Alternatively Spliced Isoforms of IRF7 Differentially Regulate Interferon Expression to Tune Response to Viral Infection

Summary

Interferon Regulatory Factor 7 (IRF7) is a master transcriptional regulator of innate immunity. IRF7 binds as a homodimer, or heterodimer with IRF3, to promoters of type I interferons (IFN-I) to drive their expression, which activates expression of antiviral genes. Here we demonstrate that alternative splicing of the first intron within the coding region of human IRF7 is regulated across immune tissues and in response to immunologic stimuli. Retention of this intron generates an alternative translation start site, resulting in an N-terminally extended form of the protein (exIRF7) compared to the canonical isoform (cIRF7). We find that exIRF7 has increased dimerization relative to cIRF7, uniquely activates expression of IFN-Is in response to dsRNA sensing, and controls viral infection to a greater extent than cIRF7. Thus, alternative splicing of IRF7 is a previously unrecognized mechanism by which human cells tune IRF7 function and the interferon response to control immune challenge.

Introduction

Alternative splicing is a mechanism of gene regulation that broadly controls protein identity through the differential inclusion or skipping of segments of pre-mRNA, which can alter the encoded protein. Such alternative splicing is often regulated by changes in the activity or expression of RNA binding proteins (RBPs) in response to cellular signals, to enable cells to modulate protein function in response to extracellular cues^1–4. In particular, work in recent years has revealed widespread changes in alternative splicing in response to viral infection and innate immune signaling^5–9. Despite this recent progress, however, we have a poor understanding of the extent to which alternative splicing plays a role in tuning the interferon response, which functions as a central component of innate immunity.

Interferon Regulatory Factor 7 (IRF7), and its homologue IRF3, are master transcriptional regulators of the innate immune response that drive expression of type I interferon genes (IFN-I) including IFNβ and IFNαs^10–12. Type I interferons (IFN-I), in turn, are secreted and bind their receptor in an autocrine and paracrine fashion to activate JAK-STAT signaling, thereby inducing the expression of numerous Interferon Stimulated genes (ISGs), many of which inhibit viral replication and activate cellular pathways to combat infection^13,14. IFNs and ISGs also play a central role in tuning inflammation and susceptibility to autoimmune diseases¹².

IRF7 and IRF3 are both comprised of an N-terminal DNA binding domain (DBD), a central transactivation domain (TAD), and a C-terminal region that undergoes extensive post-translational modifications to regulate dimerization, localization and function¹². In response to viral infection, or detection of pathogen-associated molecular patterns (PAMPs) by pattern recognition receptors (PRRs), IRF7 and IRF3 are activated by phosphorylation, which induces their homo- and hetero-dimerization and nuclear translocation^10,12. Once in the nucleus, IRF7 binds to a range of related but distinct DNA sequences, including interferon regulatory factor-binding elements (IRFEs) upstream of IFN-I mRNAs and interferon-stimulated response elements (ISREs) upstream of ISG mRNAs¹⁵. Through these elements, IRF7 activates transcription of type I IFN mRNAs in response to PAMPs, and promotes expression of ISG mRNAs both directly and via IFN-induced JAK-STAT signaling^15,16. Of note, IRF7 is typically thought to function as a heterodimer with IRF3 but can also function as a homodimer. IRF7 also associates with additional transcription factors, most notably NF-kB and ATF-2/cJun to activate the IFNβ promoter^10,15–18.

IRF3 is ubiquitously expressed across tissues and is regulated solely by phosphorylation¹⁹. By contrast, IRF7 expression is more variable: in most cell types IRF7 is expressed at only low levels under basal conditions, but it is constitutively expressed in barrier epithelial cells and some lymphoid populations to promote initial induction of IFN-I^20,21. Regardless of basal expression of IRF7, IRF7 is strongly induced by IFN-I signaling to form a positive feedback loop that amplifies IFN-I pathway activation^12,15,22. In mouse models, loss of IRF7 has a greater impact on IFN-I expression and anti-viral immunity than loss of IRF3^11,23. Moreover, humans with loss of function mutations in IRF7 are highly susceptible to viral infections^11,24,25 and autoimmune diseases^12,26. Therefore, control of both the expression and activity of IRF7 is essential for tuning proper immune function and productive anti-viral defense.

We have previously demonstrated that both adaptive immune signaling and viral infection induce changes in splicing that promote productive immune responses^6,27–29. Therefore, we explored genes involved in innate immunity for regulated alternatively splice forms that may impact innate immune signaling. Strikingly, we find that inclusion of the first intron of the human IRF7 coding region is highly variable across immune cell types and tightly regulated in response to a range of immune stimuli. Inclusion of this short (85 nt) intron alters the translation start site and results in a novel 19 amino acid N-terminal extension adjacent to the DNA binding domain of this key transcription factor, relative to the canonical isoform (cIRF7, also called IRF7-A) (Fig 1A). We find that the intron-retained extended IRF7 (exIRF7) isoform exhibits enhanced activity on the promoters of type I IFNs, particularly IFNB1 which encodes IFNβ, especially at low levels of expression. This enhanced activity is due to greater affinity of homodimerization, and heterodimerization with IRF3, both on and off DNA, and results in increased anti-viral immunity through robust IFN-I production. Together, these data reveal a previously unrecognized mechanism by which alternative splicing of the human IRF7 gene tunes the interferon response.

Figure 1. Alternative splicing of human IRF7 Intron 1 is regulated in a tissue-specific and stimulation-dependent manner.

(A) (top) Schematic of the IRF7 protein: DNA binding domain (DBD), trans-activation domain (TAD), IRF-association domain (IAD), and nuclear export signal (NES). (bottom) Schematics of the alternatively spliced variants indicating the translation start sites (green boxes) in exon 1 and intron 1 and stop codon in intron 1 (red octagon). Isoform-specific amino acids preceding the DBD are indicated in blue for exIRF7 and red for cIRF7. (B) Plot of IRF7 intron 1 retention across human cell types from public GTEx data. Blue bars indicate percent of total IRF7 expression that retains intron 1. Red area indicates percent of cIRF7 isoform. (C) RT-PCR analysis of IRF7 intron 1 alternative splicing in indicated human cell types under different stimulation conditions. (D) Quantification of n=3 RT-PCR data as in panel C (data are mean ± SD). Conditions that preferentially express exIRF7 are in blue, conditions that preferentially express cIRF7 are in red. Black bars indicate near equal expression of both isoforms.

Results

Alternative splicing of human IRF7 is regulated in a tissue- and stimulation-dependent manner.

Given the precedent that adaptive immune signaling regulates alternative splicing of genes which tune the immune response^{1,3,6,27,28,30}, we explored alternative splicing of canonical innate immune factors. We focused on IRF7 as this transcription factor plays a central role in the interferon response, and in mouse macrophages TNF-α treatment causes downregulation of IRF7 protein via retention of IRF7 intron 4, which generates a premature stop codon³¹ (Fig S1A). The initial cloning and characterization of the IRF7 gene in humans reported an isoform in which the first intron after the translation start site is retained^32,33. Retention of this first intron (intron 1) is predicted to result in translation from an in-frame intronic AUG start site to yield a unique extended N-terminus (Fig 1A, exIRF7, also called IRF7-D and IRF7H in previous literature^32,33). However, due to the limited ability to detect alternative splicing at the time, initial characterization of IRF7 indicated that the intron-retained extended isoform was not highly expressed in the cell lines tested and did not vary upon viral infection^32,33. Moreover, the murine gene does not have the intronic AUG, and thus cannot produce the extended isoform. Therefore, the bulk of published research on IRF7 through the years has focused on the fully spliced canonical isoform (cIRF7), and no studies have systematically compared the expression or function of cIRF7 with the extended exIRF7 isoform.

Strikingly, using publicly available RNA-Seq data from primary human tissues³⁴, we observe marked variability in the retention of intron 1 across primary human immune cell types, ranging from almost full inclusion in regulatory and memory T and B cells to predominantly spliced in macrophages and differentiated CD4+ T cells (Th1, Th2) (Fig 1B, blue bars, Fig 1C, S1B). By contrast, intron 1 exhibits low and consistent inclusion in non-immune primary human cells, such that the canonical isoform (cIRF7, also called IRF7A in previous literature^32,33) predominates in these non-immune tissues (Fig 1B). Importantly, unlike intron 1, IRF7 intron 4 exhibits low and consistent inclusion across all human primary cell types (Fig S1C) and variation of intron 1 inclusion is independent of transcript abundance or retention of intron 4 (Fig S1C), suggesting a distinct mode of control between these introns.

Given that monocytes differentiate to macrophages during infection and inflammation, the difference in IRF7 intron 1 retention between monocytes and macrophages suggests that this splicing event is regulated in response to immune triggers. We first confirmed the preferential retention of IRF7 intron 1 in isolated human monocytes by RT-PCR (Fig 1C,D). Consistent with immune-induced regulation of splicing, treatment of cultured primary human monocytes with poly-inosine:cytosine (pIC), a mimic of dsRNA as sensed during viral infection, leads to a clear decrease in retention of intron 1 (Fig 1C,D). By contrast, treatment of CD4 T cell-like Jurkat cells or A549 lung epithelial cells with pIC leads to a rapid and dramatic increase in IRF7 intron 1 retention, resulting in more exIRF7 mRNA over the first ~4-16 hours of stimulation, followed by a return to baseline (Fig 1C,D, S1D). Interestingly, infection of A549 cells with several RNA viruses has the opposite effect, increasing splicing rather than promoting intron retention (Fig 1C,D), suggesting that these viruses have some activity that counters the dsRNA-induced response. Treatment of Jurkat cells with PMA, which activates cells in a manner analogous to T cell signaling, also increases splicing of IRF7 intron 1 (Fig 1C,D), with kinetics similar to other PMA-induced splicing changes we have observed in past studies^35,36 (Fig S1D).

Given that pIC, PMA and viral infection all induce the transcription of IRF7 (Fig 1C), but have distinct impacts on splicing (Fig 1C,D), we conclude that changes in transcript abundance does not drive the splicing changes in IRF7 intron 1. A lack of correlation between transcription and splicing is further evident by the fact that treatment of both Jurkat and A549 with IFNα, which strongly induces expression of IRF7, results in no change in IRF7 splicing (Fig 1C,D). Together these data demonstrate that the splicing of IRF7 intron 1 is tightly and directly regulated in a signal-dependent manner across many different human cell types and immune triggers, independently of IRF7 transcription. This signal-induced regulation of intron retention is specific for human IRF7 intron 1, as we observe no change in intron 4 splicing in either Jurkat or A549 cells in response to any stimuli or virus tested (Fig S1E). Moreover, intron 1 in the murine gene shows no evidence of retention in any mouse cells, and lacks an intronic in-frame AUG. Therefore, as elaborated below (Discussion), the human and mouse systems utilize distinct strategies for the regulation of IRF7 by splicing.

JNK-induced expression of CELF2 enhances splicing of IRF7 intron 1 upon PMA stimulation

Splicing efficiency of any intron is typically regulated by the activity of RNA binding proteins (RBPs) that enhance or inhibit the association and activity of the splicing regulatory machinery (spliceosome) with the splice site sequences^37,38. Given the many different stimuli that impact the splicing versus retention of human IRF7 intron 1, we anticipate that this intron is regulated by multiple distinct RBPs and signaling pathways. As a first step to understanding the complexity of IRF7 intron 1 regulation, we focused on the enhanced splicing of this intron upon PMA stimulation of Jurkat cells, as we have previously identified several RBPs that are regulated under these conditions^27,28,35,39.

Using a standard minigene approach²⁸, we first demonstrated that IRF7 intron 1 itself, and its flanking exons, are necessary and sufficient for PMA-induced splicing. A minigene consisting of the entirety of exon 1, intron 1 and exon 2, exactly recapitulates the degree of splicing of intron 1 observed in the endogenous gene under both naïve and PMA-stimulated cells (Fig 2A, S2A; MG1). However, substitution of either of the flanking IRF7 exons with those from a heterologous unregulated b-globin gene²⁸ essentially abrogates all splicing and PMA-induced regulation of intron 1 (Fig 2A, S2A; ΔE1, ΔE2). By contrast, substitution of the IRF7 intron 1 with a b-globin intron of similar length, results in efficient splicing under all conditions (Fig 2A,S2A; BGint). Together, these data suggest that the IRF7 intron 1 is inherently inefficiently spliced, and sequences within both flanking exons promote its recognition and splicing by the spliceosome, likely through the recruitment of RBPs.

Figure 2. The JNK-induced RNA binding protein CELF2 drives PMA-induced splicing of IRF7 intron 1.

(A) (top) Schematics of minigenes with IRF7 derived sequences in light grey and β-globin derived sequences in black. (bottom) Quantification of intron retention (n=3) in naive (black) or PMA-stimulated (red) conditions from the endogenous IRF7 gene (Endo) or minigenes as shown on top. (B) Quantification of the endogenous IRF7 intron 1 retention (n=3, data are mean ± SD) in naive (black) or PMA-stimulated (red) cells depleted of indicated RBP by stable expression of shRNAs against each RBP^36,48–50. Insert shows location of CELF2 binding motifs (CUG/UG) in IRF7. (C) Quantification of intron 1 retention from the endogenous IRF7 gene under indicated conditions (n=3, data are mean ± SD). Representative gels are shown in Supplemental Figure S2.

We have previously shown that the RBPs hnRNP LL, PSF and CELF2 all exhibit increased expression and/or activity upon PMA stimulation of Jurkat cells^27,28,35,39. We therefore tested the impact of depletion of these proteins, along with additional RBPs we have shown to contribute to splicing patterns in Jurkat cells. Notably, of these proteins, the only RBP to exhibit a significant impact on IRF7 intron 1 splicing is CELF2; whose depletion blocks all PMA induction of splicing (Fig 2B,S2B). In previous work, we have demonstrated that CELF2 expression is strongly enhanced upon PMA stimulation of Jurkat cells, peaking at 24-48 hours post-stimulation, via JNK-dependent stabilization of its mRNA^28,35. The time course of PMA-induced IRF7 intron 1 splicing is consistent with this JNK-mediated induction of CELF2 (Fig S1D). As additional evidence for the role of CELF2 in regulating IRF7 splicing, we directly tested the impact of JNK-signaling, and indeed find that inhibition of JNK kinase, but not TBK1, reduces the PMA-induced impact on intron 1 splicing (Fig 2C). Taken together, these data demonstrate that JNK-induced CELF2 expression promotes splicing of IRF7 intron 1 in response to PMA. We note that IRF7 exon 1, intron 1 and exon 2 all contain CUG and UG repeats, consistent with the known CELF2 binding motif (Fig 2B insert), although exactly how CELF2 activates IRF7 intron 1 splicing remains the topic of future studies. In addition, we observe no changes in CELF2 expression upon pIC treatment of Jurkat cells, therefore additional RBPs and/or mechanisms likely contribute to active repression of IRF7 intron 1 splicing in response to innate immune signaling (see Discussion).

Isoforms generated via alternative splicing of IRF7 intron 1 exhibit distinct transcriptional activities.

Given the regulated inclusion of IRF7 intron 1, we wanted to investigate the functional impact of this alternative splicing event. Unlike most examples of intron retention, which inhibit protein production through the addition of premature stop codons, both isoforms generated through alternative splicing of IRF7 intron 1 yield an open reading frame, differing only in the N-terminal 6 or 19 amino acids immediately adjacent to the DNA binding domain. This led us to hypothesize that there may be distinct transcriptional activity between the two IRF7 isoforms.

As a first step to determining differences in function between cIRF7 and exIRF7 we used CRISPR gene editing to generate Jurkat clones that only express one isoform of IRF7. We made targeted cuts in the introns flanking the regulated first coding intron and used homology-directed repair to replace this region with either a fusion of the flanking exons (thus modeling splicing and expression of cIRF7), or a version of the gene carrying two base changes at the 5’ splice site of the intron which prevents splicing (thus enforcing exclusive expression of exIRF7). The 5’ splice site mutation used does not cause any change in the encoded amino acid, or in the presence or location of the start and stop codons (Fig 3A, see Methods), thus retaining all the translation information and potential regulation of the native gene. We confirmed two clones of each genotype by Sanger sequencing and demonstrated by RT-PCR that these clones express only the predicted RNA isoform. In contrast, wild type cells (WT) express a mixed population with near-equal abundance of both isoforms (Fig 3B, see Methods). Importantly, each of the CRISPR-edited clones express similar levels of total IRF7 protein (Fig 3B) upon stimulation with IFNα, confirming that the exIRF7 RNA isoform is translated into a stable protein. As a further control for subsequent experiments, we also generated IRF7 knockout (KO) cell lines by deleting the third coding exon by CRISPR, thus generating a premature termination codon (see Methods), and we confirmed loss of protein expression by Western blot (Fig 3B, bottom).

Figure 3. Isoforms generated through alternative splicing of IRF7 intron 1 exhibit distinct transcriptional activities in response to innate immune activation.

(A) Design of CRISPR-Cas9 genome editing to create cell lines expressing only cIRF7 or exIRF7. (B) (top) RT-PCR confirmation of IRF7 RNA isoforms in the CRISPR-modified cells (cIRF7 and exIRF7) compared to wildtype (WT) cells. (bottom) Western blot confirmation of exIRF7 and cIRF7 protein in the CRISPR-modified cells relative to WT cells and IRF7 KO cells. (C) RT-qPCR analysis of selected differentially expressed genes (DEGs) from RNA-seq (n=3, mean ± SD) under conditions of IFNα and pIC, with and without ruxolitinib treatment. Significant differences (p<0.05, unpaired student t-test) are indicated. (D-F) Scatter plots based on RNA-seq data showing DEGs in exIRF7 and cIRF7 compared to IRF7 KO cells after induction with pIC (D, F: 1 μg/ml for 6 hours) or IFNα (E: 500 U/ml for 6 hours). Panel F is the same as panel D but with different gene groupings indicated. Significant DEGs were identified with a base read minimum of 50, a minimum log2-fold change of 1 and a p-value threshold of 0.05. Additional details are in Supplemental Data 1 and Figure S3.

We first investigated how each IRF7 isoform impacts expression of type I interferons, as these are the canonical targets of IRF7 downstream of PRR engagement. Jurkat cells do not express any IFNαs tested, therefore we focused on IFNβ. As expected, we observed modest induction of IFNβ upon stimulation with pIC in IRF7KO cells, given their expression of IRF3. Strikingly, however, we observe strongly enhanced IFNβ induction in cells expressing exIRF7 relative to IRF7-KO cells, but no such induction in cells expressing cIRF7 (Fig 3C). Furthermore, by both qPCR (Fig 3C, S3A) and RNA-Seq (Fig 3D, blue and purple dots) across two independent clones of each genotype, we identify 74 additional genes that are significantly upregulated by pIC treatment in the exIRF7 cells relative to IRF7-KO cells (Supplemental Data 1). Remarkably, most of these genes exhibit little response to pIC in cells expressing cIRF7 (Fig 3D, blue dots). A general lack of responsiveness to pIC in cIRF7 expressing cells is also observed by principal component analysis, which reveals that the pIC-treated cIRF7 cells cluster with unstimulated exIRF7 and IRF7-KO cells (Fig S3B).

By contrast to the pIC-induced expression, direct treatment of the CRISPR clones with type I IFN identifies ~50 ISGs which are similarly induced in cIRF7 and exIRF7-expressing cells relative to the IRF7-KO cells (Fig 3E, panIRF7-specific; Fig S3A, STAT2). This panIRF7 activation is consistent with reports that IRF7 can cooperate or act in parallel with the JAK-STAT induced ISGF3 complex to enhance transcription of some ISGs^15,16. Importantly, this data also confirms that cIRF7 in our cells retains transcriptional activity under some conditions, namely IFN-I signaling. One factor that could explain the observed activity of cIRF7 in IFN-I treated cells, but not in response to pIC is that both cIRF7 and exIRF7 are expressed at higher levels in IFN-I treated cells than in pIC-induced conditions (Fig S3C), as will be discussed further below. Finally, as expected, we also observe a broad program of ISG induction by IFN-I that is independent of IRF7 (i.e. genes where induction in the IRF7-KO cells is similar to those in the cIRF7/exIRF7 expressing cells), presumably through the sole activity of ISGF3 (Fig 3E, IRF7-independent; Fig 3C, OAS1).

Of the 74 genes that are specifically induced in exIRF7 cells by pIC treatment, 43 (59%) are known ISGs (Supplemental Data 1). This enrichment of ISGs suggest that the program of pIC-induced gene expression in the exIRF7 cells maybe largely a downstream consequence of exIRF7-driven induction of IFNβ. As further evidence that the ISGs induced by pIC in exIRF7 cells are a secondary effect of exIRF7-induced IFNβ, 64% of the genes induced by pIC in the exIRF7-expressing cells overlap with those induced by IFNα in the IRF7-KO cells (Fig 3F, yellow dots). Moreover, treatment of cells with the JAK inhibitor ruxolitinib, blocks JAK-STAT signaling downstream of IFN-I receptor engagement, abolishes the pIC-induced expression of the ISGs OAS1 and STAT2 in the exIRF7 expressing cells (Fig 3C, S3A), but has no impact on the exIRF7-specific pIC-induction of IFNβ, or genes such as NFKBIA, which are not induced by IFNα treatment (Fig 3C, S3A). Together, these data suggest that exIRF7 uniquely regulates a small subset of genes upon pIC activation (Fig 3F, blue dots), which critically includes IFNβ. This exIRF7-enhanced transcription of IFNβ subsequently activates a robust ISG response.

Of note, the set of genes uniquely induced in exIRF7 cells by pIC also includes several repressors of transcriptional activation (NFKBIA, CXXC4, DAB2IP; Fig S3A and Supplemental Data 1) which may explain the down-regulated genes observed in pIC treated cells (Fig 3F, blue dots with negative values). As further evidence that the down-regulated genes are indirectly controlled by IRF7, genes that are repressed upon pIC in the exIRF7 cells mostly lack an ISRE consensus (GAAANNGAAA) near the transcription start site (TSS) (Fig S3D-F, light blue). Indeed, the pattern of predicted ISREs in this set mirrors that of non-responsive genes (Fig S3D-F, gray), and in the few repressed genes that do score as having an ISRE, there is typically only a single copy which is a poor match to the consensus (Fig S3F,G). By contrast, the majority of genes that are enhanced by either isoform of IRF7 in response to IFNα (pan-IRF7 enhanced) or enhanced by exIRF7 in response to pIC, have 2-4 strong matches to the ISRE consensus motifs within 200 nucleotides of the TSS (Fig S3D-G, purple).

To confirm that IRF7 isoform-specific gene expression is a general activity of IRF7 observed in other cell types, we also engineered human A549 lung epithelial cells to express single isoforms of IRF7 (Fig S4A-B). As we were unable to use the same CRISPR repair strategy to force IRF7 isoform expression in A549 cells due to inefficiency of homologous recombination, we generated IRF7-KO A549 cells (Fig S4A) and reconstituted these cells with a stable transgene expressing C-terminally 3xFLAG-tagged cIRF7 or exIRF7 (Fig S4B). We confirmed these cell lines express similar levels of cIRF7 and exIRF7 (Fig S4B) and that both isoforms are retained in the cytoplasm in uninfected cells but transit efficiently to the nucleus upon pIC treatment (Fig S4C). Importantly, A549s expressing single isoforms of IRF7 fully recapitulate the exIRF7-specific IFNβ expression pattern observed in the Jurkat CRISPR clones (Fig S4D). Several IFNαs are also induced by pIC in A549 cells, and these are also specifically regulated by exIRF7(Fig S4D). Therefore, we conclude that exIRF7 uniquely induces transcriptional activation of type I IFN genes in response to PRR activation across diverse cell types and systems.

exIRF7 preferentially dimerizes and forms specific complexes on the IFNβ promoter

To understand the molecular basis for the enhanced activity of exIRF7 relative to cIRF7 in driving expression of its critical target gene IFNβ, we first asked if the promoter sequence for IFNβ was sufficient for the isoform-specific activation of transcription using luciferase reporter assays (Fig 4A). These experiments were done in HEK293 cells using the canonical stimuli Sendai virus (SENV) since SENV is a more robust stimuli of PRRs in HEK293 cells than pIC¹⁸. We found that introduction of the minimal core promoter of IFNβ (aka PRD31) displays modest induction by cIRF7 or exIRF7; however, transfection of the full IFNβ promoter reveals that exIRF7 exhibits more robust activation than cIRF7 in response to SENV (Fig 4A,B). These results suggest that the IFNβ promoter is sufficient for the differential activity of exIRF7 and cIRF7, and provide further evidence that the difference in activity between these protein isoforms is observed across cell types and PRR stimuli.

Figure 4: exIRF7 uniquely interacts with and preferentially enhances the activation of the IFNβ promoter.

(A) Luciferase assay with cells co-expressing IRF7 isoforms, a NanoLuc internal control, and firefly luciferase under the control of an IFNβ promoter segment, before or after sendai virus (SENV) infection (500HAU/ml for 6hrs). Luciferase values are normalized to the NanoLuc control and shown relative to untreated cIRF7-expression. Each point is an independent biological replicate. Bar represents mean ± SD. Significant differences (p<0.05, unpaired student t-test) are indicated. (B) Sequence of IFNβ promoter used in reporter assays and EMSAs (panels D-E). (C) Western blot analysis of Flag-affinity purified recombinant IRF7 proteins used in gel-shift assays showing relative expression levels, phosphorylation status, and absence of known IRF7 co-factors. Proteins were purified from Jurkat cells after treatment with pIC (1 μg/ml for 6 hours). (D) Representative native gel-shift assay (EMSA) using indicated amounts of IRF7 proteins from panel C incubated with a radiolabeled DNA oligo corresponding to core IFNβ promoter as shown in panel B (green sequence). Migration of free DNA probe and protein:DNA complexes is indicated. Blue arrow highlights the unique DNA:protein complex observed with exIRF7. Replicate gels with independent protein preps shown in Supplemental Fig S5. (E) EMSA as in panel D, but with IRF7 protein held constant and antibodies to IRF7, known co-factors, and IgG control added.

Having shown that the IFNβ promoter is sufficient to reveal differential activity of exIRF7 versus cIRF7, we next investigated the binding of the IRF7 isoforms to the IFNβ promoter element by native gel analysis (EMSA). We purified C-terminally flag-tagged cIRF7 and exIRF7 from pIC treated Jurkat cells and confirmed these were both phosphorylated and free of other established activators of IFNβ, including IRF3 (Fig 4C). Consistent with previous studies^18,40, both cIRF7 and exIRF7 bind to the core IFNβ promoter (Fig 4D). However, we observe a unique, more slowly migrating species with exIRF7 that is not observed with the cIRF7 isoform (Fig 4D, blue arrow; see also Fig S5). We confirmed that this unique species observed on the IFNβ promoter corresponds to IRF7 binding, and not a residual co-factor, as this species is only super-shifted by anti-IRF7 antibodies (Fig 4E).

Given that IRF7 alone forms distinctly migrating species on the native gels, and IRF7 is known to homodimerize, we asked if the more slowly migrating EMSA species observed with exIRF7 might indicate a preference to form a dimer or higher order oligomer. Therefore, we turned to Mass Photometry, a method that uses light scattering to determine the mass of single biomolecules in their native state in solution⁴¹. Using mass photometry with our purified samples of exIRF7 and cIRF7, we indeed observe that purified exIRF7 is primarily a dimer at concentrations used in our assays, while cIRF7 is predominantly monomeric (Fig 5A). Moreover, by titrating protein in the mass photometry analysis, we determined the Kd of dimerization for exIRF7 to be ~6-fold tighter (i.e. higher affinity) than for cIRF7 (Fig. 5B). As the Mass Photometry experiment was done with purified protein in the absence of DNA, these results indicate that preferential dimerization of exIRF7 relative to cIRF7 is an intrinsic property of these proteins.

Figure 5: exIRF7 preferentially homodimerizes and binds cooperatively with IRF3.

(A) Mass photometry traces of cIRF7 (top) and exIRF7 (middle). Mass of major populations are indicated as determined by fitting guassian curves as described in Methods. Bottom plot is the overlap of the individual traces. (B) Percent dimer of exIRF7 (blue) and cIRF7 (red) as calculated by mass photometry across a range of protein concentrations (n=3, mean ± SD). **** indicates data points that differ significantly (p<0.0001, Tukey’s multiple comparison test) between isoforms. Kd calculated by curve fitting the data is shown. (C) Co-precipitation of IRF3 with Flag-tagged exIRF7 or cIRF7 from pIC stimulated Jurkat cells. IP conditions involved 150 mM NaCl and stringent washes (see Methods). Results from lower-stringency IP conditions are shown in Supplemental Figure 7C. (D) Representative native gel-shift assay (EMSA) using 200 ng of purified IRF7 protein, and/or 20 or 40 ng of purified recombinant IRF3, incubated with a radiolabeled DNA oligo corresponding to core IFNβ promoter as shown in Figure 4B. Migration of free DNA probe and protein:DNA complexes is as indicated. Blue arrow highlights the unique exIRF7 complex, the black arrow indicates the IRF3 and IRF3/IRF7 complexes. Replicate gels with independent protein preps shown in Supplemental Figure S5.

In cells, IRF7 can function on the IFNβ promoter as both a homodimer as well as a heterodimer with IRF3^18,40. Therefore, we also tested the ability of each of the isoforms of IRF7 to associate with IRF3. First, we performed co-immunoprecipitation from cells of IRF3 with each of the IRF7 isoforms. Under low stringency conditions (60-100 mM KCl) we observe interaction of IRF3 with both cIRF7 and exIRF7 following pIC treatment, consistent with previous studies¹⁷ (Fig S5C). However, under more stringent co-immunoprecipitation conditions with higher salt and more rigorous washes, we still detect an interaction of IRF3 with exIRF7, while the IRF3-cIRF7 interaction is no longer stable (Fig 5C), suggesting that IRF3 binds more strongly to exIRF7 than to cIRF7. This interaction of exIRF7 is with the active form of IRF3 as indicated by IRF3 phosphorylation (Fig 5C). Consistently, we also observe stronger cooperative binding between exIRF7 and IRF3 on IFNβ DNA than is observed for cIRF7 with IRF3 (Fig 5D, black arrow; Fig S5D).

Taken together, we conclude that the increased activity of exIRF7 in promoting expression of IFNβ relative to cIRF7 reflects the enhanced ability of exIRF7 to form the optimal enhanceosome structure through cooperative binding with IRF3 and/or itself⁴⁰. Therefore, at lower expression levels of IRF7, such as observed in the pIC treated cells (Fig S3C) or early in infection^10,42, only the exIRF7 isoform can dimerize to activate IFNβ and induce the interferon response. By contrast, we predict the cIRF7 isoform is primarily active when its expression increases beyond a certain threshold, as observed in IFN-I induced cells. Such a model is consistent with much previous work that has suggested that cIRF7 predominantly functions to amplify interferon signaling later in infection^12,15,22.

exIRF7 promotes inflammation and anti-viral immunity

Expression of IFNβ, and the resulting program of ISGs induced by IFNα/β signaling, is central to the inflammatory response and anti-viral immunity. Too much interferon can result in aberrant inflammation and autoimmunity, while too little interferon signaling increases susceptibility to infection. Therefore, the expression of type I IFNs must be tuned appropriately to maintain health. Given the unique ability of exIRF7 to drive strong IFNβ expression we hypothesized that human cells may use alternative splicing of IRF7 intron 1, in addition to transcriptional activation, to achieve appropriate IFNβ expression.

Of note, we do not detect any difference in general cell viability or growth rates for cells expressing the cIRF7 versus exIRF7 isoform in either the Jurkat or A549 background (Fig S6A). We do observe dramatic variation in IRF7 splicing across many human cancer samples (Fig S6B); although we cannot differentiate whether these variations in IRF7 isoform expression impact tumor or disease progression, or simply reflect IRF7 splicing in the tissue of origin. However, the functional relevance of IRF7 splicing is demonstrated by a naturally occurring human polymorphism in IRF7 (rs12290989 G/T), which shows strong correlation with the splicing of IRF7 intron 1 by sQTL analysis (Fig 6A). Specifically, the major allele (GG) correlates with increased intron retention and expression of exIRF7, while the minor allele (TT) correlates with splicing of intron 1 and expression of cIRF7. Interestingly, the presence of the minor allele is associated with reduced expression of type I IFNs by plasmacytoid dendritic cells (pDC) in response to viral infection⁴³, consistent with preferential expression of cIRF7 over exIRF7. Moreover, this same minor allele has been shown to be protective against the autoimmune disorder Lupus (SLE)⁴⁴, as would be anticipated from a tendency to reduced IFNs and inflammation.

Figure 6: exIRF7 promotes interferon production and anti-viral immunity.

(A) Quantification of IRF7 intron 1 retention in different human genotypes of SNP rs12290989 from the CancerSplicingQTL database. Each dot represents a single individual with the given genotype. Box plot represents average and distribution of intron retention. (B) (top) Representative images from A549 cells with the indicated genotype infected for 24 hours with GFP-labeled PIV3 at MOI of 0.25. Bar indicates 100 microns. (middle) quantification of infected (GFP+) cells over time post-initial addition of PIV3 (MOI of 0.25) for wildtype (WT, black), IRF7 KO (gray), cIRF7 expressing (red) or exIRF7 expressing (blue) A549 cells. (bottom) quantification of automated microscopy images across multiple MOI of PIV3 infection. (C) Quantification of VSV infected (GFP+) cells over time post-addition of virus (MOI of 0.01) for WT (black), IRF7 KO (gray), cIRF7 expressing (red) or exIRF7 expressing (blue) A549 cells. In B-C, data are mean ± SD.

To more directly test the relevance of IRF7 isoform expression in the control of viral infections we used the A549 cells expressing single isoforms of IRF7 (Fig S4). Parainfluenza virus (PIV3) infects respiratory epithelial cells and has been shown to be highly sensitive to IFN-I signaling^45,46. Strikingly, we observe that cells expressing exIRF7 are markedly more resistant to PIV3 infection than cells expressing no IRF7 or the cIRF7 isoform (Fig 6B, top), as would be expected from the increased expression of type I IFNs in pIC-treated exIRF7-expressing A549 cells, as shown in Fig S4D. This result is also consistent with previous reports that IRF7 promotes IFN-I expression in viral-infected A549 cells⁴². Moreover, by quantifying PIV3 infection over time in living cells (Fig 6B, middle), we observe that the primary difference in viral susceptibility between exIRF7 and cIRF7-expressing cells is reduced spread of virus from cell-to-cell, consistent with the role of type I IFNs as paracrine regulators of innate immunity. We further substantiated this difference in PIV3 viral infection in exIRF7 cells relative to cIRF7 across multiple concentrations of virus by quantitative microscopy (Fig 6B, bottom), and confirmed reduced viral titer in cells expressing exIRF7 (Fig S6C). We also show similar preferential resistance of exIRF7-expressing cells to infection and spread of another interferon-sensitive virus, vesicular stomatitis virus (VSV, Fig 6C; S6D,E), confirming that this physiologic difference is not confined to a single virus. Take together, these data confirm the physiologic relevance of the differential activity of the IRF7 isoforms in mediating an anti-viral immune response.

Discussion

The role of the transcription factors IRF7 and IRF3 in activating the expression of interferons following innate immune triggers has been studied for decades, yet the intricacies regarding how this process is regulated remain to be fully uncovered. Here we demonstrate a previously unrecognized mechanism by which alternative splicing of IRF7 tunes the transcriptional activation of type I IFNs and subsequent downstream ISGs. Specifically, we show that the N-terminally extended form of IRF7 (exIRF7), encoded by retention of the first intron within the coding region of the human IRF7 gene, has greater activity on the IFNβ promoter than cIRF7 due to its increased propensity to dimerize with itself and with IRF3. This increased dimerization allows exIRF7 to preferentially form the higher order enhanceosome complexes on the IFNβ promoter which are critical to drive transcription⁴⁰. The increased activity of exIRF7 on the IFNβ promoter, in turn, induces higher levels of IFN and ISGs to repress viral infection compared to the canonical (cIRF7) isoform. While the existence of the exIRF7 isoform, and its ability to support transcription of IFNα³³ was previously established, comparisons between the activity of exIRF7 and cIRF7 have not been explored. The discovery of enhanced activity of exIRF7 provides critical insight into the tuning of the interferon response. Nevertheless, our observations are consistent with previous work in that upon ectopic expression (Fig 4A), or after IFN stimulation (Fig 3E), when IRF7 expression is high, cIRF7 also activates the interferon pathway^12,15,22,42.

Of note, the potential to encode exIRF7 appears to be restricted to old world monkeys and Apes (which includes humans), suggesting evolutionary divergence ~21.9 million years ago. While mice do not encode the exIRF7 isoform, alternative splicing does regulate murine IRF7 expression through intron 4 retention to attenuate translation³¹; a regulatory program we do not observe in humans, at least in the conditions and cells surveyed in this study. Furthermore, the splicing factor SRSF7 enhances murine IRF7 transcription through recruiting chromatin remodeling enzymes⁴⁷. Therefore, tuning of IRF7 activity and the interferon response through the activity of the splicing machinery appears to be conserved throughout evolution, with different species adopting distinct mechanisms to establish this connection.

Previous studies have demonstrated that dimerization of cIRF7 with IRF3 is mediated by their C-termini, which is identical in both IRF7 isoforms. Alphafold does not predict any stable structure of the N-terminal peptides in either cIRF7 or exIRF7, and the N-terminus of cIRF7 was truncated in the published structure studies of IRF7 bound to DNA⁴⁰. However, the extended N-terminus is likely sufficiently flexible that it could make additional contacts with the dimeric partner. Moreover, since the N-terminus is adjacent to the DNA binding domain, we consider it likely that the extended N-terminus of exIRF7 contacts DNA, further increasing the formation of DNA-bound complexes. Future structural studies will be required to provide conclusive insight into how the N-terminal extension of IRF7 enhances dimerization and function.

Finally, we demonstrate that exIRF7 not only increases transcriptional activation of the interferon response, but also enhances antiviral defense as demonstrated by reduced susceptibility to viral infection. Many human immune cell types that control viral infections including monocytes, memory B and T cells, and Tregs preferentially express exIRF7 (Fig 1B), suggesting that these cells use exIRF7 to promote initial responsiveness of the immune system. In addition, NFKBIA, which encodes the NFkB inhibitor IkB-alpha, is one of the genes preferentially activated by exIRF7 (Fig 3, S3A). It would thus be interesting to determine if exIRF7 impacts NFkB-mediated gene expression during adaptive and innate immune responses.

In contrast to exIRF7, preferential expression of cIRF7 following immune activation of monocytes may serve to dampen the interferon response to prevent hyperinflammation. We show some viruses can reduce relative exIRF7 expression, which may be one of the many mechanisms viruses use to counter the interferon response and enhance their replication. Consistently, previous studies have suggested IRF7 primary functions as part of the amplification of the interferon response late in infection, rather than in the immediate early response^12,42. Altogether, we propose that alternative splicing of human IRF7 serves as a regulatory mechanism, separate from and complementary to transcriptional activation, that tunes the transcriptional activity of the IRF7 protein, the expression of IFN-I genes, and subsequent activation and homeostasis of the innate immune system.

Limitations of the Study

While we have confirmed that the distinct dimerization affinity of exIRF7 and cIRF7 correlate with different thresholds of transcriptional activation of type I IFNs in at least three different cell types, we have not directly tested exIRF7 and cIRF7 activity in vivo. In particular, it would be of interest to assess the contribution of exIRF7 and cIRF7 to the function of macrophages and pDCs, and to determine if the high expression of exIRF7 in Tregs impacts adaptive immunity in any way. As mice do not express exIRF7, the murine immune response has evolved separate modes of regulation. Thus, true proof of the impact of IRF7 intron 1 regulation in immunity likely will require human organoid systems or other complex models of inflammation beyond the scope of this current study.

Resource Availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Kristen W. Lynch (klync@pennmedicine.upenn.edu)

Materials availability

All unique and stable reagents generated in this study are available from the lead contact without restriction

Data and code availability

• RNA-Seq data are publicly available at GEO (https://www.ncbi.nlm.nih.gov/geo/).

• No new code is reported in this study.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR METHODS

Experimental Model and Subject Details

Cell lines and cell culture

Wild-type, CRISPR-edited, and cDNA-overexpressing stable Jurkat cells (male) were cultured in RPMI medium (Corning, 10-040-CV) supplemented with 5% heat-inactivated fetal bovine serum (FBS) (Gibco, A56697-01), and 2 mM penicillin-streptomycin (Mediatech, MT-30-002-Cl). We use a clonal population (JSL1) of Jurkat cells we derived previously⁵¹ that is authenticated by returning to source bullets every 2-3 months. The cells were maintained at 37°C in a humidified incubator with 5% CO2. A549 (male) and HEK293T (female) adherent cell lines were obtained from and authenticated by the ATCC, and grown in DMEM (Corning, 10-013-CV) supplemented with 10% FBS, under the same conditions as those described for JSL1 cells. All cells are routinely tested for mycoplasma contamination by the Penn Cell Center. Human primary monocytes and T regulatory cells (Tregs) were obtained from the Human Immunology Core at the University of Pennsylvania (donors anonymized, gender/age and other information unknown; collected under IRB 811028) and were maintained in RPMI with 10% FBS. To stimulate innate immune responses, the cells were adjusted to a density of 1 × 10⁶ cells/ml and incubated either in medium alone or with the designated treatments for the required duration. Stimulation of all cells with phorbol 12-myristate 13-acetate (PMA) and IFNα was done by adding these compounds directly to the media at the concentrations and times listed below. Stimulation of Jurkat cells and primary monocytes with polyIC was done by adding this compound to the media, thus targeting TLR3. As A549 cells lack TLR3 expression, polyIC treatment of these cells was done by transfection to activate intracellular RLRs. We detect no influence of sex or gender on our results; however, the sample size and anonymity of our cell models prevent us from making any conclusive statements.

Method Details

RT-PCR

RNA was extracted from Jurkat cells, primary monocytes, or A549 cells using TRIzol (Invitrogen, 15596018) in accordance with the manufacturer’s instructions. Low-cycle radioactive RT-PCR was conducted and analyzed as previously detailed^52–54. RNA was reverse transcribed using MMLV reverse transcriptase (Thermo, 28025013). The primer sequences for IRF7 splicing analysis are listed in the Key Resource Table/Supplementary Table 1. Three independent RT-PCR reactions were performed with 500 ng of RNA per reaction, and samples were resolved on 5% denaturing polyacrylamide gels (PAGE) using formamide buffer. Radiolabeled 32P cDNA products were detected via densitometry with an Amersham Typhoon PhosphorImager and quantified using ImageQuant software.

Key Resource Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit monoclonal IRF7	abcam	Cat# ab238137
Rabbit polyclonal phosphoIRF7	Cell signaling	Cat# 5184S
Rabbit monoclonal FLAG	Cell signaling	Cat# 14793S; RRID:AB_2572291
Rabbit monoclonal IRF3	Cell signaling	Cat# 11904S; RRID:AB_2722521
Rabbit monoclonal NFkB-p65	Cell signaling	Cat# 8242T
Rabbit monoclonal NFkB-p50	Cell signaling	Cat# 12540S
Rabbit monoclonal ATF2	Cell signaling	Cat# 35031S
Rabbit monoclonal c-JUN	Cell signaling	Cat# 9165; RRID:AB_2130165
Rabbit monoclonal HMGA1	Cell signaling	Cat# 12094S; RRID:AB_2797818
Rabbit monoclonal GAPDH	Cell signaling	Cat# 2118S; RRID:AB_561053
Anti-rabbit secondary antibody (HRP)	Cell signaling	Cat# 7074S; RRID:AB_2099233
Goat anti-rabbit Alexa488-plus	Thermo Fisher	Cat# A32731; RRID:AB_2633280
Goat anti-mouse Alexa568	Thermo Fisher	Cat# A11031; RRID:AB_144696
Bacterial and viral strains
Human Para-Influenza Virus GFP-tagged (PIV3 GFP based on the JS strain)	Lab of Sara Cherry	N/A
Sendai Virus (SeV)	Charles River Labs	Cantell Strain #10100774
Biological samples
Healthy adult Human Monocytes and T-regulatory cells	Penn Human Immunology Core	RRID:SCR_022380
HEK293T	ATCC	CRL-3216
Chemicals, peptides, and recombinant proteins
Recombinant IRF3 protein	OriGene technologies	Cat# TP309951
3XFLAG peptide	Sigma	Cat# F4799
Anti-FLAG M2 agarose affinity gel	Sigma	Cat# A2220
Poly(I:C) (HMW)	Invivogen	Cat. Code: tlrl-pic
Human IFN Alpha (Alpha 2a)	PBL assay science	Cat#111001-1
Ruxolitinib	Invivogen	Cat. Code:TLRL-RUX
Ambion TRIzol reagent	Invitrogen	Cat# 15596018
Poly(dI:dC)	Thermo Scientific	Cat# 20148E
Mass Photometry consumables	Refeyn	Product code: MP-CON-21014
30% Acrylamide/Bis 29:1	Bio-Rad	Cat# 1610156
Lipofectamine 3000	Thermo Scienific	Cat# L3000015
Lipofectamine 2000	Thermo Scientific	Cat# 11668027
cOmplete protease inhibitor	Roche	Cat# 11697498001
ECL reagent	Cytiva	Cat# RPN2106
Amersham Hyperfilm ECL	Cytiva	Cat# 95017-653
DAPI	Thermo scientific	Cat# D1306
Fetal Bovine Serum	GIBCO	Ref# A56697-01
Oligo(dT) 12-18 primer	Invitrogen	Cat# 18418012
MMLV	Thermo scientific	Cat# 28025013
Opti-MEM	Gibco	Ref# 31985
RPMI	Corning	Ref# 10-040-CV
DMEM	Corning	Ref# 10-013-CV
Penicillin-Streptomycin	Mediatech	Cat# MT-30-002-Cl
Nitrocellulose membrane	Thermo scientific	Cat# 88518
Laemmli Sample Buffer	Biorad	Cat# 1610747
NuPAGE Sample Reducing Agent	Thermo Scientific	Cat# NP0009
40% Acrylamide/Bis solution, 37.5:1	Biorad	Cat# 1610149
30% Acrylamide/Bis solution, 29:1	Biorad	Cat# 1610156
Critical commercial assays
NEBuilder HiFi DNA Assembly Mastermix	New England Biolabs (NEB)	Cat# E2621S
Slide-A-Lyzer Dialysis cassettes 10k MWCO	Thermo scientific	Cat# 66383
Amersham microspin G-25 columns	Cytiva	Cat# 27532501
Bradford reagent	Bio-rad	Cat# 500-0006
PowerUp SYBR green mastermix	Applied Biosystems Inc	Cat# A25742
KAPA HiFi HotStart ReadyMix PCR Kit	Roche	Cat# KK2602
Nano-Glo luciferase assay system	Promega	Cat# N1110
Deposited data
RNA sequencing data are available under GEO number GSE285352	This paper	GEO: GSE285352
Experimental models: Cell lines
JSL1 (Jurkat cells)	Lynch et al.51	N/A
A549	Lab of Sara Cherry	N/A
IRF7 CRISPR knock-out (in JSL1 and A549 cells)	This paper	N/A
cIRF7-FLAG overexpression clone (in JSL1 and A549 cells)	This paper	N/A
exIRF7-FLAG overexpression (in JSL1 and A549 cells)	This paper	N/A
cIRF7 CRISPR-edited (in JSL1 cells)	This paper	N/A
exIRF7 CRISPR-edited (in JSL1 cells)	This paper	N/A
Oligonucleotides
CRISPR Homologous arm up- amplification primer: Forward: TGAATTCGAGCTCGGTACCCGCTCTGTCCCCTGGGCTG	This paper	N/A
CRISPR Homologous arm up- amplification primer Reverse: TGCGGAGACGGGAAAGGCGACGTCAGGGGCGGGTCAGG	This paper	N/A
CRISPR Homologous arm down- amplification primer Forward: GCACACGTGTTCTGTCCCCCTCTTCAGGC	This paper	N/A
CRISPR Homologous arm down- amplification primer Reverse: GTCGACTCTAGAGGATCCCCCCTGGCAGGAGGAGAGGC	This paper	N/A
CRISPR Homology Directed Repair template: cIRF7 CRISPR repair template as a gBlock from IDT TCGCCTTTCCCGTCTCCGCACTTTCCGGCCTGCGCCTTCCCGCCAGGCCTGGACACTGGTTCAACACCTGTGACTTCATGTGTGCGCGCCGGCCACACCTGCAGTCACACCTGTAGCCCCCTCTGCCAAGAGATCCATACCGAGGCAGCGTCGGTGGCTACAAGCCCTCAGTCCACACCTGTGGACACCTGTGACACCTGGCCACACGACCTGTGGCCGCGGCCTGGCGTCTGCTGCGACAGGAGCCCTTACCTCCCCTGTTATAACACCTGACCGCCACCTAACTGCCCCTGCAGAAGGAGCAATGGCCTTGGCTCCTGAGAGGGCAGCCCCACGCGTGCTGTTCGGAGAGTGGCTCCTTGGAGAGATCAGCAGCGGCTGCTATGAGGGGCTGCAGTGGCTGGACGAGGCCCGCACCTGTTTCCGCGTGCCCTGGAAGCACTTCGCGCGCAAGGACCTGAGCGAGGCCGACGCGCGCATCTTCAAGGTGGGGACCCCAGCCCCGCCCCGCCCCGGGTGGGAGAGCCCGCCTAGGATTCCGCGGGCCCGGCACACGTGT	IDT/This paper	N/A
CRISPR Homology Directed Repair template: exIRF7 CRISPR repair template as a gBlock from IDT TCGCCTTTCCCGTCTCCGCAGTTTCCGGCCTGCGCCTTCCCGCCAGGCCTGGACACTGGTTCAACACCTGTGACTTCATGTGTGCGCGCCGGCCACACCTGCAGTCACACCTGTAGCCCCCTCTGCCAAGAGATCCATACCGAGGCAGCGTCGGTGGCTACAAGCCCTCAGTCCACACCTGTGGACACCTGTGACACCTGGCCACACGACCTGTGGCCGCGGCCTGGCGTCTGCTGCGACAGGAGCCCTTACCTCCCCTGTTATAACACCTGACCGCCACCTAACTGCCCCTGCAGAAGGAGCAATGGCCTTGGCTCCTGAGCGCTAAGAGCCCGGCCCACCCTCTCCAGATGCCAGTCCCCGAGCGCCCTGCAGCCGGCCCTGACTCTCCGCGGCCGGGCACCCGCAGGGCAGCCCCACGCGTGCTGTTCGGAGAGTGGCTCCTTGGAGAGATCAGCAGCGGCTGCTATGAGGGGCTGCAGTGGCTGGACGAGGCCCGCACCTGTTTCCGCGTGCCCTGGAAGCACTTCGCGCGCAAGGACCTGAGCGAGGCCGACGCGCGCATCTTCAAGGTGGGGACCCCAGCCCCGCCCCGCCCCGGGTGGGAGAGCCCGCCTAGGATTCCGCGGGCCCGGCACACGTGT	IDT/This paper	N/A
g-block- Minigene#1: GAGCTTGGTACCTGGCGTCTGCTGCGACAGGAGCCCTTACCTCCCCTGTTATAACACCTGACCGCCACCTAACTGCCCCTGCAGAAGGAGCAATGGCCTTGGCTCCTGAGAGGTAAGAGCCCGGCCCACCCTCTCCAGATGCCAGTCCCCGAGCGCCCTGCAGCCGGCCCTGACTCTCCGCGGCCGGGCACCCGCAGGGCAGCCCCACGCGTGCTGTTCGGAGAGTGGCTCCTTGGAGAGATCAGCAGCGGCTGCTATGAGGGGCTGCAGTGGCTGGACGAGGCCCGCACCTGTTTCCGCGTGCCCTGGAAGCACTTCGCGCGCAAGGACCTGAGCGAGGCCGACGCGCGCATCTTCGGATCCACTAGT	IDT/This paper	N/A
g-block- Minigene#2 (ΔExon2): AGCTTGGGTACCTGGCGTCTGCTGCGACAGGAGCCCTTACCTCCCCTGTTATAACACCTGACCGCCACCTAACTGCCCCTGCAGAAGGAGCAATGGCCTTGGCTCCTGAGAGGTAAGAGCCCGGCCCACCCTCTCCAGATGCCAGTCCCCGAGCGCCCTGCAGCCGGCCCTGACTCTCCGCGGCCGGGCACCCGCAGGCTGCTGGTGGTCTACCCTTGGACCCAGAGGTTCTTTGAGTCCTTTGGGGATCTGTCCACTCCTGATGCTGTTATGGGCAACCCTAAGGTGAAGGCTCATGGCAAGAAAGTGCTCGGTGCCTTTAGTGATGGCCTGGCTCACCTGGACAACCTCAAGGGCACCTTTGCCACACTGAGTGAGCTGCACTGTGACAAGCTGCACGTGGATCCGATCCA	IDT/This paper	N/A
g-block- Minigene#3 (ΔExon1): AGCTTGGGTACCACATTTGCTTCTGACACAACTGTGTTCACTAGCAACCTCAAACAGACACCATGGTGCATCTGACTCCTGAGGAGAAGTCTGCCGTTACTGCCCTGTGGGGCAAGGTGAACGTGGATGAAGTTGGTGGTGAGGCCCTGGGCAGGTAAGAGCCCGGCCCACCCTCTCCAGATGCCAGTCCCCGAGCGCCCTGCAGCCGGCCCTGACTCTCCGCGGCCGGGCACCCGCAGGGCAGCCCCACGCGTGCTGTTCGGAGAGTGGCTCCTTGGAGAGATCAGCAGCGGCTGCTATGAGGGGCTGCAGTGGCTGGACGAGGCCCGCACCTGTTTCCGCGTGCCCTGGAAGCACTTCGCGCGCAAGGACCTGAGCGAGGCCGACGCGCGCATCTTCGGATCCGATCCA	IDT/This paper	N/A
g-block- Minigene#4 (ΔIntron1): AGCTTGGGTACCTGGCGTCTGCTGCGACAGGAGCCCTTACCTCCCCTGTTATAACACCTGACCGCCACCTAACTGCCCCTGCAGAAGGAGCAATGGCCTTGGCTCCTGAGAGGTTGGTATCAAGGTTACAAGACAGGTTTAAGGAGACCAATAGAAACTGGGCATGTGGAGACAGAGAAGACTCTTGGGTTTCTGATAGGCACTGACTCTCTCTGCCTATTGGTCTATTTTCCCACCCTTAGGGCAGCCCCACGCGTGCTGTTCGGAGAGTGGCTCCTTGGAGAGATCAGCAGCGGCTGCTATGAGGGGCTGCAGTGGCTGGACGAGGCCCGCACCTGTTTCCGCGTGCCCTGGAAGCACTTCGCGCGCAAGGACCTGAGCGAGGCCGACGCGCGCATCTTCGGATCCGATCCA	IDT/This paper	N/A
Additional primers used for CRISPR, RT-PCR, and qPCR	This paper	See Table S1
Recombinant DNA
Plasmid: pUC19	Addgene	RRID:Addgene_50005
Plasmid: pspCas9n-BB-2A-GFP (PX458)	Addgene	RRID:Addgene_48138
Plasmid: pEFneo	Lab of Kristen Lynch	N/A
Plasmid: pEFneocIRF73XFLAG	This paper	N/A
Plasmid: pEFneoexIRF73XFLAG	This paper	N/A
Plasmid: pspCas9n-BB-2A-GFP-gRNAs	This paper	N/A
Plasmid: pUC19-cIRF7edits	This paper	N/A
Plasmid: pUC19-exIRF7edits	This paper	N/A
Plasmid: pcDNA3.1-C-(K)DYK human IRF7 transcript variant A	GenScript	Clone ID: OHu24955
Plasmid: pcDNA3.1-C-(K)DYK human IRF7 transcript variant D	GenScript	Clone ID: OHu25038
Software and algorithms
GraphPad Prism 10.0	GraphPad Software	https://www.graphpad.com/
ImageJ 1.53a	ImageJ Software	imagej.net
Amersham Typhoon Software v. 2.0.0.6	Amersham	www.cytivalifesciences.com; Serial no. 97130681
ImageQuant v. 8.1	Amersham	www.cytivalifesciences.com
AquireMP Two MP R2	Refeyn Software	https://refeyn.com
DiscoverMP Version R2	Refeyn Software	https://refeyn.com
QuantStudio 12K Flex software version 1.3	Applied Biosystems	https://www.thermofisher.com/us/en/home/brands/applied-biosystems
Molecular Devices MetaXpress version 6	Molecular Devices	https://support.moleculardevices.com
bbduk 38.79	Osti.gov	www.osti.gov/biblio/1241166
Salmon 1 version 10.0	Patro et al.55	https://github.com/COMBINE-lab/salmon
Tximport	Soneson et al.56	https://www.bioconductor.org/packages/release/bioc/html/tximport.html
DESeq2	Love et al.57	https://bioconductor.org/packages/release/bioc/html/DESeq2.html

Western Blots

Whole cell lysate was collected with RIPA lysis buffer (150 mM NaCl, 0.5% Sodium Deoxycholate, 1% NP-40, 0.1%SDS, 25mM Tris (pH 7.4)) and proteins were quantified by Bradford assay. 20 μg of total protein lysates were loaded into 10% 37.5:1 bis-acrylamide SDS-PAGE gels. SDS-PAGE gels were transferred onto nitrocellulose membranes (Thermo scientific, 88518) and targeted proteins were visualized with a chemiluminescence system (Cytiva, RPN2106) and subsequent imaging with an x-ray film (Cytiva, 95017-653) and developer. Antibodies used to detect protein expression levels are as follows: IRF7 (Abcam, ab238137), pIRF7 (Cell Signaling, 5184S), FLAG (Cell Signaling, 14793S), IRF3 (Cell Signaling, 11904S), NFkB-p65 (Cell Signaling, 8242T), NFkB-p50 (Cell Signaling, 12540S), ATF2 (Cell Signaling, 35031S), c-Jun (Cell Signaling, 9165T), HMGA1 (Cell Signaling, 12094S), and GAPDH (Cell Signaling, 2118S), anti-rabbit HRP-linked secondary antibody (Cell signaling, 7074S). All the primary antibodies were used at a 1:1000 dilution and the secondary antibody was used at a 1:3000 dilution.

Co-Immunoprecipitation

Jurkat cells stably expressing individual FLAG-tagged IRF7 isoforms were stimulated with 1 μg/mL polyIC (InvivoGen, tlrl-pic), added directly to the media, for 6 hours prior to harvesting. Whole cell lysate (WCL) was extracted with the lysis buffer: 10 mmTris-HCl-pH 8.0, 150 mm NaCl, 1 mm EDTA, 1 mm dithiothreitol, 0.5% Nonidet P-40, add 1X Protease inhibitor cocktail(Roche, 11697498001) just before use. Anti-FLAG M2 affinity agarose gel beads (Sigma, A2220) were washed 3X (1000g, 1min @ 4°C) with the lysis buffer followed by the addition of WCL and incubation with rotation at 4°C for 1 hour. The FLAG-beads were spun down (1000g, 1min @ 4°C), supernatant removed, and the beads were washed 5X (1000g, 1min @ 4°C). Finally, co-Immunoprecipitated proteins were eluted from the beads adding sample buffer (Biorad, 1610747) with reducing agent (Thermo Scientific NP0009), and boiling for 5 min followed by a quick spin and supernatant was saved. Samples were loaded into 10% 37.5:1 bis-acrylamide (Biorad, 1610149) SDS-PAGE gels. SDS-PAGE gels were transferred onto nitrocellulose membranes and targeted proteins were visualized with a chemiluminescence system and subsequent imaging with an X-ray developer. Antibodies used to detect protein expression levels are as follows: IRF7 (Abcam, ab238137), FLAG (Cell Signaling, 14793S), IRF3 (Cell Signaling, 11904S), pIRF3 (Cell Signaling, 4947S).

CRISPR designs and cloning

To generate individual cell clones that express either isoform of IRF7, we used CRISPR/Cas9-based genome editing in Jurkat cells to alter the genomic elements that encode for the regulated alternative splicing events (Supplemental Data 1). Two custom sgRNAs, sgRNA1 and sgRNA2 (Supplementary Table S1), were created to target intronic regions or the 5’UTR that surround genomic elements that encode for the alternative splicing event of interest. To knockout IRF7, two sgRNAs, sgRNA2 and sgRNA3 (Table S1), were designed targeting the introns adjacent to the constitutive exon of interest. The sgRNA sequences are listed in Table S1. The sgRNAs were designed using the Broad Institute CRISPick tool (https://portals.broadinstitute.org/gppx/crispick/public) and cloned into the pSpCas9(BB)-2A-GFP (PX458) plasmid (Addgene, 48138) containing the S. pyogenes Cas9 (SpCas9), GFP and U6 promoter sequences. For cIRF7, a repair template was constructed to fuse exons 1 and 2, mimicking the splicing outcome when Intron 1 is skipped. To generate exIRF7, the repair template included a mutation in the 5’ splice site (5’SS), changing the splice donor “G” of the 5’SS to “C” while maintaining the encoding of arginine (AGG to CGC, R-R), thereby blocking splicing and resulting in intron 1 retention (Supplemental Data 1. A). Both repair templates were synthesized as gBlocks by IDT. Upstream and downstream homology arms (~600 bp each) were PCR-amplified (with KAPA PCR kit, Roche, KK2602) and assembled into the pUC19 vector backbone using the NEBuilder HiFi DNA Assembly Kit (NEB, E2621L). The sequences for the repair templates, primers for homology arm amplification, and gBlocks are listed in the Key Resource Table. Sequences of final gene edited regions are in Supplemental Data 1. A.

Construction of plasmids

For exogenous IRF7 expression, the cIRF7 (NM_001572; 504 amino acids; GenScript Clone ID: OHu24955) and exIRF7 (NM_004031; 517 amino acids; GenScript Clone ID: OHu25038) human ORF cDNAs were cloned into the pEFneo vector. A 3xFLAG tag (DYKDHDGDYKDHDIDYKDDDDK) was added to the C-terminus of each construct for downstream applications. The final engineered plasmids contained the 3×FLAG tag fused to the C-terminal ends of the IRF7 isoforms. To construct minigenes with IRF7 and β-globin sequences, g-blocks (ordered from IDT) with the corresponding sequences with desired restriction overhangs were cloned into the parental vector pActT7 vector⁵⁴. G-block sequences are listed in the key resources table in this paper.

RNA-Seq

CRISPR-edited Jurkat suspension clones were treated with high-molecular-weight polyinosinic–polycytidylic acid (polyIC; InvivoGen, tlrl-pic) added to the media at 1 μg/mL or recombinant interferon-alpha 2a (IFN-α2a; PBL Assay Science, 11100-1) at 500 U/mL for 8 h. Both polyIC and IFNα were added directly to the suspension culture medium at the indicated final concentrations. Following an 8-hour treatment with either polyIC or IFNα, cells were collected, and RNA was isolated using TRIzol reagent (Invitrogen, 15596018) according to the manufacturer’s instructions. RNA integrity number (RIN) was measured with the Agilent bioanalyzer, and all samples had a RIN >8.0. RNA-sequencing libraries were generated by and sequenced by GeneWiz (Azenta life sciences) at a depth of 20 million reads per sample. The libraries were poly(A) selected (nonstranded) and paired-end sequenced at a 150 bp read length. Raw fastq files were trimmed with bbduk 38.79 (www.osti.gov/biblio/1241166) and then counted with salmon 1.10.0⁵⁵ in quant mode with the settings “--validateMappings --rangeFactorizationBins 4 –seqBias --gcBias --recoverOrphans”. Reads were aligned to Ensembl Homo sapiens hg38 primary cDNA transcriptome (annotation version 110). Transcript counts from salmon were collapsed to the gene level using tximport⁵⁶ and differential abundance was determined using DESeq2⁵⁷. Differential expression analysis was quantified by DESeq2. In Figure 3, differential expression of a gene is considered significant with a pValue < 0.05, log2 fold-change > 1 (stimulated/unstimulated) and base mean reads > 50. The RNA-sequencing data generated for this study is available in GEO (GSE285352). See Supplemental Data 1. B and C for complete set of differentially expressed genes and RNA-Seq analysis.

ISRE analysis

Transcriptional start sites (TSS) were extracted from Ensembl using the biomaRt package⁵⁸. Next, the regions around the TSS were cross referenced against the JASPAR database⁵⁹ to note the location of all ISRE (i.e., predicted STAT1:STAT2) binding sites. To generate the logos, the ISRE sites from each gene group were extracted and visualized using ggseqlogo⁶⁰.

RT-qPCR

Cells were cultured as described above and stimulated with polyIC or IFNα for the indicated durations. For experiments involving Ruxolitinib (1 μg/mL final concentration used; InvivoGen, TLRL-RUX), the inhibitor was added directly to the media 1 hour prior to stimulation with polyIC or IFNα. Total RNA was extracted from cells using TRIzol, following the manufacturer’s protocol (Invitrogen, 15596018). cDNA was synthesized from 250 ng of RNA using oligo(dT) primers (Invitrogen, 184118012) and MMLV reverse transcriptase (Thermo, 28025013). Quantitative PCR was performed using SYBR qPCR Master Mix (Applied Biosystems, A25742) on the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems). Data acquisition was managed with QuantStudio 12K Flex software version 1.3 (Applied Biosystems) and analyzed using the ΔCT method. Actb, encoding β-actin, was used as the reference gene for normalization. The sequences of the gene-specific primers utilized for PCR are listed in the Table S1.

Cell transfection

All transfections in JSL1 cells were done by electroporating 10–20 million cells. To generate cIRF7 and exIRF7 CRISPR-engineered cells, Jurkat cells were co-transfected with PX458-sgRNA plasmids (RRID: Addgene_48138) targeting intronic regions flanking the splicing event and pUC19-repair template plasmids (RRID: Addgene_50005) containing homologous arms and repair templates (5 μg each). To generate IRF7 knockout (KO) cells, PX458-sgRNA plasmids targeting the introns flanking the constitutive exon of interest were co-transfected into cells (5 μg each). After electroporation, edited Jurkat cells were rested for 48hrs and sorted into 96-well plates (FACS Jazz) by gating on doublet exclusion and GFP+ expression. One to three cells were deposited per well to allow the growth of single-cell colonies. Individual clones were screened by PCR amplification of the targeted genomic regions, analyzed for band size on a 1.5% EtBr agarose gel, confirmed through genomic DNA sequencing, and validated by Western blot analysis. All Jurkat cells were maintained in RPMI medium supplemented with 5% heat inactivated FBS. Stable expression of cIRF7 and exIRF7 with C-terminal 3xFLAG tags was achieved by integrating pEFneo cDNA expression constructs (refer to the above subheading “construction of plasmids” for details) into JSL1 cells. Linearized plasmids (10μg) were electroporated into 10–20 million cells, and neomycin was used as a selective marker for stable integration.

To generate IRF7 KO clones in A549 cells, the same PX458-sgRNA plasmids used for JSL1 cells were employed, and cells were sorted and selected using the same method. Unlike JSL1 cells, A549 cells were transfected with Lipofectamine 3000 (Thermo fisher, L3000015) instead of electroporation, using the same plasmid amounts. In this IRF7 KO background, stable expression of cIRF7 and exIRF7 with C-terminal 3xFLAG was achieved by transfecting linearized pEFneo cDNA constructs (10 μg) with Lipofectamine 3000, followed by neomycin selection for stable integration.

FLAG-affinity purification of IRF7 isoforms

Jurkat suspension cells stably expressing individual C-terminal 3×FLAG-tagged IRF7 isoforms were cultured in RPMI (Corning, 10-040-CV) supplemented with 5% heat-inactivated fetal bovine serum (FBS; Gibco; A56697-01) and penicillin–streptomycin (Mediatech, MT-30-002-Cl; 2 mM as supplied) in roller bottles under gentle rotation (Corning, 06-415-6) at 37 °C in 5% CO₂. Each isoform-expressing clone was expanded to ~5 L total culture volume, adjusted to 1 × 10⁶ cells/mL, and stimulated with polyIC (InvivoGen, tlrl-pic) at 1 μg/mL for 6 h before harvest. Cells were collected by centrifugation (3,000 rpm, 15 min, 4 °C), washed once with 200 mL ice-cold PBS, and pelleted again (3,000 rpm, 15 min, 4 °C). Pellets were resuspended in Buffer A (10 mM Tris-HCl pH 7.5, 1.5 mM MgCl₂, 10 mM KCl), incubated on ice for 10 min, and centrifuged (4,000 rpm, 15 min, 4 °C). Pellets were then resuspended in 2× packed-cell-volume (PCV) Buffer A, Dounce-homogenized (10 strokes), and centrifuged (3,000 rpm, 10 min, 4 °C) followed by 16,500 rpm for 20 min at 4 °C. The resulting pellet was extracted in 0.7× PCV Buffer C (20 mM Tris-HCl pH 7.5, 1.5 mM MgCl₂, 420 mM NaCl, 25% glycerol, 0.2 mM EDTA), Dounce-homogenized (10 strokes), rocked 30 min at 4 °C, and centrifuged (16,500 rpm, 30 min, 4 °C). The supernatant (nuclear extract) was mixed with 2× volumes Buffer F (60 mM Tris-HCl pH 7.5, 30 mM NaCl, 1 mM EDTA, 1.5% Triton X-100) and applied by gravity flow to anti-FLAG M2 affinity agarose beads (Sigma, A2220). Columns were washed with TBS (100 mM Tris-HCl, 150 mM NaCl) and IRF7-FLAG proteins were eluted in five 1.5-mL fractions of 3×FLAG peptide (500 μg/mL; Sigma, F4799) in TBS. Eluates were dialyzed in Slide-A-Lyzer cassettes 10 kDa MWCO (Thermo Scientific, 66383), against ice-cold BC-100 buffer (100 mM KCl, 20 mM Tris-HCl pH 7.5, 0.2 mM EDTA, 20% glycerol) at 4 °C with gentle stirring (~60 rpm). Purified, dialyzed proteins were snap-frozen on dry ice and stored at −80 °C.

Electrophoretic Mobility Shift Assay

The FLAG-purified recombinant IRF7 proteins were used for DNA binding via gel-shift assays (EMSA) using a 32P-labeled double-stranded oligonucleotide corresponding to the PRD3,1 region of the IFNβ promoter (5’-GAAAACTGAAAGGGAGAAGTGAAAGTG-3’). The radiolabeled primers were purified using G-25 microspin columns (Cytiva, 27532501). The EMSA binding buffer consisted of 20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 20 mM KCl, 1 mM MgCl₂, 10% glycerol, 5 mM DTT, 0.5% NP-40, 10 μg of BSA, and 62.5 μg/mL Poly(dI-dC) (Thermo Scientific, 20148E) to minimize nonspecific binding. For each binding reaction, 1pmol of the radiolabeled probe was added. Reactions were incubated at 30°C for 20 minutes, then immediately placed on ice. For supershift assays, recombinant proteins in the binding buffer were pre-incubated with 0.2 μg of antibodies against IRF7 (Abcam, ab238137), IRF3 (Cell Signaling, 11904S), NF-kB (Cell Signaling, 8242T), or rabbit IgG secondary antibody (Cell Signaling, 7074S) prior to adding the radioactive probe. Following the binding reactions, protein-DNA complexes were separated on 5% acrylamide/bis-acrylamide gels (29:1 cross-link ratio, Bio-Rad, 1610156) prepared in 0.5× TBE buffer. Gels were run at 200 volts for 2 hours, dried, and exposed to film at −80°C for 2 hours. The complexes were visualized using an X-ray developer.

Mass Photometry

For Mass Photometry experiments, we used the same FLAG-purified batch of cIRF7 and exIRF7 proteins used in the previously described gel-shift assays. Measurements were carried out in silicone gaskets placed on mass photometry coverslips (MP consumables, Refeyn RD501078). Immediately prior to mass photometry measurements, protein stocks were added to a buffer (20 mM Tris-HCl (pH 7.5), 20 mM KCl, 1 mM MgCl₂, and 5 mM DTT). The reactions were incubated at 30°C for 20 minutes, then immediately placed on ice. The final working concentrations of protein was 20 nM. For each measurement, 20μL of sample (protein+ buffer) was added to a silicone gasket placed on a coverslip. Following autofocus stabilization, movies of either 60 or 120s duration were recorded. Data acquisition was started ≤5 s after the addition of proteins. Each sample was measured at least three times independently (n ≥ 3). All experiments were conducted using the same mass photometry instrument, the TwoMP mass photometer (Refeyn). Data acquisition was performed using AcquireMP software (Refeyn), and subsequent analysis was carried out with DiscoverMP software (Refeyn).

Immunofluorescence

A549 cells expressing IRF7 isoforms were fixed with 4% formaldehyde for 20 minutes at room temperature and then washed with PBS to remove residual formaldehyde. Permeabilization was performed in a permeabilization buffer (PBS containing 0.15% Triton-X100) for 5 minutes at room temperature and then washed twice with PBS. Blocking was performed using a blocking buffer (PBS containing 10% normal goat serum (NGS)) for 20 minutes at room temperature. Primary antibodies were diluted in blocking buffer and incubated with the cells overnight at 4°C. Following primary antibody incubation, cells were washed three times with PBS. Fluorophore-conjugated, species-specific secondary antibodies were diluted in blocking buffer and applied to the cells for 1 hour at room temperature. Cells were then washed three times with PBS and stained with Hoechst 3342 (Thermo scientific, D1306) in PBS for fifteen minutes at room temperature. Cells were then washed three additional times with PBS and visualized using a Zeiss LSM 980 confocal unit (50 μm pinhole size) attached to a Zeiss Axio Observer 7 inverted microscope. Images were acquired using a Plan Apochromat 63x/1.40 oil DIC M27 immersion objective, with a final pixel size of 0.071 μm/pixel. Green Fluorescence and Hoechst fluorescence were collected by illuminating the sample with a solid-state visible light laser at 488 nm and 405 nm, respectively. A 32-channel GaAsP-PMT detector was used to detect the emitted signal for Alexa Fluor 488, and a multialkali-PMT detector was used for Hoechst. The images were captured bidirectionally, with a pixel dwell time of 0.55 μs, a 1 AU pinhole and a repeat per line averaging using the mean intensity. ZEN Blue (v. 3.5, Zeiss) was used to acquire the data which was saved in the czi file format.

Luciferase Reporter Assay

HEK293T cells were seeded at a density of 1 × 10⁴ cells/well in 96-well plates containing DMEM supplemented with 10% FBS. Cells were co-transfected for 48 hours with 100ng of plasmids encoding 40ng of IRF7 isoforms (pEFneo-cIRF7 or pEFneo-exIRF7) or an empty vector, 10ng of pNL1.1.TK [Nluc/TK] to drive NanoLuc activity, and 50ng Firefly luciferase constructs with IFN-β promoter segments cloned upstream of the minimal promoter in the pGL4.26 luc2-minP-Hygro vector, or PsiCHECK2 (as a control). Transfections were performed using Lipofectamine 2000 (Thermo Fisher, 11668027) and Opti-MEM media (Gibco, 31985) according to the manufacturer’s protocol. For assays involving Sendai Virus (SeV) stimulation, cells were infected with 500 HA units/mL of SeV (Charles River Labs, 10100774) in the media for 8 hours. Luciferase activity was measured on a Tecan Spark multimode plate reader using the Nano-Glo^® Luciferase Assay System (Promega, N1110) as per the manufacturer’s instructions. Luciferase activity was normalized to uninfected cells expressing pEFneo-cIRF7.

Viral infections and quantitation

For automated microscopy of viral infections, human parainfluenza virus type 3 (HPIV3-GFP) was propagated in LLC-MK2 cells. A549 cells (IRF7-KO, exIRF7 and cIRF7) at a density of 2x10⁴ per well were plated in 96-well black tissue culture plate and infected with HPIV3-GFP at different MOIs for 24 hours. Cells were fixed using 4% formaldehyde for 15 mins at room temperature then washed with PBS three times and blocked in blocking buffer (PBST (0.1% TritonX-100) + 2% Bovine Serum Albumin (BSA)) and stained with Hoechst 3342 (Thermo scientific, D1306) for nuclei in blocking buffer for 1h. Cells were washed three times with PBS-T and imaged on a Molecular Devices ImageXpress Micro 4 imaging system. Images were quantified using Molecular Devices MetaXpress version 6 modules and percentage of infected cells was calculated.

Quantification and statistical analysis

All experiments were performed at least three times unless otherwise indicated, and representative results are shown in the figures. Student’s t-test were used to compare between two group (treatments vs. control). One-way ANOVA models were used to compare continuous outcomes across multiple experimental groups. Tukey’s multiple comparison test was performed to determine the significance of the differences between the dissociation constants (Kd) of protein isoforms in the Mass Photometry experiment. All the error bars in the bar graphs indicate Standard Deviation (SD). For all analyses, statistical analysis was performed using GraphPad Prism 10.0 program (GraphPad Software) and Microsoft excel. Throughout manuscript, statistical significance is defined as p < 0.05 by statistical tests specified above. In all figures, significance is indicated as follows: ns = not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).

Supplementary Material

Supplemental Data

Supplemental Data S1: Validation and complete RNA-Seq analysis of Jurkat CRISPR clones. (A) CRISPR clone validation, (B) Gene expression following pIC stimulation, (C) Gene expression following IFN stinulation.