The SARS-CoV-2 subgenome landscape and its novel regulatory features

•Dynamic subgenome landscapes of SARS-CoV-2 in two host cells are constructed•Bidirectional and successive template switching diversify sgRNA biogenesis•Several key determinants governing efficacy discovered•Canonical TRS-independent RNA-RNA interaction mediates switches Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is currently a global pandemic. CoVs known to generate negative subgenomes (subgenomic RNAs [sgRNAs]) through transcription-regulating sequence (TRS)-dependent switching, but the dynamic coronaviral regulatory rules remain unclear. Here, using next-generation sequencing (NGS) short-read Nanopore long-read poly(A) RNA cell types at multiple time points after infection with SARS-CoV-2, we identified hundreds constructed subgenomes. Interestingly, could occur bidirectional manner, diverse generated from template-switching events. The majority result interactions, including seed compensatory modes, terminal pairing status as determinant. Two switch modes also responsible for biogenesis. Our findings reveal landscape its features, providing molecular basis understanding biogenesis developing novel anti-viral strategies. recent outbreak (SARS-CoV-2; referred human [HCoV-19]; Zhou et al., 2020Zhou P. Yang X.-L. Wang X.-G. Hu B. Zhang L. W. Si H.-R. Zhu Y. Li Huang C.-L. al.A pneumonia associated new probable bat origin.Nature. 2020; 579: 270-273Crossref PubMed Scopus (12911) Google Scholar; Liu 2020Liu Q. Chen J. Xiang R. Song H. Shu S. Liang You al.Detection Covid-19 Children Early January 2020 Wuhan, China.N. Engl. Med. 382: 1370-1371Crossref (440) 2020aChen Xu K. Ye G. Wu Sun Z. F. Zhong al.RNA based mNGS approach identifies individual cases Wuhan outbreak.Emerg. Microbes Infect. 9: 313-319Crossref (366) Scholar) has turned into pandemic, causing more than million deaths October (Dong 2020Dong E. Du Gardner An interactive web-based dashboard track COVID-19 real time.Lancet Dis. 20: 533-534Abstract Full Text PDF (5770) Scholar). an enveloped virus ?30,000-nt-long positive-sense genome belongs genus Betacoronavirus, which shows 50% 77.5% identity Middle East (MERS-CoV) SARS-CoV, respectively (Zhou genomic (gRNAs) have 5? cap structure 3? tail, whose end large open reading frames (ORF1a/1b) encode 16 viral nonstructural proteins (nsps) occupying two-thirds genome. Polyproteins 1a/1ab (pp1a/1ab) translated directly gRNA ?1 ribosomal frameshifting (Perlman Netland, 2009Perlman Netland Coronaviruses post-SARS: update on replication pathogenesis.Nat. Rev. Microbiol. 2009; 7: 439-450Crossref (1164) CoV (one-third size) contains genes encoding several main structural proteins, spike protein (S), envelope (E), membrane (M), nucleocapsid (N), various accessory 2020bChen Guo D. Emerging coronaviruses: Genome structure, replication, pathogenesis.J. Virol. 92: 418-423Crossref (1967) genomes hallmark process transcription, facilitated replication-transcription complex (RTC) RNA-dependent polymerase (RdRP) activity (Snijder 2016Snijder E.J. Decroly Ziebuhr Nonstructural Proteins Directing Synthesis Processing.Adv. Virus Res. 2016; 96: 59-126Crossref (349) Scholar), complicated that other viruses. negative-strand synthesized RdRP starting positive (+)gRNAs, continuous synthesis generates full-length complementary (?)gRNAs, whereas discontinuous jumping produces (?)subgenomic (sgRNAs) common ends (Hussain 2005Hussain Pan Peng Tien Identification subgenomic noncanonical transcription initiation signals coronavirus.J. 2005; 79: 5288-5295Crossref (175) Positive-sense progeny gRNAs sgRNAs these intermediates templates (Sola 2015Sola I. Almazán Zúñiga Enjuanes Continuous Discontinuous Coronaviruses.Annu. 2015; 2: 265-288Crossref (359) step, called “template switch,” mediated (TRS) body (TRS-B) leader (TRS-L) upstream ORF1ab resulting fusion leader-body sequences. Thiel 2003Thiel V. Ivanov K.A. Putics Á. Hertzig T. Schelle Bayer Weißbrich Snijder Rabenau Doerr H.W. al.Mechanisms enzymes involved SARS expression.J. Gen. 2003; 84: 2305-2315Crossref (667) Scholar eight SARS-CoV (Thiel our subsequent study 10 sgRNAs, Recently, variants HCoV-229E were reported be characterized (Viehweger 2019Viehweger A. Krautwurst Lamkiewicz Madhugiri Hölzer M. Marz Direct nanopore provides insights enables modification analysis.Genome 2019; 29: 1545-1554Crossref (115) Kim 2020Kim Lee J.Y. J.S. J.W. V.N. Chang Architecture Transcriptome.Cell. 181: 914-921.e10Abstract (1168) high-resolution map transcriptome modifications Vero E6 cells. However, unclear genomes, whether events happen positive-strand largely unknown. TRSs comprise conserved 6- 7-nt core surrounded variable Different been previously, CUAAAC transmissible gastroenteritis (TGEV) (Zúñiga 2004Zúñiga Sola Alonso Sequence motifs regulation synthesis.J. 2004; 78: 980-994Crossref (171) ACGAAC CUUUAGA equine torovirus (Stewart 2018Stewart Brown Dinan A.M. Irigoyen N. Firth A.E. Transcriptional Translational Landscape Equine Torovirus.J. 2018; 24Crossref (16) It hypothesized formation duplex between TRS-L downstream TRS-B sequences determines features TRS-like elements CoVs, not yet defined. Previous studies, work mainly used RT-PCR coupled clone characterize junction, low throughput unable detect Northern blotting generally validate specific sizes detailed unknown, resolution limited. Using technologies, detected people (Liu assembled (Chen transcriptomes samples affected individuals (Xiong 2020Xiong Cao Jiang Tang al.Transcriptomic characteristics bronchoalveolar lavage fluid peripheral blood mononuclear patients.Emerg. 761-770Crossref (731) NGS high-throughput short reads capacity quantify gene expression splicing junctions, it difficult assemble RNAs. Viehweger direct full HCoV-229 without amplification. We devised integrative multi-strategic (RNA-seq), Pacific Biosciences (PacBio) sequencing, construct transcriptional (Wang 2019Wang Zheng X. Qin Wen Multi-strategic RNA-seq analysis reveals cotton.Nat. Commun. 10: 4714Crossref (51) In this study, employed techniques systematically (1) profiles (2) different points. further investigated junction sites those found major interactions. Moreover, exists during classes provide view uncover their To explore subgenomes, first verified presence SARS-CoV-2-infected northern (Figure S1). hybrid poly(A)-selected technologies analyze local simultaneously sgRNAs. enriched total extracted monkey (ATCC number CRL-1586) Caco-2 infected (WIV04, IVCAS 6.7512), libraries duplicates Illumina MinION platforms, respectively. sequenced paired-end 150-bp mode, 1A). investigate RNAs, performed assays SARS-CoV-2. chose potential differences hosts species tissue origins. ratio relatively consistent, around 0.1%, 7%, 4 h, 6 12 24 h reaches same level 48 80%–90% (Table fractions 0.02%, 1.2%, 21% respectively; smaller consistent rate long contain tails median length 50 nt 1B), similar report (Kim all mapped (+) genome, suggesting (?)gRNA or (?)sgRNAs do tails. infection, 1,248 nt, 22 close 30,000 covering whole next developed toolkit identify robust across Template “jumping” junctions 1C), exon-exon pre-mRNA splicing. ensure correct localization required flanking both sides least 20-nt exact matches analyzed infection. verify robustness compared counts observed replicates 1 significant highly reproducible 1D). total, 45,343 combined data replicates. remove noise uneven abundance, statistical scoring effect background (Figures S2A-S2D; STAR Methods) obtained 100 1D, red points; Table S2). 141 embedded methods S2E; S2), 31 overexpressed significantly 1E), reliability predicted included previously fusions SARS-CoV. quantified levels N showed highest level, increased direction 1F). This indicates intermediate can serve shorter (see multi-switches below). These canonical abundant represent 57.8% (99,548/172,107) reads. Beyond many well body-body earlier increasing 1G S2F) 1H S2G). For each sample, counted numbers termed annotated group. 1I). complete took advantage only considered extending Internal must datasets By definition, 433 208 clusters merging neighboring within 5 subsequently classified 3 groups 2A; S3). group leader/S-N, representing joining S N; second group, ORF1ab/S-N, represents linking positions third S-N/S-N, internal S-to-N regions. latter non-canonical classification leader/S-N named junction. distribution shown Figure 2B, strong (with support) site largest marked lines asterisks, overall structures illustrated 2C, S3. corresponding clusters, events, sample 2D, S4A, requirements noticed some had gaps switching. suggests may independent multi-switch leader+S-N leader+2×S-N, S-N/S-N junction(s), 2C 2E). read single-switch much larger leader+2×S-N Leader/S-N dominant form share junctions. As 2F, seven bi-switch four tri-switch 28,525–28,576, support Generally, higher one parental sgRNA, originating it. correlation 0.89 2G). evident S4B). theory function additional results show levels, useful resource studying functions mechanisms. switches, examined base pairings 9 almost 3A, left). expected, already TRS motif (ACGAAC) ACGAAC/AAGAAC (anti-TRS-B). Surprisingly, extensive 7–12 consecutive pairs beyond anti-TRS-B center). general (JSs), denote left right segments JSs UL (upstream left) UR right), respectively, DL (downstream DR RNAhybrid program (Rehmsmeier 2004Rehmsmeier Steffen Hochsmann Giegerich Fast effective prediction microRNA/target duplexes.RNA. 1507-1517Crossref (1873) find optimal minimum free energy (MFE) segments, anti-DR (containing anti-TRS-B) anti-DL. stronger anti-DL right). Intriguingly, tendency closer 3A). analogy microRNA (miRNA)-mRNA base-pairing (Bartel, 2018Bartel D.P. Metazoan MicroRNAs.Cell. 173: 20-51Abstract (1926) defined patterns: mode (6 bp 1- JSs) outside region), 3A. positive-to-positive assumed (+)sgRNAs then copied switches. negative-to-negative exists, occurring while generating (+)sgRNAs. discriminate processes sets either UR-DR pair (UR::anti-DR mode) UL-DL (anti-UL::DL mode). posited relative strength over indicate mode. tested hypothesis; discriminated MFE (method details). 3B, top example 64-28254, lower pair, switched (?)sgRNA synthesis. contrast, bottom 76-26480, supporting occurs (+)sgRNA checked above high (top junctions) (bottom random S5A). evidence exist differentiated considering positioning MFEs modes. patterns mediate involving anti-DL, (20 JSs, Figures 3A 3B). small shifts, method maximal connected subgraph assigning count leader-type groups—UR-DR strong, uncertain—based difference 3C). There about 50%, 17.5%, 32.5% uncertain groups, sorted high-expression especially S5B. 78 134 3D). Although ones, they frequently samples, 3E). supported (Nanopore-seq) data, correlations S5C). cases, mapping influence inferred S5D). Empirically, minimap2 (Li, 2018Li Minimap2: pairwise alignment nucleotide sequences.Bioinformatics. 34: 3094-3100Crossref (3495) prefers setting overestimated. determining switch. factor because betw