Triggering the Interferon Antiviral Response Through an IKK-Related Pathway
The IKK-related kinases IKKepsilon and TBK1 are identified as components of the virus-activated kinase activity that phosphorylates the C-terminal serine cluster of IRF-3 and IRF-7, linking pathogen sensing to type I interferon gene induction and to establishment of an antiviral state.
- Senior authors
- John Hiscott
- Correspondence
- John Hiscott; Rongtuan Lin
Research areas & themes
Citation
Sharma S, tenOever BR, Grandvaux N, Zhou G-P, Lin R, Hiscott J. Triggering the Interferon Antiviral Response Through an IKK-Related Pathway. Science 2003, volume 300, issue 5622, pages 1148-1151.
DOI 10.1126/science.1081315. PMID 12702806.
One-sentence contribution
The IKK-related kinases IKKepsilon and TBK1 are identified as components of the virus-activated kinase activity that phosphorylates the C-terminal serine cluster of IRF-3 and IRF-7, linking pathogen sensing to type I interferon gene induction and to establishment of an antiviral state.
Executive summary
Rapid induction of type I interferon on virus infection requires the coordinated activation of several transcription factors. By 2002 the routes to nuclear factor kappa B and to ATF-2/c-Jun were reasonably well described, but the kinase activity that phosphorylates the C-terminal serine and threonine cluster of IRF-3 and IRF-7, referred to at the time as the virus-activated kinase, had not been assigned to a defined enzyme. This work asked which member of the IKK family carries that activity. The authors expressed IKKalpha, IKKbeta, IKKepsilon and TBK1 in human cells and tested extracts against glutathione S-transferase fusions carrying the IRF-3 and IRF-7 C termini. Only IKKepsilon and TBK1 phosphorylated the IRF substrates, and a substrate in which the serine and threonine cluster was substituted with alanine was not phosphorylated. IKKepsilon expression produced slower migrating IRF-3 forms recognised by an antibody specific for phosphorylated Ser396, drove nuclear accumulation of IRF-3 and IRF-7 fluorescent fusions, and generated IRF-containing protein-DNA complexes. In reporter assays only IKKepsilon and TBK1, among the kinases tested, activated the IFNA4 and IFNB promoters, while all of them activated a nuclear factor kappa B reporter. Silencing IKKepsilon and TBK1 in A549 lung epithelial cells blocked virus-induced IRF-3 phosphorylation and ISG56 induction, and IKKepsilon expression suppressed vesicular stomatitis virus replication in a manner dependent on functional IRF-3. The work supplies the kinase step that had been missing from the interferon induction pathway.
Scientific context
Detection of an invading virus initiates signalling cascades that converge on the interferon beta enhancer, where nuclear factor kappa B, ATF-2/c-Jun and the interferon regulatory factors must act together. The paper states the gap plainly. The pathways activating nuclear factor kappa B and ATF-2/c-Jun had been well characterised, whereas the route to IRF-3 and IRF-7 activation remained undelineated and was described by the authors as a critical missing link. It was already established that IRF-3 and IRF-7 are activated by phosphorylation of a C-terminal domain by an uncharacterised virus-activated kinase, that this modification permits dimerisation, nuclear translocation and interferon gene activation, and that IRF-3 acts early while IRF-7 amplifies expression of interferon genes not induced in the first phase. A two-hybrid screen reported earlier had found an interaction between IRF-3 and the C-terminal domain of IKKalpha, which motivated a survey of IKK family members as candidate virus-activated kinases.
Central question
Which kinase or kinases constitute the virus-activated kinase activity that phosphorylates the C-terminal regulatory region of IRF-3 and IRF-7, and is that activity required for virus-induced interferon gene expression and for establishment of the antiviral state?
Experimental strategy
The design moves from biochemistry to cell biology to loss of function. First, candidate IKK family kinases were expressed in HEK293 cells and their activity tested directly against recombinant IRF-3 and IRF-7 C-terminal fragments, with an alanine-substituted substrate and a catalytically inactive kinase mutant as specificity controls, and with an IkappaBalpha fragment as a substrate that reports the classical IKK activity. Second, the consequences of kinase expression for the transcription factors themselves were read out at three levels that correspond to the accepted steps of IRF activation, namely site-specific phosphorylation detected with a phosphospecific antibody, subcellular localisation of fluorescent IRF fusions, and sequence-specific DNA binding measured by bandshift with antibody supershift for identity. Third, promoter reporters for IFNA4 and IFNB were used to ask whether these kinases are sufficient to drive interferon transcription, with a nuclear factor kappa B reporter run alongside to separate an IRF-specific effect from general IKK signalling. Finally, RNA interference in A549 lung epithelial cells tested necessity in a physiological infection setting, and plaque assays with vesicular stomatitis virus tested whether the pathway confers protection.
Key findings
- Whole-cell extracts from cells expressing IKKepsilon or TBK1, but not IKKalpha or IKKbeta, phosphorylated a glutathione S-transferase fusion carrying IRF-3 residues 380 to 427 in vitro, and a version of that substrate with alanine substitutions at Ser396, Ser398, Ser402, Thr404 and Ser405 was not phosphorylated. A catalytically inactive IKKepsilon(K38A) blocked the phosphorylation (Figure 1A). The observation is direct evidence that these two kinases act on the known IRF-3 regulatory cluster.
- IKKepsilon transcribed and translated in vitro and then immunoprecipitated phosphorylated both the IRF-3 and the IRF-7 C-terminal fragments, while IKKalpha, IKKbeta and IKKepsilon all phosphorylated an IkappaBalpha fragment (Figure 1B). The authors read this as direct action of IKKepsilon on the IRF substrates rather than action through an intermediate kinase in the extract.
- Inducible IKKepsilon expression generated slower migrating IRF-3 species detected by an antibody specific for phosphorylation at Ser396, a site previously shown to be critical for physiological IRF-3 activation (Figure 2A).
- IRF-3 and IRF-7 fluorescent fusions were predominantly cytoplasmic when expressed alone. Co-expression of IKKepsilon moved approximately 35 percent of IRF-3 and 95 percent of IRF-7 into the nucleus, while IKKepsilon(K38A) left under 5 percent nuclear. TBK1 also drove nuclear localisation (Figure 2B). The quantitative difference between the two factors is the basis for the authors' later suggestion that IRF-7 may be the preferred IKKepsilon substrate.
- IKKepsilon expression produced IRF-7 and IRF-3 containing protein-DNA complexes on an IRF-7 binding site and on the ISG15 interferon-stimulated response element, confirmed by antibody supershift (Figure 2C).
- Expression of IKKepsilon with IRF-7 stimulated the IFNA4 reporter about 2000-fold and IKKepsilon stimulated the IFNB reporter about 40-fold. Sendai virus infection alone or with IKKepsilon gave about 60-fold IFNB stimulation, and this was blocked by a dominant negative IRF-3 (Figure 3, A and B).
- Of IKKalpha, IKKbeta, IKKepsilon, TBK1 and NIK tested against the IFNA4 promoter with IRF-7, only IKKepsilon and TBK1 activated it, whereas every one of them gave 10 to 40-fold stimulation of a nuclear factor kappa B reporter (Figure 3D and supplementary figures). The separation is the paper's argument that the IRF branch is specific to the IKK-related kinases.
- RNA interference against TBK1 and IKKepsilon in A549 cells eliminated IKKepsilon protein and reduced TBK1 by 70 to 75 percent. In these cells virus-induced phosphorylation of endogenous IRF-3 was inhibited, ISG56 induction was reduced, and IRF-7-driven IFNA4 reporter activation decreased (Figure 4, A and B). This is the loss-of-function evidence that the two kinases are required, not merely sufficient.
- IKKepsilon expression reduced vesicular stomatitis virus titres by four logs to 10^6 plaque-forming units per millilitre, and co-expression of dominant negative IRF-3 restored titres to 10^10, with immunoblotting showing IKKepsilon-induced ISG56 and suppressed viral nucleocapsid protein (Figure 4, C and D). The authors state that an identical experiment with TBK1 gave similar results, citing unpublished data.
Mechanistic model
The data support a model in which virus sensing engages IKKepsilon and TBK1, which phosphorylate the C-terminal serine and threonine cluster of IRF-3 and IRF-7, permitting nuclear translocation, sequence-specific DNA binding and transcription of interferon genes, whose products then induce interferon-stimulated genes such as ISG56 and restrict virus replication. The kinase-to-substrate step is demonstrated biochemically and the requirement in infected cells is demonstrated by silencing. What the study does not establish is how virus detection reaches these kinases. The upstream sensor and the adaptor architecture that activates IKKepsilon and TBK1 after infection are outside the scope of the experiments, and the paper does not address them. Nor does the study resolve whether IKKepsilon and TBK1 act on the IRFs within a single complex or in separate complexes, or whether either kinase phosphorylates the IRFs directly in infected cells as opposed to in the reconstituted assays. The authors propose, on the basis of expression patterns and the localisation data, that TBK1 and IRF-3 may serve an early ubiquitous phase of the response while IKKepsilon, for which IRF-7 may be the preferred substrate, mediates amplification, and they further propose a possible functional link to the IKKalpha and IKKbeta complexes through TANK and IKKgamma/NEMO. Both of these are author interpretation rather than demonstrated results in this paper.
Conceptual or technical advance
Assigning the virus-activated kinase activity to IKKepsilon and TBK1 converts an operationally defined activity into named enzymes that can be expressed, mutated, silenced and, in principle, inhibited. It also draws a functional line within the IKK family, separating a nuclear factor kappa B arm served by IKKalpha and IKKbeta from an interferon regulatory factor arm served by the IKK-related kinases, with the promoter reporter comparison providing the evidence for that separation. The phosphospecific Ser396 readout and the kinase-dead and alanine-substituted controls together give a tractable assay system for the step, and the demonstration that IKKepsilon-driven restriction of vesicular stomatitis virus depends on functional IRF-3 makes the pathway testable as an antiviral effector axis rather than only a transcriptional one.
Relationship to the broader research program
This paper predates the tenOever laboratory and comes from doctoral training in the Hiscott laboratory at the Lady Davis Institute and McGill University, where Benjamin tenOever was one of three authors marked as having contributed equally and was not the senior author. Correspondence was handled by John Hiscott and Rongtuan Lin. Read against the later independent work, the paper sits at the origin of a continuing interest in how virus detection is converted into a transcriptional antiviral program and in how that program is measured. The interferon-stimulated gene readouts and the IRF-3 dependent logic used here recur in the later laboratory's work on innate immune signalling and on transcriptional profiling of infected tissue, though establishing those continuities is category 3 synthesis and requires the later papers to be set alongside this one.
Related publications
- Blanco-Melo and colleagues, 2020, conceptual extension. That study reads out the type I interferon response as a transcriptional program during SARS-CoV-2 infection, the same output whose induction step is defined here, although no direct methodological line is asserted from this paper.
The list is deliberately short. This record was built from one paper read in isolation, and the reference list of this report points to work by other laboratories rather than to the tenOever corpus.
Limitations and boundaries
Much of the evidence rests on ectopic expression of kinases and transcription factors in HEK293, Vero and COS-7 cells, together with recombinant fragment substrates, so the results establish sufficiency and direct biochemical capability rather than the stoichiometry or the complex composition that operates in an infected cell. The necessity argument comes from a single loss-of-function setting, RNA interference in A549 lung epithelial cells with partial TBK1 knockdown of 70 to 75 percent, and no genetic knockout is presented here. Two viruses are used, Sendai virus and vesicular stomatitis virus, both RNA viruses assayed over hours in cell culture, so nothing is shown for other virus classes, for primary cells or tissues, or for animal infection. The claim that TBK1 behaves like IKKepsilon in the virus replication experiment is supported by data not shown. The authors themselves flag that the response may be cell-type specific, since TBK1 is ubiquitously expressed while IKKepsilon expression is inducible in lymphoid and other cell types, and they present the division of labour between the two kinases and the possible link to the classical IKK complex as suggestions rather than findings. The pathway upstream of the kinases is not addressed.
Audience summaries
25 words
Two related kinases, IKKepsilon and TBK1, phosphorylate the interferon regulatory factors IRF-3 and IRF-7, switching on interferon genes and the cellular antiviral state after virus infection.
75 words
Cells infected by a virus switch on interferon genes, but the enzyme that activates the key transcription factors IRF-3 and IRF-7 had not been identified. Testing the IKK kinase family, the authors found that IKKepsilon and TBK1, and not the classical IKKalpha or IKKbeta, phosphorylate the regulatory tail of both factors, move them into the nucleus and turn on interferon promoters. Silencing the two kinases blocked the response in lung epithelial cells.
150 words
Induction of type I interferon requires phosphorylation of a C-terminal serine and threonine cluster in IRF-3 and IRF-7 by an activity that had been defined only operationally as the virus-activated kinase. Surveying the IKK family, the authors show that IKKepsilon and TBK1, but not IKKalpha or IKKbeta, phosphorylate recombinant IRF-3 and IRF-7 C termini in vitro, with loss of phosphorylation on an alanine-substituted substrate and with a kinase-dead IKKepsilon acting as a block. In cells, IKKepsilon generates Ser396-phosphorylated IRF-3, drives nuclear accumulation of both factors, and produces IRF-containing DNA complexes. Only the two IKK-related kinases activate IFNA4 and IFNB reporters, while all family members tested activate a nuclear factor kappa B reporter, separating the two arms. Silencing IKKepsilon and TBK1 in A549 cells blocks virus-induced IRF-3 phosphorylation and ISG56 induction, and IKKepsilon restricts vesicular stomatitis virus in an IRF-3 dependent manner. The upstream sensing step is not addressed.
Documented publication relationships
- Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19 — conceptual extension.
- Multiple Functions of the IKK-Related Kinase IKKε in Interferon-Mediated Antiviral Immunity — predecessor.
- Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19 — methodological foundation.