Murine Hepatitis Virus Exoribonuclease nsp14 Is Required for the Biogenesis of Viral Circular RNAs

Circular RNA (circRNA), a newly identified important component of the transcriptome, is formed by covalently bonded single-stranded RNA through back splicing (1) or other unknown mechanisms. It has been reported that circRNA plays important biological functions including microRNA (miRNA) sponges, parental gene expression regulators, and the translation template (2–5). Viral circRNAs have been recently identified from cells infected with different DNA viruses, such as Epstein-Barr virus (EBV) (6, 7), Kaposi’s sarcoma-associated herpesvirus (KSHV) (8), human papillomaviruses (HPVs) (9), and human cytomegalovirus (HCMV) (10). Coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARSCoV-2), has become a worldwide pandemic and poses a high threat to global health. We unprecedentedly identified viral circRNAs from cells that were infected by different coronaviruses, including SARS-CoV-2, SARS-CoV, and Middle East respiratory syndrome (MERS)-CoV (11). However, the biogenesis of circRNAs from coronavirus is still unknown. Murine hepatitis virus (MHV), a betacoronavirus, has been used in a mouse model to study human coronaviruses (12). Gribble et al. (13) reported that an RNA proofreading exoribonuclease, nsp14-ExoN, encoded by the MHV genome contributes to RNA recombination. In their study, deep transcriptome sequencing (RNA-seq) was performed on murine DBT cells infected either with wild-type MHV (MHV-WT) or with nsp14-ExoN inactive mutant MHV (MHV-ExoN2) (Table 1). We hypothesized that nsp14-ExoN may mediate the biogenesis of MHV circRNAs. To systematically test this hypothesis, we analyzed RNA from MHV-WTor MHV-ExoN2-infected cells and virioncontaining supernatants. Currently used methods of identifying circRNA are established on the examination of the back-splicing junction (BSJ) (14). ViReMa is one such tool that can quickly and sensitively identify viral RNA splicing junctions, including forward-splicing junctions (FSJs) and BSJs from next-generation sequencing data (15). Therefore, we applied ViReMa to assess the abundance of MHV-derived BSJs and FSJs and mapped the breakpoints to their respective genomic locations (Fig. 1A to D). In summary, there are two significant hot spots of FSJs and BSJs: (i) distant splicing between the 39 and the 59 ends of the genome corresponding to the N gene and the untranslated region (UTR) and (ii) local splicing in regions of MHV. The statistical analysis of genome coverage shows that about 23,043,062 to 56,476,922 nucleotides of each sample were mapped to the MHV genome (Table 1), suggesting that the numbers of MHV RNA molecules in the samples are comparable. To assess the effect of nsp14-ExoN loss of function, we calculated the junction Editor Haidong Gu, Wayne State University Copyright © 2023 Yang et al. This is an openaccess article distributed under the terms of the Creative Commons Attribution 4.0 International license. Address correspondence to Qiyi Tang, qiyi.tang@howard.edu, or Hua Zhu, hua.zhu@rutgers.edu. The authors declare no conflict of interest. Published 15 May 2023

Murine hepatitis virus (MHV), a betacoronavirus, has been used in a mouse model to study human coronaviruses (12). Gribble et al. (13) reported that an RNA proofreading exoribonuclease, nsp14-ExoN, encoded by the MHV genome contributes to RNA recombination. In their study, deep transcriptome sequencing (RNA-seq) was performed on murine DBT cells infected either with wild-type MHV (MHV-WT) or with nsp14-ExoN inactive mutant MHV (MHV-ExoN 2 ) ( Table 1). We hypothesized that nsp14-ExoN may mediate the biogenesis of MHV circRNAs. To systematically test this hypothesis, we analyzed RNA from MHV-WT-or MHV-ExoN 2 -infected cells and virioncontaining supernatants. Currently used methods of identifying circRNA are established on the examination of the back-splicing junction (BSJ) (14).
ViReMa is one such tool that can quickly and sensitively identify viral RNA splicing junctions, including forward-splicing junctions (FSJs) and BSJs from next-generation sequencing data (15). Therefore, we applied ViReMa to assess the abundance of MHV-derived BSJs and FSJs and mapped the breakpoints to their respective genomic locations ( Fig. 1A to D). In summary, there are two significant hot spots of FSJs and BSJs: (i) distant splicing between the 39 and the 59 ends of the genome corresponding to the N gene and the untranslated region (UTR) and (ii) local splicing in regions of MHV. The statistical analysis of genome coverage shows that about 23,043,062 to 56,476,922 nucleotides of each sample were mapped to the MHV genome ( Table 1), suggesting that the numbers of MHV RNA molecules in the samples are comparable.
To assess the effect of nsp14-ExoN loss of function, we calculated the junction frequency by normalizing to the genome coverage, which has been described by Gribble et al. (13). The results of the statistical analysis show that the loss of MHV nsp14-ExoN led to significant decreases in BSJ and FSJ frequencies in infected cells but not in the viral supernatant (Fig. 1E). These data indicate that nsp14-ExoN may mediate the formation of both FSJs and BSJs of MHV.
In our previous study, we found that CIRI2, as a circRNA prediction pipeline, is more reliable for coronavirus circRNA identification than other RNA prediction algorithms (11). Therefore, we also applied CIRI2 to identify MHV circRNAs. To improve the identification accuracy, biological replicates of the same condition were pooled. As shown in Fig. 1F to I, the CIRI2 pipeline identified 96, 50, 131, and 68 MHV circRNAs in the MHV-WT-and MHV-ExoN 2 -infected cell monolayers and and MHV-WT and MHV-ExoN 2 viral supernatant, respectively. By normalizing the genome coverage, the number of MHV circRNAs in MHV-ExoN 2 -infected cells (Fig. 1J) is significantly lower than that in MHV-WT-infected cells. These results indicate that nsp14-ExoN is required for the biogenesis of MHV circRNAs. Interestingly, we noticed that most of the circRNAs of MHV-WT-and MHV-ExoN 2 -or MHV-infected cell monolayers and viral supernatants were different (Fig. 1K). Only 5 circRNAs were found in both MHV-WT-or MHV-ExoN 2 -infected cells and the supernatant (Fig. 1L, red). These data suggested that nsp14-ExoN may serve as a viral proof factor to correct the breakpoint of MHV BSJs.
To experimentally confirm viral circRNAs from MHV-infected cells, we extracted total RNA from murine DBT cells that were infected with the MHV-A59 strain at 48 h postinfection (hpi). circRNAs have a covalently closed configuration and are hence more resistant to exoribonuclease RNase R than linear RNAs (16). Thus, we treated total RNA with RNase R for 30 min, and agarose gel electrophoresis showed that most rRNAs were degraded after RNase R treatment ( Fig. 2A). Divergent primers were designed to amplify the targeted BSJs of the most abundant MHV-encoded circRNAs in the region of 29,000 to 31,300 (see Data Set S1 in the supplemental material). To determine whether the inverse PCR products were from BSJs rather than nonspecific PCR products, we gel purified candidate BSJ amplicons based on molecular weight (Fig. 2B) and performed subcloning and Sanger sequencing for each candidate (Fig. 2C and Data Set S1). Consequently, linear viral RNA derived from ORF1a/b was degraded by RNase R. In contrast, a well-characterized mouse circRNA circHIPK3, BSJ#1 (30946j30787), BSJ#4 (31271j30793), BSJ#7 (30201j29788), BSJ#9 (31174j30775), and BSJ#11 (30889j30453), still remained. These data confirmed that MHV-infected cells generate virus-specific circRNAs.
Our overarching findings of this study include that MHV encodes circRNAs and nsp14-ExoN is important for the biogenesis of MHV circRNAs. Since nsp14 is required for RNA recombination (13), we speculate that MHV circRNAs may participate in viral

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, DOCX file, 0.02 MB.

ACKNOWLEDGMENTS
This study was supported by the National Institute on Minority Health and Health Disparities of the National Institutes of Health under award number G12MD007597 (Q.T.), NIH/NIAID grant SC1AI112785 (Q.T.), and Huazhong University of Science and Technology Union Shenzhen Hospital (Nanshan Hospital) (no. NY2021008).
We declare no competing financial interests.