Discontinuous template switching generates coronavirus subgenomic RNAs from the 3' viral genome end by 5' to 3'...

In coronavirus (CoV)-infected cells, several structural and accessory proteins are synthesized from subgenome RNAs (sgRNA) containing a common genomic 5'-leader followed by a given open reading frame (ORF). We report that the abundance of these sgRNAs varies with distance from the 3'-end of the genome. Thus, there are more sgRNAs encoding nucleocapsid (N) than spike (S), presumably the results from discontinuous 5'-3' transcription template switch mediated by the viral replication and transcription complex (RTC). We optimized the circular polymerase extension reaction (CPER) methodology to generate infectious double-stranded circular cDNA (ds-circDNA) containing the mNeonGreen (NG) reporter in accessory ORFs of human CoVs OC43 and SARS-CoV-2. In each CoV, we found that levels of sgRNAs and NG expression increased with 3' proximal genomic NG location. By reinfection of HCT-8 cells with the same MOI 0.01, however, we found that the slow-growing OC43 NG-ns2 virions exhibited equal infectivity and productivity as the fast-growing OC43 NG-ns12.9 virions. Introduction of point-mutations into the mapped TRS_B motif for synthesis of the OC43 ns12.9 and M sgRNAs led to disrupt TRS_B-TRS_L cross-interactions and block production of the corresponding sgRNAs and infectious virions. Together, our data indicate that, by using an optimized CPER approach for positional NG insertions, we demonstrated the progressional reduction of the RTC-mediated template switch (read-through individual TRS_B) efficiency in production of the corresponding sgRNAs in an order from the 3' viral genome end.IMPORTANCEThe mechanism for sgRNA synthesis in the coronavirus life cycle is poorly understood. The current model suggests discontinuous template-switch transcription mediated by viral replication and transcription complex (RTC) for synthesis of individual sgRNAs to translate corresponding structural and accessory proteins but lacks experimental data support. This report provides the first experimental evidence that, in both hCoV-OC43 and SARS-CoV-2, viral RTC synthesizes its sgRNAs by long-range base-pairing between a distal transcription regulatory body sequence (TRS_B) upstream of each structural/accessory ORF and the transcription regulatory leader sequence (TRS_L) from the viral genome 5'-UTR, leading to the production of viral sgRNAs in abundance order from the viral genome 3' -end, with more N sgRNAs but less S sgRNAs. Our data support a "first-come, first-serving" model in TRS_B-TRS_L cross-interaction and read-through TRS_B process to mediate discontinuous transcription switch in coronavirus sgRNA synthesis in a 5'-3' transcription direction from the 3' viral genome during coronavirus infection.