Coronavirus endoribonuclease targets viral polyuridine sequences to evade activating host sensors

Abstract

Coronaviruses (CoVs) are positive-sense RNA viruses that can emerge from endemic reservoirs and infect zoonotically, causing significant morbidity and mortality. CoVs encode an endoribonuclease designated EndoU that facilitates evasion of host pattern recognition receptor MDA5, but the target of EndoU activity was not known. Here, we report that EndoU cleaves the 5'-polyuridines from negative-sense viral RNA, termed PUN RNA, which is the product of polyA-templated RNA synthesis. Using a virus containing an EndoU catalytic-inactive mutation, we detected a higher abundance of PUN RNA in the cytoplasm compared to wild-type-infected cells. Furthermore, we found that transfecting PUN RNA into cells stimulates a robust, MDA5-dependent interferon response, and that removal of the polyuridine extension on the RNA dampens the response. Overall, the results of this study reveal the PUN RNA to be a CoV MDA5-dependent pathogen-associated molecular pattern (PAMP). We also establish a mechanism for EndoU activity to cleave and limit the accumulation of this PAMP. Since EndoU activity is highly conserved in all CoVs, inhibiting this activity may serve as an approach for therapeutic interventions against existing and emerging CoV infections.

Keywords: EndoU; coronavirus; endoribonuclease; interferon; nsp15.

Fig. 1.

Evaluating the accumulation of an epitope recognized by K1 antibody in virus-infected AML12 hepatocytes. AML12 hepatocytes were infected with wild-type (WT) or EndoUmut MHV at an MOI of 0.1. Cells were fixed at 8 hpi and stained with K1 anti-dsRNA antibody, anti-nsp2/3, and Hoescht 33342 nuclei stain. (A) Subcellular localization of dsRNA and nsp2/3 foci was visualized. (B) Foci for (Left) dsRNA and (Right) nsp2/3 were quantified using Imaris software from 50 individual cells. (C) The median fluorescent intensity was calculated for each individual dsRNA foci and compared between WT and EndoUmut infections. Values were analyzed by Student t tests. Data are representative of three independent experiments and presented as individual cell points with mean ± SD; n.s., not significant.

Fig. 2.

RNA-seq analysis of input viral RNA and RNA immunoprecipitated with anti-dsRNA antibody K1. IFNAR^−/− BMDMs were infected with WT or EndoUmut virus at an MOI of 1. At 6 hpi, RNA was purified, mixed with anti-dsRNA antibody, precipitated with protein G beads, and purified off the beads. Input RNA and immunoprecipitated RNA samples were evaluated by RNA-seq. (A) Summary of RNA reads mapped to MHV-A59 genome. Values in tables are the means of three biological triplicates. (B–E) Total reads mapped to the viral genome. Reads were mapped to the positive-sense (+) RNA from (B) input RNA and (C) immunoprecipitated RNA. Reads were mapped to the negative-sense (−) RNA from (D) input RNA and (E) immunoprecipitated RNA. Read counts were quantified for each nucleotide of the genome, then averaged into bins of 200 nucleotides for three biological triplicates. The black (WT) and red (EndoUmut) lines represent the mean of each bin, and shaded regions are the SD from the mean. Data are representative of two independent experiments.

Fig. 3.

Quantifying viral RNA immunoprecipitated with antibody K1. IFNAR^−/− BMDMs, C57BL/6 BMDMs, and AML12 cells were infected with WT or EndoUmut virus (EUmut) at an MOI of 1. At 6 hpi, RNA was collected and processed for dsRNA immunoprecipitation (IP) with anti-dsRNA antibody or an isotype control. (A) CoV RNA immunoprecipitated using K1 Ab was quantified using primers to the nucleocapsid (N) gene by qPCR. (B) CoV RNA from input RNA was quantified by measuring the N gene expression. Viral RNA was normalized to 18s rRNA and set relative to WT. Values were analyzed by Student t tests. Data are representative of three independent experiments and presented as mean ± SD. n.s., not significant.

Fig. 4.

Quantifying PUN RNAs from virus-infected cells. IFNAR^−/− BMDMs and AML12 cells were infected with WT or EndoUmut virus at an MOI of 1, and RNA was purified from cell lysates. (A) Schematic of cDNA and qPCR design. The cDNA was generated using cDNA primers specific to the negative-sense RNA, random hexamers for total RNA, or oligo-dT primers for positive-sense RNA. The qPCR was performed with either primer set 1 or primer set 2 for each polyU qPCR. Nucleotide number where negative-sense (−) cDNA primer and probe bind to viral RNA are labeled. (B) The qPCR of cDNA synthesized with no primers or negative-sense cDNA primers. (C) PolyU qPCR of negative-sense RNA (Left) or PolyA qPCR primed with random hexamers (Middle) or oligo-dT primers (Right) from AML12 cells at 8 hpi. (D) PolyU qPCR of negative-sense RNA (Left) or PolyA qPCR primed with random hexamers (Middle) or oligo-dT primers (Right) from IFNAR^−/− BMDMs at 6 hpi. Set 2 is normalized to set 1 and is presented as mean ± SD. Values were analyzed by Student t tests. Data are representative of three independent experiments. ND, not detected; n.s., not significant.

Fig. 5.

Evaluating the length of polyU extensions on PUN RNA. AML12 cells were infected with WT or EndoUmut virus at an MOI of 1. At 8 hpi, RNA was purified from cell lysates, and polyU nested PCR was performed. (A) Schematic of nested PCR. Briefly, cDNA was generated with a strand-specific primer for negative-sense (−) RNA or an oligo-dT anchor primer for positive-sense RNA, and then nested PCR was performed. (B) PolyU or (C) PolyA PCR products separated on a 10% polyacrylamide gel and stained with SYBR Green II. (D) PolyU PCR products were purified from the polyacrylamide gel in B and sequenced with MiSeq Next-Gen Sequencing. Graph of read counts that contain a specific nucleotide (nt) length of polyU extensions (Left). Graph of proportion of reads that contain a specific length of polyU extensions (Right). Data are representative of three independent experiments.

Fig. 6.

Evaluating the abundance and length of PUN RNA during PEDV infection. PK1 or Vero cells were infected with WT or EndoUmut PEDV at an MOI of 0.1. RNA was purified at 24 hpi. (A) PolyU qPCR quantified in PK1 cells. (B) PolyU qPCR quantified in Vero cells. Set 2 is normalized to set 1 and is presented as mean ± SD. (C) PolyU or (D) PolyA nested PCR products from PK1 Cells. (E) PolyU PCR products from PK1 cells were purified from the polyacrylamide gel in C and sequenced with MiSeq Next-Gen Sequencing. Graph of read counts that contain a specific nucleotide (nt) length of polyU extensions (Left). Graph of proportion of reads that contain a specific length of polyU extensions (Right). Values were analyzed by a Student t test. Data are representative of two independent experiments.

Fig. 7.

Determining whether PUN RNA is an MDA5-dependent PAMP. RNA was in vitro-transcribed from DNA constructs and transfected into AML12 cells. At 8 hpt, RNA was purified from cell lysates, and IFNβ1 gene expression was measured by qPCR. (A) Schematic diagram of RNA products from in vitro transcription. (B) IFNβ1 gene expression induced by coronaviral RNA termini. (C) IFNβ1 gene expression by PUN RNA constructs. (D) IFNβ1 gene expression by PUN RNA constructs with varying polyU lengths. MDA5-knockdown AML12 cells (MDA5-KD) were generated by CRISPR/Cas9 transduction. (E) Western blot of WT and MDA5-KD AML12 cells for MDA5 and Actin. (F) WT and MDA5-KD AML12 cells were infected with WT or EndoUmut MHV at an MOI of 1. IFNβ1 expression was measured at 16 hpi. (G) In vitro transcribed PUN RNA was transfected into WT or MDA5-KD AML12 cells. IFNβ1 expression was measured at 8 hpt. IFNβ1 gene expression is normalized to 18s rRNA and set relative to mock. Values were analyzed by Student t tests. Data are representative of three independent experiments and presented as mean ± SD.

Fig. 8.

Evaluating EndoU activity on PUN RNA. RNA was cleaved by EndoU and separated by gel electrophoresis. (A) Sequences and length of RNAs 1 to 4. (B) EndoU cleavage of RNA 1 and RNA 2 performed for stated times and separated on a 10% polyacrylamide gel. (C) EndoU cleavage of RNA 3 and RNA 4 for 30 min and separated on a 10% polyacrylamide gel. (D) EndoU cleavage of in vitro transcribed PUN RNA (N5) for 45 min and separated on a 1% agarose gel. (E) RNA treated by EndoU cleavage was transfected in AML12 cells, and, at 8 hpt, RNA was purified from cell lysates, and IFNβ1 gene expression was measured by qPCR. IFNβ1 gene expression is normalized to 18s rRNA and set relative to mock. Values were analyzed by Student t tests. Data are representative of three independent experiments and presented as mean ± SD. n.s., not significant.

Fig. 9.

Model depicting EndoU cleavage of PUN RNAs. This model depicts how EndoU activity limits the generation of PUN RNA, which can act as a PAMP. We found that PUN RNAs with variable lengths of polyU sequences are generated in the absence of EndoU activity. We predict that these PUN RNAs can fold back and generate stem−loop structures by hybridizing with an A/G-rich domain located within the PUN RNA or on adjacent RNAs. This stem−loop structure may be recognized as dsRNA by host PRRs, thus stimulating the host innate immune response. The function of EndoU during replication is to reduce the length of polyU sequences, thus limiting the potential for generating PAMPs.