Virology tidbits

Thursday, 28 April 2016

MERS-CoV and antiviral singling: role of orf4b and M protein

Coronaviruses (CoV) are positive sense RNA viruses with a genome size of 29-32 kb with four genera (Alpha-, Beta-, Gamma- and Delta CoV) belonging to the family of the Coronaviridae within the order of Nidovirales, with the Betacoronaviruses further divided into four lineages (A-D).

Table: Classification of known CoV

Most CoV identified until now are causing severe disease only in animals including agricultural important animals such as chicken, cattle, and pigs. To date only six human CoV (HCoV) have been identified, namely HCoV-229E, HCoV-OC43, HCoV-NL63, HCoV-HKU1, Severe Respiratory Syndrome (SARS)-CoV and most recently Middle Eastern Respiratory Syndrome (MERS)-CoV, although a SARS-like CoV, WIV1-CoV, replicates in primary human epithelial cells at low levels as well as in mice expressing the human SARS-CoV receptor ACE2 albeit at lower levels compared to SARS-CoV.

In contrast to SARS-CoV and MERS-CoV, the infection with HCoV-229E, HCoV-OC43, HCoV-NL63 and HCoV-HKU1only causes mild upper respiratory tract infections, whereas the infection with MERS-CoV causes severe lower respiratory infection and renal failure which may lead to the death of infected individuals.

Although the natural host for MERS-CoV has not been conclusively identified, bats might act as a natural reservoir of MERS-CoV (similar to Hendra and Nipah Virus) as evidenced by studies indicating the presence of MERS-like CoV in bats from Saudi Arabia, Europe, Africa and Asia. Furthermore, MERS-CoV can enter and replicate both in cell lines derived from bats and humans and the closely related Bat CoV-HKU4 can enter cells by binding to DPP4, the receptor for MERS-CoV. In Jamaican fruit bats (Artibeus jamaicensis) experimentally infected with MERS-CoV, clinical signs are absent despite viral replication and shedding in the respiratory and intestinal tract for up to 9 days p.i. as well as the induction of an antiviral response (see below for discussion).

Despite the ability of MERS and MERS-like CoV to infect both human and bat cell lines, bats are unlikely to be the source for the current outbreak of MERS in the Arabian Peninsula based on findings that livestock –in particular dromedary camels- have a high seroprevalence for MERS-CoV. In addition, recent studies showed that DPP4 is expressed at high levels in the upper respiratory tract of dromedary camels (but not humans) and that sequences of MERS-CoV isolated from dromedary camels are highly similar to sequence from human derived viral isolates. Finally, camel derived MERS-CoV can efficiently replicate in human cells.

MERS-CoV and the Interferon response

The polycistronic single stranded positive sense viral RNA of MERS-CoV encodes for two polyproteins, orf1a and orf1b, within the 5’ end which are cleaved into a total of 16 non-structural proteins (nsp) including the viral RNA dependent RNA polymerase whereas the 3’ end of the viral RNA encodes for the structural proteins, namely spike (S), membrane (M), envelope (E) protein as well as lineage specific proteins.

Figure: Processing of the orf1ab into several nsp

Following viral entry and release of the viral RNA, MERS-CoV replicates in the cytoplasm of cell, although the viral N protein is imported into the nucleus and localizes to the nucleolus. During viral replication, replication centers containing viral proteins and dsRNA are formed in the ER-Golgi intermediate compartment (ERGIC) in a process that might involve the induction of autophagy and/or viral induced rearrangements of the ER membrane that promote the formation of vesicular structures.

Figure: Outline of CoV replication cycle

The infection of a variety of cell lines with SARS-CoV and MERS-CoV has been reported to induce the expression of type I interferons (IFN) including IFN-α and IFN-β, suggesting that both SARS-CoV and MERS-CoV induce antiviral signaling via the recognition of viral RNA by PRR and PAMPS. Furthermore, both SARS-CoV and MERS-CoV express proteins that counteract the IFN pathway and thus contribute to the severity of infection. In general, MERS-CoV RNA is recognized by viral sensors of the retinoic acid-inducible gene-I (RIG-1)-like receptor (RLR) family melanoma differentiation gene 5 (MDA5) that upon activation by viral dsRNA intermediates, induce the nuclear translocation of both IFN regulatory factor (IRF)-3 and nuclear factor κB (NF-κB), both of which induce the expression of IFN-β. RIG-1 signaling induced by the viral RNA as well as by Poly (I:C) however can be inhibited by several viral proteins, including ORF4a, ORF4b, ORF5 and M. In the case of ORF4a, the viral protein binds dsRNA and thus inhibits the activation of RIG-1 and MDA-5, whereas the nuclear localization of ORF4b inhibits the induction of IFN-β as well as the expression of RIG-I, MDA5, MAVS, IKKε, and TBK-1 by inhibiting both (nuclear) IRF-3 and IRF-7 whereas in the cytoplasm ORF4b inhibits antiviral signaling by binding to TANK binding kinase 1 (TBK1) and IκB kinase epsilon (IKK-ε) thus suppressing the interaction between mitochondrial antiviral signaling protein (MAVS) and IKK-ε. In addition to inhibiting IRF-3 by binding to TBK1, both MERS-CoV and ORF4b protein also inhibits antiviral signaling by inhibiting the IFN inducible oligoadenylate synthetase (OAS)-RNase L pathway that is activated by the viral dsRNA intermediate via degradation of 2′,5′-oligoadenylate (2-5A), the activator of RNase L, thus inhibiting RIG-1 and MDA-5 induced signaling. Interestingly, ORF4b localises to the nucleolus in ORF4b transfected cells, but so far the importance for nucleolar localisation of ORF4b for antiviral signaling has not been demonstrated. In contrast to ORF4b, ORF5 –like ORF4a and M- can only be detected in the cytoplasm of infected cells. Similar to ORF4b however, the expression of ORF5 inhibits the induction of IFN-β upon infection with Sendai Virus (albeit to a lesser effect) probably by inhibiting IRF-3 and -7 but not NF-κB.

Activated RNaseL cleaves viral and cellular ssRNA including ribosomal RNA preferentially at UU and UA dinucleotide residues, therefore not only inhibiting replication of the viral genome but also the expression of viral proteins as well as processing viral RNA to be recognized by intracellular RNA sensing proteins thus activating antiviral signaling pathways.

In the case of CoV, the Mouse Hepatitis Virus (MHV) NS2 protein encoded by ORF2a has been identified as an inhibitor of the OAS-RNaseL pathway by inhibiting the conversion of intracellular ATP to 2’,5’- OAS via the 2’,5’ Phosphodiesterase domain (PDE) following the detection of viral dsRNA intermediates or synthetic Poly (I:C) RNA. Alignment of MERS-CoV, BtCoV-HKU5 and BtCoV-SC2013 orf4b with MHV nsp2 lead to the identification of a 2’,5’- PDE domain that is preceded by a nuclear localization signal that is catalytically active in a fluorescence resonance energy transfer (FRET) based RNaseL activity assay with similar kinetics to MHV nsp2. As expected, all orf4b proteins localize to the nucleus and the cytoplasm if expressed in murine L2 cells infected with MHV chimeras that express a catalytically inactive mutant of nsp2 and catalytically active orf4b derived from MERS-CoV or BatCoV-SC2013; in case of the MHV-BtCoV-HKU5 orf4b chimera, orf4b predominantly localises to the nucleus which does not affect viral replication but might not inhibit the activation of RNaseL, indicating that only cytoplasmic orf4b can inhibit RNaseL activation. In Calu-3 cells, a human airway epithelial cell line, infected with MERS-CoV, the deletion of nsp3-5 or the deletion of orf4b does not activate the degradation of rRNA to the full extent when compared to wt MERS-CoV probably due to the alternative pathways that inhibit OAS dependent activation of RNAseL such as the orf4a dependent sequestering of viral dsRNA intermediates.

In addition to the localisation of orf4b, the presence of a second His residue within the PDE is important for the ability of orf4b to regulate the activity of RNaseL since substitution of the H182 (MERS-CoV) or H 186 (BtCoV SC2013) with R result in decreased ability of orf4b to inhibit RNaseL.

Figure: N-terminal NLS in orf4b is required for nuclear localisation of MERS-Cov
and BtCoV off 4b

Figure: Cytoplasmic localisation and His at position of 182 or 186 of MERS-CoV and BtCoV orf4b
are required for inhibition of RNaseL

The induction of antiviral signaling by via OAS/RNase L dependent pathways has also been demonstrated for West Nile Virus (WNV), Influenza A Virus (IAV) and Vaccinia Virus (VACV). OAS3 knockout cells infected with WNV, IAV, VACV or treated with Poly(I:C) exhibit only minimal 2’-5’ oligoadenylate synthase activity and infected A549, hTERT-HME and HT1080 OAS3 knockout cells display higher viral titres than the corresponding wt cells. Interestingly, probably because of increased expression of IFN-λ1, trophoblasts of the human placenta express higher levels of OAS-1,-2, and -3 compared to JEG-3 cells, suggesting that increased OAS levels contribute to the resistance of trophoblasts to ZIKV resistance as well.

Figure: orf4b degrades 2',5' OAS and prevents activation of RNaseL

More recently, the viral M protein of both SARS-CoV and MERS-CoV but not HCoV-HKU1 has been demonstrated to form a complex with the TRAF3 adapter protein, thus inhibiting the formation of the TRAF3-TANK-TBK1/ IKK-ε complex via the N terminal domain of the viral M protein. Consequently, the expression of MERS-CoV or SARS-CoV inhibits RLR signaling induced by Sendai Virus or Poly (I:C) and further analysis using RIG-1 and NF-κB reporter plasmids showed that the expression of MERS-COV M preferentially inhibits RIG-1 mediated activation of IRF-3, thus preventing the induction of IFN-β. Similar to SARS-CoV M, the interaction between TRAF3 and MERS-CoV M is mediated by the N-terminal transmembrane domain 1 whereas the two other transmembrane domains as well as the C-terminal domain are dispensable. Both the SARS-CoV and MERS-CoV derived M and ORF 4b protein therefore inhibit antiviral RIG-1 by targeting the formation of the TRAF3-TANK-TBK1/ IKK-ε complex, thus inhibiting IRF-3 dependent signaling. From an experimental standpoint it would be interesting if a recombinant HCoV-HKU1 expressing MERS-CoV or SARS-CoV derived M exhibits increased viral titres following infection of primary human cells compared to wt virus. Also, it remains to be seen if the expression of MERS-CoV in Artibeus jamaicensis or in bat derived cell lines inhibits the antiviral signaling in a similar way. So far, the infection of Artibeus jamaicensis with MERS-CoV has been demonstrated to increase the expression of Mx-1, ISG56 and RANTES moderately suggesting that MERS-CoV does inhibit antiviral signaling in bats, thus explaining the absence of clinical signs in infected bats despite the presence of low viral titres in the duodenum of bats. The individual role of ORF4a, ORF4b, ORF5 and M has not been assessed in bat or camel derived cell lines. Indeed, the importance of the genetic and functional diversity of proteins involved in the induction of antiviral signaling pathways has recently been highlighted in studies comparing RIG-1 from waterfowl and mammals in their capacity to facilitate an immune response to Influenza A virus. In short, duck derived RIG-1 had a weaker ability to induce IFN compared to goose derived RIG-1 and both duck and goose derived RIG-1 are more effective than pigeon RIG-1 in (chicken) DF-1 cells infected with Influenza A virus.

Figure: MERS-CoV M, orf4a and orf4b: points of action

So far it has not been shown if the expression of mutant orf4b in the context of Coronavirus infection not only abrogates immune signaling and thus promotes viral replication solely by inhibiting apoptosis but also abrogates the induction of RNaseL induced autophagy and degradation of viral replication centers. If so, the number of viral RC should be increased in cells expressing mutant orf4b and/or mouse embryonic fibroblasts derived from RNaseL -/- mice.

Wednesday, 13 April 2016

Zika Virus and induction of apoptosis: combination of IFN-β and downregulation of gene expression?

As described in a previous post, Zika virus (ZIKV) infection of primary human skin fibroblasts induces the upregulation of TLR3 mRNA as well as inducing the expression of MDA-5 and RIG-1, components of the antiviral response induced not only ZIKV but by RNA viruses in general. As a result of the activation of RIG-1 and MDA-5, the expression of both Interferon-α (IFN-α) and Interferon-β (IFN-β) is increased. In mouse models such as A129, AG129 or Ifnar1 -/- mice that are deficient for either the receptor for IFN-α or both for IFN-α and IFN-β, higher viral loads are established in a number of tissues compared to wt mice, suggesting that the induction of IFN-α and IFN-β by ZIKV limits viral replication. Interestingly, in human neural progenitor cells (hNPC) infected with ZIKV the expression of IFNAR-1 (the human equivalent of Ifnar-1) is downregulated whereas the expression of the IFN-γ receptor (IFNGR-1) is upregulated. The induction of the type I IFN response can however also contribute to the severity of viral infections as shown for SARS-CoV in mice. In infected mice, the delay of the induction of type I IFN by SARS-CoV contributes to the accumulation of pathogenic inflammatory monocyte derived macrophages, resulting in elevated levels cytokines and chemokines as well as vascular leakage and the impairment of virus specific T cell responses, thus contributing to decreased survival of infected mice compared to Ifnar-1 -/- mice. Furthermore, blocking the IFNAR-1 receptor and deleting IFN-β in mice promotes the clearance of lymphocytic choriomengitis virus (LCMV) in mice whereas the activation of IFN-α inhibits early dissemination of LCMV. In contrast to SARS-CoV and LCMV however, Ifnar1 -/-, A129 and AG129 mice show a decreased survival rate, indicating that IFN-α is necessary for viral clearance and survival of infected mice.

The infection of hNPC with ZIKV MR766 induces the downregulation of a number of genes promoting cell survival including those encoding for anti-apoptotic proteins, cell cycle regulators, factors involved in the DNA damage response pathway and autophagy as well inducing apoptosis that is preceded by an arrest in G2/M phase of the cell cycle.

In addition, human neurospheres and brain organoids infected with ZIKV MR766 exhibit extensive caspase-3 dependent cell death as early as 3 days p.i. which is in agreement with previous results that showed that 56 hrs p.i. the expression of Caspase-3 is upregulated, suggesting that in human neuronal cells ZIKV induces Caspase-3 dependent apoptosis. None of these experiments however determined the contributing factors, i.e. the pathway of apoptosis induction.

Figure: Changes of gene expression in hNPC infected with ZIKV compared to Mock infected
cells; RECQL4, IFNAR-1, IRF-7, IFIT-2

The infection of human lung epithelial A549 cells with ZIKV PF/25013/18 results in effective viral replication as measured by immunofluorescence for the viral E protein and dsRNA, flow cytometry for E protein positive cells, and determination of viral titres that showed maximum titres within 48 hrs p.i. concomitant with the induction of PARP cleavage and decreased cell viability. RT-PCR analysis of ZIKV infected cells revealed that ZIKV infection induces a 200-fold increase of both IFIT-1 (ISG56) and IFIT-2 (ISG54), both of which are regulated by IRF-3 and IRF-7 as well as STAT-1/-2 (in addition to A549 cells, ZIKV also upregulates IFIT-2 and IRF-7 expression in hNPC). Since IFIT-2 induces the depolarization of mitochondria and thus contributes to the induction of Caspase-3 and -9 dependent apoptosis, it might be possible that the formation of the complex consisting of IFIT-1 and IFIT-2 is induced by ZIKV and induces apoptosis in a Bak and Bax dependent manner. Unfortunately, the current data do not contain experiments using either Bak -/- Bax -/- cells nor cells that are deficient for either IFIT-1/-2 or cells treated with siRNA targeting Bak, Bax, IFIT-1 or IFIT-2. Downstream of IFIT-1 and IFIT2, the infection of A549 cells with ZIKV triggers the production of mitochondrial reactive oxygen species (ROS) as detected by MitoSOX without increasing cytoplasmic ROS levels similar to cells infected with DENV by increasing the expression of mitochondrial SOD2 although in the case of DENV infected cells cytoplasmic ROS levels are increased as the result of DENV induced activation of the ER stress response. If the downregulation of RecQL4 expression in ZIKV infected cells contributes to the activation of SOD2 is a contributing factor is not known, but possible. Alternatively, ZIKV proteins localizing to the ER might induce the ER stress response and thus induce the expression of mitochondrial ROS independent of IFIT-1/-2.

Also it remains to be seen if the expression of IFN-β in ZIKV infected cells can be induced via the relocalisation of STING to perinuclear punctae.

Figure: Induction of STING signalling by delocalisation of STING to perinuclear punctae

The infection of primary human skin fibroblasts with ZIKV induces the expression of pro-inflammatory cytokines which can also be detected in patients infected with ZIKV. Accordingly, the infection of A549 cells induces the expression of IL-1β, IL-6 and MCP-1 at 24-48 hrs p.i.. IFN-β expression and secretion can be detected at early times post infection (12 hrs p.i., increasing between 18-24 hrs p.i.), suggesting that the induction of IFN-β increases the expression of both IFIT-1 and IFIT-2 contributing to the induction of apoptosis. Interestingly, pretreatment of A549 cells with IFN-β not only reduces viral titres but also reduces caspase-3 activity suggesting that other factors than the IFIT-1/IFIT-2 complex trigger apoptosis. Since IFN-β not only triggers an antiviral response and apoptosis but also autophagy, pretreatment of A549 cells might induce the formation of autophagosomes that not only promotes the degradation of viral particles upon entry but also promotes the degradation of viral RNA in a process called RNautophagy. Further studies are however needed to verify and explore this pathway. Since the infection of mouse embryonic fibroblasts (MEF) with West Nile Virus (WNV) or Japanese Encephalitis Virus (JEV) activates the inflammasome, pretreatment of A549 cells might prime cells to MEFV dependent inactivation of the inflammasome and thus apoptosis following inflammasome activation by ZIKV.

Figure: Induction of SOD2: dependent on downregulation of RECQL4 and/or induction of
ER stress response

In conclusion, the infection of A549 cells with ZIKV results in productive infection and triggers an antiviral response that includes the expression of chemokines, IFN-β, and IFIT-1 as well as IFIT-2, similar to observations from ZIKV patients or ZIKV infected hNPC. Similar to hNPC, neurospheres, and brain organoids, ZIKV triggers caspase dependent apoptosis probably in a IFIT-1/-2 independent manner, although further experiments are warranted. Pretreatment and treatment up to 2 hrs p.i. of A549 cells with IFN-β results in decreased viral titres and decreased caspase-3 activation suggesting that IFN-β has a pro-survival rather than a pro-apoptotic role in ZIKV infection, a notion that is supported by observations from existing mouse models. The increase of mitochondrial ROS by ZIKV might induce DNA damage that due to the potential inhibition of the DNA damage response by ZIKV might induce cell cycle arrest and subsequent apoptosis as well MAVS/RIG1 mediated induction of IFN expression and/or MAPK/NFκ-B mediated increase of chemokines.