RIG-I-dependent sensing of poly(dA:dT) through the induction of an RNA polymerase III-transcribed RNA intermediate

Abstract

RNA is sensed by Toll-like receptor 7 (TLR7) and TLR8 or by the RNA helicases LGP2, Mda5 and RIG-I to trigger antiviral responses. Much less is known about sensors for DNA. Here we identify a novel DNA-sensing pathway involving RNA polymerase III and RIG-I. In this pathway, AT-rich double-stranded DNA (dsDNA) served as a template for RNA polymerase III and was transcribed into double-stranded RNA (dsRNA) containing a 5'-triphosphate moiety. Activation of RIG-I by this dsRNA induced production of type I interferon and activation of the transcription factor NF-kappaB. This pathway was important in the sensing of Epstein-Barr virus-encoded small RNAs, which were transcribed by RNA polymerase III and then triggered RIG-I activation. Thus, RNA polymerase III and RIG-I are pivotal in sensing viral DNA.

Figure 1. Poly(dA-dT) triggers type I IFN induction via RIG-I in human cells

a, 293T cells were reverse transfected with three individual siRNAs targeting the indicated genes. 48h after transfection cells were transfected with an pIFN-β reporter plasmid in conjunction with poly(dA-dT) or SEV. After an additional period of 24h transactivation of the IFN-β promoter was measured. Three individual siRNAs were used per gene and each was tested in triplicate. Data are represented as mean values normalized to the set of control siRNAs. b, MoDCs were electroporated with siRNA directed against RIG-I or control siRNA. 48h after electroporation, cells were stimulated with the indicated stimuli and IFN-α or TNF-α production was assessed after 24h by ELISA. c, Murine bone marrow-derived dendritic cells (conventional dendritic cells, cDCs) were electroporated with siRNA targeting RIG-I or a control siRNA. 48h after electroporation cells were stimulated and IFN-β induction was assessed 5h after stimulation by real-time PCR. Mean values ± SEM of one representative experiment out of two (a) or three (b, c) experiments are depicted.

Figure 2. Poly(dA-dT) triggers the formation of an endogenous RIG-I stimulatory RNA

a, Human PBMCs were transfected with 3pRNA or poly(dA-dT) and IFN-α production was assessed at the time points indicated. b and c, RNA isolated from 293T cells stimulated with SEV or poly(dA-dT) was transfected into chloroquine-treated PBMCs. In addition, PBMCs were stimulated directly with 3pRNA, poly(dA-dT) or SEV. 24h after stimulation IFN-α production was assessed. d, RNA was isolated from 293T cells 0, 4, 9, 12 and 24h after poly(dA-dT) transfection. The obtained RNA was then used to stimulate chloroquine-blocked PBMCs, whereas directly transfected poly(dA-dT) served as a control. e, RNA from poly(dA-dT) transfected monocytes and L929 cells was purified and used to stimulate PBMCs as described above. f, Human MoDCs or murine cDCs were electroporated with siRNA targeting RIG-I or a control siRNA and after 48h cells were stimulated as indicated. IFN-α production or IFN-β induction was assessed 24h or 5h later as described. Mean values ± SEM of one representative experiment out of two (a, d, e), three (f) or four (c) experiments are depicted.

Figure 3. Non poly(dA-dT) double stranded DNA triggers type I IFN induction via an RNA independent pathway

a and b, double stranded DNAs of different length derived from pcDNA3 by PCR were transfected into chloroquine-blocked PBMCs (200ng/transfection in 96 well) or 293T cells and type I IFN induction was measured 24h after stimulation by ELISA or via reporter gene activity. c, RNA from 293T cells that had been transfected with the above dsDNAs was isolated and transfected into chloroquine-blocked PBMCs. IFN-α production was measured 24h after transfection. Mean values ± SEM of one representative experiment out of four (a, c) or three (b) experiments are depicted.

Figure 4. Poly(dA-dT) triggers the formation of a small stimulatory RNA species independent of the OAS/RNase L pathway

a, RNA was isolated from poly(dA-dT) transfected 293T cells and fractionated using different concentrations of ethanol during the initial binding step of the RNA in a silica matrix based spin column purification system. The obtained RNA was analyzed on an agilent 6000 pico chip and tested for IFN-α production in chloroquine-treated PBMCs. Direct stimulation with 3pRNA or poly(dA-dT) served as a control. b, 293T cells were reverse transfected with three individual siRNAs targeting the indicated genes. 48h after transfection cells were transfected with a pIFN-β reporter plasmid in conjunction with poly(dA-dT) or SEV. After an additional period of 24h transactivation of the IFN-β promoter was measured. Three individual siRNAs were used per gene and were each tested in triplicate. Data are represented as mean values normalized to the set of control siRNAs. c, RNA was isolated from murine embryonic fibroblasts transfected with poly(dA-dT) and tested in chloroquine-treated PBMCs for IFN-α production. Direct transfection of poly(dA-dT) served as a control. Mean values ± SEM of one representative experiment out of four (a) or two (b, c) experiments are depicted.

Figure 5. Poly(dA-dT)-induced RNA is a 5′ triphosphate, double stranded RNA species that is devoid of guanosine

RNA from poly(dA-dT) transfected 293T cells was isolated and treated with RNA 5′ polyphosphatase (a), RNase III (b), RNase T1 under denaturing conditions (c) or RNase A at low salt concentration (d). Poly(dA-dT) or 3pRNA were treated the same way. Chloroquine-treated PBMCs were then transfected with the respective nucleic acids and IFN-α production was assessed 24h later. In addition, the above nucleic acids were analyzed on an agilent small RNA chip. Mean values ± SEM of one representative experiment out of two (a, d) or four (c, b) experiments are depicted.

Figure 6. RNA polymerase III transcribes AT rich DNA and is required for poly(dA-dT)-mediated type I IFN induction

a, The synthetic dsDNA template AT+30N or AT70 was transfected into 293T cells, whereas non transfected cells were used as a control. 8 h after transfection total RNA was isolated, polyadenylated and reverse transcribed as depicted in supplementary Fig. 6. The obtained cDNA was then used for PCR using a specific upstream primer for either 5S rRNA or the 30N transcript and a downstream primer binding within the RT primer. A standard PCR was done with primer pairs for β2 microglobulin. A representative PAGE analysis is shown for a conventional PCR amplification for all three primer combinations. In addition a real-time PCR analysis is shown for the AT+30N transcript normalized to β2 microglobulin expression. b, 293T cells were transfected with (AT)35, AT+30N, AT+6T+30N dsDNA or left untreated. RNA was isolated and treated as above and 3′ RACE real time PCR was performed for the 30N tag. The 30N tag expression data were normalized to β2 microglobulin expression, c, 293T cells were treated with ML-60218 for 2h at the indicated concentrations and transfected with AT+30N dsDNA. 16h later RNA was isolated and the expression of the 30N tag normalized to β2 microglobulin expression was assessed. d, 293T cells were treated with ML-60218 for 2h at the indicated doses. Subsequently cells were transfected with an IFN-β reporter plasmid in conjunction with poly(dA-dT) or SEV. After an additional period of 24h transactivation of the IFN-β promoter was measured. In addition, 293T cells were reverse transfected with two individual siRNAs targeting the indicated genes. After 2 days cells were transfected with AT+30N and after an additional period of 16h transcription of 5S rRNA (e) and AT+30N (f) was determined via 3′ RACE PCR. g, 293T cells were reverse transfected with four individual siRNAs targeting the indicated genes. 48h after transfection cells were transfected with an IFN-β reporter plasmid in conjunction with poly(dA-dT) or SEV. After an additional period of 24h transactivation of the IFN-β promoter was measured. Four individual siRNAs were used per gene and were each tested in triplicates. Data are represented as mean values normalized to the set of control siRNAs. Mean values ± SEM of one representative experiment out of two (c, e, f), three (a, b, g) or four (d) experiments are depicted.

Figure 7. EBV encoded EBER RNAs are transcribed by RNA Pol III and trigger activation of RIG-I

a and b, Mutu III cells were treated with ascending doses of ML-60218 for 24h and IFN-α production was assessed by ELISA (a, left panel) and cell viability was assessed using Calcein AM staining (a, right panel), while EBER RNA and 5S rRNA transcription was measured by real-time PCR (b). c, The genomic locus that encodes for EBER-1 and 2 RNA, EBER-1 RNA, EBER-2 RNA was PCR-amplified and transfected into IFN-primed 293T cells (all 120ng or as indicated). PCR-amplified EGFP served as a negative control, whereas poly(dA-dT) was used as a positive control. Subsequently transactivation of the IFN-β promoter was measured after 24h. d, IFN-primed 293T cells transfected with the full length EBER RNA gene locus were treated with ascending doses of ML-60218 and after 24h EBER-1, EBER-2, 5S rRNA and Cyclophilin B transcription was determined. e, In addition, transactivation of the IFN-β promoter was assessed in IFN-primed 293T cells that had been transfected with either poly(dA-dT), the EBER RNA gene locus or RIG-I (all 120ng). Data were normalized to the positive control RIG-I. f, 293T cells were reverse transfected with two individual siRNAs targeting the indicated genes. 48h after siRNA transfection cells were transfected with an IFN-β reporter plasmid in conjunction with the EBER RNA gene locus, poly(dA-dT) (all 120ng) or SEV. After an additional period of 24h transactivation of the IFN-β promoter was measured. Two individual siRNAs were used per gene and were each tested in triplicates. Mean values ± SEM of one representative experiment out of two (b, d, f), three (a, e) or four (c) experiments are depicted.