Phenotypic Mutation 'Long_lost' (pdf version)
Mutation Type splice site (6 bp from exon)
Coordinate55,605,910 bp (GRCm38)
Base Change T ⇒ C (forward strand)
Gene Irf9
Gene Name interferon regulatory factor 9
Synonym(s) p48, Isgf3g, Irf-9
Chromosomal Location 55,603,571-55,610,030 bp (+)
MGI Phenotype PHENOTYPE: Mice homozygous for disruptions in this gene display an apparently normal phenotype. However, antivirus response induced by Ifn alfpha and Ifn gamma are impaired. [provided by MGI curators]
Accession Number

NCBI RefSeq: NM_001159417 (variant 1), NM_001159418 (variant 2), NM_008394 (variant 3); MGI:107587

Mapped Yes 
Amino Acid Change
Institutional SourceBeutler Lab
Gene Model predicted gene model for protein(s): [ENSMUSP00000120359 ] [ENSMUSP00000120525 ] [ENSMUSP00000119477 ]   † probably from a misspliced transcript
SMART Domains Protein: ENSMUSP00000120359
Gene: ENSMUSG00000002325

IRF 5 117 1.19e-53 SMART
low complexity region 158 182 N/A INTRINSIC
low complexity region 185 194 N/A INTRINSIC
IRF-3 211 377 1.13e-59 SMART
Predicted Effect probably null
SMART Domains Protein: ENSMUSP00000120525
Gene: ENSMUSG00000002325

low complexity region 34 58 N/A INTRINSIC
IRF 71 183 1.19e-53 SMART
low complexity region 224 248 N/A INTRINSIC
low complexity region 251 260 N/A INTRINSIC
IRF-3 277 443 1.13e-59 SMART
Predicted Effect probably null
SMART Domains Protein: ENSMUSP00000119477
Gene: ENSMUSG00000002325

IRF 23 135 1.19e-53 SMART
low complexity region 176 200 N/A INTRINSIC
low complexity region 203 212 N/A INTRINSIC
IRF-3 229 395 1.13e-59 SMART
Predicted Effect probably null
Predicted Effect probably benign
Meta Mutation Damage Score 0.9755 question?
Is this an essential gene? Probably nonessential (E-score: 0.144) question?
Phenotypic Category
Phenotypequestion? Literature verified References
dsDNA induced type I IFN production - increased
IgE response to a Cysteine Protease (Papain) - decreased
impaired response to dsDNA- type I IFN production by macrophages
LPS-induced Necroptosis - decreased
Macrophage necroptosis: low
Candidate Explorer Status CE: not good candidate; Verification probability: 0.203; ML prob: 0.264; human score: -2
Single pedigree
Linkage Analysis Data
Alleles Listed at MGI

All Mutations and Alleles(19) : Chemically induced (other)(1) Gene trapped(15) Targeted(3)

Lab Alleles
AlleleSourceChrCoordTypePredicted EffectPPH Score
IGL01285:Irf9 APN 14 55607601 missense probably damaging 0.98
IGL02283:Irf9 APN 14 55607739 missense probably damaging 1.00
IGL02317:Irf9 APN 14 55607739 missense probably damaging 1.00
IGL02407:Irf9 APN 14 55605221 missense possibly damaging 0.92
R0233:Irf9 UTSW 14 55606094 missense probably benign 0.00
R0233:Irf9 UTSW 14 55606094 missense probably benign 0.00
R1959:Irf9 UTSW 14 55607717 missense possibly damaging 0.93
R2324:Irf9 UTSW 14 55605910 splice site probably null
R4669:Irf9 UTSW 14 55605766 missense probably benign
R4882:Irf9 UTSW 14 55609039 utr 3 prime probably benign
R5393:Irf9 UTSW 14 55606457 unclassified probably benign
R6072:Irf9 UTSW 14 55605827 missense probably damaging 1.00
R6277:Irf9 UTSW 14 55607652 missense probably benign 0.04
R6337:Irf9 UTSW 14 55606342 missense possibly damaging 0.62
R6545:Irf9 UTSW 14 55605227 missense probably damaging 1.00
R6993:Irf9 UTSW 14 55608957 missense probably benign 0.06
R7956:Irf9 UTSW 14 55609024 missense probably benign 0.00
R8145:Irf9 UTSW 14 55605798 nonsense probably null
R8326:Irf9 UTSW 14 55605753 missense probably benign 0.02
RF014:Irf9 UTSW 14 55605877 nonsense probably null
Mode of Inheritance Autosomal Semidominant
Local Stock Sperm
Last Updated 2019-09-04 9:45 PM by Katherine Timer
Record Created 2015-07-31 4:05 PM by Bruce Beutler
Record Posted 2016-09-26
Phenotypic Description

Figure 1. Long_lost mice exhibit increased type I interferon production after dsDNA exposure. ELISA was utilized to determine IFN levels. Normalized data are shown. Abbreviations: WT, wild-type; REF, homozygous reference mice; HET, heterozygous variant mice; VAR, homozygous variant mice. Mean (μ) and standard deviation (σ) are indicated.

Figure 2. Long_lost mice exhibit increased rates of macrophage necroptosis after treatment with LPS. Log normalized data are shown. Abbreviations: WT, wild-type; REF, homozygous reference mice; HET, heterozygous variant mice; VAR, homozygous variant mice. Mean (μ) and standard deviation (σ) are indicated.

The Long_lost phenotype was identified among G3 mice of the pedigree R2324, some of which showed an increase in type I interferon production after dsDNA exposure (Figure 1) as well as an increased rate of macrophage necroptosis after treatment with lipolysaccharide (Figure 2).

Nature of Mutation

Figure 3. Linkage mapping of the increased rate of macrophage necroptosis after LPS stimulation using an additive model of inheritance. Manhattan plot shows -log10 P values (Y-axis) plotted against the chromosome positions of 36 mutations (X-axis) identified in the G1 male of pedigree R2324. Normalized phenotype data are shown for single locus linkage analysis without consideration of G2 dam identity.  Horizontal pink and red lines represent thresholds of P = 0.05, and the threshold for P = 0.05 after applying Bonferroni correction, respectively.

Whole exome HiSeq sequencing of the G1 grandsire (R2324) identified 36 mutations. Both of the above anomalies were linked by continuous variable mapping to a mutation in Irf9:  a T to C transition at base pair 55,605,910 (v38) on chromosome 14, or base pair 1,926 in the GenBank genomic region NC_000080 within the splice donor site of intron 4 of Irf9, six nucleotides from exon 4. The strongest association was found with an additive model of linkage to the normalized LPS-induced necroptosis value, wherein 10 variant homozygotes and 26 heterozygous mice departed phenotypically from 23 homozygous reference mice with a P value of 3.123 x 10-7 (Figure 3).  A substantial semidominant effect was observed in both of the assays. 


The effects of the mutation at the cDNA and protein level have not examined, but the mutation is predicted to result in skipping of the 184-base pair exon 8 (out of 10 total exons). Skipping of exon 8 would result in a frame shifted protein product beginning after amino acid 126 of the protein, which is normally 465 amino acids in length, and terminating after the inclusion of 33 aberrant amino acids.

             <--exon 3          <--exon 4 intron 4-->         exon 5-->        <--exon 6
122  ……-A--A--I--F--K- ……-A--G--T--L--P--                     -T--N--H--G-……-M--A--L--*-

          correct            deleted                                   aberrant


Genomic numbering corresponds to NC_000080. The donor splice site of intron 4, which is destroyed by the Long_lost mutation, is indicated in blue lettering and the mutated nucleotide is indicated in red. 

Protein Prediction

In mice and humans, interferon regulatory factor (IRF)-9 (also known as p48 or ISG3γ) is one of nine members of the IRF family of transcription factors, which regulate the transcription of type I interferons (IFN-α/β) and IFN-inducible genes during immune system development, homeostasis, and activation by microbes [reviewed by (1;2)]. An additional IRF, IRF10, has been identified in chickens (3).


Figure 4. Domain structure of the IRF9 protein. The DNA binding region (DBD) is characterized by the presence of five highly conserved tryptophans (W81, W96, W109, W128, and W146). Abbreviation: IAD1 = IRF association domain. The Long_lost mutation is within the splice donor site of intron 4 of Irf9, six nucleotides from exon 4.

As in the other IRFs, the N-terminal half of IRF8 (residues 75-182) serves as the DNA binding region (4-6), and is characterized by the presence of five highly conserved tryptophans (residues 81, 96, 109, 128, and 146 in mouse IRF9) each separated by 10-20 amino acids (Figure 4) (4). The DNA binding region bears similarity to that of the c-Myb oncoprotein that also contains a tryptophan cluster (7), but not to any other transcription factor classes. IRF proteins share sequence and structural homology in their DNA binding regions, and all bind to a similar DNA motif (A/G NGAAANNGAAACT) called the IFN-stimulated response element (ISRE) (8) or IFN regulatory element (IRE) (9) that is present in the regulatory regions of interferons and interferon-stimulated genes (ISGs). 


Figure 5. DNA binding domain of IRF4. β-strands and α-helices are labeled and shown in teal and pink, respectively. The long loops are also labeled. UCSF Chimera structure is based on PDB:2IRF. Click on the image to view the structure rotate.

Crystal structure analyses of IRF DBDs suggest that they comprise a four-stranded antiparallel β-sheet (β1-β4), three helices (α1-α3), and three long loops (L1-L3) connecting β2 to α2, α2 to α3, and β3 to β4, respectively (Figure 5) (10-14). IRF DBDs bind to DNA over a 12 base pair (bp) stretch with protein-DNA contacts in the major and minor grooves determining specificity. The structure of the IRF9 DBD has not been analyzed, but crystallization of the DNA binding domains of PU.1 and IRF4 in complex with DNA suggests that the DNA in the complex contorts into an S shape in order to juxtapose PU.1 and IRF4 and allow electrostatic and hydrophobic interactions across the central minor groove. The IRF4 and PU.1 DBDs bind to opposite faces of the DNA in a head-to-tail orientation. Loop L3 of the IRF4 DBD is actually a short 310 helix. The PU.1 α3 recognition helix lies in the major groove perpendicular to the DNA axis, while the IRF4 α3 recognition helix is tilted in the major groove with its axis almost parallel to the DNA backbone. IRF4 interacts with PU.1 in this complex with Asp117, Leu116, Val111, and His56, residues that are not conserved in IRF9. Amino acids in the IRF4 DBD contacting the DNA include Trp54, Ala57, Arg64, Lys94, Thr95, Asn102 and Lys123. His56, which also interacts with PU.1, specifies adenines found commonly upstream of the IRF GAAA core sequence. The outer three base pairs of the IRF core (GAAA) are recognized by Arg98 and Cys99. Lys103 at the end of the recognition helix is also important for DNA recognition (12)


The C-terminal halves of all IRF family members contain either an IRF association domain 1 (IAD1) or an IAD2, with which they bind to other IRFs, other transcription factors, or self-associate. These interactions allow the IRFs to modulate their activity and target a variety of genes. The IAD1 is approximately 177 amino acids in length, and is conserved in all IRFs except IRF1 and IRF2. IAD2 domains are found only in IRF1 and IRF2 (5;6;15)


The Long_lost mutation is predicted to result in skipping of the 184-base pair exon 8, resulting in a frame shifted protein product beginning after amino acid 126 of the protein and terminating after the inclusion of 33 aberrant amino acids (at amino acid 159). Amino acid 159 is within the DBD.


IRF9 is ubiquitously and constitutively expressed.

Figure 6. Diagram of the type I interferon receptor complex composed of IFNAR1 and IFNAR2 and type I IFN signaling pathways. The extracellular regions of the IFNAR receptor subunits consist of Fibronectin III domains, which are labeled as subdomain (SD) 1-4 in IFNAR1. The intracelluar regions of the IFNAR receptor subunits can interact with several factors that are important from type I IFN signaling. IFNAR1 and 2 contain proline-rich regions that bind to JAK tyrosine kinases (TYK2 binds to IFNAR1 and JAK1 binds to IFNAR2.) Phosphorylation of the IFNAR intracellular domains by these kinases leads to the activation of STATs, which are then able to form homo- and heterodimeric complexes. The interaction of SOCS1 with IFNAR1 inhibits STAT activation, while UBP43 binds to the JAK1 binding site. The STAT1/STAT2 heterodimer forms the ISGF3 complex with IRF9 that translocates into the nucleus where it activates genes containing the IFN-stimulated response element (ISRE). STAT1/STAT1 homodimers can activate genes containing the IFN-γ-activated sequence (GAS). The JAK kinases also activate other substrates, which initiate many pathways. For example, phosyphorylation of insulin receptor substrate (IRS) proteins leads to PI3K activation, wh ich has multiple downstream effectors. Phosphorylation of guanine nucleotide exchange factor (VAV) leads to activation of the small G protein, RAC1, which in turn activates MAP kinase pathways leading to p38 activation. Other factors that are activated in response to type I IFNs include the transcription factor NF-κB, which is downstream of PI3k or TRAF2, and AP1, which is activated by the ERK pathway. Please see text for further details. This image is interactive. Click on the image to view mutations within the pathway (red) and the genes affected by these mutations (black). Click on the mutations for more specific information.

Type I IFNs are a critical class of cytokines that have antiviral, growth-inhibitory and immunomodulatory functions, and comprise 13 IFN-α subtypes, as well as IFN-β, w, ε, ω, κ and others in some animal species [reviewed in (16)]. Type I IFN signaling modulates the expression of hundreds of ISGs, accounting for the diverse biological properties of these cytokines and their highly pleiotropic and diverse effects.  Type I IFNs induce ISGs by activating several signal transduction pathways including the classical JAK/STAT signaling pathway (Figure 6). Binding of IFN-β to the type I IFN receptor (see the records for macro-1 and macro-2) stimulates the Janus activating kinases (JAK)– signal transducer and activator of transcription (STAT) pathway (see the record for domino). Phosphorylation of critical tyrosine residues on the IFNAR receptor results in recruitment of signal transducing molecules such as STAT1 and STAT2 to the complex (17-19). Phosphorylation of these signal-transducing molecules by the IFNAR-associated protein kinases (IRAKs) results in formation of STAT1/STAT2 heterodimers and STAT1 homodimers that dissociate from the receptor and translocate into the nucleus to form transcriptional complexes with other factors. STAT1/STAT2 heterodimers, along with IRF-9, forms the IFN-stimulated gene factor 3 (ISGF3) complex that binds to ISRE of type I IFN-inducible genes to initiate transcription. The interaction of ISGF3 with the ISRE on target genes is mediated by IRF9, but STAT1/STAT2-mediated DNA contacts are necessary for stabilization of the IRF9-ISRE interaction. ISGF3 regulates the transcription of IRF-7 (20;21), providing the type I IFN system a positive feedback loop that allows massive amplification of the type I IFN response. STAT1 homodimers stimulate the transcription of genes containing the IFN-γ-activated sequence (GAS) (22). In addition to the positive feedback loop mediated by ISGF3-induction of IRF-7, constitutively low levels of IFN-α/β have been found to be expressed in several cell types and appear to be a prerequisite for enhanced production of type I IFNs in response to stimuli. These low levels of IFN-α/β do not significantly activate downstream signaling events through IFNAR, but they do maintain tyrosine phosphorylation on IFNAR1, allowing for more efficient recruitment of STAT1 (23)


A STAT2/IRF9-dependent, STAT1-independent signaling pathway stimulates a subset of ISGs independent of ISGF3 (24). STAT2 is capable of forming homodimers when phosphorylated in response to IFNα (25). The STAT2 homodimers subsequently interact with IRF3 to form an ISGF3-like complex, STAT2/IRF9. The STAT2/IRF9 complex and ISGF3 can regulate the expression of a common set of ISGs, but with different kinetics. The STAT2/IRF9 complex stimulated the expression of the ISGs in a more prolonged manner compared to the early and transient response mediated by ISGF3. Furthermore, the STAT2/IRF9 complex regulated the expression of ISGs that did not have an ISRE. Similar to ISGF3, STAT2/IRF9 could trigger an antiviral response to encephalomyocarditis virus (EMCV) and vesicular stomatitis Indiana virus (VSV).


An ISGF3 complex comprised of IRF9 with an unphosphorylated STAT1 and unphosphorylated STAT2 (designated as U-ISGF3) is upregulated in response to IFN-λ and IFN-β during chronic hepatitis C virus (HCV) infection (26). Also, high levels of U-ISGF3 confer unresponsiveness to IFN-α therapy. The U-ISGF3 promotes the expression of a subset of ISGs and restricts HCV chronic replication in hepatocytes. For example, the U-ISGF3 stimulates the expression of ISG15, leading to increased expression of USP18, a negative regulator of the response to exogenous IFN-α.


IRF9 has additional functions outside of the immune system, including the regulation of oncogenesis, cell proliferation, and apoptosis (27-30).


Hepatic insulin resistance and steatosis

Irf9-deficient mice exhibit increased body weights, aggravated insulin resistance, hepatic steatosis, and inflammation compared to wild-type mice after chronic high-fat diet feeding (31). Adenoviral-mediated hepatic IRF9 overexpression in diet-induced and ob/ob mice improved insulin sensitivity and attenuated hepatic steatosis and inflammation. IRF9 binds to peroxisome proliferator-activated receptor α (PPARα) to activate PPARα target gene expression in the liver, thus attenuating hepatic insulin resistance and steatosis (31).


Cardiac hypertrophy

IRF9 is a negative regulator of cardiac hypertrophy (32). Irf9-deficient mice exhibit cardiac hypertrophy after pressure overload. The mice had increased cardiomyocyte size, fibrosis, and reduced cardiac function. Transgenic mice that overexpressed IRF9 in the heart showed reduction in the hypertrophic response. IRF9 competes with p300 for binding to myocardin, a coactivator of serum response factor (32). Interaction of IRF9 with myocardin suppresses the transcriptional activity of myocardin.

Neointima formation

After injury, vascular smooth muscle cells proliferate, migrate, and secrete extracellular matrix, forming a neointima. Neointima formation is a pathophysiological process that occurs during severe vascular pathologies such as atherosclerosis, in-stent restenosis, and vein bypass graft failure (33). IRF9 functions in neointima formation (34). In Irf9-deficient mice, proliferation and migration of vascular smooth muscle cells (VSMC) was inhibited and intimal thickening was attenuated after injury. In contrast, a gain-of-function IRF9 promoted VSMC proliferation and migration. IRF9 inhibited the transcription of the neointima formation modulator SIRT1.


Ischemia/reperfusion (I/R) injury

Cerebral ischemia/reperfusion (I/R) injury triggers several cell death mechanisms, including excitotoxicity, oxidative stress, inflammation, and apoptosis-like cell death (35). IRF9 mediates neuronal death in male mice (36). Loss of IRF9 expression diminished post-stroke neuronal death and neurological deficits. In contrast, neuron-specific overexpression of IRF9 sensitized neurons to cell death. In the neurons, IRF9 inhibited SIRT1 deacetylase activity, subsequently promoting the acetylation of p53 and activation of p53-mediated cell death signaling. SIRT1 is a member of the class III deacetylases (NAD+-dependent). The SIRT enzymes function in several cellular pathways, including metabolism, stress, and genome stability. P53 is a substrate of SIRT1; the p53-SIRT1 axis mediates cell senescence and tumor progression. Furthermore, SIRT1 reduces p53-dependent stress-induced apoptosis (37).


IRF9 is upregulated in ischemic heart tissue and mouse hearts after I/R injury (38). Loss of IRF9 expression protects the heart against I/R-induced cardiomyocyte death, development of inflammation, and loss of heart function (38). Cardiomyocyte-specific overexpression of IRF9 aggravated myocardial reperfusion injury and inflammation. Similar to that observed in cerebral I/R injury, IRF9 negatively regulated the SIRT1-p53 axis after cardiac I/R (38).


Irf9 deficient mice exhibited reduced necrotic area, immune cell infiltration, inflammatory cytokine levels, and hepatocyte apoptosis after liver I/R (39). In contrast, injuries were augmented in IRF9 overexpressing mice after liver I/R. Similar to that in cerebral and cardiac I/R injury, IRF9 induces hepatocyte apoptosis after liver I/R injury by reducing SIRT1 expression and increasing acetyl-p53 levels.


IgG autoantibody production

Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by the production of high-titer isotype-switched IgG autoantibodies that recognize nuclear autoantigens and subsequent causes destruction of organ systems such as skin, joints, and kidneys. The development of IgG autoantibodies directed towards SLE autoantigens is IRF9- and STAT1-dependent (40). Irf9-deficient mice treated with pristine to induce lupus produced higher titers of IgM autoantibodies, indicating that IRF9 may function in isotype switching (40). Furthermore, B cells from the Irf9-deficient mice displayed defects in TLR7 expression and responses to TLR7 ligands (40). The Irf9-deficient mice were capable of isotype switching to all IgG subtypes in response to immunization with OVA in CFA, indicating that B cell activation, signaling, maturation, affinity maturation, and isotype switching were not affected upon loss of IRF9 expression.



IRF9 deficiency protects against inflammatory responses in a mouse model of dextran sodium sulfate (DSS)-induced colitis (41). Irf9-deficient mice exhibited a delay in weight loss of up to two days compared to wild-type controls and lost less weight than wild-type mice after treatment with 2% DSS. Furthermore, during colon inflammation, IRF9 induces CXCL10 expression, which subsequently recruits CXCR3+ inflammatory cells. IRF9 is proposed to participate in a noncanonical complex with STAT1 independently of IFN-I and IFN-III receptor signaling. Participation in this purported noncanonical complex promotes the procolitogenic activity of IRF9.

Putative Mechanism

Type I IFN signaling is a major mechanism of necroptosis. IFN-I-induced necroptosis after LPS, polyI:C, or TNF-α stimulation occurs through both Toll/IL-1 receptor domain-containing adaptor inducing IFN-β (TRIF)-dependent or –independent mechansims, subsequently leading to persistent phosphorylation of receptor-interacting protein 3 (Rip3) kinase and necroptosis (42). IRF9 is necessary for sustained Rip3 activation and necroptosis. IRF9-STAT1- and STAT2-deficient macrophages were resistant to necroptosis. IFN-β–induced macrophage necroptosis occurs through tonic ISGF3 signaling, which leads to persistent expression of STAT1, STAT2, and IRF9. The phenotypes observed in the Long_lost mice indicate loss of IRF9Long_lost function.

Primers PCR Primer

Sequencing Primer

PCR program

1) 94°C 2:00
2) 94°C 0:30
3) 55°C 0:30
4) 72°C 1:00
5) repeat steps (2-4) 40x
6) 72°C 10:00
7) 4°C hold

The following sequence of 412 nucleotides is amplified (chromosome 14, + strand):

1   aggcttgggc actgtttaag gaaaagcaca aagatgggga cataggacac cccgctgtct
61  ggaagactcg cctacgctgt gccctcaaca agagttccga atttgaggag gttcccgaga
121 gaggtcgtat ggatgttgct gaaccctaca aagtatatcg aatactgcca gcaggaaccc
181 tccctagtga gtgctctctt gccgctcctg gaaggggcag actgacaagg atagagataa
241 gagccctccc tctgtgggct tccaggcctt cttgggctct ctagtcctgc ccagcttttt
301 gttgtttcag atactcactc ccccaccccc tttgtagaag acctaacttt cttctgcctg
361 caacatttcc acagaccaac cacggaacca gaaatcacca tgcaagcgaa gt

Primer binding sites are underlined and the sequencing primers are highlighted; the mutated nucleotide is shown in red.

Science Writers Anne Murray
Illustrators Diantha La Vine, Peter Jurek, Katherine Timer
AuthorsDoan Dao*, Ying Wang*, Hexin Shi, Zhao Zhang, Lei Sun, Jianhui Wang, Ashley Leach, and Bruce Beutler. (*equal)