Pseudo-killer, Qu’estquec’est?

Abstract

Necroptosisis mediated by engagement of RIP-kinases and a downstream pseudokinase, MLKL. In this issue of Immunity, Murphy et al. (2013) show that it operates at or close to the final execution mechanism of the death process.

“Fa, fa, fa, fa, fa, fa, fa, fa, fa, fa”

David Byrne

Active cell death, the phenomenon where a cell responds to signals by engaging molecular pathways that lead to its demise, is best known as the process of apoptosis, in which caspases orchestrate the cell’s quietus. In the past decade, however, we have realized that there are other, non-apoptotic forms of active cellular suicide, including an intriguing and potentially important process of regulated necrosis called “necroptosis.” While caspases act as the “executioners” of apoptosis, the molecules involved in actually killing a cell that dies by necroptosis have been elusive. In this issue of Immunity, we move a step closer to unmasking this killer, a pseudokinase, that hides behind the moniker, Mixed Lineage Kinase- Like, or MLKL (Murphy, et al., 2013).

Necroptosis is best known as a response to ligation of a death receptor, one of a subset of members of the tumor necrosis factor receptor family, including TNFR1. It can also be engaged by cell surface Toll-like receptors that engage the signaling molecule, TRIF (Kaczmarek, et al., 2013). In both cases, and in striking contrast to apoptosis, necroptosis occurs when such signals are paired with inhibition or ablation of a caspase, caspase-8. This is because caspase-8, in the form of a heterodimer with another caspase-8-like molecule, c-FLIPL, actively antagonizes the signals that trigger necroptosis (Green, et al., 2011). These include two kinases, Receptor Interacting Protein kinase-1 (RIPK1) and RIPK3, both of which are activated in the process and form an amyloid structure upstream of the final death knoll (Li, et al., 2012). The kinase activity of both RIPK1 and RIPK3 are required for necroptosis, and inhibition of RIPK1 with an inhibitor, necrostatin, blocks the process. In turn, RIPK3 then phosphorylates MLKL, which is required for the ensuing necrotic death (Sun, et al., 2012; Zhao, et al, 2012) (see Figure 1). The requisite role of MLKL in necroptosis is clearly indicated by ablation of the gene in mice, as shown by Murphy, et al (2013) and another study (Wu, et al., 2013), and like RIPK3 ablation, the animals show no developmental or homeostatic phenotype in the absence of overt stress but lack cell death induced by TNF plus caspase inhibitor.

Figure 1. MLKL acts downstream of RIPK3 in necroptosis.

Ligation of death receptors or cell surface TLRs (via the adapter, TRIF) engage RIPK1, which then forms an amyloid-like complex with RIPK3. RIPK1 also recruits FADD, cFLIPL (FLIP) and caspase-8; the protease activity of the complex prevents necroptosis. Disruption or blockade of the protease activity allows the formation of the RIPK3 necrosome, which then phosphorylates MLKL. This probably induces MLKL “activation” via disruption of the K219-Q343 interaction in the pseudokinase domain. Active MLKL then promotes necroptosis and possibly other inflammatory events. (The author thanks Dr. Tudor Moldoveanu, who assisted with the image of the MLKL structure).

Necroptosis has been implicated in several pathological processes, including ischemia or reperfusion injury of brain, heart, and kidney, as well as pancreatitis, based on animal models in wild-type versus RIPK3-deficient mice. With the availability of MLKL-deficient animals, it will now be possible to further explore pathologies that may be associated with necroptosis, and indeed MLKL-deficient animals were resistant to pancreatitis induced by cerulean (Wu, et al., 2013). As definitive roles for MLKL in pathology are revealed, effective inhibitors of its function would be of value as therapeutics. Although such therapeutics are not yet available, the identification of necrosulfonamide as an MLKL inhibitor (Sun, et al., 2012) provides evidence that MLKL function is, in principle, drug-able.

By sequence homology, MLKL is a member of the TKL family of kinases, but while it conserves the ATP-binding VAIK motif (and binds ATP) it lacks the DFG and HRD domains needed for catalysis, and is therefore a pseudokinase (Boudeau, et al, 2006). Pseudokinases are known to have a number of functions, acting as partners or scaffolds for regulation (positive or negative) of bona fide kinases, phosphatases, histone acetyl-transferases, transcription factors, and other functions. Among the pseudokinases, MLKL is most closely related to IRAK2 (Boudeau, et al, 2006), which participates in TLR signaling to NF-kB. Intriguingly, it is also phylogenically related to the RIP-kinases, including RIPK1 and RIPK3. RIPK1, in addition to its kinase-dependent function in necroptosis, also has a kinase-independent scaffolding function that, similarly to IRAK2, promotes TLR- and TNFR1-mediated NF-kB activation.

Murphy, et al (2013) resolved the structure of MLKL and found a critical interaction between K219 of the VAIK motif and Q343in the activation domain where there are also three phospo-sites targeted by RIPK3. In contrast, in kinases, the corresponding lysine normally associates with what would be E239 in MLKL. The K-Q interaction is highly conservedin MLKL from different species (as K-Q or K-E), although inlemursthere is an E219-K343 “swap,” altogether suggesting that this hydrogen bond interaction is critically important for MLKL function. Strikingly, they found the mutation K219M results in an MLKL protein that directly causes necrosis, even in cells lacking RIPK3. They suggest that RIPK3 phosphorylation of MLKL promotes dissociation of K219-Q343 to activate the pro-necroptotic function of MLKL (Figure 1, inset). If so, then MLKL is likely to be the downstream effector of RIPK3 in the death process, and may well act as more than a regulator of RIPK3 and perhaps more than scaffold.

What then, does “activated” MLKL do to promote necroptosis? Although studies have implicated reactive oxygen, elevated metabolism, or other unknown effects involving mitochondria (Green, et al., 2011), a definitive demonstration that one or more of these represent the executioner for this form of death is lacking. Similarly, the mitochondrial fission protein DRP1 and the mitochondrial outer membrane protein PGAM5 have been proposed as agents of necroptosis (Wang, et al., 2012), but again neither definitive proof of their involvement nor a mechanism whereby they cause cell death has been described. Murphy, et al (2013) failed to observe an effect of silencing PGAM5, and while this cannot be taken as evidence that it is not involved in necroptosis, it precluded further explorations of its function in this study. Although MLKL binds to ATP it cannot hydrolize it, and mutation of the binding site to ablate ATP binding did not destroy the pro-necroptotic function of this molecule. At this point, the final effector of necroptosis downstream of MLKL has been elusive. We know that cells die in an active manner, and we know that RIPK3-phosphorylated MLKL functions in the process, but we do not know what ultimately kills the cell.

Further, we do not even know if the primary function of MLKL activation by RIPK3 is the execution of cell death, per se. It is axiomatic that such a pathway as this did not evolve as a mechanism to promote organismal diseases such as ischemia and reperfusion injury (obviously, such a function cannot be selected). However, active cell death is a fundamental strategy for a multicellular organism to resist intracellular infection; cells die prior to replication of the parasite, and indeed, necroptosis has been implicated in defense against some viruses. Perhaps not surprisingly, then, activation of RIPK3 is associated with inflammatory responses (Kaczmarek, et al., 2011), and in one case the response has been shown to be dependent on MLKL but possibly not cell death (Kang, et al., 2013). Necroptosis might then be somewhat analogous to the cell death process of pyroptosis, where caspase activation performs both pro-inflammatory cytokine production and orchestrates the death of the cell.

Nevertheless, the finding that the mutant MLKL K219M protein kills cells independently of RIPK3 raises an intriguing question: Might there be kinases other than RIPK3 capable of phosphorylating and “activating” MLKL to promote necrosis? A survey of RIPK3 phospho-targets suggests that the specificity of the kinase is similar to that of MAP kinases (Wu, et al., 2012), and therefore it is not ridiculous to think that a kinase with this specificity may in some cases activate the cell death function of MLKL. In this regard it is interesting that Jun-kinase (JNK) has long been associated with cell death, although the mechanisms have been obscure. That said, it is clear that JNK does not substitute for RIPK3 in necroptosis induced by TNFR or TLR ligation, and therefore a role for JNK in MLKL activation is, at best, a long shot. But in any case, if MLKL can mediate necrotic death independently of the RIP kinases, we may have to modify our definitions of necroptosis (perhaps as “necrotic cell death dependent on MLKL?”) and may well find that its roles in health and disease are even more extensive than currently thought. David Byrne once imagined a deranged murderer who felt he was launching into glory (“Je me lance, vers la gloire”). We have seen the face of MLKL, and while its function remains a mystery, we are coming to know, and will someday cage, this killer.

This is a commentary on article Murphy JM, Czabotar PE, Hildebrand JM, Lucet IS, Zhang JG, Alvarez-Diaz S, Lewis R, Lalaoui N, Metcalf D, Webb AI, Young SN, Varghese LN, Tannahill GM, Hatchell EC, Majewski IJ, Okamoto T, Dobson RC, Hilton DJ, Babon JJ, Nicola NA, Strasser A, Silke J, Alexander WS. The pseudokinase MLKL mediates necroptosis via a molecular switch mechanism. Immunity. 2013 Sep 19;39(3):443-53.