Abstract

   The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has
   caused the global pandemic, resulting in great fatalities around the
   world. Although the antiviral roles of RNA interference (RNAi) have
   been well studied in plants, nematodes and insects, the antiviral roles
   of RNAi in mammalians are still debating as RNAi effect is suspected to
   be suppressed by interferon (IFN) signaling pathways in most cell
   types. To determine the role of RNAi in mammalian resistance to
   SARS-CoV-2, we studied the profiling of host small RNAs and SARS-CoV-2
   virus-derived small RNAs (vsRNAs) in the early infection stages of Vero
   cells, an IFN-deficient cell line. We found that host microRNAs
   (miRNAs) were dysregulated upon SARS-CoV-2 infection, resulting in
   downregulation of microRNAs playing antiviral functions and
   upregulation of microRNAs facilitating viral proliferations. Moreover,
   vsRNA peaked at 22 nt at negative strand but not the positive strand of
   SARS-CoV-2 and formed successive Dicer-spliced pattern at both strands.
   Similar characteristics of vsRNAs were observed in IFN-deficient cell
   lines infected with Sindbis and Zika viruses. Together, these findings
   indicate that host cell may deploy RNAi pathway to combat SARS-CoV-2
   infection in IFN-deficient cells, informing the alternative antiviral
   strategies to be developed for patients or tissues with IFN deficiency.

   Keywords: SARS-CoV-2, microRNAs, virus-derived small RNAs, immunity,
   IFN

Introduction

   The global pandemic of severe acute respiratory syndrome coronavirus 2
   (SARS-CoV-2) has caused over 572,239,451 infections and 6,390,401
   fatalities around the world since late 2019 (data collected from WHO
   website on July 25^th, 2022). As an enveloped, positive sense,
   single-stranded RNA betacoronavirus, SARS-CoV-2 is a member of
   Coronavirvidae family, which consists of two other members, severe
   acute respiratory syndrome (SARS-CoV) emerged in 2002 and Middle East
   respiratory syndrome (MERS) in 2012 ([45]1, [46]2). All these three
   coronaviruses are highly pathogenic and can cause severe respiratory
   diseases in human. However, different with SARS-CoV and MERS, which
   lead to 10% mortality for SARS-CoV and 37% for MERS, SARS-CoV-2 spreads
   more efficiently with a relatively lower level of mortality of 2%~3%
   ([47]3, [48]4). SARS-CoV-2 shares the same receptor
   angiotensin-converting enzyme 2 (ACE2) with SARS-CoV, using spike (S)
   protein for ACE2 binding as well as host serine protease TMPRSS2 for S
   protein primming to achieve viral and cellular membrane fusion and
   finally cell entry ([49]5). Symptoms accompanied as viral invasion are
   not limited to respiratory failure, but in multiple organ systems,
   including brain ([50]6, [51]7), liver ([52]8) and gastrointestinal
   tract ([53]9). The cytokine storm stimulated by viral infection is one
   of the main death causes of SARS-CoV-2 ([54]10). Besides, SARS-CoV-2
   infection might induce the new-onset diabetes, suggesting the potential
   health risk exists upon viral infection ([55]11).

   Mammalian cells have evolved multiple antivirus mechanisms, among which
   the interferon (IFN) signaling-mediated innate immune responses are
   characterized as critical and effective pathway for antiviral defense.
   In these pathways, viral RNA is recognized by several receptors such as
   retinoic acid-inducible gene I (RIG-I) ([56]12) and melanoma
   differentiation-associated gene 5 (MDA5) ([57]13) and subsequently
   triggers the producing and secretion of type I IFNs (mainly IFN-α and
   IFN-β) ([58]14), followed by activation of various
   interferon-stimulated genes (ISGs) against viral invasion ([59]15,
   [60]16). RNA interference (RNAi) is another crucial antiviral mechanism
   in eukaryotes. It regulates gene expression through 20-30 nucleotide
   (nt) small RNAs (sRNAs) processed from double-stranded or base-paired
   RNA ([61]17, [62]18). sRNAs can be classified into microRNAs (miRNAs),
   small interference RNAs (siRNAs) and piwi-interacting RNAs (piRNAs)
   based on their differences in biogenesis pathways ([63]19). The
   accumulation of partial host miRNAs is significantly altered to
   modulate the interactions between hosts and viruses. Some viruses also
   encode miRNAs to facilitate their infections or proliferations ([64]20,
   [65]21). RNAi plays major roles in plant resistance to viruses
   ([66]22). Viral RNAs and DNAs in plants and invertebrates are targeted
   by vsRNAs to hinder the proliferations of viruses ([67]18). However,
   the antiviral capacity of vsRNAs in mammalian remains much to debate
   ([68]23, [69]24). Studies have demonstrated the antiviral ability of
   RNAi in undifferentiated cells such as oocytes, embryonic
   teratocarcinoma cell lines and mouse embryonic stem cells (mESCs) but
   not in highly differentiated somatic cells ([70]25–[71]27). This might
   be caused by the existence of IFN pathway that represses RNAi by
   inhibiting siRNA processing ([72]28, [73]29). Prior studies suggest
   that suppression of IFN pathway is able to activate RNAi antivirus
   function in differentiated mouse cells ([74]28, [75]30).

   Recent studies found that SARS-CoV-2 genome encodes miRNAs and inhibit
   host immune response by targeting the IFN signaling pathway as well as
   metabolic pathways ([76]31–[77]33), which have been reported to be
   impaired in SARS-CoV-2 patients ([78]34). However, most of the studies
   are performed in cell lines expressing interferon. The accumulation of
   SARS-CoV-2 vsRNAs and their possible regulatory roles in
   host-SARS-CoV-2 interaction when IFN crippled still remains unclear. To
   understand the function RNAi exerts in response to SARS-CoV-2
   infection, we set out to construct the sRNA libraries and characterize
   the host miRNAs and viral sRNAs during the early stages of viral
   infection in Vero cells, which are uncapable of producing IFN-αand
   IFN-β ([79]35). We found that the accumulation of host miRNAs related
   to pathological processes or antiviral defense is modulated upon the
   infection of SARS-CoV-2, indicating that host miRNAs may play important
   roles in host-SARS-CoV-2 interaction. We observed characteristics of
   Dicer processed vsRNAs at negative strand of SARS-CoV-2, including a
   predominant 22-nt peak, preference of uracil at 5’ terminal, and
   successive 22-nt reads. We then characterized the accumulation of viral
   sRNAs in other cell lines infected with distinct viruses. We found that
   these characteristics at negative strand of viral genome were also
   identified in IFN-deficient Vero cells infected with Sindbis virus
   (SINV) ([80]36) as well as IFN-deficient neural stem cells infected
   with Zika virus (ZIKV) ([81]37–[82]39). Together, these results suggest
   that SARS-CoV-2 transcripts in Vero cells were likely to be processed
   into vsRNAs by Dicers and may play similar important antiviral roles as
   in plants and invertebrates. These observations will lead to
   development of alternative antiviral strategies against SARS-CoV-2 in
   patients with IFN-deficiency or tissues that lack IFN producing.

Results

Host sRNAs in Vero cells react to SARS-CoV-2

   To get a comprehensive profiling of sRNAs in IFN-deficient cell lines,
   we sequenced and analyzed sRNAs in Vero cells infected with SARS-CoV-2
   at a multiplicity of infection of 0.1 (MOI = 0.1). Control and
   SARS-CoV-2 infected Vero cells were harvested at 6 and 12 hours post
   infection (hpi). sRNA libraries of two biological replicates were
   constructed and sequenced from both control and SARS-CoV-2 infected
   Vero cells. After trimming adapters and filtering low-quality reads,
   ~20-32 million clean reads were obtained for each library ([83] Table
   S1 ). We evaluated the reproducibility of libraries using 2,652 known
   miRNAs obtained from miRbase ([84]http://www.mirbase.org). The
   correlation analysis showed that miRNA transcription of two biological
   replicates were highly correlated (R ≥ 0.99, P < 2.2e-16, [85]Figure S1
   ). We then processed clean reads with small noncoding RNA analysis
   pipeline from SPORTS1.1, using the human (hg38) and viral
   ([86]NC_045512.2) genomes as references ([87] Figure 1A and [88]Table