UV light potentiates STING-dependent innate immune signaling UV light potentiates STING (stimulator of interferon genes)-dependent innate immune signaling through deregulation of ULK1 (Unc51-like kinase 1)

The mechanism by which ultraviolet (UV) wavelengths of sunlight trigger or exacerbate the symptoms of the autoimmune disorder lupus erythematosus is not known but may involve a role for the innate immune system. Here we show that UV radiation potentiates STING (stimulator of interferon genes)-dependent activation of the immune signaling transcription factor interferon regulatory factor 3 (IRF3) in response to cytosolic DNA and cyclic dinucleotides in keratinocytes and other human cells. Furthermore, we find that modulation of this innate immune response also occurs with UVmimetic chemical carcinogens and in a manner that is independent of DNA repair and several DNA damage and cell stress response signaling pathways. Rather, we find that the stimulation of STING-dependent IRF3 activation by UV is due to apoptotic signaling-dependent disruption of ULK1 (Unc51-like kinase 1), a proautophagic protein that negatively regulates STING. Thus, deregulation of ULK1 signaling by UV-induced DNA damage may contribute to the negative effects of sunlight UV exposure in patients with autoimmune disorders. The symptoms of a number of autoimmune diseases, including systemic lupus erythematosus (SLE), have long been known to be triggered or exacerbated by exposures to ultraviolet (UV) wavelengths of sunlight that damage genomic DNA and other cellular biomolecules (1-3). This damage and associated cell death is thought to result in the release of potential autoantigens, including DNA and other nuclear proteins, which are taken up by macrophages to initiate autoantibody development or targeted by the adaptive immune system. In addition, DNA from dying cells as well as DNA from viruses and microbial pathogens are known to be strong immune stimulants that can accumulate in the cytosol of infected cells and activate an intracellular signal transduction 1 http://www.jbc.org/cgi/doi/10.1074/jbc.M115.649301 The latest version is at JBC Papers in Press. Published on March 19, 2015 as Manuscript M115.649301 Copyright 2015 by The American Society for Biochemistry and Molecular Biology, Inc. by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling cascade that produces various immune system modulators and pro-inflammatory cytokines (4, 5). This pathway is critically dependent on a protein known as STING (stimulator of interferon genes; also known as TMEM173, MITA, MYPS, and ERIS) (6-8), which responds not to DNA directly but to the cyclic dinucleotide 2’3’-cGAMP that is produced upon the stimulation of the enzyme cGAS (cyclic GMP-AMP synthase) by cytosolic DNA (9-15). The binding of STING to 2’3’cGAMP and related cyclic dinucleotides that are produced by microbial pathogens (16-19) causes a conformational change in STING (20, 21) that is thought to allow it to function as a scaffold protein to mediate the phosphorylation and activation of IRF3 (interferon regulatory factor 3) by the upstream kinase TBK1 (TANK-binding kinase 1) (22). This phosphorylation event induces IRF3 homo-dimerization and is required for IRF3 to function as a transcription factor for a variety of gene targets, including type I interferons, proinflammatory cytokines, and pro-apoptotic factors (23-26). The major model for UV stimulation of immune responses involves the abnormal clearance of DNA or other cellular constituents that are released from dying cells and then taken up by macrophages and other immune cells (2731). However, there may be additional mechanisms that take place in viable and nonimmune cells, such as keratinocytes, that modulate immune signaling within the skin and throughout the body. In this report, we show that UV and UVmimetic chemical carcinogens stimulate STINGdependent IRF3 activation in response to cytosolic DNA and cyclic dinucleotides in keratinocytes and other human cells. Moreover, we find that this effect is independent of DNA repair and several major UV-dependent DNA damage response and cell stress signaling pathways. Instead, abrogation of the negative STING regulator and proautophagic factor ULK1 (Unc51-like kinase 1) (32) by UV-induced apoptotic signaling is responsible for potentiating the cellular innate immune responses to cytosolic DNA and cyclic dinucleotides. These data therefore provide a novel mechanism by which UV and related environmental genotoxins may trigger or exacerbate autoimmunity in susceptible individuals. EXPERIMENTAL PROCEDURES Cell Lines—Human THP-1 monocytes, HaCaT keratinocytes, and HEK293T cells were cultured at 37°C in a 5% CO2 humidified incubator in either RPMI1640 or Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin. HEK293T cells stably expressing HA-tagged STING were generated by transfection of pUNO1hSTING-HA3x (Invivogen) and selection with blasticidin. To expose cells to UV radiation, cells were placed under a GE germicidal lamp that emits primarily 254-nm UV light (UV-C) connected to a digital timer and irradiated with the indicated fluences of UV light (typically 100 J/m). Other genotoxic compounds were added directly to tissue culture media, as indicated. Fractionation of cells to yield cytosolic and nuclear fractions was performed as previously described (33). Chemicals and Reagents—The 45-nt-long sense and anti-sense strands of inteferon-stimulatory DNA (ISD) oligonucleotides (23) were synthesized by Sigma, resuspended in 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 100 mM NaCl, and then annealed by heating to 90°C for 5 min and slow cooling to room temperature. Benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE) and N-acetoxy-2-acetylaminofluorene (AAF) were obtained from the NCI Chemical Carcinogen Reference Standard Repository (Midwest Research Institute, MO). Cycloheximide, the p38 inhibitor SB202190, the Jun kinase (JNK) inhibitor SP600125, the MEK1/2 inhibitor U0126, the ATM (ataxia telangiectasia-mutated) kinase inhibitor KU55933, the DNA-PK (DNA-dependent protein kinase) inhibitor NU7026, the Chk1 (checkpoint kinase 1) inhibitor PF-477736, lipopolysaccharide, and poly-inosinic-poly-cytidilyic acid (pIC) were purchased from Sigma. The ATR (ataxiatelangiectasia-mutated and Rad3-related) inhibitor VE-821 was purchased from Selleckchem. 2’3’cGAMP (cyclic GMP-AMP), 3’3’-cGAMP, and cdi-GMP (cyclic di-GMP) were from Invivogen. ISD and cGAMP Transfections—ISD was transfected into THP-1 cells (typically 400,000 cells per ml in 2 ml of culture medium in a 3.5 cm plate) or HaCaT (90% confluent) using a 1:1 ratio of DNA (typically 5 μg) to Lipofectamine 2000 (5 μl) as recommended by Invitrogen. Where 2 by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling indicated, cells were irradiated with UV radiation 20-30 min prior to transfection with ISD. Transfection with 2’3’-cGAMP, 3’3’-cGAMP, and cyclic di-GMP (c-di-GMP) was performed in a similar manner using Lipofectamine 2000. Unless otherwise indicated, cells were harvested 3.5 to 4 h following transfection, pelleted by centrifugation (1,600 x g, 4 min), washed with PBS, and then lysed for 10 min on ice in 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.5% NP-40 containing 0.1 mM PMSF, 1 mM DTT, 10 mM NaF, 1 mM Na2VO3, and 10 mM glycerophosphate. Soluble lysates were prepared by centrifugation at 14,000 x g for 10 min at 4°C, separated by SDS-PAGE, and transferred to nitrocellulose membranes for immunoblotting. Dimerization of IRF3 was performed using native PAGE as previously described (34). Immunoblotting—Antibodies used for immunoblotting included anti-phospho-RPA2 (Ser33) (catalog no. A300-246A) from Bethyl Laboratories; anti-XPB (catalog no. sc-293), antiChk1 (sc-8408), anti-IRF3 (sc-9082), and anti-HA (sc-805) from Santa Cruz Biotechnology; and antiphospho-IRF3 (Ser396; #4947), anti-phosphoTBK1 (Ser172; #5483), anti-TBK1 (#3504), antiSTING (#13647), anti-phospho-ULK1 (Ser555; #5869), anti-ULK1 (#8054), anti-phospho-AMPKalpha (Thr172; #2535), anti-AMPK-α (#5832), anti-phospho-LKB1 (Ser428; #3482), antiAMBRA1 (#12250), anti-PARP (#9542), antiphospho-Chk1 (Ser345; #2348), anti-cleaved Caspase-3 (#9661), anti-phospho-Chk2 (T68; #2661), anti-phospho-Chk1 (Ser296; #2349), antiphospho-p44/p42 MAPK (ERK1/2) (Thr202/Tyr204; #4370), anti-phosphoMAPKAPK-2 (Thr334; #3041), anti-phospho-cJun (Ser63; #9261), and anti-phospho-DNA-PK (Ser2056; #4215) from Cell Signaling Technology. Secondary antibodies included horseradish peroxidase-linked anti-mouse and anti-rabbit IgG (catalog nos. NA931V and NA934V) from GE Healthcare. Chemiluminescence was visualized with Clarity Western ECL Substrate (Bio-Rad) or Amersham ECL Advance Substrate (GE Healthcare) using a Molecular Imager Chemi-Doc XRS+ system (BioRad). Chemiluminescent signals were quantified using ImageQuant software (GE Healthcare). For each treatment condition, the phospho-IRF3 (IRF3-P) signals were normalized to total IRF3 levels and then normalized to the highest IRF3P/IRF3 signal ratio for each blot (which was set to an arbitrary value of 100). All experiments were repeated at least twice, and the average and standard error of the mean calculated for each treatment. RNA Interference—Lentiviral DNA particles were generated in HEK293T cells by co-transfection of the appropriate pLKO.1 vector with the packaging plasmid psPAX2 and the envelope plasmid pMD2.G with Lipofectamine 2000. Empty and XPA shRNA-containing pLKO.1 plasmids were from the Open Biosystems TRC1 shRNA library (35). THP-1 cells were infected with lentivirus in the presence of 8 μg/ml polybrene, incubated for 24 h, and then selected with puromycin. siGENOME Non-Targeting siRNA Pool #1 and ON-TARGETplus Human ULK1 siRNA SMARTpool (Dharmacon) were transfected into THP-1 cells using Lipofectamine RNAiMax (Invitrogen). Cells were transfected twice with siRNAs (40 nM) over a period of 48 hours. Cells were pelleted and resuspended at 400,000 cells per ml in fresh medium prior to transfection with ISD. Measurement of excision repair—UV irradiated cells were lysed using a modified method of Hirt (33, 36-38). Briefly, following centrifugation and washing of UV-irradiated cells with cold PBS, the cell pellets were resuspended in 10 mM Tris-HCl pH 8, 10 mM EDTA, and 1% SDS. NaCl was added to a final concentration of 1 M, and cells were incubated on ice for 1 h. Hirt lysates were then prepared by centrifugation (14,000 x g, 30 min, 4°C) and digested with proteinase K for 30 min at 60°C. Following phenol-chloroform extraction and ethanol precipitation, the samples were resuspended in 10 μl of 10 mM Tris-HCl pH 8.5. Half of the resuspended DNA was 3’-end labeled in a 10 μl reaction containing 6 U of terminal deoxynucleotidyl transferase (TdT; New England Biolabs), 0.25 mM CoCl2, and [α-P]-3’deoxyadenosine 5’-triphosphate (cordycepin 5’triphosphate; PerkinElmer Life Sciences) in 1X TdT buffer (NEB) for 1 h at 37°C. Following phenol-chloroform extraction and ethanol precipitation, the radiolabeled excised oligonucleotides were separated on ureacontaining polyacrylamide gels and then detected with a phosphorimager. Radiolabeled oligonucleotides of known length were resolved on all gels as size markers. Excision repair activity 3 by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling was quantified using ImageQuant 5.2 software (GE Healthcare) as previously described (33). RESULTS UV radiation does not directly activate IRF3 Though exposure to UV wavelengths of sunlight is known to exacerbate the symptoms of the autoimmune disorder SLE (1-3), the molecular mechanisms that are responsible for this effect and the relative roles of the innate and adaptive immune systems are not clear. Because the protein STING has emerged as a critical regulator of innate immunity (6-8), we therefore considered the possibility that UV light may specifically modulate STING-dependent innate immune signaling. To test this hypothesis, THP-1 monocytes were exposed to increasing amounts of UV radiation and then harvested to examine STING activation. In response to specific immune stimuli, such as cytosolic DNA, activated STING functions as a scaffold or adaptor protein to facilitate the phosphorylation of IRF3 on Ser396 by the upstream kinase TBK1 (22). As a positive control for activation of the STING-IRF3 pathway in these experiments, we transfected cells with a 45-bp-long dsDNA termed interferon-stimulatory DNA (ISD) for its ability to induce interferon expression in responsive cells (23). As shown in Figure 1A, though ISD induced a robust response (lane 6), IRF3 phosphorylation was not observed in the UV-irradiated cells even at high doses of UV that saturate the ability of the nucleotide excision repair system to remove UV photoproducts from genomic DNA (Figure 1A, bottom panel). We next performed time course experiments to monitor the kinetics of IRF3 phosphorylation following UV irradiation. Though UV induced rapid phosphorylation of the DNA damage checkpoint kinase Chk1, there was no evidence of IRF3 phosphorylation at any time point following irradiation (Figure 1B). Similarly, when we treated THP-1 cells with the UV-mimetic N-acetoxy-2-acetylaminofluorine (AAF), we did not observe detectable levels of IRF3 phosphorylation (data not shown). These results indicate that the exposure of cultured human cells to UV alone is not sufficient to activate IRF3. Because IRF3 phosphorylation is essential for downstream innate immune signaling responses to cytosolic DNA (23), we conclude that UV does not directly induce a STING/IRF3mediated innate immune response. UV stimulates TBK1-IRF3 signaling in response to cytosolic DNA In addition to sunlight exposure, viral and microbial infections are additional risk factors for SLE (39, 40) and may result in the release of pathogen-derived DNA into the cytosol of infected cells. Whether exposures to UV wavelengths of sunlight may work in concert with microbial infections to impact the symptoms of autoimmune disorders is not known. We therefore sought to mimic such a scenario by introducing DNA into the cytosol of UV-irradiated cells. THP-1 cells were irradiated with UV and then transfected with increasing amounts of ISD. As shown in Figure 2A, under conditions where limiting amounts of ISD were introduced into cells, UV exposure enhanced IRF3 phosphorylation and to a lesser extent the phosphorylation of the upstream IRF3 kinase TBK1, which can undergo STING-dependent autophosphorylation (22). Furthermore, the level of IRF3 phosphorylation potentiated by UV in ISD-transfected cells was dependent on UV dose and became saturated at approximately 100 J/m (Figure 2B), which is a dose that saturates nucleotide excision repair capacity in THP-1 cells (Figure 1A). Moreover, time course experiments showed that UV affected the amplitude of IRF3 phosphorylation but not the general kinetics of the response following ISD administration (Figure 2C). Additional experiments aimed at determining how the timing of UV exposure relative to the exposure of the cytosol to DNA affected this response showed that UV exposures up to 1 h before or after transfection with ISD led to maximal stimulation of IRF3 phosphorylation but that prolonged incubation times of greater than 1-2 h following irradiation and prior to ISD transfection instead inhibited TBK1-IRF3 signaling in response to ISD (data not shown). These results indicate that an optimal window of exposure exists under which UV can enhance TBK1-IRF3 signaling in response to cytosolic DNA. A quantitative analysis of the response of THP-1 monocytes to UV under optimized conditions of ISD concentration, UV dose, and 4 by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling treatment time is presented in Figure 2D. These data show that the phosphorylation of IRF3 was completely ISD-dependent and was stimulated 3to 5-fold by UV exposure. To confirm these observations in another cell line of physiological relevance to sunlight UV exposure, we irradiated HaCaT keratinocytes with UV and then transfected the cells with ISD. As shown in Figure 2E, IRF3 phosphorylation was enhanced 4to 5fold by UV in ISD-transfected keratinocytes but was not impacted by UV in the absence of ISD. Phosphorylation of IRF3 is associated with protein dimerization and entry into the nucleus, where it acts as a transcription factor for genes involved in immunity, inflammation, and apoptosis (23-26, 34). We therefore used nativePAGE to monitor the dimerization of IRF3 and observed that IRF3 dimerization was ISDdependent and further stimulated by UV (Figure 2F). Similarly, the amount of IRF3 found in the nuclear fraction of cells was increased in ISDtransfected cells exposed to UV (Figure 2G). Thus, using several different biochemical readouts for IRF3 activation, we conclude that UV stimulates IRF3 activation in a manner that is dependent upon the introduction of cytosolic DNA into cells. UV-mimetics stimulate cytosolic DNA-dependent TBK1-IRF3 signaling UV radiation results in the formation of photoproducts in genomic DNA that interfere with gene transcription and other aspects of DNA metabolism. However, UV also causes damage to other cellular biomolecules that may modulate the cellular response to cytosolic DNA. Thus, to determine whether damage to cellular DNA is responsible for potentiating TBK1-IRF3 signaling in response to transfected DNA, we treated cells with the UV-mimetic chemical carcinogens BPDE and AAF. As shown in Figure 3A, BPDE and AAF stimulated ISD-dependent IRF3 phosphorylation to a similar extent as UV. Moreover, the effects of AAF and BPDE were found to be dose-dependent (Figure 3B and C). UV, AAF, and BPDE all induce formation of “bulky adducts” in DNA that are removed from the genome by the nucleotide excision repair system (41-43), which indicated that a cellular response to these types of lesions is responsible for the effect of UV and UV-mimetic agents on the stimulation of cytosolic DNA-dependent TBK1IRF3 signaling. To examine whether the repair of these forms of DNA damage is required to potentiate the cellular response to cytosolic DNA, we generated THP-1 cell lines that stably expressed shRNAs against the essential nucleotide excision repair factor XPA. Though XPA protein expression was reduced by greater than 95% in two independent cell lines (Figure 4A), which reduced the rate of nucleotide excision repair by 4to 5-fold (Figure 4B), XPA knockdown did not affect the stimulation of ISD-dependent IRF3 activation by UV (Figure 4C, D). We conclude that an intact nucleotide excision repair system is not required for UV and UV-mimetics to potentiate TBK1-IRF3 signaling in response to cytosolic DNA. These lesions also activate a number of cell signaling pathways that govern various cellular responses to DNA damage and associated genomic stress (44, 45). However, the use of chemical inhibitors of several DNA damage response kinases and mitogen-activated protein (MAP) kinases that are known to be activated following UV irradiation showed that the kinase activities of ATR, Chk1, ATM, DNA-PK, p38, JNK, and MEK1/2 were not responsible for the effect of UV on STING-dependent IRF3 activation by cytosolic DNA (data not shown). Furthermore, additional experiments demonstrated that UV potentiated the cytosolic DNA response in nocodazole-arrested mitotic THP-1 cells and in serum-starved quiescent HaCaT cells (data not shown), which indicates that the effect of UV on IRF3 activation is independent of cell cycle phase. UV stimulates TBK1-IRF3 signaling in response to cyclic dinucleotides To better understand the mechanism by which UV and UV-mimetic chemical carcinogens impact IRF3 activation, we next determined whether UV affected the cellular response to other activators of innate immune signaling. We first investigated the cyclic dinucleotide 2’3’-cGAMP, which is produced by the enzyme cGAS upon binding to cytosolic DNA (10, 11, 13, 14). Though STING has a low affinity for DNA, its activation in response to cytosolic DNA is thought to occur instead through binding to the intermediate signaling molecule 2’3’-cGAMP (17), which alters 5 by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling the conformation of STING and allows it to mediate phosphorylation of IRF3 by TBK1 (5, 20, 21). Exogenous 2’3’-cGAMP can therefore be introduced into cells to directly activate the STING-TBK1/IRF3 pathway and bypass the need for cGAS or cytosolic DNA (9). Thus, we next asked whether UV affects the innate immune response to 2’3’-cGAMP in a similar manner as for cytosolic DNA. Non-irradiated and UV-irradiated THP-1 cells were therefore transfected with increasing amounts of cGAMP and then analyzed for stimulation of TBK1-IRF3 signaling. As shown in Figure 5A, the level of IRF3 phosphorylation was stimulated by UV at each concentration of 2’3’cGAMP. Thus, UV stimulates IRF3 activation following the administration of both cytosolic DNA (ISD) and 2’3’-cGAMP. Other pathogen-derived cyclic dinucleotides, including 3’3’-cGAMP and c-diGMP, are released into the cytosol of mammalian cells upon infection with specific microbes and have been shown to bind and stimulate STING (16, 18, 19, 46). As shown in Figure 5B, we found that UV stimulated the level of IRF3 phosphorylation approximately 2to 2.5-fold in response to both of these cyclic dinucleotides. Because IRF3 activation by cytosolic DNA and cyclic dinucleotides is dependent upon the expression of STING, we next examined the protein levels of STING in different human cell lines. Consistent with previous reports (10), STING protein expression was readily detectable in THP-1 monocytes but was nearly absent in HEK293T cells (Figure 5C). Though less abundant than in THP-1 cells, STING protein was clearly evident in HaCaT keratinocytes, which is supported functionally by the ability of HaCaT cells to respond to cytosolic DNA (Figure 2E). Furthermore, when we transfected UV-irradiated HaCaT keratinocytes with low amounts of 2’3’cGAMP, we observed that UV exposure promoted cGAMP-dependent IRF3 phosphorylation (Figure 5D). We conclude that UV stimulates IRF3 phosphorylation in response to both cytosolic DNA and cyclic dinucleotides in keratinocytes and other human cells. Transfection of HEK293T cells (which express very low levels of STING) with 2’3’cGAMP led to only a small induction of IRF3 phosphorylation. However, when we stably expressed an HA-tagged STING construct in HEK293T cells (Figure 5E), we observed a much more robust response to 2’3’-cGAMP that was further enhanced by exposure of the cells to UV. We conclude that UV enhances IRF3 activation in response to both cytosolic DNA and cyclic dinucleotides and that this response is dependent on the expression of STING. UV does not stimulate TBK1-IRF3 signaling in response to RNA or LPS TBK1 and IRF3 are also utilized in a STING-independent, but MAVS-dependent signaling pathway that responds to viral RNA and which similarly results in IRF3 phosphorylation and activation (47). Thus, to determine whether the effect of UV on TBK1-IRF3 signaling is specific for the cytosolic DNA/cyclic dinucleotide-STING-dependent pathway, we transfected non-irradiated and UV-irradiated cells with ISD, 2’3’-cGAMP, or the viral RNA mimic pIC and then monitored the phosphorylation status of TBK1 and IRF3. As an additional control, we also added the immune trigger and bacterial wall component lipopolysaccharide (LPS) to the culture medium of non-irradiated and UV-irradiated cells. As shown in Figure 6A, IRF3 phosphorylation was enhanced by UV in cells that were transfected with ISD or cGAMP but not in cells that were stimulated with pIC or LPS. Because the kinetics of innate immune signaling may be different with pIC and LPS than with DNA and cGAMP, we also performed time course experiments following stimulation with pIC and LPS. Though pIC induced a time-dependent increase in IRF3 phosphorylation in non-irradiated cells, the response was partially abrogated in UVirradiated cells (Figure 6B). IRF3 phosphorylation was not observed at any time point following LPS administration, either in the absence or presence of UV (Figure 6C). These results lead us to conclude that the effect of UV on TBK1-IRF3 signaling is specific to the STING-dependent pathway that responds to cytosolic DNA and cyclic dinucleotides. UV modulates LKB1-AMPKα-ULK1 signaling and impacts ULK1 protein levels We next considered that UV may affect a pathway that directly controls STING activity. Interestingly, STING was recently shown to be 6 by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling negatively regulated by a signaling cascade comprising the kinases LKB1-AMPKα-ULK1 (liver kinase B1, AMP-activated protein kinase, and Unc51-like kinase 1, respectively), which ultimately results in the phosphorylation of STING by ULK1 (32) and which interferes with the activation of IRF3. Thus, the activation of ULK1 through LKB1 and AMPK is thought to dampen or turn off TBK1-IRF3 signaling in response to cytosolic DNA and cyclic dinucleotides (32). We therefore next examined the status of the LKB1-AMPKα-ULK1 signaling cascade following ISD transfection in the absence and presence of UV irradiation by monitoring the phosphorylation status of the pathway kinases. Consistent with a previous report (32), we observed that ISD induced a modest reduction in AMPKα phosphorylation under our conditions (Figure 7A, lanes 1-4). In contrast, the phosphorylation status of AMPKα increased in UV-irradiated cells regardless of ISD transfection (Figure 7A, lanes 5-12). Moreover, UV also induced a dramatic reduction in the phosphorylation status of LKB1. These results indicate that UV alters LKB1-AMPKα-ULK1 signaling and may prevent it from responding normally to ISD (Figure 7A, lanes 5-12). However, we also noted a dramatic loss in total ULK1 protein levels in UV-irradiated cells (Figure 7A), such that ULK1 protein levels decreased by nearly 80% within 6 h following UV irradiation (Figure 7B). The loss of ULK1 following UV would be expected to prevent ULK1 from phosphorylating and negatively regulating STING. Thus, loss of this negative regulator should result in increased phosphorylation of IRF3 by TBK1. To determine whether UV potentiation of IRF3 activation in response to cytosolic DNA is dependent on ULK1, we next used RNA interference to reduce ULK1 expression in THP-1 cells. Consistent with a recent report (32), knockdown of ULK1 led to elevated IRF3 phosphorylation following ISD transfection (Figure 7C, compare lanes 2 and 6). Furthermore, though UV stimulated the effect of cytosolic DNA on IRF3 phosphorylation in cells transfected with a control siRNA (Figure 7C, lanes 2 and 3), UV did not affect the response in cells depleted of ULK1 (Figure 7C, lanes 6 and 7). Quantitation of this experimental approach is provided in Figure 7D and shows that ISDinduced IRF3 phosphorylation administration is elevated in both the absence and presence of UV when ULK1 is depleted from the cells. To ensure that the signaling pathway leading to ISD-induced IRF3 activation is not simply saturated under our experimental conditions in ULK1-depleted cells, we transfected varying amounts of ISD into cells depleted of ULK1 via RNA interference and then monitored TBK1 and IRF3 phosphorylation. As shown in Figure 7E, UV failed to enhance the effect of ISD on IRF3 activation at each concentration of ISD in cells transfected with ULK1 siRNAs. These results demonstrate that UV potentiation of IRF3 activation by cytosolic DNA only takes place when ULK1 is present in cells. UV-induced apoptotic signaling destabilizes ULK1 and AMBRA1 We next sought to determine the mechanism by which ULK1 protein levels decrease upon UV exposure. Treatment with cycloheximide to block new protein synthesis showed that ULK1 protein has a relatively long half-life in THP-1 cells (Figure 8A), which indicates that the loss of ULK1 protein upon UV irradiation may be through an active posttranslational or proteolytic process and not likely not due to simple inhibition of ULK1 transcription or translation. ULK1 protein stability and kinase activity have been shown to be influenced by the autophagy regulatory protein AMBRA1 (autophagy/beclin-1 regulator 1) (48). Interestingly, we observed that AMBRA1 protein levels decreased following UV irradiation (Figure 8B). Thus, the decrease in ULK1 protein levels and increase in STING activity may be associated with the UV-mediated loss of AMBRA1. AMBRA1 has been reported to be degraded by a caspaseand calpain-mediated pathway in response to apoptotic stimuli (49-51), and we observed that the loss of ULK1 and AMBRA1 following UV irradiation coincided with the activation of apoptotic signaling, as measured by cleavage of poly-ADP ribose polymerase (PARP) and Caspase 3 (Figure 8B). This correlation between activation of apoptotic signaling and loss of ULK1 and AMBRA1 was similarly observed upon treatment of THP-1 cells with the UV-mimetic AAF (Figure 8C). 7 by gest on N ovem er 8, 2017 hp://w w w .jb.org/ D ow nladed from UV light potentiates STING-dependent innate immune signaling We therefore next examined how inhibition of apoptotic signaling affected the levels of ULK1 and AMBRA1 in UV-irradiated cells. As shown in Figure 8D, inhibition of apoptotic signaling prevented the loss of ULK1 and AMBRA1 protein following UV irradiation. To determine whether apoptotic signaling and the subsequent loss of ULK1 and the ULK1 regulator AMBRA1 were responsible for the UVmediated enhancement of IRF3 phosphorylation following ISD transfection, we treated cells with caspase and calpain inhibitors prior to UV irradiation and transfection with ISD. As shown in Figure 8E, UV was unable to stimulate IRF3 phosphorylation when apoptotic signaling was inhibited. These results indicate that the activation of apoptotic signaling and the resulting loss of the negative STING regulator ULK1 are responsible for the stimulation of the cytosolic DNA response by UV radiation. The ultimate impact of this deregulation by UV is a potentiation of the STINGIRF3-dependent innate immune response to ISD and cyclic dinucleotides. DISCUSSION In this study, we uncovered a novel mechanism by which UV radiation and related chemical carcinogens specifically enhance the STING-dependent innate immune response that activates the transcription factor IRF3 following exposure to cytosolic DNA and cyclic dinucleotides. A schematic summarizing our findings is presented in Figure 9. Cytosolic DNA or cyclic dinucleotides that arise in cells following viral or microbial infection bind to and stimulate STING, which allows STING to function as an adaptor for phosphorylation and activation of IRF3 by TBK1. These same stimuli also activate a negative regulatory pathway that leads ULK1 to phosphorylate and inactivate STING to dampen the immune response (32). Importantly, we discovered that high levels of UV radiation, which saturate DNA repair capacity (Figure 1A), disrupts this negative regulation by promoting the destruction of ULK1 and the ULK1 regulator AMBRA1 in a caspase/calpain-dependent manner (Figure 8). The result of this deregulation is an increase in STING-IRF3 activation following UV irradiation. Because IRF3 regulates the expression of a wide variety of genes, including those involved in immune signaling, inflammation, and apoptosis, future work will need to characterize how UV radiation modulates the transcription program that is controlled by IRF3. Our observation that UV impacts an innate immune response in human keratinocytes has implications for understanding how sunlight UV exposures trigger or exacerbate disease symptoms in patients with SLE. This issue is particularly relevant in light of reported cases of SLE patients who have experienced serious skin flare-ups of cutaneous lupus and even life-threatening lupus nephritis following sunburn or tanning (52, 53). Furthermore, our results show that UV exposure modulates a canonical response to pathogenic immune stimuli (cytosolic DNA and cyclic dinucleotides) but that UV alone does not apparently induce the STING-dependent innate immune pathway. Experimental models of autoimmunity should therefore consider the possibility that these two risk factors (infection and sun exposure) work in concert to impact the immune system.