The astrovirus N-terminal nonstructural protein anchors replication complexes to the perinuclear ER membranes

An essential aspect of positive-sense RNA virus replication is anchoring the replication complex (RC) to cellular membranes. Positive-sense RNA viruses employ diverse strategies, including co-translational membrane targeting through signal peptides and co-opting cellular membrane trafficking components. Often, N-terminal nonstructural proteins play a crucial role in linking the RC to membranes, facilitating the early association of the replication machinery. Astroviruses utilize a polyprotein strategy to synthesize nonstructural proteins, relying on subsequent processing to form replication-competent complexes. This study provides evidence for the perinuclear ER membrane association of RCs in five distinct human astrovirus strains. Using tagged recombinant classical human astrovirus 1 and neurotropic MLB2 strains, we establish that the N-terminal domain guides the ER membrane association. We identified di-arginine motifs responsible for the perinuclear ER retention and formation of functional RCs through mutational analysis of the N-terminal domain in replicon and reverse genetics systems. In addition, we demonstrate the association of key components of the astrovirus replication complex: double-stranded RNA, RNA-dependent RNA polymerase, protease, and N-terminal protein. Our findings highlight the intricate virus-ER interaction mechanism employed by astroviruses, potentially leading to the development of novel antiviral intervention strategies.


Introduction
Astroviruses are a family of small non-enveloped (+)ssRNA viruses that infect a wide range of mammalian and avian species.The Astroviridae family includes two genera: Mamastrovirus and Avastrovirus.Humans are susceptible to the classical (HAstV1-8) and genetically divergent non-classical VA/HMO and MLB clades of astroviruses, which cause mild to severe diseases depending on the age and health of an individual [1,2].Classical astroviruses were discovered in the 1970s and represent one of the leading causes of gastroenteritis in children, the elderly, and those who are immunocompromised.Two groups of non-classical astroviruses were identified much later and are often associated with neurological diseases, such as meningitis and encephalitis [3].Despite the high zoonotic potential and non-gastrointestinal tropisms [2,4,5], there are still significant gaps in our understanding of the biology of human astroviruses and disease progression.Besides, astrovirus infections are often under-reported despite their high prevalence [6].Currently, no vaccines or drugs against astroviruses are available.Therefore, it is important to investigate their replication mechanisms to effectively control future outbreaks.
Replication of (+)ssRNA viruses requires RdRp and other nonstructural proteins that form the viral replication complex.These viruses exploit host membranes to assemble replication complexes (RCs), protect double-stranded RNA (dsRNA) intermediates, and segregate replicating RNAs from translating viral mRNAs [12][13][14].In astroviruses, the components of the replication complex contain RdRp, protease, and VPg-three well-characterized enzymatic units [7,11,15].However, membrane-anchoring and retention strategies are still unknown.The predicted transmembrane (TM) domain located at the N-terminal part of the astrovirus nonstructural protein suggests its involvement in membrane association.Recently, ER-derived membranes were implicated in the formation of double-membrane vesicles during astrovirus infection [16].However, the exact location, ER membrane specificity, and viral proteins responsible for this remain to be characterized.In addition to the TM domain, the N-terminal domain of the astrovirus polyprotein contains a putative RNA helicase motif [1].However, its poor conservation and incomplete motif integrity raise questions regarding the functional significance of the proposed helicase.
Here, we demonstrate the function of the N-terminal protein in two astrovirus genotypes, HAstV1 and MLB2.This small membrane protein drives the formation of replication complexes in tight association with perinuclear ER membranes, which is a key feature in a wide range of astroviruses (HAstV1, HAstV4, MLB1, MLB2, and VA1).We also found that ER retention and the astrovirus replication complex's function depend on the di-arginine motif located in the N-terminal protein.In the future, an improved understanding of astrovirus replication complex formation may represent a promising therapeutic strategy for controlling astrovirus infections in young children, immunocompromised individuals, and farm animals.

Results
The putative helicase domain is not conserved and is dispensable for HAstV1 replication RNA viruses harbor several cis-acting elements that play essential roles in viral RNA replication, translation and assembly of virions [17].During viral genome replication, these structured RNA elements require RNA helicases or chaperones to facilitate the unwinding and remodeling of double-stranded RNA structures.Helicases use energy derived from NTP hydrolysis to catalyze dsRNA unwinding and contain several conserved motifs, including NTP-binding Walker A and Walker B motifs [18].Several RNA viruses encode NTPase/RNA helicases, which assist in the unwinding of dsRNA replicative intermediates during virus replication [19].Interestingly, astrovirus genomes do not have features that can be attributed to a fully functional NTPase/helicase.The only Walker A-like motif can be found in classic human astroviruses, but not in non-classical MLB and VA genotypes (Fig 1A ), suggesting poor conservation and questioning the presence of virus-encoded helicases in astroviruses.In addition, this domain is predicted to be attached to the membrane and separated from the rest of the replication module (Fig 1B ), which is an unusual localization for RNA-processing enzymes.To address this question, the putative Walker A motif in HAstV1 was mutated from GKT to GAT in the context of replicons and infectious viruses.In the replicon system, a minor reduction in activity was observed (Fig 1C), and in the GKT-to-GAT recombinant virus, no significant differences were observed (Fig 1D), whereas mutation of functional GKT/GKS motifs in other viruses is usually lethal [20,21].This further indicates that the GKT motif is unlikely to play a functionally important role in the HAstV1 replication.In addition, several viral NTPases/helicases, including enterovirus 2C ATPase [22], are inhibited by guanidine hydrochloride (GuHCl), resulting in NTPase-specific inhibition of replication [22].Using HAstV1-based and enterovirus-based replicon systems, we examined the ability of GuHCl to inhibit replication.Unlike previously characterized inhibition in enteroviruses, astrovirus replication was not strongly affected by GuHCl treatment (Fig 1E and 1F).The 20-30% decrease in astrovirus replicon activity is at least partially caused by the decrease in cell confluency in the presence of GuHCl at later time points (Fig 1E), whereas enterovirus replication was efficiently inhibited to ~0.1% (Fig 1F ), further confirming that GuHCl-sensitive replicase components are absent in astroviruses.Considering the poor conservation (Fig 1A ) and lack of functional effect (Fig 1C -1E) of the putative Walker A motif in HAstV1-based assays, it is plausible to suggest that cellular helicases can be recruited by astrovirus replication complexes (RCs) as an alternative strategy utilized by several virus families [23].Therefore, the remaining GKT motif may represent an evolutionarily lost NTPase/ helicase rather than a functional unit within the astrovirus genome.

Role, accumulation and localization of NTD in astrovirus replication
To investigate the role(s) of the NTD in virus replication, we engineered HA-tagged astroviruses using classical HAstV1 [8] and neurotropic MLB2 [24] reverse genetics systems by placing an HA-tag sequence in the predicted disordered region between the folded N-terminal part of the protein and the first transmembrane helix.If N-terminal cleavage occurs at the predicted signal peptidase (SPase) cleavage site, the molecular weight of the N-terminal HAtagged astrovirus protein is predicted to be 21-22 kDa (Fig 2A).MLB2-HA virus was successfully rescued, and growth kinetics was assessed in Huh7.5.1 cells, displaying a minor delay in growth (Fig 2B ), which is common for tagged viruses with small genomes.The rescue of HAstV1-HA was successful following the transfection in BSR cells but not on passaging in Caco2 cells (Fig 2C).This could indicate a defect specific to infection in Caco2 cells and potential differences in antiviral responses between different cell lines.The accumulation of ~14 kDa HA-tagged product was detected in HA-tagged but not in wt astrovirus-infected cells (Fig 2D and 2E).Due to the differences in predicted and observed sizes of HA-tagged proteins, we analyzed the same samples in Bis-Tris acrylamide gels, where ~20 kDa HA-tagged products were detected, confirming their aberrant mobility in Tris-glycine acrylamide gel systems (Fig 2F) [25].Capsid accumulation showed a slight delay in HA-tagged MLB2 samples when compared to the wt virus, consistent with the delay in virus growth (Fig 2D).Next, we analyzed the same viral products using immunofluorescence of the virus-infected cells at 24 hpi.Interestingly, we observed a distinct perinuclear localization of HA-tagged protein for both viruses.In contrast, the capsid protein was dispersed throughout the cytoplasm (Fig 2G and 2H), most likely due to its excess and less likely membrane association.Notably, the cytoplasmic distribution of the capsid was more evident in Huh7.

Astrovirus replication complexes are associated with endoplasmic reticulum membranes, close to the nuclear periphery
Similar to other (+)ssRNA viruses, the replication of human astroviruses is associated with the host endoplasmic reticulum (ER) by forming double-membrane vesicles [16].However, the mode of RC formation and its retention within the ER membranes still needs to be determined.To systematically investigate the cellular localization of astrovirus replication sites, we used five available astrovirus strains (HAstV1, HAstV4, MLB1, MLB2 and VA1) and infected two different cell lines (Huh7.5.1 and Caco2) that support selected astrovirus replication and spread.Staining with anti-dsRNA antibody, a hallmark of (+)ssRNA replication sites, revealed perinuclear localization of RCs in Caco2 (   4).These results suggest the specific targeting of RCs to the perinuclear ER membranes and the possible involvement of NTD in this process (Fig 2G and 2H).

During astrovirus infection, the NTD is colocalized with replicating RNA and perinuclear ER membranes
To investigate the localization of NTD during astrovirus infection, Caco2 and Huh7.5.1 cells were infected with HAstV1-HA and MLB2-HA, respectively.As expected, a strong overlap between dsRNA-and NTD-specific signals was observed (mean PCC > 0.7, Fig 5), confirming the previous observations.Consistent with the staining obtained with wt astrovirus strains (Figs 3 and 4), HA-tagged NTD-specific localization followed the same overlapping perinuclear ER-(mean PCC > 0.6) and weak lamin-specific (mean PCC < 0.5) pattern (Fig 5).

The NTD is colocalized with protease and RdRp during astrovirus infection
First, we developed a set of antibodies for specific immune detection of the following astrovirus proteins: VA1 capsid protein (CP), HAstV1 protease and RdRp, and MLB2 protease.The folded region of the indicated proteins possessing a C-terminal 8×His-tag was used for bacterial expression and affinity purification, resulting in homogeneous proteins of the expected sizes (Fig 6A ).Each purified recombinant protein was used for the production and affinitybased purification of antibodies [24].VA1 CP-specific antibody was used to titrate VA1 stocks and identify VA1-infected cells by immunofluorescence (Figs 3, 4 and 6B), whereas antibodies against nonstructural proteins were used to visualize replication complex-specific components (Fig 6B and 6C).
To identify the components of RCs, we performed several pairwise immunofluorescent stainings, where a combination of antibodies would allow such an approach (e.g.combination of mouse and rabbit-derived antibodies).The pattern of VA1 and MLB2 capsid proteins was distinct from dsRNA-specific staining (PCC < 0.6, Fig 6B ), following the same trend as NTD-CP co-staining (Fig 2G and 2H).
Out of several nonstructural proteins, RdRp and protease are usually found within RCs of (+)ssRNA viruses [26], performing RNA synthesis and cleavage of the nonstructural polyprotein, respectively.Due to the availability of antibody combinations, we could test several possibly co-localized pairs of proteins.The strong co-localization is observed between all three replication components: dsRNA, protease and RdRp (PCC > 0. To further investigate co-localization of dsRNA, protease, RdRp and NTD, we adapted a proximity ligation assay (PLA) that permits the detection of protein-protein or protein-RNA interactions in situ (at distances < 40 nm).In agreement with previous observations (Fig 6C ), we detected positive PLA signals for RdRp-dsRNA (in HAstV1 and HAstV1-HA infected cells) and protease-dsRNA (in HAstV1, HAstV1-HA, MLB2 and MLB2-HA infected cells), indicating proximity of dsRNA, RdRp and protease.In concordance with immunostaining of HA-tagged viruses (Fig 5), the specific NTD-protease/NTD-RdRp PLA signal was detected in HAstV1-HA-infected Caco2 cells, and NTD-protease PLA foci were observed both in HAstV1-HA and in MLB2-HA-infected cells (Fig 6D and 6E).

PLOS PATHOGENS
Taken together, we confirm the association of the key components of RNA RCs: dsRNA, protease, RdRp, and NTD (Figs 5 and 6B-6E).
Overall, these results confirmed the proximity of the astrovirus NTD with RCs, raising an important question: what drives the NTD to the perinuclear ER membranes and how does this affect astrovirus replication?
Di-arginine motifs in NTD are responsible for perinuclear ER localization and astrovirus replication ER targeting of RCs in (+)ssRNA viruses can be achieved through various mechanisms [27].In astroviruses, the NTD, followed by transmembrane helices, represents a unique combination to target the nonstructural polyprotein to the correct position within the infected cell.We explored the power of molecular mimicry used by many viruses to identify the potential residues responsible for ER targeting that were predicted using the ELM web server (http://elm.eu.org/) [28,29].The search revealed the presence of conserved di-arginine motifs (Fig 7A ), which were defined by two consecutive arginine residues (RR) or with a single residue insertion (RXR).This motif is characteristic of several membrane proteins with ER localization and allows for correct folding and membrane association.The functional motif needs to be exposed to the cytoplasm and requires distinct proximity to the TM region, thus fulfilling the criteria for a predicted astrovirus nonstructural polyprotein (Figs 1B and 7A) [30].
The structure of astrovirus NTD is unknown, so we performed a structure prediction using the AlphaFold 3 server [31][32][33] and mapped identified di-arginine residues and the position of the HA-tag in HAstV1 and MLB2 NTD (Fig 7B and 7C).Both di-arginine motifs were mapped to the surface-exposed cytoplasmic domain of the NTD, suggesting that these residues could have a functional impact on the properties of NTD.The transmembrane helix was predicted at the end of NTD and should not be affected by mutations or the inserted HA-tag (Fig 7B and  7C).
To investigate the function of the di-arginine motifs of NTD in the virus life cycle, arginine-to-alanine mutations were introduced into replicons and full-length infectious clones of HAstV1 and MLB2 astroviruses (Fig 7D).Replication was drastically decreased in all mutant replicons (>95%, Fig 7E and 7F) and all di-arginine mutant viruses were not viable (Fig 7G and 7H), confirming the critical role of both di-arginine motifs in the virus life cycle.
To link the replication defect to the specific ER localization performed by the NTD, the same subset of mutations (Fig 7A) was introduced into a mammalian expression vector that encodes only 187 (MLB2) or 190 (HAstV1) amino acid residues of the HA-tagged NTD (Fig 8A ).All NTD variants containing mutated N-terminal di-arginine motifs showed reduced amounts of protein (Fig 8B and 8C), suggesting that the stability of the protein can be dictated by the correct ER targeting mediated by the N-terminal di-arginine (NR) motif.Poor expression of NR and 4R mutants could be explained by protein degradation due to mislocalization and/or misfolding of the protein and thus precluded their usage in the following immunofluorescent-based studies.
The perinuclear localization of overexpressed wt NTD-HA (Fig 8D and 8E, top panels) recapitulated the NTD-HA localization in virus-infected cells (Fig 5 ), suggesting the role of this domain in the correct anchoring of astrovirus RCs.Perinuclear ER staining was drastically altered in C-terminal arginine-to-alanine (CR) mutants, resulting in a diffuse ER pattern in cells expressing both HAstV1 and MLB2 NTD-HA protein, despite the TM helix and N-terminal di-arginine motif remaining unmodified.The changes in the overlap signal between NTD-HA and ER were also analyzed using Manders's coefficient (M 2 ), which further confirmed the redistribution of NTD-HA signal in the population of analyzed cells (Fig 8D and  8E).
Taken together, these results (Figs 7 and 8) provide evidence for the distinct features of Nand C-terminal di-arginine motifs and the functional role of the NTD in the membrane association of astrovirus RC, a prerequisite for efficient replication.The functional significance of other viral/host proteins in this process remains to be characterized.

Discussion
In this study, we investigated the role of the astrovirus N-terminal protein in ER membrane tethering and the formation of functional RCs.We demonstrate that the putative helicase is unlikely to be a functional unit within astrovirus RCs.Instead, the di-arginine signature-driven ER membrane localization and replication represents a key role of the N-terminal protein in the astrovirus replication cycle.
Numerous positive-sense RNA viruses encode RNA helicases, while others depend on cellular counterparts in their absence [19,23].Interestingly, astrovirus genomes encode a single

PLOS PATHOGENS
Walker A-like motif that can be found in classic human astroviruses but not in the newly emerging human (Fig 1A) and avian astroviruses [34].Consistent with poor conservation, mutation of the GKT motif was not detrimental to HAstV1 replication and life cycle (Fig 1).The absence of evidence of a functional helicase in the genome of astroviruses suggests dependence on host proteins with NTPase/helicase activity.Notably, proteomic analysis of HAstV8-infected Caco2 cells showed enrichment of membrane-only fractions with the cellular RNA helicase DDX23, detected alongside viral RdRp and protease, established components of RCs [35].Additionally, siRNA-mediated depletion of DDX23 significantly decreased virus replication [35], indicating the host helicase dependence of astrovirus replication.In a recent transcriptomic analysis of HAstV1-infected Caco2 cells, cellular helicases HELZ2 and DDX58 transcripts were found to be upregulated in infected cells [16], providing more candidates for the host helicases that can be involved in the RNA replication process.
The correct formation of (+)ssRNA virus RCs is mediated by the recruitment of cellular membranes that prevent immune detection of viral RNA, separate the processes of replication and translation, and increase the local concentration of active replication components, both viral and cellular [12][13][14].Viruses employ a range of strategies to ensure the correct association of replication-competent virus-host replication machinery.Co-opting the ER for this purpose has been described for numerous RNA viruses including the Flaviviridae, Coronaviridae, Picornaviridae [36] families, and has also been suggested for astroviruses [16,35,37].We confirm this localization and identify perinuclear ER membranes as the preferential RNA replication site (Fig 3), similar to several other RNA viruses [38][39][40].Besides ER, many RNA viruses have evolved to recruit alternative subcellular membranes, such as mitochondrial, Golgi, endosomal, and lysosomal membranes [41,42].
The proposed involvement of the N-terminal part of the nonstructural polyprotein in the positioning of the whole replication complex is a convenient strategy to ensure translocation across the ER membrane which can be mediated by the N-terminal signal/sorting peptide [43].In astroviruses, the N-terminal protein, followed by transmembrane helices, is ideally positioned to target the nonstructural polyprotein to cellular membranes.Of the multiple strategies of ER membrane integration [30,43], we found that astroviruses most likely use diarginine-based ER-sorting motifs [44] to anchor and assemble RCs.This strategy is used by several mammalian (lip35, GABA B , Kir6.2) [30], plant (AtGCSI) [45], and viral (hepatitis B virus S) [46] proteins for membrane trafficking [30].We demonstrated that two predicted pairs of N-terminal di-arginine motifs are essential for virus replication and are responsible for the perinuclear ER localization of the N-terminal protein (Figs 6-8).This links two crucial functions together: the correct localization of the RC components and their activity (Fig 9).As many viruses exploit the ER during infection, pharmacological strategies aimed at disrupting virus-ER traffic/interactions or ER morphogenesis should, in principle, lead to the generation of broad-spectrum antiviral targets.
Perinuclear localization of replication complexes has been reported for several RNA viruses.In particular, perinuclear ER-specific replication is a hallmark of the replication of brome mosaic virus [38,47], flaviviruses [39,48], and several other virus families.The mechanisms of perinuclear ER targeting represent a diverse set of approaches based on molecular mimicry of ER-residing proteins and recruitment of related ER-associated proteins [49].We demonstrate a similar strategy employed by astroviruses thus expanding the list of viral ER hijackers.An outstanding question is how host factors cooperate with virus-encoded di-arginine motifs to construct the ER-derived replication organelle and how virus assembly and maturation are linked to the modified cellular membranes.
Altogether, our findings provide insight into astrovirus replication and the involvement of the N-terminal nonstructural protein in the correct positioning of RCs.Future studies will address the exact host components involved in the formation of active replication organelles and describe the topology and functionality of viral nonstructural proteins.Understanding virus replication mechanisms will potentially lead to the development of future therapeutics.

Cells
BSR cells (single clone of BHK-21 cells) were maintained at 37˚C in DMEM supplemented with 5% fetal bovine serum (FBS), 1 mM L-glutamine and antibiotics.Caco2 and Huh7.5.1 cells (Apath, Brooklyn, NY) were maintained in the same media supplemented with 10% FBS and non-essential amino acids.All cells were tested mycoplasma negative throughout the work (MycoAlert Mycoplasma Detection Kit, Lonza).

Plasmids and cloning
Reverse genetics and replicons of the human astroviruses MLB2 [24] and HAstV1 [8,50] were previously described.HA-tagged human astroviruses (MLB1 and HAstV1) were generated by site-directed mutagenesis.The coding sequence of the HA-tag was inserted in the nsP1a as shown in Fig 2A .The resulting infectious clones were designated as HAstV1-HA and MLB2-HA.
For mammalian expression of the MLB2 and HAstV1 nsP1a N-terminal domain, the coding sequence of the HA-tagged N-terminal domain from corresponding HA-tagged virus was inserted into vector pCAG-PM [8] using AflII and PacI restriction sites.The resulting constructs-designated pCAG-HAstV1-NTD-HA and pCAG-MLB2-NTD-HA-were confirmed by sequencing.All mutations were introduced using site-directed mutagenesis and confirmed by sequencing.

Purification of His-tagged astrovirus proteins and generation of specific antibodies
The VA1 CP, HAstV1 Pro, HAstV1 RdRp and MLB2 Pro proteins were produced in Rosetta 2 (DE3) cells (Novagen) cultured in 2×YT media with overnight expression at 18˚C induced with 0.4 mM IPTG.The proteins were purified first by immobilized metal affinity chromatography using PureCube Ni-NTA resin and then by affinity chromatography using amylose resin (NEB).MBP fusion tag was removed by the cleavage with TEV protease (produced inhouse).Proteins were further purified by heparin chromatography using HiTrap Heparin HP or anion-exchange chromatography using HiTrap Q-HP 5 ml column (Cytiva) and, finally, by size exclusion chromatography using a Superdex 200 16/600 column (Cytiva).Protein solution in 50 mM Na-phosphate pH 7.4, 300 mM NaCl, 5% glycerol was concentrated to 2 mg/ml and used for immunization.
Antibodies against indicated proteins were generated in rabbits using 5-dose 88-day immunization protocol.Sera were used for specific affinity purification, followed by purification of specific IgG fractions (BioServUK Ltd).

Virus growth curves
Multistep growth curves were performed using an MOI of 0.1, with infections performed in triplicates.Equal amounts of media-derived samples were collected at 0, 24, 48, 72 and 96 hpi and titrated.The titers were determined as infectious units per ml (IU/ml).

Electroporation of plasmid DNA
To analyze overexpressed proteins, electroporation of Huh7.5.1 cells was performed using 3 μg of plasmid DNA in complete media at 220 V and 975 μF using a Bio-Rad Gene Pulser, in the presence of 5 mM NaBes (N,N-Bis(2-hydroxyethyl)-2-aminoethanesulfonic acid sodium salt) and 0.2 mg/ml salmon sperm carrier DNA.After electroporation, cells were split between immunofluorescent and western blotting analyses and incubated for 24 hours.

Proximity ligation assay
Caco2 and Huh7.5.1 cells were seeded on IBIDI wells and infected with MOI 1 for 24 hours (HAstV1 and HAstV1-HA) or MOI 0.3 for 30 hours (MLB2 and MLB2-HA).The cells were then fixed with 4% PFA for 30 minutes and permeabilized with 1% (Caco2) or 0.25% (Huh7.5.1) Triton X-100 in PBS.The cells were incubated with combinations of mouse-and rabbit-derived antibodies, followed by incubation with two PLA probes (DUO92101, Sigma Aldrich) for 1 hour at 37˚C, ligation for 30 minutes, and signal amplification for another 100 minutes at 37˚C.The wells were then covered with Duolink In Situ Mounting Medium with DAPI and imaged under Zeiss LSM700 microscope using an oil-immersion 63× objective.

Replicon assays
Linearized replicon-encoding plasmids were used to generate T7 RNAs using T7 mMESSAGE mMACHINE Transcription kit (ThermoFischer, AM1344) according to the manufacturer's instructions, purified using Zymo RNA Clean & Concentrator kit and quantified.Cells were transfected in triplicate with Lipofectamine 2000 reagent (Invitrogen), using previously described reverse transfection protocol [24].Three independent experiments, each in triplicate, were performed to confirm the reproducibility of the results.

Protein structure prediction
The NTD 3D structure was predicted using AlphaFold 3, a neural network-based model that predicts protein three-dimensional structures from sequence, even where no similar structure is known [31][32][33].The confidence of prediction was quantified by pLDDT, the predicted local distance difference test on the Cα atoms, with >70% for most NTD regions.The alignment of two structures and amino acid residue positions were visualized using PyMOL (https://pymol.org/).Membrane topology was predicted using several online servers to ensure consistency.
Fig 3A) and Huh7.5.1 (Fig 3B) cells.The dsRNA-specific signal overlapped with ER-specific staining only very close to the nuclear periphery, but

Fig 8 .
Fig 8. Di-arginine motifs define the efficient expression and perinuclear ER localization of NTD.(A) Schematic representation of CAG promoter containing mammalian expression vector used to over-express HA-tagged NTD in Huh7.5.1 cells.(B-C) Huh7.5.1 cells were electroporated with pCAG plasmids expressing wt and mutant NTD-HA and analyzed by western blotting with anti-HA and anti-tubulin antibodies.Lysates obtained from infected cells were used as a control.(D-E) Huh7.5.1 cells were electroporated with pCAG plasmids expressing wt and mutant NTD-HA and analyzed by immunofluorescence.Representative confocal images of fixed and permeabilized cells visualized for HA-tag (NTD, magenta) and ER (calnexin, green).Nuclei were stained with Hoechst (blue).Scale bars are 25 μm.Averaged intensity profiles of NTD (magenta), ER (green) and nuclear staining (blue) were obtained using ImageJ software, along a straight line spanning from the nucleus to the plasma membrane of 17-20 transfected cells.The changes in the overlap signal between NTD-HA and ER were compared using Manders coefficient (M 2 ).https://doi.org/10.1371/journal.ppat.1011959.g008

Fig 9 .
Fig 9.The proposed mechanism of NTD-directed formation of astrovirus replication complex.The translation of astrovirus nonstructural polyprotein begins with the NTD, that is anchored to ER membranes through cooperative interactions of charged arginine residues.Non-structural proteins including viral RdRp, protease, and NTD, form the replication complex in tight association with perinuclear ER membranes and other virus/host proteins.Newly synthesized viral genomic and subgenomic RNAs (gRNA and sgRNA) are used for translation, replication, and/or packaging.https://doi.org/10.1371/journal.ppat.1011959.g009 or Manders coefficient (Fig8).