Virology tidbits: 2016

Zika Virus (ZIKV) is a mosquitoe-borne Flavivirus (MBFV), related to other Flavivirus’ that are transmitted by mosquitoes such as Dengue Virus (DENV), Yellow Fever Virus (YFV) or Tick Borne Encephalitis Virus (TBEV) and as such the viral genome is released into the cytoplasm following viral entry and translated into a single polyprotein that is cleaved by both viral and cellular proteases into the 3 structural and 7 non-structural proteins.

One of the characteristic features of positive strand RNA viruses is the interaction of proteins that are involved with viral replication with cellular membranes thus leading to the formation of viral replication centers (RC) in infected cells as discussed in extensio before for Coronavirus’ and Chikungunya Virus (CHIKV). In general, the viral RC contain not viral proteins such as the RNA dependent RNA Polymerase and viral helicase but also dsRNA, thus shielding viral RNA from being recognized by components of the cellular antiviral signalling pathway that recognize viral dsRNA such as RIG-1 and MDA-5 as well as providing a scaffold for the assembly for viral particles. In the case of DENV, WNV, and TBEV, Electron Tomography (ET) studies revealed the presence of different virus induced structures derived from the ER such as cytoplasmic vesicles containing viral proteins and dsRNA that appear to be derived from invaginations of the ER as well as vesicular structures that contain viral proteins within the ER lumen that appear as invaginations of the ER bearing pore-like connections to the cytoplasm as indicated by the co-localisation of either DENV NS1 and NS5 protein or TBEV NS1 protein as well as dsRNA with the ER marker Protein Disulfide Isomerase-1 (PDI), indicating that assembly of the viral RC does take place at the ER but not within the ER of infected cells or BHK-21 cells transfected with a TBEV replicon respectively. As indicated by specific labeling of newly synthesized viral RNA using a MS2 tag, RNA synthesis takes place within ER derived vesicles that are ER invaginations which are connected to the cytoplasm via pore forming channels which – in WNV infected C6/36 cells- are positive for NS3 and NS5. Nascent viral RNA is probably transferred from the sites of viral RNA synthesis to the assembly site by the viral C protein. It should be noted that vesicular structures in TBEV infected BHK-21 cells do not stain positive for dsRNA, NS3 or NS5 proteins but for the viral E and NS1 protein indicating that these vesicular structures might indeed either fuse with viral RNA containing vesicles or that viral RNA is transferred.

Like other mosquitoe borne Flavivirus’, the infection of mosquitoe salivary gland lineage cells (C6/36) cells that are derived from Aedes Albopictus as well as human neuroblastoma (SK-N-SH) cells with ZIKV Paraiba 2015 induce a rearrangement and expansion of the ER with extensive localisation of the viral E protein throughout the cytoplasm and the ER at 72 hrs p.i. with dsRNA present within the ER, indicating that similar to WNV MRC MRM61C infected BHK-21 and DENV-2 infected C6/36 cells, ZIKV RNA is synthesized at the nuclear membrane and within ER invaginations both in C6/36 and SK-N-SH cells with virions intermixed with dsRNA positive vesicles thus indicating that both viral RNA synthesis and assembly of virions are taking place in close proximity during ZIKV replication.

Interestingly, the expansion of the ER, the levels of the viral E protein, the amount of dsRNA staining as well as viral titres despite being similar in pattern and amount in in C6/36 and SK-N-SH cells are delayed by 24 hrs in C6/36. One factor might be that in C6/36 cells ZIKV proteins might inhibit antiviral signalling more efficiently than in mammalian cells or that viral entry in C6/36 cells differs from mammalian thus leading to an activation of antiviral signalling pathways that delay viral replication. Interestingly, CCL-125 cells that are derived from Aedes Agypti do not support the replication of ZIKV Paraiba 2015, suggesting that indeed either viral entry or viral replication is restricted by host specific factors although female Aedes Agypti mosquitoes support the replication of ZIKV BRPE243/2015 and ZIKV /SPH/2015.

In accordance with results obtained in from primary human neuronal cells and from foetal mice brains, the infection of both SK-N-SH and C6/36 with ZIKV Paraiba 2015 induces caspase-3 dependent apoptosis as early as 48 hrs pi. which is preceded by the release of cytochrome-c and mitochondrial depolarisation. In contrast to SK-N-SH cells, C6/36 cells only exhibit a small percentage of cells undergoing apoptosis as indicated by the absence of TUNEL positive cells, indicating that ZIKV might persist in these cells. Considering that the infection of CCL-115 cells with ZIKV Paraiba induces low levels of viral titres, the inhibition of ZIKV replication in CCL-115 cells whilst not complete might prevent apoptosis (maybe by inducing mitophagy of damaged mitochondria) of infected cells and thus allow the transfer of viral particles to humans.

Following the infection with either DENV-2, Semliki Forest Virus (SFV), Bunyamwera Virus or Chikungunya Virus (CHIKV), mosquitoe salivary gland cells produce cytokines that promote the migration of monocyte derived macrophages not only in infected mosquitoes but also following a blood meal in hosts, thus facilitating the infection of those cells, suggesting that despite low viral titres the amount of viral particles might be sufficient to infect human cells, aided by cytokines.

Role of p53 in activating ZIKV induced apoptosis

In ZIKV Paraiba 2015 infected C6/36 and SK-N-SH cells the ER is enlarged indicating ER stress, although it is not clear if the enlargement is dependent on the induction of the UPR by viral proteins such as the viral E, NS4a or prM proteins (similar to DENV) resulting in the formation of autophagosomes (as observed in Varicella Zoster Virus (VZV) infected cells) or not.

As discussed before in extensio, the expression of viral proteins derived from a variety of positive strand RNA viruses including CHIKV and Japanese Encephalitis Virus (JEV) induces the ER stress response that can result in increased formation of autophagosomes in infected cells in a Beclin-1 dependent manner as well as to apoptosis in a p53 independent manner.

Furthermore, ER stress promotes the expression of the p53 isoform p53/47 thus inducing a G2 arrest by preventing the expression of p21 in HCT-116 cells, leading to a stabilisation of 14-3-3σ. Indeed, the infection of human Neural Progenitor Cells (hNPC) with ZIKV MR 766 results in a G2 arrest, suggesting that ZIKV induced ER stress induces a G2 arrest. p53 can however also be activated in ZIKV infected cells following the induction of genotoxic stress as a result of stalled replication forks. In this model, the expression of viral proteins induces the downregulation of genes associated with the repair of damaged DNA as well as of genes associated with DNA replication as discussed in a previous post and which has been demonstrated for ZIKV MR 766 infected hNPC and foetal brains infected with ZIKV SZ01. In ZIKV H/PF/2013 infected hNPC and human i90c16 (induced pluripotent human neural stem cell that are derived from IMR-90 human lung fibroblast cells) (nuclear) Ser-15 phosphorylated p53 and active Caspase-3 can be detected as early as 24 hrs p.i. suggesting that ZIKV indeed does induce the activation of p53 in neural cells. If however p53 is also induced in mosquitoe cells and if the activation of p53 also induces DNA damage repair response or if the DDR is abrogated (and induces apoptosis) has not been determined. ZIKV H/PF/2013 infected i90c16 cells do however display an increase in γH2AX (H2AX-Ser19) positive foci indicative of replicative stress as well as indicating failure to induce the repair of ds/ss DNA breaks, which is supported by findings that mouse foetal brains infected with ZIKV Mex 1-144 exhibit a decrease in Ki67 positive cells, a decrease of Histone H3-P (Ser-10) positive cells as well as a decrease in CldU positive cells with an increased cell cycle length while the percentage of TUNEL positive cells increases slightly, suggesting that ZIKV Mex 1-144 and ZIKV H/PF/2013 induce replicative stress which induces apoptosis in a subset of infected neural cells in the developing foetal brain whilst inhibiting the progression of the cell cycle including mitotic entry. It should be emphasised however that majority of infected neural cells might not undergo apoptosis as indicated by observations that the infection of foetal mice at embryonic day (E) 17.5 with ZIKV Mex 1-144 the majority of infected cells is TUNEL negative; these results are consistent with previous observations that ZIKV associated intrauterine growth restriction is more pronounced in foetal mice that were infected at E 14.5 or earlier.

In addition to genotoxic stress, ZIKV might activate p53 by sequestering Mdm2/HDM2 to the nucleolus similar to WNV Capsid protein since ZIKV antigen(s) have been described to localise to structures tentatively identified as the nucleolus and/or nuclear speckles in infected Vero cells and the DENV Capsid protein co-localises with DAXX in the nucleus, an interaction that has been linked to the induction of apoptosis by DENV C protein . In this scenario, the expression of the viral Capsid protein promotes the stabilisation of p53 by sequestering Mdm2 independent of p19^Arf(as well as by binding inhibitor of the serine/threonine phosphatase PP2A (I(2)(PP2A)), thus inducing the expression of pro-apoptotic genes including Bax and subsequently promoting apoptosis via the mitochondrial pathway. In addition, the cellular nucleolar helicase DDX56 relocalises in WNV infected cells from the nucleolus to sites of viral replication, indicating that the infection of cells with WNV does indeed cause nucleolar stress and thus activation of p53.

The question remains however if in addition to genotoxic stress and the nucleolar localisation of the viral Capsid protein the induction of ER stress by the formation of viral RC contributes to the activation of p53 or if the expression of the viral NS4A/NS4B proteins promote the degradation of the cellular Filamin A (FLNA) protein by increasing the nuclear levels of p62/SQSTM-1. In the latter scenario, the inhibition of ER stress induced autophagy by NS4A/NS4B via the inhibition of PI-3K prevents the degradation of p62/SQSTM-1 by selective autophagy, facilitating the proteasomal degradation of nuclear FLNA and thus prevent the recruitment of Rad51 to sites of DNA damage and subsequent activation of the ATR dependent pathway of DNA repair whilst promoting the initiation of (error prone) non-homologues end joining (NHEJ) dependent DNA repair. Promoting NHEJ in the absence of apoptosis therefore might lead to the accumulation of mutations that resemble those observed in genetic cases of microcephaly as described in a recent study that compared data from ZIKV H/PF/2013 infected human NPC with three different mouse models of microcephaly.

Figure: ZIKV and p62/SQSTM-1 dependent selective autophagy

The induction of ER stress by viral proteins such as NS4A, NS4B, E or prM or the formation of viral RC therefore might induce apoptosis either in a p53 dependent or independent pathway (via CHOP), promote the proteasomal degradation of FLNA (via p62/SQSTM-1) and might prevent DNA repair by facilitating the translocation of p53 into the cytoplasm by Glycogen Synthase Kinase-3 β (GSK3β) dependent phosphorylation of p53 at Ser-376. In contrast, the localisation of the viral Capsid protein to the nucleus/nucleolus might activate p53 by sequestering Mdm2/HDM2 into the nucleolus and/or nuclear speckles, thus promoting the initiation of the DNA damage response (which in the presence of other viral proteins might be inhibited downstream). In ZIKV infected cells these processes might be present at different times post infection and/or dependent on the cell line/ cell type.

ZIKV and G1 cell cycle arrest: inhibition of CDC-7?

The induction of p53 by ZIKV following the induction of the ER stress response may not only induce a G2 arrest but also delay the progression from G1 to S phase which is regulated by two kinases, CDK-2 and CDC-7. As discussed previously, the infection of foetal mouse brains with ZIKV SZ 01 decreases the expression of genes related to the initiation and progression of DNA replication and the infection of hNPC with ZIKV MR 766 decreases the expression of CDK-2, suggesting that ZIKV infection can indeed at least delay the onset and certainly the progression of S phase which –as described above- is in accordance with recent results obtained from ZIKV Mex 1-14 infected mouse brains.

CDC-7 is the catalytic subunit of the Dbf4b dependent Ser/Thr kinase whose expression is inhibited by p53 via p53 dependent upregulation of miR-192/215 as well as the E3 ubiquitin ligase Fbxw7β following the treatment of IMR90 fibroblast cells with Doxorubicin, low doses of Actinomycin D or γ-irradiation. The activation of p53 in ZIKV infected cells therefore might delay the progression of G1 to S phase by at least two mechanisms; first by downregulating the expression of CDK-2 and second by inhibiting CDC-7 via the expression of miR-192/195 and Fbxw7β.

ZIKV and CD95/Fas: sensitising of infected cells to extrinsic apoptosis?

The DENV Capsid protein not only localises to the nucleus, but also co-localises with Death Associated Protein-6/DAXX in PML nuclear bodies and induces the expression of CD95/ TNFRSF6/Fas, thus inducing apoptosis of HepG2 cells expressing wt DENV Capsid protein but not DENV Capsid protein variants that do not localise to the nucleus due to K73A/K74A or R85A/K86A mutations via the induction of Fas dependent activation of Caspase-3 and -9 following treatment with an anti-Fas/CD95 antibody.

It is therefore possible that the infection of neuronal and non-neuronal cells with ZIKV may trigger the expression of CD95 at the cell surface thus sensitizing infected cells to apoptosis induced by the activation of CD95/FasR either in an autocrine manner and/or by FasL expressing activated T-lymphocytes.

Figure: ZIKV Capsid protein: Relocalising DAXX to nuclear bodies?

In summary, ZIKV may induce the activation of p53 by various pathways, including the induction of genotoxic stress as a result of stalled replication forks, the induction of nucleolar stress by relocalising or sequestering nucleolar proteins and the activation of the ER stress response as a result of the formation of viral replication centers, thus resulting in extending the length of the cell cycle by prolonging S phase and delaying the progression of G1 to S phase followed by an arrest in G2 phase of the cell cycle without progressing to Prophase. The localisation of the viral Capsid protein, similar to the Capsid protein of WNV and DENV, might contribute to the activation of intrinsic and extrinsic apoptotic pathways by not only redistributing nucleolar proteins such as Mdm2/HDM2 but also by sequestering DAXX in PML bodies respectively thus sensitizing infected cells to CD95L induced apoptosis. Further research however is needed to elucidate the contribution of individual viral proteins to these processes and the involvement of additional cellular proteins. One possibility is that ZIKV infection prevents the degradation of FAP-1 by inhibiting p62/SQSTM-1 dependent selective autophagy of FAP-1 thus preventing Fas induced apoptosis even in the presence of FasL whereas late in infection intrinsic apoptosis is activated by NS4B as well as the ER stress response and p53. The expression of the viral NS4A and NS4B might not only inhibit PI3K mediated regulation of autophagy but also promoting PI3K mediated apoptosis induced by p53 by preventing Mdm2 mediated degradation of p53, thus enhancing the effects of Mdm2 sequestration by the viral Capsid protein. Further experiments are however warranted to confirm this hypothesis.

Figure: Inhibition of Fas dependent apoptosis by stabilising Fap-1?

In addition to investigating the individual contribution of viral proteins to ZIKV induced activation of p53 and ZIKV induced apoptosis, differences between mosquitoe and human derived cell lines have to be examined as well since the infection of CCL-125 cells with ZIKV Paraiba 2015 does not produce high viral titres and infected cells do not undergo viral induced apoptosis, suggesting that either p53 is not activated in CCL-125 cells or that anti-apoptotic pathways are activated.

Figure: ZIKV and the induction of p53

Virology tidbits

Virology tidbits

Wednesday, 14 December 2016

ZIKV replication and the induction of p53: a tale of three places ?