ENPP1’s regulation of extracellular cGAMP is a ubiquitous mechanism of attenuating STING signaling

Significance The immune system strikes a careful balance between launching a robust response to threats and avoiding overactivation. The molecule cGAMP is an immunotransmitter that activates innate immunity and signals extracellularly, where it is subject to degradation by the enzyme ENPP1. Here, we engineer ENPP1 to lose activity toward cGAMP but not other substrates, thus creating a biochemically precise tool to understand how ENPP1 regulates extracellular cGAMP and thus innate immunity. We uncover that ENPP1's degradation of extracellular cGAMP has a long evolutionary history, and that this mechanism is critical for controlling diverse immune threats, including viral infection and inflammation.

Kinetic analysis and expression of ENPP1 guanosine-binding pocket mutations. Supplementary Fig. 2 Kinetic analysis and expression of ENPP1 zinc binder mutations. Supplementary Fig. 3 Kinetic analysis and expression of ENPP1 H362X mutations. Supplementary Fig. 4 Bacterial NPP selectively cleaves 2'-5' linkages in cyclic dinucleotides using a conserved histidine. Supplementary Fig. 5 Enpp1 H362A mice do not exhibit the severe systemic calcification seen in ENPP1-null humans and mice. Supplementary Fig. 6 Enhanced extracellular cGAMP signaling confers resistance to HSV-1 infection. Supplementary Fig. 7 HSV-1 infection of Enpp1 asj and Enpp1 H362A mice. Supplementary Fig. 8 Enhanced extracellular cGAMP exacerbates radiation-induced inflammation.   Fig. 1 Kinetic analysis and expression of ENPP1 guanosine-binding pocket mutations. a Schematic illustration of mutations (red X's) of the nucleophile T238 (ball and stick), leading to no substrate hydrolysis. b Expression of ENPP1 WT and ENPP1 T238A mutation, representative of 3 independent experiments. c TLC cGAMP degradation assay for ENPP1 mutations expressed in cell lysate. Representative data shown from 3 independent reactions. d Luciferase-based ATP degradation assay for ENPP1 mutations expressed in cell lysate. n = 3 independent reactions, mean ± SD. e Schematic illustration of mutations (red X's) of the nucleotide-binding site (blue), leading to no substrate hydrolysis. f Mouse ENPP1 (gray cartoon) in complex with pApG (gray ball and sticks) (PDB:6aek). Guanosine-adjacent residues are colored salmon. g Expression of ENPP1 guanosine-binding pocket mutations, representative of 2 independent experiments. h Heat map showing the initial velocity relative to WT of guanosine-binding pocket mutations for the substrates cGAMP and ATP at pH 7.5. i-n cGAMP degradation kinetics of ENPP1 guanosine-binding pocket mutations at pH 9 (i-k) and pH 7.5 (l-n). Cell lines were transfected with the indicated ENPP1 mutations. Lines in (i) and (l) represent linear fits of kinetic data during the linear portion of the reaction. Time = 0 to 1 h for (i) and time = 0 to 6 h for (l). Data is representative of 2 independent experiments. These linear fits were plotted in bar graphs in (k) and (n) showing the best fit slope ± SD of the fit and were used to make heat maps shown in Fig. 1d and (h). TLCs in (j) and (m) depict reaction progress after 24 hours. o-r ATP degradation kinetics of ENPP1 guanosine-binding pocket mutations at pH 9 (o-p) and pH 7.5 (q-r). Cell lines were transfected with the indicated ENPP1 mutations. Lines in (o) and (q) represent linear fits of kinetic data during the linear portion of the reaction. Time = 0 to 20 min for (o) and time = 0 to 90 min for (q). n = 4 independent reactions, mean ± SD shown with some error bars too small to visualize. These linear fits were plotted in bar graphs in (p) and (r) showing the best fit slope ± SD of the fit and were used to make heat maps shown in Fig. 1d and (h).

Supplementary Fig. 2 Kinetic analysis and expression of ENPP1 zinc binder mutations. a
Expression of ENPP1 zinc-binding residue mutations, representative of 2 independent experiments. b Heat map showing the initial velocity relative to WT of zinc-binding residue mutations for the substrates cGAMP and ATP at pH 7.5. The mean initial velocity was calculated from a linear fit of the degradation reactions during early time points. n = 3 independent reactions. c-f cGAMP degradation kinetics of ENPP1 zinc-binding residue mutations at pH 9 (c-d) and pH 7.5 (e-f). Cell lines were transfected with the indicated ENPP1 mutations. Lines in (c) and (e) represent linear fits of kinetic data during the linear portion of the reaction. Time = 0 to 45 min for (c) and time = 0 to 6 h for (e). Data is representative of 2 independent experiments. These linear fits were plotted in bar graphs in (d) and (f) showing the best fit slope ± SD of the fit and were used to make heat maps shown in Fig. 1g and (b). g-j ATP degradation kinetics of ENPP1 zinc-binding residue mutations at pH 9 (g-h) and pH 7.5 (ij). Cell lines were transfected with the indicated ENPP1 mutations. Lines in (g) and (i) represent linear fits of kinetic data during the linear portion of the reaction (time = 0 to 20 min). n = 3 independent reactions, mean ± SD with some error bars too small to visualize. These linear fits were plotted in bar graphs in (h) and (j) showing the best fit slope ± SD of the fit and were used to make heat maps shown in Fig. 1g and (b).

Supplementary Fig. 3 Kinetic analysis and expression of ENPP1 H362X mutations. a
Expression of ENPP1 H362X mutations (where X represents the indicated amino acid), representative of 2 independent experiments. b-c cGAMP degradation kinetics of ENPP1 H362X mutations at pH 9. Cell lines were transfected with the indicated ENPP1 mutations. b TLC showing cGAMP degradation after 24 hours at pH 9. c Linear fits of kinetic data during the linear portion of the reaction (time = 0 to 30 minutes). d ATP degradation kinetics of ENPP1 H362X mutations at pH 9, organized by amino acid class. Cell lines were transfected with the indicated ENPP1 mutations. Lines represent linear fits of kinetic data during the linear portion of the reaction (time = 0 to 15 min). n = 3 independent reactions, mean ± SD with some error bars too small to visualize. e ATP activity rates for ENPP1 H362X mutations plotted from linear fits shown in (d). n = 3 independent reactions, best fit slope ± SD of the fit from (d). f TLC of cGAMP degradation comparing recombinant ENPP1 WT to ENPP1 H362A . n = 3 independent reactions. g Kinetic analysis of GTP, UTP, and CTP monitoring pyrophosphate production using recombinant purified ENPP1. n = 2 independent reactions, mean ± SD. Fig. 4 Bacterial NPP selectively cleaves 2'-5' linkages in cyclic dinucleotides using the conserved histidine. a 998 eukaryotic, 1000 bacterial, and 584 archaeal NPP protein sequences were downloaded from Uniprot and pairwise aligned using MUSCLE alignment. The histidine corresponding to H362 in mouse ENPP1 was identified and the percent conservation was determined for Eukaryota, Bacteria, and Archaea. b Signal peptides (red boxes) or lipoprotein signal peptides (blue boxes) for selected species predicted by SignalP5.0. No signal peptide was found for B. cereus. c TLCs showing the degradation of 3'3'-cGAMP (top) and 2'3'-cGAMP (bottom) by 1.5 µM Xac NPP WT over the indicated times.  Fig. 5. Enpp1 H362A mice do not exhibit the severe systemic calcification seen in ENPP1-null humans and mice. a A single-guide RNA (sgRNA) was designed to target the region near H362A in exon 9 of Enpp1. The sgRNA and Cas9 were introduced into 4T1 cells through lentiviral transduction. The cells were then sequenced to determine editing efficiency at the intended cleavage site. There was no evidence of editing in cells transduced with a control sgRNA, while there was a large deletion in cells transduced with the Enpp1 sgRNA. b Diagram indicating the sequence of the donor ssDNA used to generate the H362A point mutation through homologous recombination. Silent mutations were introduced downstream of H362 to prevent Cas9 from recognizing the cleavage site following successful recombination. c Genomic DNA sequencing from one of the Enpp1 H362A mice. The sequencing indicated that this mouse harbored a homozygous H362A point mutation in Enpp1, as well as the point mutations indicated in (b). d A cGAMP ELISA was performed to measure basal cGAMP in the spleen, kidney, liver, and lung. There was no detectable cGAMP in any of the kidneys, so they were omitted from analysis. n = 3 mice per genotype. p values were calculated by unpaired t test; *p < 0.05. e RT-qPCR was performed to measure basal Ifnb1 in the spleen, kidney, liver, and lung. n = 5 mice per genotype. p values were calculated by unpaired t test. f Ex vivo liver lysate (1 mg/mL) ATP degradation at pH 9 assessed by luciferase assay after 20 minutes in Enpp1 H362A , Enpp1 asj , and Enpp1 H362A mice. n = 4 Enpp1 H362A , 4 Enpp1 asj , and 4 Enpp1 H362A mice. g Ex vivo plasma ATP degradation at pH 9 assessed by luciferase assay after 45 minutes in Enpp1 WT , Enpp1 asj , and Enpp1 H362A mice. n = 4 Enpp1 WT , 4 Enpp1 asj , and 5 Enpp1 H362A mice. Data are shown as the mean ± SD. p values were calculated by unpaired t test with Welch's correction. ****p < 0.0001. h-i In vitro ATP degradation comparing overexpressed ENPP1 WT and ENPP1 H362A as cell-surface proteins from cell lysate (h) and as secreted protein from cell supernatant (i) at pH 9. Fig. 6 Enhanced extracellular cGAMP signaling confers resistance to HSV-1. a-c Mice were inoculated with 2.5 x 10 7 PFU/mouse HSV-1 through intravenous injection. The mice were euthanized at 6 or 12 hpi and organs were isolated for plaque assays or total RNA isolation. n = 6 (6 hpi) or 8 (12 hpi) infected mice per genotype (as previously described in Fig. 5). a Plaque assays of spleen and kidney lysates from Enpp1 WT and Enpp1 H362A mice. b-c RT-qPCR was performed to determine the expression levels of Cxcl10 and Tnfa in indicated organs of Enpp1 WT and Enpp1 H362A mice. Cytokine transcript levels were normalized to the average of 2 uninfected controls per genotype. d-h Mice were inoculated with 2.5 x 10 7 PFU/mouse HSV-1 through intravenous injection. The mice were euthanized at 6 days post infection (dpi) and organs were isolated for total RNA isolation. n = 6 infected Enpp1 WT mice and n = 5 ENPP1 H362A mice. One lung ENPP1 H362A sample was excluded from d-h as an outlier based on the ROUT method (Q = 1%). n.d. = not detected. Data are shown as the mean ± SD. p values were calculated using the non-parametric Mann-Whitney test. *p < 0.05, **p < 0.01, ***p < 0.001; p value is shown if between 0.05 and 0.1.

Supplementary Fig. 7 HSV-1 infection of Enpp1 asj and Enpp1 H362A mice. a-f
Mice were inoculated with 2.5 x 10 7 PFU/mouse HSV-1 through intravenous injection. The mice were euthanized at 12 hpi and organs were isolated for plaque assays or total RNA isolation. n = 8 infected Enpp1 WT mice (replotted from Fig. 5) and n = 10 infected ENPP1 asj mice. RT-qPCR was performed to measure expression of HSV-gB (a), Ifnb1 (b), Il6 (c), Cxcl10 (d), and Tnfa (e). Cytokine transcript levels were normalized to the average of 2 uninfected controls per genotype. f Plaque assays of spleen and kidney lysates from Enpp1 WT and Enpp1 asj mice. g Plasma ATP was measured from uninfected mice or mice infected with 2.5 x 10 7 PFU/mouse HSV-1 through intravenous injection. Plasma was collected from infected mice at 6 and 12 hpi. h Enpp1 H362A mice were injected with 2.5 x 10 7 PFU/mouse HSV-1 and euthanized at the indicated time points. Plasma was collected from each mouse, and cGAMP concentration was determined by cGAMP ELISA. A cGAMP standard curve was created in 50% mouse plasma (left). None of the infected plasma samples (right) gave readings above the published limit of detection (85 pg/mL), suggesting the absence of any cGAMP. For RT-qPCR data, transcript levels were normalized to the average of 2 uninfected mice per genotype. Data are shown as the mean ±  Fig. 7, separated into mice receiving 8 Gy (a) and 9 Gy (b). n = 6 mice per genotype (a) or 5 mice per genotype (b). p value was calculated using a log rank (Mantel-Cox) test.

Synthesis and purification of cGAMP and [ 32 P]-cGAMP
To enzymatically synthesize cGAMP, 1 μM purified sscGAS was incubated with 20 mM Tris-HCl pH 7.4, 2 mM ATP, 2 mM GTP, 20 mM MgCl2, and 100 μg/mL herring testis DNA (Sigma) for 24 h. The reaction was then heated at 95°C for 3 min and filtered through a 3-kDa filter. cGAMP was purified from the reaction mixture using a PLRP-S polymeric reversed phase preparatory column

Enzyme activity assays
a) cGAMP activity assays (20 L total) were composed of the following: cell/organ lysate (50%) or recombinant ENPP1 (1 -10 nM), cGAMP (1 to 5 M, with trace [ 32 P]-cGAMP spiked in), and buffer (standard assay buffer unless otherwise noted was 100 mM Tris pH 9 or pH 7.5, 150 mM NaCl, 500 M CaCl2, 10 M ZnCl2). At indicated times, 1 L aliquots of the reaction were quenched by spotting on HP-TLC silica gel plates (Millipore). The TLC plates were run in mobile phase (85% ethanol, 5 mM NH4HCO3) and exposed to a phosphor screen (GE BAS-IP MS). Screens were imaged on a Typhoon 9400 scanner. b) ATP activity assays (10 L total in a 384 well PCR plate) were composed of the following: cell/organ lysate (0.1-1%, depending on ENPP1 expression level) or recombinant ENPP1 (1 -3 nM), 1 M ATP (Sigma) and buffer (standard assay buffer unless otherwise noted was 100 mM Tris pH 9 or pH 7.6, 150 mM NaCl, 500 M CaCl2, 10 M ZnCl2). Reactions were started at indicated times and ended simultaneously by heating at 95 C for 10 minutes. Reactions (5 L) were transferred to a white 384 well plate, mixed with CellTiterGlo (5 L), and luminescence was read after 15 minutes on a Tecan Spark plate reader. c) GTP, CTP, and UTP activity assays were monitored by coupling pyrophosphate production to ATP production. To convert pyrophosphate into ATP, 5 μL of each plasma sample was added to a reaction mixture consisting of 16 μM adenosine phosphosulfate, 80 mM MgSO4, 50 mM HEPES, and 0.1 μL ATP sulfurylase (MCLab).
The reaction mixture was then incubated at 37°C for 10 min followed by 90°C for 10 min to inactivate the enzyme. In order to measure ATP, 25 μL of the reaction mixture was added to 25 μL of CellTiter-Glo (Promega). Luminescence was measured after 10 min using a 0.5 s integration time. For measurement of kcat/Km for recombinant ENPP1 with the substrate ATP, the commercial AMPGlo kit (Promega) was used according to the manufacturer's instructions. ZnCl2 (40 μL total). Reaction progress was measured by coupling the reaction to alkaline phosphatase (1 U/μL of FastAP (ThermoFisher) per 40 μL reaction) and then measuring phosphate production with malachite green (Millipore Sigma).

Recombinant mouse ENPP1 purification
Procedures for culturing and transfecting Expi293F cells (Thermo Fisher) were based on the manufacturer's instructions. One day prior to transfection, the cells were split to 310 6

Preparing ENPP1-transfected cell lysates
Plasmids containing Flag-ENPP1 WT or ENPP1 mutations were transfected into 293T ENPP1 -/cells with polyethylenimine (PEI) at ratio of 1 g plasmid to 3 g PEI per well of a 12 well plate.
After 24 hours, cells were lysed for western blotting and activity assays. For western blotting, cells were lysed on the plate in 150 L of Laemmli sample buffer. For activity assays, cells were washed off the plate in 1 mL of PBS, centrifuged at 1000 x g for 10 minutes, and lysed in 100 L of lysis buffer (10 mM Tris pH 9, 150 mM NaCl, 10 M ZnCl2, 1% NP-40). Lysates were stored at -20 C. All diffracted Xac NPP T90A crystals were in the P212121 orthorhombic space group as has been reported previously (2). Regardless of cGAMP co-crystallization and/or varying soaking regimes, these orthorhombic crystals were bound to AMP rather than cGAMP or GMP. Supplementary Table 1 presents data collection, refinement, and structure quality check parameters. Supplementary Table 2 presents ligand-protein interactions. Data was reduced with XDS (6), scaled with SCALA (7), and analyzed with different computing modules within the CCP4 suite (8). Graphic renderings were prepared with PyMOL (9). The final refined structure shows an excellent agreement with reference protein data as shown by Ramachandran statistics as surveyed with Molprobity (10). The structure has been validated and deposited with the RCSB Protein Data Bank (11) with codes 7MW8 (Xac NPP T90A with pApG) and 7N1S (Xac NPP H214A apo).

Computational modeling of substrates
A monomer of wild-type cGAMP-bound mENPP1 (PDB: 6AEK) was isolated and mutated on Pymol to generate mENPP1 H362A . The molecular coordinates of adenosine, including the alpha phosphate, were isolated together with Zn1009 and N-epsilon of H362, which was replaced by a gamma phosphate oxygen to constitute an ATP structure with the free oxygens of both beta and gamma phosphates bonded to Zn resembling an approximately octahedral geometry, similar to previously proposed coordination between ATP and another 2+ ion, Mg 2+ (12). The ligand was reassigned and used to replace cGAMP and Zn1009 on the mENPP1 H362A structure on Pymol.
The composite structure was then minimized in Schrödinger Maestro using standard protein preparation workflow using all default settings at pH 7.4, with missing side chains filled against the corresponding FASTA sequence and using the OPLS3e force field. Similar minimization procedures were performed on Xac NPP H214A , derived from the wild-type, pApG-bound structure (PDB: 7MW8, this paper). The bound pApG was used to prepare a receptor grid (101010 Å 3 ), which was used to dock (using Glide at standard precision) all possible conformations of 2'3'-cGAMP and 3'3'-cGAMP, generated at pH 7.4 using LigPrep.  Table 3). The sgRNA was then complexed with Alt-R S.p. Cas9 nuclease (Integrated DNA Technologies) as a ribonucleoprotein (RNP) particle. We then designed a donor sequence based on mouse Enpp1 to serve as the template for homologous recombination (Supplementary Table 3). The 100 nucleotide-long donor sequence contained blocking mutations near the PAM sequence to prevent repeated editing (14).

Generation and characterization of the transgenic
The donor sequence was then synthesized as single-stranded DNA (Integrated DNA Technologies). The RNP particles and donor template were microinjected into the pronuclei of one-cell embryos from C57BL/6 mice, which were then implanted into pseudopregnant mice. As the initial litter of mice likely consisted of chimeras, the F1 generation was crossed with each other to generate a non-chimeric F2 generation. The

c) Histology sectioning and staining.
Organs were harvested at 20 weeks of age and fixed in 4% buffered Formaldehyde solution (pH 6.9) for 72 hours before transfer into 70% ethanol. Samples were submitted to Stanford Animal Histology Services for paraffin embedding, cutting and Alizarin Red staining. Imaging was done on a Zeiss AxioImager microscope in the Stanford Cell Sciences Imaging Facility. d) Plasma chemistry. Blood was collected through terminal cardiac puncture into heparin-coated microtainers (BD). The blood was then spun at 2,000 x g for 15 min and the resulting plasma layer was collected. Plasma phosphate was measured using a malachite green phosphate assay kit (Sigma-Aldrich) according to the manufacturer's instructions; each sample was diluted 1:250 in water. Plasma calcium was measured using a commercial colorimetric assay (Stanbio) according to the manufacturer's instructions; each sample was diluted 1:4 in water. Plasma pyrophosphate was measured using a previously published method (15). To convert pyrophosphate into ATP, 5 μL of each plasma sample was added to a reaction mixture consisting of 16 μM adenosine phosphosulfate, 80 mM MgSO4, 50 mM HEPEs, and 0.5 μL ATP sulfurylase (MCLab). The reaction mixture was then incubated at 37°C for 10 min, followed by 90°C for 10 min to inactivate the enzyme. In order to measure ATP, 25 μL of the reaction mixture was added to 25 μL of CellTiter-Glo (Promega). Luminescence was measured after 10 min using a 0.5 s integration time. e) In vivo cGAMP metabolism. Mice were injected subcutaneously with 5 mg/kg cGAMP diluted in 100 μL PBS. After 30 min, the mice were anesthetized with isoflurane and 50 μL of blood was collected retro-orbitally and immediately supplemented with ~20 M of ENPP1 inhibitor STF-1623 to prevent cGAMP degradation. The blood was placed in heparin-coated microtainers (BD) and spun at 2,000 x g for 15 min and the resulting plasma layer was collected. The plasma was processed for LC-MS/MS by mixing plasma (7 L) with acetonitrile containing 2 M of internal standard cyclic GMP-[ 13 C10, 15 N5]AMP (20 L), centrifuging at 16,000 x g for 15 min, and then adding 23 L of the mixture to 15 L of water containing 0.1% formic acid. cGAMP was analyzed on a Q-Exactive FT mass spectrometer (Thermo) equipped with a Vanquish UHPLC.
Samples were injected onto a Biobasic AX LC column (5 μm, 50 × 3 mm; Thermo Scientific). The mobile phase consisted of 100 mM ammonium carbonate (A) and 0.1% formic acid in acetonitrile (B). The initial condition was 90% B, maintained for 0.5 min.
The mobile phase was ramped to 30% A from 0.5 min to 2.0 min, maintained at 30% A from 2.0 min to 3.5 min, ramped to 90% B from 3.5 min to 3.6 min, and maintained at 90% B from 3.6 min to 5 min. The flow rate was set to 0.6 mL min -1 . Quantification was performed with TraceFinder 4.1 software (Thermo Fisher).

HSV-1 purification
The HSV-1 KOS strain was purchased from ATCC. The day prior to infection, Vero cells were plated in five T175 tissue culture flasks (Corning) at a density of 8 x 10 6 cells/flask so they would be 80-100% confluent on the day of infection. Cells were infected with HSV-1 at MOI 0.01 in 5 mL/flask of serum-free DMEM for 1 hour with gentle rocking every 15 minutes. Media was collected 48 hours post infection when the cells displayed >90% CPE and centrifuged at 600 x g for 10 min to pellet debris. Clarified media was then centrifuged at 48,000 x g for 30 min to pellet virus. The pelleted virus was gently washed with PBS and resuspended in 2 mL of PBS, aliquoted, snap frozen, and stored at -80 °C until further use. Plaque assay was performed to determine titer (usually ~1 x 10 9 pfu/mL).

Mouse BMDM isolation and infection
The bone marrow from mouse hind limb femurs and tibias was flushed by removing the end cap of the bones and centrifuging (16). Red blood cells were lysed by resuspending the pellet in red cell lysis buffer (155 mM NH4Cl, 12 mM NaHCO3, and 0.1 mM EDTA) and incubating for 5 min at room temperature. The cells were diluted in PBS, pelleted by centrifuge, and plated on one 10-cm dish per mouse in DMEM supplemented with 10% FBS, 1% P/S, and 10% conditioned L929 media. Cells were washed after 24 hours and allowed to differentiate for 5-7 days. To prepare for infection assay, BMDMs were plated at 80-90% confluence in 12-well plates. They were infected with various MOI of HSV-1 in 100 μL for 1 hour with periodic shaking. Viruscontaining media was aspirated and replaced with fresh media.

RT-qPCR
Total RNA was isolated from cells and tissues using TRIzol (Invitrogen) by following the manufacturer's protocol. Tissue samples were homogenized in TRIzol prior to RNA isolation. To obtain cDNA, 20 uL reverse transcriptase (RT) reactions were set up containing 1 μg total RNA, 100 pmol random hexamer primers, 0.5 mM dNTPs, 20 U RNaseOUT, 1x Maxima RT buffer, and 200 U Maxima RT (Thermo Scientific). RT reactions were incubated for 10 min at 25 °C, 15 min at 50 °C, then 5 min at 85 °C. To measure transcript levels, 10 μL qPCR reactions were set up containing 0.7 μL cDNA, 100 nM qPCR primers (Supplementary Table 3), and 1x AccuPower GreenStar master mix (Bioneer) or 1x PowerTrack SYBR Green master mix (Thermo Scientific).
Reactions were run on a ViiA 7 Real-Time PCR System (Applied Biosystems) using the following program: ramp up to 50°C (1.6°C/s) and incubate for 2 min, ramp up to 95°C (1.6°C/s) and incubate for 10 min, then 40 cycles of the following: ramp up to 95°C (1.6°C/s) and incubate for 15 s, then ramp down to 60°C (1.6°C/s) and incubate for 1 min. Transcript levels for each gene were normalized to Actb transcript levels.
14. cGAMP ELISA for plasma cGAMP measurement 2.5 x 10 7 PFU of HSV-1 was diluted in 100 L PBS and injected intravenously into the tail vein of each mouse. After the indicated timepoints, the mice were euthanized in a CO2 chamber and the blood was collected through cardiac puncture into heparin-coated microtainers (BD). The blood was then spun at 2,000 x g for 15 min and the resulting plasma layer was collected. A commercial cGAMP ELISA (Cayman Chemical) was used to determine the cGAMP concentration in each sample following the manufacturer's specifications. Each sample was diluted 1:2 in the provided buffer and the standard curve was generated in buffer mixed with 50% mouse plasma from uninfected Enpp1 H362A mice.

Total body irradiation mouse model
Male and female 8-12-week-old mice were irradiated with either 8 or 9 Gy using a 225 kVp cabinet X-ray irradiator with a 0.5 mm Cu filter (IC-250, Kimtron Inc.). Mice were anesthetized with a mixture of 80 mg/kg ketamine (VetaKet) and 5 mg/kg xylazine (AnaSed) prior to irradiation. The mice were weighed daily and were euthanized if they met the humane endpoint of greater than 20% weight loss for two consecutive days. 50 µL of blood was withdrawn retroorbitally 5 days after irradiation for IFN-β ELISA analysis. The blood was spun at 2,000 x g for 15 min and the resulting plasma layer was collected. A commercial high-sensitivity IFN-β ELISA kit (PBL Assay Science) was used to determine the IFN-β concentration in each sample following the manufacturer's specifications. Each sample was diluted 1:10 in the provided buffer and the standard curve was generated in buffer mixed with 10% mouse plasma from healthy mice. Finally, spleens were harvested at endpoint for RT-qPCR analysis. PDB code 7MW8 7N1S a Ratio of the volume of the asymmetric unit to the molecular weight of all protein in the asymmetric unit b Value in parentheses is for the highest-resolution shell: 1.90 -2.00 Å. c Reliability factor for symmetry-related reflections calculated as: Rmerge = Σhkl Σj=1 to N | Ihkl -Ihkl (j) | / Σhkl Σj=1 to N Ihkl (j), where N is the redundancy of the data. In parentheses, the cumulative value at the highest-resolution shell d Ratio of mean intensity to the mean standard deviation of the intensity over the entire resolution range e Fraction of measured reflections to possible observations at the resolution range f Number of measurements of individual, symmetry unique reflections g Average deviation between the observed and calculated structure factors calculated as: Rwork = Σhkl ||Fobs| -|Fcalc|| / Σhkl |Fobs|, where the Fobs and Fcalc are the observed and calculated structure factor amplitudes of reflection hkl. Rfree is equal to Rfactor but for a randomly selected 5.0 % subset of the total reflections that were held aside throughout refinement for cross-validation h Correlation coefficient between observed and calculated structure factor amplitudes i According to Molprobity for non-proline and non-glycine residues