 and terminates its transcriptional activity; dephosphorylation of IRF-3 also prevents it from being degraded by the proteasomal machinery. In another pathway, IRF-3 is activated by phosphorylation of different Ser residues by TBK1, interacts with BAX, translocates to mitochondria and triggers cellular apoptosis.

PP2A comprises a large family of hetero-trimeric proteins (Shi, 2009), of which only one member has now been studied in the context of IRF-3 deactivation. It remains to be seen whether other members of this family share this property and whether the functions of other phospho-proteins of the IRF-3 signaling pathways are regulated by PP2A as well. The PP2A/RACK1 complex binds IRF-3 in the cytoplasm whereas phospho-IRF-3 functions in the nucleus; so the burning question is what makes active IRF-3 translocate to the cytoplasm? Unfortunately, this issue was not addressed experimentally. Finally, PP2A has been reported to interact with more than 50 proteins, thus affecting many cell functions including other transcription factors, such as p53 and NF-κB (Li et al., 2012; Reid et al., 2013). Indeed, as noted by the authors, in mice harboring PP2A-less macrophages, LPS-mediated TLR4 signaling caused stronger activation of NF-κB as well. Consequently, attributing specific cellular functions of PP2A to only one target, IRF-3, is tenuous, unless appropriate genetic evidence is provided. To solidify the conclusions made by Long et al, one needs to identify a PP2A mutant that is deficient in its interaction with only IRF-3 and replace the wild type gene with the mutant gene in cells and mice, before testing their biological functions. Notwithstanding the above caveats, the paper by Long et al identifies a major regulator of IRF-3 activation.