Virology tidbits

Wednesday, 15 April 2015

Porcine Circovirus: Autophagy, Nucleolus, and Apoptosis

Porcine circovirus type 2 (PCV2) is a small non-enveloped single-strand (ss) DNA virus with a genome of 1768 bp in length. Although the infection of pigs with PCV2 by itself only causes a relatively mild diseases, symptoms -including Postweaning Multisystemic Wasting Syndrome (PMWS), congenital tremors, Porcine Dermatitis and Nephropathy Syndrome, reproductive failure, proliferative and necrotizing pneumonia, enteritis, exudative epidermitis, and porcine respiratory disease complex- are alleviated upon co-infection with other porcine viruses such as PRRSV or porcine parvovirus, Mycoplasma hyopneumoniae or following immunostimulation with Interferon-α/-γ (IFN-α/γ). This is reflected by low virus yields in PK-15 cells infected with PCV2 which is increased following the treatment with IFN-γ or agents inhibiting the acidification of endosomes and lysosomes such as Chloroquine, Monesin, or NH₄Cl, suggesting that the release of the genome from internalised virus is inhibited by the acidic pH of the endosome and that the expression of viral genes is stimulated by IFN-α/γ; indeed the PCV2 genome contains a interferon-stimulated response element (ISRE)-like sequence is responsive to both IFN-α and IFN-γ.

Prototype Circovirus particle

The genome itself encodes at least four ORF’s, with ORF1 encoding the replication proteins (Rep and Rep’), ORF2 the Capsid protein (Cap), ORF3 activating the NF-κB pathway by facilitating the ubiquitin-mediated proteasomal degradation of regulator of G protein signalling 16 (RGS16), thus promoting the expression and secretion of IL-6 and -8, and ORF4 being a regulator of both ORF1 and ORF3 expression, with latter being less well characterised.

PCV genome

PCV2: Autophagy, nuclear egress, non-lytic spread and apoptosis

As discussed before, the induction of autophagosome formation by viral proteins can support the formation of viral replication centers that function as the site for the replication of the genome (as discussed in extensio for Coronaviruses) whilst also can be part of the antiviral response by facilitating the degradation of viral components including viral RNA and/or stimulating the immune response by increasing the presentation of viral proteins via MHC Class II on the cell surface. Paradoxically, autophagy can also promote viral replication and in particular egress of viral proteins via the ESCRT pathway, allowing viral egress without cell lysis. Indeed, Hepatitis A virus (HAV) has been shown to exist in extracellular vesicles, whose formation is dependent on two proteins, Alix and VPS4B, that are components of the exosome pathway, but independent of Tsg101 or Beclin-1, suggesting that HAV induces the formation of autophagosome or autophagosome-like particles that contain viral particles that are exported via the ESCRT pathway. Indeed, Alix has been demonstrated to links autophagy to the ESCRT pathway probably by binding a complex consisting of ATG12 and ATG3. In this scenario, autophagosomes containing ATG12 and ATG3 bind to Alix at sites of intraluminal vesicle formation thus allowing membrane curvature and exosome formation. Although in the case of HAV a functional link between Alix and the ATG12-ATG3 complex has not been shown, the formation of virus like particles by a retroviral protein, Murine Leukaemia Virus (MLV) Gag, has recently been shown to be dependent on the formation of ATG12-ATG3-Alix complex. In the case of Poliovirus, the inhibition of autophagy by siLC3 decreases the spread of viral particles in cell culture whereas the stimulation of autophagy by Rapamycin, Loperamide, or Nicarpidine increases cell-to-cell transmission of viral particles without cell lysis. Contrary to these results however, the induction of autophagy can also be detrimental for viral replication, especially if induction of autophagy causes apoptosis early in the replication cycle. A number of viruses therefore encode proteins that prevent the fusion of mature autophagosomes with the lysosomes, such as the M2 protein of Influenza A virus (IAV) or ICP 34.5 of Herpes Simplex Virus (HSV)-1. Stimulating autophagy in cells infected with HSV-1 indeed decreases viral titres (without affecting cell viability), whereas in IAV infected cells (massive) induction of autophagy decreases cell viability.

In the case of PCV2 infected PK-15 cells, treatment of cells with either Chloroquine or NH₄Cl increases viral replication, which is being attributed to facilitating the release of the viral genome. Both reagents however also prevent the degradation of the mature autophagosome suggesting that autophagy -or to be precise autophagic flux- might inhibit viral replication. Indeed, in the studies published both reagents were applied not only during viral entry but also over the whole course of the experiments, making it difficult to separate early from later stages of viral replication.

Analyzing the formation of autophagosome by determining the levels of LC3-II however indicates that PCV2 induces the formation of (LC3-II positive) mature autophagosomes and increasing autophagic flux as measurement of p62/SQSTM-1 levels by 24 hrs p.i. , suggesting that PCV2 indeed does induce autophagy. Inhibiting either the formation of autophagosomes with siATG5 or 3-Methyladenine (3-MA), or the fusion of mature autophagosomes with Chloroquine decreases levels of viral DNA and viral titres, suggesting that autophagy is indeed required for efficient viral replication. In contrast to IAV, the increase in autophagy however is not associated with decreased cell viability. Closer examination of the viral proteins revealed that only the Capsid protein but not the proteins derived from ORF1 or ORF3 (with ORF4 not being examined) induces autophagy. Paradoxically, the Capsid protein also induces apoptosis in PK-15 cells which have been pre-treated with IFN-γ, suggesting

the expression of the Capsid protein might sensitize cells to IFN-γ induced apoptosis.

Since so far the precise mechanism of autophagy induction by the Capsid protein has not been determined. The PCV2 Cap localises both to the nucleoplasm and the nucleolus by interacting with NPM-1/B23 in transfected HEK 293T cells. Similar to the coronaviral N protein, the localisation of Cap might induce the redistribution of nucleolar protein(s) and thus nucleolar stress, which might induce autophagy and apoptosis in a p53 dependent manner, thus sensitizing cells expressing Cap to IFN-γ induced apoptosis. The observed subnucleolar localisation of the Capsid protein in infected cells might be necessary for the formation of viral particles which egress from the nucleus in a process involving components of the autophagy machinery akin to HSV-1, thus explaining the dependent of PCV2 replication on an intact autophagy pathway. Expressing viral proteins from other viruses, such as PRRSV or Coronavirus’, that induce the formation of autophagosomes in cells infected with PCV2 might therefore increase nuclear egress by recruitment components of the autophagy machinery in close proximity to the nuclear membrane, especially in a situation in which autosis occurs as a result of increased ER stress induced autophagy. Alternatively, the expression of viral proteins inducing the formation of autophagosomes might facilitate the release of PCV virions from the infected cell using the autophagy and ESCRT pathway in a Alix dependent manner similar to MLV Gag. So far however, none of these hypotheses has been tested.

In the context of viral infection, autophagy is induced via the AMPK/ERK/TSC2, mTOR pathway, namely by inhibiting mTOR by activating the ERK-1/-2 pathway in TSC-2 dependent manner. It might be possible however that either the replication of viral DNA or the presence of other pathogens induces the activation of ATM and thus inducing autophagy in a ATM dependent manner. Again, further studies are needed. Activation of ATM might also exacerbate nucleolar stress by re-localising E2F1, thus linking viral induced activation of ATM to nucleolar stress induced autophagy and sensitizing infected cells to IFN-α/γ treatment. Alternatively, it might be possible that the nucleolar accumulation of the Capsid protein is required for the export of the viral RNA, similar in function to the HTLV-1 Rex protein whose nucleolar localisation is required for the export of tax and rex mRNA, and in infected cells both functions may not be mutually exclusive.

Model of PCV2 ORF3 and Capsid protein mediated
induction of autophagy, cell cycle delay, and apoptosis

Additionally, the ORF3 protein of both PCV1 and 2 have been demonstrated to induce caspase dependent apoptosis as a result of induction of porcine p53, whereas the expression of ORF4 counteracts ORF3 induced apoptosis by regulating the expression of ORF3. The ability of ORF4 to inhibit ORF3 mediated apoptosis and thus promote viral replication is particular evident early in the infection by restricting ORF3 mediated via inhibition of ORF3 expression. In addition, ORF4 also regulates the expression of ORF1 -and thus Rep and Rep’- as evidenced by the increase in ORF1 levels and increased viral replication in cells infected with a ΔORF4 virus. The expression of ORF4 therefore has both a beneficial effect by preventing apoptosis as well as a inhibitory effect on viral replication by regulating the expression of ORF1.

PCV2 ORF3 itself is a protein of 37 kDa in size, localised both in the cytoplasm and the nucleolus. Both nucleolar and cytoplasmic PCV2 ORF3 co-localises and interacts with the p53 binding domain of porcine ubiquitin E3 ligase Pirh2, thus degrading Pirh2 and activating p53, subsequently inducing p53 dependent apoptosis. Since ORF3 also activates the NF-κB pathway, it might also be possible that ORF3 inhibits ASK1 induced apoptosis, thus contributing to the antiapoptotic signalling induced by PCV2 via activation of Akt and subsequent phosphorylation of ASK1. It might therefore be possible that the nucleolar -but not cytoplasmic- localisation of ORF3 is necessary for promoting apoptosis (or vice versa). Interestingly, PCV1 ORF3 localises exclusively to the cytoplasm in a pattern resembling the ER, indicating that nuclear localisation of PCV1 ORF3 is not necessary to induce apoptosis. Further studies comparing ORF3 derived from PCV1 and 2 are therefore warranted. In addition, the recent identification of novel Circoviruses in mosquitoes and bats might offer further insights into the biology of Circoviruses’.

Tuesday, 31 March 2015

Viruses and the Nucleolus: Coronavirus N, Cytokinesis and Autophagy

The nucleolus is a subnuclear structure not surrounded by a membrane, which is disassembled during mitosis and reformed during late telophase via the interaction of nucleolar proteins with loci of ribosomal DNA. At this point, the inhibition of CDK1 leads to the resumption of RNA Polymerase I (RNA Pol I) dependent transcription of rDNA, resulting in the expression of the pre-rRNA (18S, 5.8S, and 28S, with internal and external transcribed spacers) and the recruitment of the processing machinery, forming the prenucleolar bodies (PNBs) and subsequently processing factors as nucleoli mature.

During interphase -in particular during late G1 and early S phase- multiple nucleoli can be observed within the nucleus whereas in middle/late S phase these commonly fuse to become one or two nucleoli. Since large nucleoli are associated with (high) rates of cell proliferation due to an increase in RNA Pol I dependent transcription of rDNA, nucleolar size can be linked to cancer. This is also reflected by changes in the localisation pattern of nucleolar proteins such as Fibrillarin, a factor involved in the processing of newly synthesised pre-rRNA, whose localisation changes upon treatment with Actinomycin D. Since Actinomycin D inhibits RNA Pol I activity, Actinomycin D treated cells exhibit a weaker signal for Fibrillarin similar to cells treated with ALLN or the small molecule inhibitor 10058F4, indicating that Fibrillarin is associated with sites of active RNA Pol I transcription and that the perinuclear region is disassembled following the inhibition of RNA Pol I.

Domains and marker proteins of a prototype nucleolus

In general, the nucleolus can be subdivided into three different regions, the granular component (GC), the Fibrillar Centre (FC), and the Dense Fibrillar Centre (DFC), the latter being the “core” formed around the nucleolar organizing region (NOR) containing the 18S, 5.8S, and 28S rDNA loci which in the human genome are localised on the p- arms of the five acrocentric chromosomes. Transcription by RNA Pol I occurs at the interface of the FC and DFC following the assembly of rDNA transcription factories in a c-Myc dependent manner. Traditionally, the nucleolus has been associated sole with the biogenesis of ribosomes and while this is certainly one of the main functions, in recent years multiple pathways -including apoptosis, autophagy, and DNA repair- have been identified which depend on nucleolar integrity. Nucleolar disruption induced by various stressors including DNA damage and inhibition of Pol I dependent transcription induces the activation of p53 via p14^Arfmediated disruption of the MDM2/p53 complex - and subsequent induction of p53 responsive genes- as well as inducing non-canonical autophagy in B23/Nucleophosmin (NPM) dependent manner but independent of nucleophagy. Since (activated) p53 not only induces the expression of proapoptotic genes such PUMA and NOXA or the Cyclin E inhibitor p21^Waf1 but also DRAM-1 and Sestrin-2, it might be possible that the induction of p53 following nucleolar stress induces autophagy in a DRAM-1 dependent pathway. Since the expression of p21^Waf1 has also been linked to oncogene induced apoptosis and oncogene induced senescence (OIS), inhibition of the nucleolar stress response pathway by Akt and mTORC1 mediated phosphorylation of PRAS40 has been demonstrated to promote the inhibition of p53 by stabilising a complex consisting of the ribosomal protein L11, HDM2 and p53, thus preventing the induction of OIS by preventing the expression of p21^Waf1 as well as inhibiting autophagy.

Degradation of p53 in unstressed nucleoli by HDM2

Stabilisation of p53 following nucleolar stress induces apoptosis, senescence,
and autophagy via p14 which can be inhibited by nuclear PRAS40

Viral proteins and the nucleolus: Coronavirus Nucleocapsid protein as a case study

Viral proteins localise to the nucleolus either via nucleolar localisation sequences (NoLS) that are usually part of a NLS, via binding of nucleolar shuttle proteins (such as Coilin or Argonaute 4), or by binding nucleolar proteins.

Table: Examples of viral proteins localising to the nucleolus

A number of viral proteins have been shown to localise to the nucleolus or to interact with nucleolar proteins, included -but not limited to- viral proteins derived from HIV (Rev, Gag, Tat), Adenovirus (V, VI, and the viral genome), Herpesvirus Saimiri and KSHV (ORF57), Influenza A Virus (NS1), and both the major and minor Capsid proteins from Potato Leafroll virus (PLRV), indicating that the nucleolar localisation of viral proteins is a common feature. In addition, the nucleolar architecture is altered in cells infected with Human Cytomegalovirus (HCMV) as such that Nucleolin expression is not only induced by HCMV but also relocalised to the nucleolar periphery in close proximity to the viral replication centres thus maintaining the architecture of the replication centres.

In the case of plant viruses, it has been demonstrated that the recruitment of nucleolar proteins is required for the viral replication, viral movement, and the assembly of viral RNP particles. The ORF3 long distance movement protein localises to the nucleolus via binding to nuclear (and highly dynamic) Cajal Body (CB) as indicated by the reorganisation of CB into structures that fuse with the nucleolus, leading to the recruitment of Fibrillarin. Fibrillarin in turn is required for the assembly of infectious viral particles in the cytoplasm, which is similar to PLRV. Interestingly, both the Herpesvirus Saimiri and KSHV ORF57 proteins induce the nucleolar redistribution of the human TREX (transcription/export) proteins that are involved in the export of mRNA, thus inducing the export of viral mRNA suggesting that the nucleolar localisation of viral proteins in general is essential for viral replication.

In the case of the Coronaviridae, the nucleocapsid (N) protein derived from a variety of different members, including IBV, MHV, TGEV, PEDV, and SARS-CoV, has been shown to localise to the nucleolus as indicated by co-localisation with nucleolar proteins, in particular Nucleolin and Fibrillarin. Since the N protein from both MHV and IBV does not co-localise with B23, it can be assumed that the N protein does localise within the DFC or the FC of the nucleolus.

CoV N protein localises to the nucleolus in LLCPK (TGEV) or Vero (MHV, IBV, SARS-CoV) cells
and induces incomplete cytokinesis

Relocalisation of Fibrillarin in Vero cells expressing
MHV or IBV N

Delocalisation of Fibrillarin in various cell lines expressing
SARS-CoV N

The expression of N derived from IBV, MHV, TGEV, and SARS-CoV has also been shown to induce aberrant cytokinesis. If this feature however is related to the nucleolar localisation of N or not is at present not known. One possible scenario might be that the nucleolar localisation of N causes nucleolar stress as indicated by the relocalisation of Fibrillarin within the nucleolus. The induction of nucleolar stress induces the activation of p53 that can induce autophagy via DRAM-1 and/or Sestrin-2 and thus might promote the degradation of proteins required for the degradation of the midbody such as active RHOA. So far however this has not been demonstrated.

The activation of p53 however might also the induction of the intrinsic apoptotic pathway and the expression of SARS N has been demonstrated to induce the cleavage of Caspase-3 in the absence of apoptosis. Since activated Caspase-3 cleaves Beclin-1, Atg4D and Atg5 –thus inhibiting the formation of the autophagosome- only a subset of cells expressing N might exhibit an increase in autophagy. Apoptosis itself might be antagonized by binding PARP-1; indeed, PRRSV N protein has been shown to bind PARP-1 although the functional consequences are not known.

The expression of IBV (top) or MHV N sequesters B23
and induces incomplete cytokinesis

Model: induction of nucleolar stress induces multiple competing
pathways

Alternatively the sequestration of B23 by the coronaviral N protein might prevent the phosphorylation of B23 by Polo-like Kinase (Plk)-1 and thus induce mitotic defects including cytokinesis failure.

It would be interesting to investigate if the defects of cytokinesis can be alleviated by the expression of other Coronavirus proteins, in particular those whose expression inhibits the degradation of autophagosomes, namely nsp-6 and nsp-3/-4, or in the presence of autophagy inhibitors such Vps34 inhibitors or Chloroquine. Since only a subset of N expressing cells are undergoing aberrant cytokinesis and not all cells expressing N exhibiting a cytokinesis defect also exhibit nucleolar localisation of N, aberrant cytokinesis might also be dependent on other factors. Clearly, further experiments are warranted to elucidate the mechanism.