Invasion and Propagation of White Spot Syndrome Virus: Hijacking of the Cytoskeleton, Intracellular Transport Machinery, and Nuclear Import Transporters

ABSTRACT

The pathogenesis of white spot syndrome virus (WSSV) is largely unclear. In this study, we found that actin nucleation and clathrin-mediated endocytosis were recruited for internalization of WSSV into crayfish hematopoietic tissue (Hpt) cells. This internalization was followed by intracellular transport of the invading virions via endocytic vesicles and endosomes. After envelope fusion within endosomes, the penetrated nucleocapsids were transported along microtubules toward the periphery of the nuclear pores. Furthermore, the nuclear transporter CqImportin α1/β1, via binding of ARM repeat domain within CqImportin α1 to the nuclear localization sequences (NLSs) of viral cargoes and binding of CqImportin β1 to the nucleoporins CqNup35/62 with the action of CqRan for docking to nuclear pores, was hijacked for both targeting of the incoming nucleocapsids toward the nuclear pores and import of the expressed viral structural proteins containing NLS into the cell nucleus. Intriguingly, dysfunction of CqImportin α1/β1 resulted in significant accumulation of incoming nucleocapsids on the periphery of the Hpt cell nucleus, leading to substantially decreased introduction of the viral genome into the nucleus and remarkably reduced nuclear import of expressed viral structural proteins with NLS; both of these effects were accompanied by significantly inhibited viral propagation. Accordingly, the survival rate of crayfish post-WSSV challenge was significantly increased after dysfunction of CqImportin α1/β1, also showing significantly reduced viral propagation, and was induced either by gene silencing or by pharmacological blockade via dietary administration of ivermectin per os. Collectively, our findings improve our understanding of WSSV pathogenesis and support future antiviral designing against WSSV.

IMPORTANCE As one of the largest animal DNA viruses, white spot syndrome virus (WSSV) has been causing severe economical loss in aquaculture due to the limited knowledge on WSSV pathogenesis for an antiviral strategy. We demonstrate that the actin cytoskeleton, endocytic vesicles, endosomes, and microtubules are hijacked for WSSV invasion; importantly, the nuclear transporter CqImportin α1/β1 together with CqRan were recruited, via binding of CqImportin β1 to the nucleoporins CqNup35/62, for both the nuclear pore targeting of the incoming nucleocapsids and the nuclear import of expressed viral structural proteins containing the nuclear localization sequences (NLSs). This is the first report that NLSs from both viral structure proteins and host factor are elaborately recruited together to facilitate WSSV infection. Our findings provide a novel explanation for WSSV pathogenesis involving systemic hijacking of host factors, which can be used for antiviral targeting against WSSV disease, such as the blockade of CqImportin α1/β1 with ivermectin.

INTRODUCTION

White spot syndrome virus (WSSV) disease has caused huge economic losses in crustacean farming, particularly in the shrimp and crayfish industries, and certainly also poses ecological risks for aquaculture (1). As the sole member of the genus Whispovirus of the Nimaviridae family, WSSV is an enveloped, double-stranded DNA virus with a large genome of ~300 kbp packaged within the capsid (2). Intriguingly, the assembly of WSSV progeny virions, particularly the encapsulation of the nucleocapsid, i.e., the enclosure of the viral genome, with viral envelope proteins is completed entirely in the crustacean cell nuclei. This process is unlike the process for other typical enveloped DNA viruses, such as the insect virus Autographa californica nucleopolyhedrovirus, whose nucleocapsid acquires an envelope via budding from the plasma membrane (3, 4), and human herpesvirus, whose nucleocapsid obtains an integrated envelope via budding from cytoplasmic organelles (4). Currently, only 30% of the open reading frames (ORFs) of the WSSV genome are functionally annotated to encode structural proteins and nonstructural proteins. For example, each WSSV particle contains more than 40 structural proteins, among which some major structural proteins have been identified, including the capsid proteins VP664 and VP15 and the envelope proteins VP28, VP26, VP24, and VP19. Although the known capsid of WSSV contains relatively few proteins compared with the envelope (5), the capsid proteins involved in WSSV infection are still poorly characterized. For example, VP664, the most abundant component in the capsid, is distributed around the periphery of the nucleocapsid; hence, it is speculated that VP664 plays a role in the nuclear targeting as well as the disassembly of the incoming nucleocapsid to release the WSSV genome into the cell nuclei, but experimental evidence is still lacking (6).

Usually, actin assembly provides a force to drive vesicle internalization and actin polymerization during clathrin-mediated endocytosis (CME), mediated by the Arp2/3 complex, which plays a critical role in the intracellular motility and vesicular trafficking of pathogens (7). We previously found that the actin cytoskeleton is involved in the invasion of WSSV (8); furthermore, endocytosed WSSV virions were delivered into endosomal vesicles; in this process, the viral envelope is detached in a reduced pH environment, after which the nucleocapsid enclosing the viral genome moves from the endosome into the cytoplasm in the early stage of infection (9). It is known that microtubules can transport intracellular virions in vertebrates (10) with both the envelope within endosomes and with the nucleocapsid after transport from the endosome to the perinuclear membrane. However, very little is known about the intracellular transport of WSSV after its internalization or about nucleocapsid transport after its exit from endosomes; in particular, little is known about how the viral genome is delivered into crustacean cell nuclei after its exit from the nucleocapsid for subsequent viral gene replication.

Upon analyzing the WSSV structural proteins deduced from the viral genome, we noticed that some of the major capsid proteins, such as VP664 and VP15, and a few envelope proteins contained nuclear localization signal sequences (NLSs); in contrast, some of the capsid proteins, and most of the known envelope proteins, did not contain any NLS. NLSs are recruited to introduce the genomes of DNA viruses, via nuclear pore complexes (NPCs), into the host cell nucleoplasm for replication in vertebrates (11). Meanwhile, the nuclear transport of proteins carrying NLS through nuclear envelope-embedded NPCs is mediated by members of the importin superfamily, such as importin α and importin β (11, 12); as an adaptor molecule, importin α binds to the cargoes or macromolecules with NLS and in turn binds to importin β, which is responsible for docking the imported cargo to a distal site on the nuclear pore by binding with nucleoporins to facilitate nuclear translocation in a manner assisted by the small GTPase Ran via binding to a class of importin β-like proteins (13–15). Consequently, the uncoating of the viral genome from the nucleocapsid, which is composed of capsid proteins with NLS taken as nuclear import cargoes, is triggered by the protein interactions among the viral nucleocapsid, importin β, Ran, and NPCs (16). In particular, the nuclear localization of viral proteins is dependent on both their own NLSs and cellular transport proteins (17). However, whether these NLSs of WSSV can be recognized by importin receptors for nuclear targeting of the incoming nucleocapsid or nuclear import of the expressed WSSV structural proteins either with or without NLS is truly an enigma in crustaceans, particularly in terms of the encapsulation of the WSSV nucleocapsid with an envelope for progeny assembly, which is achieved solely in crustacean cell nuclei.

To reveal the pathogenesis of WSSV, in this study, the key aspects of both endocytosis and intracellular transport of WSSV via the cytoskeleton, the delivery of the viral genome into the host cell nuclei, and the nuclear import of expressed viral structural proteins, via NPCs mediated by importins, were determined in hematopoietic tissue (Hpt) cell cultures from the red claw crayfish Cherax quadricarinatus, one of the crayfish cell models suitable for WSSV progeny assembly in vitro at present (18, 19). Furthermore, the possibility of anti-WSSV protection via importin blockade was assessed in vivo in crayfish. Our findings improve our understanding of WSSV pathogenesis and support future anti-WSSV target designing.

RESULTS

Internalization of WSSV is facilitated by actin nucleation promoted by the CqArp2/3 complex in Hpt cells.

As indicated by a monoclonal antibody against the viral envelope protein VP28, WSSV virions colocalized considerably with protruding filopodia (Fig. 1A, a) and the cytoplasmic skeleton of F-actin (Fig. 1A, b) at 0.5 h postinfection (hpi) in Hpt cells. Thus, imaging of both the virus and cells indicated that the actin cytoskeleton indeed participated in WSSV entry into host cells. Actin polymerization initiated by the Arp2/3 complex is a prerequisite for remodeling of the actin cytoskeleton (20), and a WD repeat-containing protein gene was found to be upregulated upon WSSV challenge in crayfish Hpt cells, which was further characterized to be CqArpc1A containing five WD40 repeat-containing amino acid motifs in our previous study (21). We then treated Hpt cells with the pharmacological inhibitor CK666 to prevent the activating conformational change of the Arp2/3 complex for depolymerization of the actin cytoskeleton. As shown in Fig. 1B, a, the actin cytoskeleton was strongly depolymerized by CK666, indicating that the Arp2/3 complex is critical for the polymerization of the actin cytoskeleton in crayfish Hpt cells. Intriguingly, WSSV entry was significantly reduced in Hpt cells in which the Arp2/3 complex was blocked with CK666 at 1 hpi compared with the control treatment (Fig. 1B, b; P < 0.05), suggesting the key role of the Arp2/3 complex in actin activity in terms of WSSV invasion. To confirm this result at the molecular level, we then cloned the full-length cDNA sequences of Arp2/3 complex subunits, including CqArpc1A (GenBank accession no. OM962866), CqArpc3 (GenBank accession no. OM962868), CqArpc4 (GenBank accession no. OM962867) and CqArpc5 (GenBank accession no. OM962869), from red claw crayfish for further study. Considering that CqArpc1A appeared to be the most abundant transcript present in Hpt cells while CqArpc3, CqArpc5, and CqArpc4 were present at lower levels (Fig. 1C), CqArpc1A was selected as the representative CqArp2/3 complex for further investigation in our present study. To reveal the function of CqArpc1A, we then prepared a recombinant rCqArpc1A protein fused with a glutathione S-transferase (GST) tag (Fig. 1D, a), which exhibited binding activity with cytoskeletal components such as Cqβ-Actin in Hpt cell lysates (Fig. 1D, b), suggesting that CqArpc1A, and possibly the CqArp2/3 complex, regulates cytoskeletal nucleation in a manner similar to that in vertebrates (20). To further elucidate whether the assembly of actin filaments was indeed affected by the CqArp2/3 complex, the gene expression of CqArp2/3 complex subunits, including CqArpc1A, CqArpc3, CqArpc4, and CqArpc5, was silenced by RNA interference (RNAi) using mixed double-stranded RNAs (dsRNAs) targeting the endogenous mRNA sequences of CqArpc1A, CqArpc3, CqArpc4, and CqArpc5. As shown in Fig. 1E, a, gene expression of the endogenous CqArp2/3 subunits was significantly reduced (P < 0.01); this resulted in the clear depolymerization of F-actin (as expected) in Hpt cells compared to the control cells, which showed bundles of cytoskeletal filaments (Fig. 1E, b). Importantly, gene silencing of CqArp2/3 complex subunits resulted in clearly reduced WSSV entry into Hpt cells at 1 hpi (Fig. 1E, c; P < 0.05), which was consistent with the pharmacological inhibition caused by CK666 (Fig. 1B, b; P < 0.05). Together, these data strongly suggest that polymerization of the actin cytoskeleton, promoted by the CqArp2/3 complex, is necessary for the efficient entry of WSSV into crayfish Hpt cells.

FIG 1.

Internalization of WSSV in Hpt cells is significantly reduced by disrupting actin nucleation via inhibiting CqArp2/3 complex. (A) Microfilament was recruited for WSSV entry into Hpt cells. WSSV colocalized with protruding cytoskeletal filopodia (a) and F-actin filaments (b) at 0.5 hpi. Colocalization (yellow) is indicated by white arrows. (B) Internalization of WSSV was significantly reduced by depolymerizing F-actin. (a) F-actin formation was disrupted by blockade of CqArp2/3 complex with pharmacological inhibitor CK666 in Hpt cells. Control cells were treated by dimethyl sulfoxide (DMSO). (b) Internalization of WSSV was significantly reduced by inhibiting CqArp2/3 complex with CK666. (C) Characterization of CqArpc1A from red claw crayfish. Gene expression of four subunits of CqArp2/3 complex was determined in vitro in Hpt cells, in which the relative gene expression of CqArpc1A was significantly higher than that of CqArpc3, CqArpc4, and CqArpc5, as determined by qRT-PCR. (D) CqArpc1A bound to Cqβ-Actin. (a) rGST-CqArpc1A protein of about 65 kDa (indicated by the arrow) was purified from E. coli. (b) rCqArpc1A bound to Cqβ-Actin from crayfish Hpt. Binding protein by GST pulldown assay is indicated by an arrow. (E) Internalization of WSSV was significantly reduced by depolymerizing F-actin via gene silencing of CqArp2/3 complex. (a) Gene expression of CqArp2/3 complex was significantly reduced by RNAi. (b) F-actin formation was blocked by gene silencing of CqArp2/3 complex, including CqArpc1A, CqArpc3, CqArpc4, and CqArpc5. (c) WSSV entry into Hpt cells was significantly reduced by gene silencing of CqArp2/3 complex. (Upper) Protein band intensities of three independent experiments were calculated using the Quantity One program. Hpt cells were infected by WSSV for 0.5 h followed by immunostaining of virions (red) with anti-VP28 antibody and labeling of F-actin (green) by phalloidin. Endogenous gene expression of CqArp2/3 complex, including CqArpc1A, CqArpc3, CqArpc4, and CqArpc5, was silenced together with the mixed dsRNA of four subunits of CqArp2/3 in Hpt cell cultures followed by phalloidin staining of F-actin (green). Cell nucleus was labeled with DAPI (blue). Gene silencing efficiency of CqArp2/3 complex, including CqArpc1A, CqArpc3, CqArpc4, and CqArpc5, was determined by qRT-PCR in Hpt cells, in which the relative transcript levels were all normalized to those of their own controls treated by dsGFP accordingly. Significant differences are indicated with different letters, which were analyzed using Student's t test: *, P < 0.05; **, P < 0.01.

Formation of WSSV-containing CME vesicles induced by viral infection is associated with the actin cytoskeleton.

Preliminarily, the early infection stages of WSSV were determined by a time series of transmission electron microscopy (TEM) analysis, in which most of the virions were entering Hpt cell at 0.5 hpi and then wrapped into smaller endocytic vesicles at 1 hpi after internalization (8, 9). To further address how WSSV, after internalization, is intracellularly transported, we investigated the intracellular trafficking of WSSV by fluorescent imaging of both virion and Hpt cell. Interestingly, cell imaging analysis revealed that the formation of endocytic vesicles, with an enlarged size, containing WSSV virions was clearly induced by viral infection at an early infection stage from 0.5 (data not shown) to 1 hpi, showing more fluorescent vesicle signals at 1 hpi (Fig. 2A) than at 0.5 hpi (data not shown) in Hpt cells. In the cells, the induced vesicles were labeled by the membrane-impermeable dye FilmTracer FM 1-43 (FM1-43), and the virions were immunostained with the anti-VP28 antibody. In contrast, no obviously observable vesicles with both enlarged sizes and fluorescent signals were found in the control Hpt cells lacking WSSV challenge (Fig. 2A). Additionally, WSSV-containing vesicles were shown to be clearly colocalized with the actin cytoskeleton component F-actin by immunofluorescence staining (Fig. 2A), implying that the actin cytoskeleton is recruited for the formation and cytoplasmic trafficking of these endocytic vesicles enclosing virions whose formation is induced by WSSV infection at an early stage of infection. This finding suggests that the actin cytoskeleton is involved not only in the extracellular uptake of WSSV but also in the intracellular transport of WSSV-containing endocytic vesicles after their detachment from the host cell membrane.

FIG 2.

Formation of virion-containing CME vesicles is induced by WSSV infection. (A) Formation of virion-containing endocytic vesicles was induced by WSSV challenge in Hpt cells. The endocytic vesicles (indigo) containing WSSV (red), with an enlarged size, were colocalized with F-actin cytoskeleton (green). No obvious endocytic vesicle with an enlarged size was observed in control cells lacking WSSV infection. The images of endocytic vesicles containing internalized WSSV virion were captured at 1 h after viral challenge, as determined by immunofluorescence assay. (B) Endocytic WSSV virions were enclosed in CME vesicles. Colocalization between CqCLC with WSSV (orange) and CqCLC with endocytic vesicles (indigo) was present in cytoplasm of Hpt cells.

To further verify whether these WSSV-containing vesicles indeed originated from endocytic vesicles at the molecular level, the subcellular colocalization of the key components of CME, such as clathrin, with WSSV was determined. As shown in Fig. 2B, clathrin light chain (CqCLC) immunostained with a monoclonal antibody was found to be clearly colocalized not only with WSSV, as indicated by the anti-VP28 antibody staining, but also with enlarged WSSV-containing vesicles, as indicated by FM1-43 staining. Imaging of the host cells further confirmed that WSSV enters Hpt cells through the CME pathway, in which clathrin recruitment for the formation of endocytic vesicles is strongly induced by WSSV infection, demonstrating the essential role of endocytic vesicles in both the cellular entry of WSSV into host cells and the subsequent trafficking of endocytic virions within the cytoplasm toward the cell nuclei.

The cytoplasmic trafficking of WSSV virions is dependent on microtubules.

To gain more insight into the intracellular transport of WSSV, the cytoplasmic localization of WSSV at an early infection stage of 1.5 hpi was determined primarily by TEM analysis in Hpt cells. The internalized WSSV virions were found to be enclosed within vesicle structures in the cytoplasm (Fig. 3A, a), which were subsequently verified to be endosomes, as indicated by immunostaining with a monoclonal anti-RabGEF1 antibody that also colocalized with WSSV, as indicated by immunostaining with an antibody against the viral envelope protein VP28 (Fig. 3A, b). Generally, a fusion event between the viral envelope and endosomal membrane is a prerequisite for the subsequent penetration of the nucleocapsid into the cytosol (22). Next, WSSV-containing endosomal vesicles were examined by TEM at 2 hpi in Hpt cells. As shown in Fig. 3B, a, the WSSV particles were clearly enclosed in the endosomal vesicles, in which the virions interacted with the luminal face of the endosomal vesicle membrane. Importantly, the WSSV virions were also shown to be undergoing a fusion process between the envelope and the endosomal vesicle membrane; consequently, the edge of the viral envelope gradually faded (Fig. 3B, b), which is consistent with our previous report that viral fusion occurs within endosomes, as determined by immunofluorescence analysis (9). In particular, the nucleocapsid emerged from endosomes and was finally released into the cytoplasm (Fig. 3B, c) after complete fusion between the viral envelope and endosomal membrane. These findings together indicate that the detachment of the WSSV envelope from the nucleocapsid by fusion with the endosomal membrane is crucial for viral penetration and the subsequent transport of the nucleocapsid toward the perinuclear area of the host cell.

FIG 3.

Cytoplasmic trafficking of WSSV in Hpt cells depends on endosomal vesicles and microtubules. (A) Endocytic virions were delivered into endosomal vesicles in Hpt cells. (a) WSSV virions were localized in the intracellular vesicle as examined by TEM. Black arrowheads indicate virions in vesicle. C, cytoplasm; N, nucleus. (b) WSSV virions were localized in endosomal vesicle as determined by immunofluorescence assay. Colocalization (yellow) of virions and endosomal vesicle was indicated by white arrows. (B) WSSV nucleocapsids exited from endosome into cytoplasm after fusion between viral envelope and endosomal membrane. (a) Fusion between WSSV envelope and endosomal membrane was examined by TEM. Endocytic WSSV virions were enclosed within endosome in which the exposed viral envelope interacted with the luminal face of endosomal membrane, as indicated by white arrowheads. Viral envelope, nucleocapsid, and endosomal membrane are indicated by dashed line with yellow, pink, and red color accordingly. (b) Undergoing fusion between viral envelope and endosomal membrane. As fusion goes on, viral envelope gradually faded. Nucleocapsids prepenetrated from endosome are indicated by black arrowheads. (c) Naked WSSV nucleocapsid in cytoplasm, after penetration from endosome, is indicated by yellow arrowhead. (C) Separation of WSSV envelope from nucleocapsid within endosome in Hpt cell was inhibited by depolymerizing microtubules. (a) Microtubules were depolymerized by ND. Nonbundled microtubules were present in WSSV-infected Hpt cells after depolymerization with ND. (b) WSSV virions were trapped in endosomes after depolymerizing microtubules. Aggregated endosomes, containing WSSV virions indicated by yellow arrows, were obviously induced after depolymerization of microtubules by ND. However, no virion signal was present in endosomes, as indicated by white arrows, in control cells. (c) Dissociation of WSSV envelope from nucleocapsid was strongly inhibited by depolymerizing microtubules. Hpt cells were challenged with WSSV for 0.5 h, which were then treated with ND to depolymerize the microtubules, in the presence of CHX to block the expression of viral proteins. Extensively colocalized fluorescence of WSSV envelope and nucleocapsid, with fluorescent signals (yellow) indicated by yellow arrows, was present in Hpt cells after depolymerizing microtubules. WSSV envelope and nucleocapsid were immunostained with antibody against VP28 (red) and VP664 (green), respectively. (D) Transport of incoming WSSV nucleocapsids toward perinuclear area was substantially reduced by depolymerizing microtubules. Most viral nucleocapsid VP664 signals (green) were distributed in the rim of the nucleus, as indicated by white circle in control cells treated with DMSO, but they were rarely present in the periphery of nucleus, i.e., mostly located in cytoplasm relatively close to cell membrane, in Hpt cells with depolymerized microtubules (pink) by ND in the presence of CHX. (E) WSSV replication was significantly reduced by depolymerizing microtubules. Both gene transcript (a) and protein expression (b) of WSSV were significantly decreased by depolymerizing microtubules in Hpt cells treated with ND at 0.5 hpi after viral infection. The data are presented as the means ± SD. **, P < 0.01. WSSV virion and endosome were immunostained with anti-VP28 antibody (red) and anti-RabGEF1 (green) antibody, respectively. DNA of Hpt cell was labeled with DAPI (blue). Microtubules (pink) were stained by tubulin-tracker. Microtubules were depolymerized by ND in Hpt cells, while control cells were treated by DMSO. For WSSV infection, Hpt cells were infected by WSSV for 0.5 h followed by ND treatment for microtubule depolymerization, in which microtubules were imaged at 3 hpi.

To reveal whether microtubule-dependent endosomal transport affects WSSV infection, Hpt cells were infected with WSSV for 0.5 h, after which microtubules were depolymerized with the chemical inhibitor nocodazole (ND) and the expression of WSSV proteins was prevented by cycloheximide (CHX). As shown in Fig. 3C, a, microtubules were substantially unbundled in Hpt cells after depolymerization with ND but not in control cells, which showed clear microtubule distribution with bundles. Intriguingly, the endocytic WSSV virions immunostained with an anti-VP28 antibody were found to be strongly trapped in the accumulated endosomes at 3 hpi in Hpt cells after depolymerization of microtubules with ND (Fig. 3C, b). In contrast, no viral envelope signals, such as VP28 fluorescence, were present in the control cells without inhibitor treatment, because fusion between the envelope and endosomal membrane was already completed at 3 hpi, leading to further degradation, as suggested in our previous report (9). Furthermore, the dissociation of the WSSV envelope from the nucleocapsid was also determined after depolymerization of microtubules in Hpt cells. Undetached WSSV virions with envelopes, as indicated by the substantial colocalization and accumulation of envelopes and nucleocapsids immunostained with VP28 and VP664 antibodies, respectively, were clearly present at 3 hpi in Hpt cells after depolymerization of microtubules with ND (Fig. 3C, c). In contrast, the nucleocapsids that penetrated without accumulating, as indicated with an anti-VP664 antibody, were shown to be dispersed in the cytoplasm of control Hpt cells lacking inhibitor treatment at 3 hpi. Taken together, these data demonstrate that the endocytosed WSSV virions are indeed delivered into endosomes, where the dissociation of the viral envelope from the nucleocapsid before the exit of the nucleocapsid from the endosome to the cytoplasm can be significantly blocked by depolymerizing microtubules at the early infection stage in Hpt cells. These findings strongly suggest that the intracellular transport of WSSV virions, via hijacking of the endosome system, is dependent on polymerized microtubules.

To reveal the further transport of unenveloped nucleocapsids to the perinuclear area and the release of the viral genome into the cell nucleus, we investigated the cytoplasmic transport of nucleocapsids at 3 hpi, when the nucleocapsid has completed its exit from endosomes into the cytoplasm (9) after depolymerizing microtubules with ND in Hpt cells in the presence of CHX to prevent the expression of WSSV proteins, considering that microtubules have been shown to be recruited for WSSV infection (8). The subcellular localization of incoming nucleocapsids was analyzed with confocal fluorescence microscopy, and the capsid proteins immunostained with the anti-VP664 antibody clearly accumulated at the nuclear rims in the control Hpt cells treated with dimethyl sulfoxide (DMSO) at 3 hpi (Fig. 3D). In contrast, the perinuclear targeting of the WSSV nucleocapsid was substantially reduced by depolymerization of microtubules with ND in Hpt cells; in these cells, the incoming nucleocapsid was mainly dispersed in the cytoplasm relatively close to the Hpt cell membrane (Fig. 3D). This finding implies that the detainment of the viral nucleocapsid in the cytoplasm, i.e., the lack of transport to the perinuclear area, is caused by the depolymerization of microtubules. Meanwhile, both the gene transcription, as indicated by assessment of VP28 gene expression from the early infection time of 3 hpi to the very late infection stage of 24 hpi (Fig. 3E, a; P < 0.01), and the protein synthesis of WSSV at 24 hpi, as determined by Western blotting against the VP28 protein (Fig. 3E, b; P < 0.01), were substantially reduced in Hpt cells. This result was caused by the inaccessibility of the viral genome in the Hpt cell nucleus after depolymerization of the microtubules by ND at 0.5 hpi. Taken together, these data demonstrate that microtubules are recruited for the cytoplasmic trafficking of nucleocapsids toward the perinuclear area for subsequent efficient invasion in crayfish Hpt cells.

Nuclear pores are required for both delivery of the incoming genome and import of the expressed structural protein of WSSV into the host cell nucleus.

To reveal how the WSSV genome was delivered into the cell nucleus, we primarily examined the ultrastructure of the nuclear membrane area in Hpt cells by TEM analysis. As indicated in Fig. 4A, a, the nuclear pore showed perforation, with a typically static channel embedded in the nuclear double membrane, through which receptor-cargo complexes are believed to be dynamically transported across the nuclear envelope. Interestingly, nucleocapsids with an electron-dense DNA core, i.e., nucleocapsids after exit from the endosome into the cytoplasm, were captured near the nuclear pores at 2.5 hpi in Hpt cells (Fig. 4A, b), which suggested that the nucleocapsids enclosing the viral genome were successfully transported to the nuclear pores and then prepared for the subsequent release of the viral genome into the host cell nucleus at an early stage of infection. Intriguingly, capsids lacking the central electron-dense mass, i.e., those without viral genomic DNA, were also found within the typical nuclear basket, a cage-like structure on the karyoplasmic face of NPCs (23), at 3 hpi but not at the perinuclear membrane if compared to that at 2.5 hpi, strongly indicating that the WSSV genome had already been released into the Hpt cell nucleus through the nuclear pores (Fig. 4A, c). Furthermore, the incoming nucleocapsid component was also confirmed to be localized within the nuclear pore area, as determined by immunogold labeling against VP664 at 3 hpi in Hpt cells in the presence of CHX to block the expression of viral proteins (as determined by immunoelectron microscopy [IEM]) (Fig. 4A, d). These data together demonstrate for the first time, at the ultrastructural level by TEM imaging that the WSSV genome is released into crustacean cell nuclei via the nuclear pores.

FIG 4.

Nuclear pores are required for both delivery of the incoming genome and import of the expressed structural proteins of WSSV into Hpt cell nucleus. (A) WSSV genome was released into Hpt cell nucleus through nuclear pores. (a) Nuclear pore of Hpt cell examined by TEM. (b) WSSV nucleocapsids located close to nuclear pore or nuclear membrane in Hpt cell cytoplasm. Nucleocapsids enclosing WSSV genome are indicated by white arrowheads. (c) Capsids, after releasing viral genome into Hpt cell nucleus, located in nuclear baskets. Capsids free of viral genome within nuclear baskets are indicated by white arrowheads, in which the viral genome was supposed to have been released into cell nucleus through nuclear pores at an early infection stage of 3 hpi. (d) Localization of incoming nucleocapsid within nuclear pore as analyzed by IEM. Nucleocapsids, labeled by anti-VP664 antibody, are indicated by white arrowheads within nuclear pore at an early infection stage of 3 hpi in Hpt cells, with the presence of CHX to block the expression of viral proteins. (B) Expressed capsid protein VP664 entered Hpt cell nucleus through nuclear pores for viral assembly, as determined by immunogold labeling. (a to c) Expressed VP664 protein distributed in the cell cytoplasm (a), within nuclear pore as boxed (b), and in cell nuclei (c). VP664 is indicated by white arrowheads. (C) Expressed envelope protein VP28 entered Hpt cell nucleus through nuclear pores for viral assembly, as determined by immunogold labeling. (a to c) Expressed VP28 protein located near nuclear pore (a), within nuclear pore (b), and in nuclei (c). VP28 is indicated by white arrowheads. Nuclear pores in Hpt cells are indicated by black arrowheads. For immunogold labeling, capsid protein VP664 was probed by anti-VP664 antibody followed by a secondary antibody with 12-nm colloidal gold anti-rabbit IgG, and envelope protein VP28 was probed by anti-VP28 antibody followed by a secondary antibody with 25-nm colloidal gold anti-mouse IgG conjugate. C, cytoplasm; N, nuclei.

By using immunoelectron microscopy analysis, we found that the key structural proteins of WSSV, such as the nucleocapsid protein VP664 and the envelope protein VP28, were expressed and present in the Hpt cell cytoplasm at the early infection stage of 6 hpi (data not shown). Their levels gradually increased in the cytoplasm over time (data not shown). To reveal the possible mechanism of import of these expressed structural proteins into the host cell nucleus, we selected 8 hpi as the detection time, since the presence of expressed viral proteins was relatively lower at 6 hpi (data not shown), which was not a suitable time for analysis of distribution by immunoelectron microscopy in the Hpt cell cytoplasm. Interestingly, expressed capsid proteins, such as VP664 labeled with the anti-VP664 antibody conjugated to gold particles of 12 nm, were found in the cytoplasm (Fig. 4B, a), within the nuclear pores (Fig. 4B, b), and inside the nucleus (Fig. 4B, c) in Hpt cells, suggesting that the expressed capsid proteins, such as VP664, were imported into the host cell nucleus through the nuclear pores for the subsequent assembly of progeny virions. Similarly, de novo-generated VP28 immunostained with the anti-VP28 antibody conjugated with 25-nm gold particles was also shown to be located in the cytoplasm (Fig. 4C, a), within the nuclear pores (Fig. 4C, b), and inside the nucleus (Fig. 4C, c) in Hpt cells at 8 hpi, suggesting that the envelope protein VP28 is imported into the Hpt cell nucleus through the nuclear pores for progeny virion assembly. Taken together, these results certainly indicate that the representative structural components of WSSV, i.e., both the major nucleocapsid proteins, such as VP664, and major envelope proteins, such as VP28, are imported into the Hpt cell nucleus for assembly of WSSV progeny via the nuclear pores.

Nuclear pore targeting of the incoming nucleocapsids is significantly inhibited by pharmacological suppression of CqImportin α1/β1 activity.

To reveal how the nuclear import of WSSV components, i.e., both viral genome and the expressed viral proteins as shown above (Fig. 4), is regulated at the molecular level, the temporal expression patterns of the related molecules, such as the nuclear transporters (CqImportin α1, GenBank accession no. OM962861; CqImportin β1, GenBank accession no. OM962860) together with an associated GTP energy source molecule (CqRan, GenBank accession no. OM962862) and the NPC molecules, including CqNucleoporin 35 (CqNup35, GenBank accession no. OM962863), CqNucleoporin 37 (CqNup37, GenBank accession no. OM962864) and CqNucleoporin 62 (CqNup62, GenBank accession no. OM962865), were determined by quantitative reverse transcription-PCR (qRT-PCR), considering that expression of all these molecules showed upregulation in crayfish Hpt cells after WSSV challenge (21) and in our other transcriptome libraries of crayfish (unpublished data). Additionally, both the gene and protein expression of WSSV were gradually increased in Hpt cells after viral infection (data not shown); in particular, the gene expression of crayfish CqImportin α1, CqImportin β1, CqRan, CqNup35, CqNup37, and CqNup62 was significantly induced by WSSV challenge, especially in the early infection stages (i.e., from 1 to 6 hpi) in Hpt cells (data not shown). Therefore, CqImportin α1, CqImportin β1, CqRan, CqNup35, CqNup37, and CqNup62 were likely to be recruited for WSSV infection, probably via regulation of the nuclear targeting and nuclear import of WSSV, which need investigation. In vertebrates, importin activity can be blocked by specific pharmacological inhibitors, such as ivermectin (IVM), which inhibits importin α1/β1-mediated nuclear import by blocking importin α1/β1 heterodimer formation (24), and importazole (IPZ), which inhibits importin β1/RanGTP-mediated nuclear import by interfering with the binding between RanGTP and importin β1 (25). Preliminarily, to reveal whether IVM and IPZ are capable of binding to crustacean importins with specific blocking activity and engaging in competitive protein binding inhibition (24, 25), the recombinant GST-tagged CqImportin α1 (rCqImportin α1) protein was expressed and purified for IVM-protein binding analysis by biolayer interferometry (BLI) assay. Similar to the inhibitor binding affinity in vertebrates, IVM showed a strong dose-dependent inhibitor binding affinity, with a fitted dissociation constant (K_D) of 24.37 ± 2.33 μM, which was close to the steady-state K_D value of 24.35 ± 0.37 μM for rCqImportin α1 yielded by fitting of all concentrations and their associated residuals using globally fitted data in Octet System Data Analysis software (Fig. 5A, a and b). However, IVM had no obvious binding affinity for the negative-control protein recombinant GST (rGST); the curves did not show any significant shifts with all concentrations tested (Fig. 5A, c). This result strongly demonstrates the direct binding affinity of IVM for CqImportin α1 but not for the negative-control protein rGST. Meanwhile, the His-tagged rCqImportin β1 protein was also expressed and purified for an IPZ-protein binding assay as described above. Similarly, IPZ exhibited a strong concentration-dependent inhibitor binding affinity, with a K_D of 92.34 ± 6.48 μM, for recombinant rHis-CqImportin β1 (Fig. 5B, a). This value was close to the K_D value of 92.0 ± 0.99 μM in a steady-state analysis (Fig. 5B, b). However, IPZ had no obvious binding affinity for the negative-control protein rHis-green fluorescent protein (GFP) (Fig. 5B, c). Furthermore, the binding of CqImportin α1 to CqImportin β1 or CqRan to CqImportin β1 was strongly blocked by the presence of IVM (Fig. 5C, a) or IPZ (Fig. 5C, b), respectively, via competitive inhibition (24, 25). These data together suggest that the activity of crustacean CqImportin α1/β1 and CqImportin β1/Ran can be specifically blocked by IVM and IPZ accordingly, which is consistent with findings in vertebrates. Considering the possible viral cargo-CqImportin α1/β1 ternary complex for nuclear pore targeting of the WSSV nucleocapsid, these two inhibitors can be applied to Hpt cells to disrupt crayfish importins. This finding certainly benefited our next investigation to further reveal how crustacean importins affect WSSV infection, because it allowed us to block the activity of the importins with specific inhibitors. Next, crayfish Hpt cell viability under exposure to different concentrations of IVM or IPZ inhibitors was determined to exclude cytotoxicity based on the 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) colorimetric method. A concentration of 8 μM for IVM or IPZ, which lacked a clear cytotoxic effect on Hpt cell viability (data not shown), was then used to block protein activity against crayfish CqImportin α1/β1 or CqImportin β1/Ran, respectively, in the following experiments.

FIG 5.

Nuclear pore targeting of the incoming nucleocapsids is significantly inhibited by pharmacological suppression of CqImportin α1/β1 activity. (A) IVM bound to rCqImportin α1. (a) Direct binding of IVM to CqImportin α1 was in a dose-dependent manner of inhibitor. rGST-CqImportin α1 labeled with biotin was captured on Super Streptavidin biosensors for binding affinity test with gradient concentrations of IVM. (b) Steady-state analysis of IVM binding to rGST-CqImportin α1 was determined by ForteBio Octet. (c) IVM did not bind to control protein rGST. (B) IPZ bound to rCqImportin β1. Binding of IPZ to biotin-labeled rHis-CqImportin β1 exhibited a dose-dependent manner of inhibition, as determined by dynamic analysis with ForteBio Octet (a) and steady-state analysis (b), while control protein rHis-GFP showed no significant binding affinity (c). (C) Binding inhibition of CqImportin α1 or CqRan to CqImportin β1. (a) Binding of rCqImportin α1 to rCqImportin β1 was blocked by IVM. (b) Binding of rCqRan to rCqImportin β1 was blocked by IPZ. Binding inhibition of rCqImportin α1 or rCqRan to rCqImportin β1 was in a dose-dependent manner of IVM or IPZ, respectively. rHis-CqImportin β1 was captured on anti-penta-HIS biosensors immobilized with anti-His antibody for binding examination with rGST-CqImportin α1 or rCqRan in the presence of gradient concentrations of IVM or IPZ. (D) Nucleocapsids strongly accumulated in the perinuclear area after inhibition of CqImportin α1/β1 in Hpt cells. Blockade of CqImportin α1/β1 with IVM led to a substantial accumulation of nucleocapsid in perinuclear area (a); a relative quantification of the integrated optical density of fluorescent VP28 and VP664 is shown in panel b. Blockade of CqImportin β1/Ran with IPZ also resulted in a substantial accumulation of nucleocapsid in perinuclear area (c); a relative quantification of the integrated optical density of fluorescent VP28 and VP664 is shown in panel d. DMSO-treated cells served as a control group. (e) Gene silencing of CqImportin α1, CqImportin β1, or CqImportin α1/β1 resulted in significant accumulation of nucleocapsids in the perinuclear area. Control cells were treated by GFP dsRNA. Relative quantification of the integrated optical density of fluorescent VP28 or VP664 in panel e is shown in panel f. Results were determined in Hpt cells in the presence of CHX for blocking the expression of viral proteins. The kinetic curves were acquired by fitting data to the 1:1 binding model analyzed by ForteBio Octet software as described in Materials and Methods. The K_D values were calculated by ForteBio DataAnalysis software 11. The value on the y axis represents thickness (nanometers) of the biolayer on biosensor surface, which was increased after the binding of IVM and IPZ to the captured rCqImportin α1 and rCqImportin β1 or the binding of rCqImportin α1 and rCqRan to the captured rCqImportin β, respectively, on the biosensor surface. Viral components were immunostained by both anti-VP28 and VP664 antibodies. Cell nucleus DNA was stained by DAPI (blue). All data are presented as the means ± SD from biological triplicates. Significant differences were analyzed using Student's t test: **, P < 0.01; *, P < 0.05; ns, no significant difference. CqImportin α1 and CqImportin β1 were abbreviated to Imp α1 and Imp β1 in all images.

After the activity of the CqImportin α1/β1 or CqImportin β1 pathway was artificially blocked with the specific inhibitor IVM or IPZ, respectively, for 1 h in Hpt cells, cells were subjected to WSSV challenge in the presence of CHX to prevent the expression of viral proteins. As shown in the control Hpt cells treated with DMSO (Fig. 5D, a to c), a relatively large amount of fluorescence indicating incoming WSSV envelope protein (from immunostaining with an anti-VP28 antibody) was present in the Hpt cell cytoplasm at 1 hpi, but fluorescence was minimal for capsid proteins immunostained with an anti-VP664 antibody, suggesting that the WSSV envelope had not been detached from the nucleocapsid in the very early stage of viral infection, in agreement with our previous report (9). Gradually, a much stronger fluorescent signal for the incoming nucleocapsid appeared in the Hpt cell cytoplasm, becoming distributed toward the cell nuclear membrane area, at 2 hpi than at 1 hpi, indicating increased detachment of the viral envelope with infection time. In addition, the fluorescent signal of the envelope protein VP28 gradually decreased (Fig. 5D, a to c) due to protein degradation via the autophagic-lysosome pathway, as we have previously described (9). Importantly, the distribution of WSSV nucleocapsids, as indicated by fluorescent VP664, was significantly increased at the perinuclear area at 3 hpi, and the nucleocapsids accumulated around the nuclear membrane, forming a larger cluster in Hpt cells treated with either IVM or IPZ than in control Hpt cells without pharmacological blocking (Fig. 5D, a to c). Furthermore, according to statistical analysis of the relative fluorescence intensity of the envelope protein VP28 or the nucleocapsid protein VP664, the fluorescent signal of the WSSV envelope was found to gradually decrease, while the nucleocapsid protein signal increased, with increasing infection time (Fig. 5D, b to d). In particular, the fluorescent signal of the viral capsid protein VP664 significantly accumulated around the perinuclear membranes of Hpt cells after treatment with either IVM (75.5% increase) or IPZ (71.6% increase), respectively, at 3 hpi compared with the levels in negative controls. This effect might have been caused by aggregation of nucleocapsids that were incapable of releasing the WSSV genome properly into the Hpt cell nucleus, which then led to inefficient and delayed degradation of these nucleocapsid components during our observations (Fig. 5D, b to d; P < 0.01). Moreover, blocking CqImportin α1/β1 with IVM much more efficiently reduced nuclear targeting of incoming WSSV nucleocapsids than blocking CqImportin β1/Ran using IPZ (Fig. 5D, a, b, c, and d). This finding implies that the CqImportin α1/β1 contributes dominantly to the docking of incoming nucleocapsids to the nuclear pores for the subsequent release of the viral genome into the Hpt cell nucleus. Furthermore, the accumulation of incoming nucleocapsids of WSSV after inhibition of CqImportin α1/β1 activity was also confirmed at the molecular level by gene silencing of CqImportin α1 and CqImportin β1: the accumulated fluorescent signal of the VP664 protein at 3 hpi in Hpt cells was significantly increased by 53.4%, 58.1%, and 71.0%, respectively, by gene silencing of CqImportin α1, CqImportin β1, and CqImportin α1/β1 compared with the signal after mock treatments with GFP dsRNA (Fig. 5D, e and f). This finding was similar to that of the inhibitory effect caused by pharmacological inhibition shown above (Fig. 5D, a, b, c, and d). Together, these data strongly demonstrate that inhibition of crustacean nuclear importins, i.e., crayfish CqImportin α1/β1, can result in significant accumulation of incoming WSSV nucleocapsids around the perinuclear area in Hpt cells, which leads to blockade of the subsequent release of the WSSV genome from the nucleocapsids into the host cell nucleus.

WSSV replication is strongly decreased by suppression of CqImportin α1/β1 and CqNup35.

To identify whether WSSV replication was indeed affected by artificial manipulation of importin activity as described above, both the gene expression and protein expression of WSSV were examined by qRT-PCR and Western blot analysis, respectively, in Hpt cells. This analysis was conducted under conditions of reduced delivery of the viral genome into the cell nucleus at the early infection stage after inhibition of CqImportin α1/β1 activity either by pharmacological blocking or by gene silencing. As shown in Fig. 6A, compared with that in control Hpt cells, the transcription of the WSSV gene VP28 was significantly reduced, with approximately 85.5%, 84.8%, 96.3%, and 98.7% decreases in the IVM-treated groups (Fig. 6A, a; P < 0.01) and approximately 82.8%, 80.9%, 77.8%, and 68.3% decreases in the IPZ-treated groups (Fig. 6A, b; P < 0.01) at the selected time points of 3, 6, 12, and 24 hpi, respectively. Meanwhile, the gene expression of CqImportin α1, CqImportin β1, and CqRan was significantly reduced by gene silencing at all selected time intervals from 3 to 24 hpi (Fig. 6A, c, d, and e). The transcript levels of CqImportin α1 (approximately 90% decreased; P < 0.01), CqImportin β1 (approximately 92% decreased; P < 0.01), or CqRan (approximately 95% decreased; P < 0.01) were significantly reduced by either single gene silencing or double gene silencing, respectively, compared to the levels in cells mock treated with GFP dsRNA (Fig. 6A, c, d, and e). As a result, the VP28 gene transcript levels were obviously suppressed at 3, 6, and 12 hpi in Hpt cells after gene silencing of CqImportin α1 or CqImportin β1 or after double gene silencing of the CqImportin α1/β1 complex compared to controls (Fig. 6A, f; P < 0.01). This effect was in agreement with the inhibitory effect on WSSV gene transcription caused by the pharmacological inhibitors described above (Fig. 6A, a and b). In addition, the gene silencing of CqRan also led to the markedly reduced VP28 gene expression of WSSV at 3, 6, 12, and 24 hpi (Fig. 6A, g; P < 0.01). The expressed protein abundance of structural proteins for progeny virion assembly, such as VP28, was also clearly decreased at 12 hpi in Hpt cells. Approximately 88.6% and 79.9% decreases were observed with IVM and IPZ treatments, respectively; approximately 49.9%, 58.9%, 79.2%, and 42.5% decreases were present after gene silencing of CqImportin α1, CqImportin β1, both CqImportin α1 and CqImportin β1, or CqRan, respectively (Fig. 6A, h, i, j, and k; P < 0.01 or P < 0.05). Furthermore, blocking the CqImportin α1/β1 with IVM exhibited much greater efficiency in reducing WSSV infection, both in terms of gene transcription and in viral protein expression, than blocking CqImportin β1/Ran with IPZ under our experimental conditions (Fig. 6A, a, b, h, and i). Similarly, double gene silencing of the CqImportin α1/β1 also resulted in much more efficient inhibition of WSSV infection than single gene silencing of CqImportin α1 or CqImportin β1 (Fig. 6A, f to j). In particular, reduced WSSV gene expression was present from the early infection stage of 3 hpi to the middle infection stage of 12 hpi after single gene silencing and was further extended to the very late infection stage of 24 hpi by double gene silencing of both CqImportin α1 and CqImportin β1 (Fig. 6A, f). Together, these data strongly suggest that the dominant contribution of CqImportin α1/β1 to targeting of WSSV nucleocapsids to nuclear pores at an early infection stage (like that of viral gene transcription at 3 hpi) for further release of the viral genome into the host cell nucleus results in significantly reduced viral replication. Additionally, gene silencing of CqNup35 decreased gene expression by approximately 74.6% at all selected time periods from 3 to 24 hpi in Hpt cells (Fig. 6B, a; P < 0.01); importantly, CqNup35 gene silencing also led to the markedly reduced VP28 gene expression of WSSV at 3, 6, and 12 hpi (Fig. 6B, b) and decreased VP28 protein expression at 12 hpi (Fig. 6B, c), suggesting the key role of CqNup35 in WSSV infection for efficient nuclear import of the viral genome as well as expressed structural proteins. Unfortunately, we did not succeed in gene silencing of either CqNup37 or CqNup62 at present, by which the direct effect on WSSV replication needs further study. These results together prove that the nuclear transporters efficiently benefit WSSV infection in crustacean.

FIG 6.

WSSV replication is strongly decreased by suppression of CqImportin α1/β1 and CqNup35. (A) WSSV replication was significantly reduced by inhibition of CqImportin α1/β1 or CqImportin β1/Ran with chemical inhibitors or gene silencing. WSSV gene transcription was strongly reduced by blockade of CqImportin α1/β1 with IVM (a) or of CqImportin β1/Ran (b) with IPZ, respectively, in Hpt cells. Gene expression of CqImportin α1 (c), CqImportin β1 (d), and CqRan (e), respectively, was significantly reduced by RNAi in Hpt cells. WSSV gene transcription was substantially decreased by singe gene silencing of CqImportin α1 or CqImportin β1, double silencing of both CqImportin α1 and CqImportin β1 (f) or CqRan gene silencing (g) in Hpt cells. (h to k) WSSV replication was strongly reduced by blockade of CqImportin α1/β1 or of CqImportin β1/Ran. Expression of viral protein VP28 at 12 hpi was strongly decreased by IVM (h) or IPZ (i) treatment or gene silencing of CqImportin α1, CqImportn β1, both CqImportin α1 and CqImportn β1 (j), or CqRan (k). DMSO or GFP dsRNA-treated cells were used for control groups accordingly. (B) WSSV replication was significantly reduced by genes silencing of CqNup35. (a) Gene expression of CqNup35 was significantly reduced by gene silencing. Both gene transcription (b) and protein expression (c) of VP28 at 12 hpi were decreased after gene silencing of CqNup35. The Hpt dissected from three crayfish was defined as one biological replicate, and one representative result of the Western blot assay was presented accordingly. Significant differences were analyzed using Student's t test: **, P < 0.01; *, P < 0.05; ns, no significant difference.

Nuclear pore targeting of the incoming nucleocapsid is mediated by CqImportin α1/β1-Ran docked with the nucleoporins CqNup35/62.

Concerning the nucleocapsid, the most abundant component is the VP664 protein, which has a large molecular size of 664 kDa, accounts for more than 50% of the WSSV capsid proteins, and is regularly distributed around the periphery of the capsid (6, 26); additionally, both bipartite NLS with two stretches and monopartite NLS with one stretch of basic amino acids, which are evolutionarily molecular features for nuclear targeting signals (27) recruited for several virus infections in vertebrates (28–30), were predicted to exist in VP664 by NLS Mapper analysis (Table 1). To address how nuclear delivery of the incoming viral genome enclosed in the nucleocapsid is molecularly regulated, in terms of the possible binding to WSSV nucleocapsid for nuclear targeting, crayfish importin α1 was cloned for recombinant protein preparation followed by protein interaction analysis with VP664, considering that the gene transcript of CqImportin α1 was upregulated upon WSSV challenge in Hpt cells (data not shown). Similar to vertebrate importins, CqImportin α1 is composed of a typical N-terminal importin β1 binding domain (IBB) with a predicted internal NLS at amino acids 16 to 48 showing nuclear translocation activity (data not shown), a central region with eight helical Armadillo (ARM) repeat domains believed to bind to the cargo NLS, and an atypical ARM repeat at the C terminus proposed to bind with cellular apoptosis susceptibility protein (31). As the intact VP664 protein contains 6,077 amino acids and is difficult to recombinantly express at present, a truncated VP664^2160–2714 construct of residues 2,160 to 2,714 with a Myc-tag containing predicted bipartite NLS (residues 2,465 to 2,484) and a monopartite NLS (residues 2,465 to 2,474) (Table 1) were then constructed for a protein binding test with CqImportin α1 by coimmunoprecipitation (co-IP) assay. Meanwhile, wild-type CqImportin α1, ARM repeat domain, and CqImportin α1^ΔARM constructs lacking the ARM repeat domain all fused with the Flag tag were constructed and cotransfected with the truncated VP664^2160–2714 construct described above for overexpression in Human Embryonic Kidney 293T (HEK293T) cells. The cells were then examined with co-IP and Western blotting. As shown in Fig. 7A, VP664^2160–2714 exhibited clear binding affinity for both CqImportin α1 and the ARM repeat domain; however, no binding affinity was found for CqImportin α1^ΔARM lacking the ARM repeat domain. These data together suggest that the WSSV nucleocapsid is capable of binding to CqImportin α1, at least via the affinity of NLS-containing VP664 to ARM repeat domain in CqImportin α1.

TABLE 1.

The predicted NLSs in WSSV capsid proteins^a

Capsid protein	NLS	Sequence
VP15	Bipartite NLS1	11RRGSKKRSTTAGRISKRRSPSMKKRAGK38
	Bipartite NLS2	11RRGSKKRSTTAGRISKRRSPSMKKRAGKKSST42
	Monopartite NLS1	23RISKRRSPSM32
	Monopartite NLS2	30PSMKKRAGKK39
VP60	Monopartite NLS	362SNNKRKRNT370
VP95	Monopartite NLS	597PISKRPRKD605
VP136	Monopartite NLS	110SSSKRKRGSG119
VP190	Monopartite NLS	415RKRTKRQKTS424
VP664	Bipartite NLS	2465RKCKRRTSEDQCAFVKRVVR2484
	Monopartite NLS	2465RKCKRRTSED2474
VP51	No	No
VP76	No	No

^aPredicted amino acid sequences of bipartite NLS, monopartite NLS, or both bipartite and monopartite NLS were found in six viral capsid proteins, including VP15, VP60, VP95, VP136, VP190, and VP664, from eight annotated capsid proteins, but not in VP51 and VP76, deduced from WSSV genome. Predicted NLS amino acids are indicated with underlining in the selected capsid proteins.

FIG 7.

Nuclear transporter CqImportin α1/β1, CqRan, and nucleoporin CqNup35/62 are recruited for nuclear pore targeting of the incoming nucleocapsid enclosing WSSV genome. (A) The major capsid protein VP664 is recruited for nuclear pore targeting of the incoming nucleocapsids. VP664 protein bound with CqImportin α1 via binding to the ARM repeat domain. The truncated segments of Myc-tagged VP664^2160–2714 bound with Flag-CqImportin α1 and Flag-ARM, respectively, but not with Flag-CqImportin α1^ΔARM lacking the ARM repeat domain were determined by co-IP assays. (B) CqImportin α1 bound, via its IBB domain, with CqImportin β1. Myc-CqImportin β1 bound to Flag-CqImportin α1 or Flag-IBB domain of CqImportin α1; however, Myc-CqImportin β1 did not bind with Flag-CqImportin α1^ΔIBB lacking IBB domain. (C) CqImportin β1 bound, via its IBN domain, with CqRan. Myc-CqRan bound to Flag-CqImportin β1 or Flag-IBN domain of CqImportin β1 but not to the Flag-CqImportin β1^ΔIBN lacking IBN domain. (D) The formation of CqImportin α1/β1-Ran complex and the binding of CqImportin β1 to CqNup35/62. (a to c) Binding affinity between rCqImportin β1 and rCqImportin α1 or rCqRan. rHis-CqImportin β1 protein was captured on anti-penta-HIS biosensors immobilized with an anti-His tag antibody for binding examination with a gradient concentration of rGST-CqImportin α1 (a), rGST-CqRan (b), or rGST (c), respectively, by ForteBio Octet. K_D values were calculated by the ForteBio Data Analysis software 11. (d) Simultaneous binding of rCqImportin β1 to both rCqImportin α1 and rCqRan as identified by protein pulldown assay. rGST-CqImportin α1 and rGST-CqRan were pulled down by rHis-CqImportin β1 simultaneously via His pulldown assay. (e and f) CqImportin β1 bound to CqNup35/62 as examined by co-IP assay. Flag-CqImportin β1 bound to HA-CqNup35 and HA-CqNup62, respectively, by co-IP assays. For protein binding assay, the bound protein bands are indicated by black arrows. rGST was taken as control accordingly.

To elucidate whether a viral cargo-CqImportin α1/β1 complex can form for the nuclear targeting of the WSSV nucleocapsid, the recombinant proteins of both CqImportin β1 and the associated factor CqRan, both showing upregulation upon WSSV challenge in red claw crayfish, were prepared for functional study. In addition to an importin β N-terminal domain (IBN_N), CqImportin β1 contains a huntingtin, elongation factor 3, regulatory subunit A of protein phosphatase 2A, and target of rapamycin (HEAT) domain (data not shown), which is believed to be involved in the recognition of IBB (32). In addition, a small GTPase domain of the Ras superfamily is present in CqRan (data not shown); this domain is believed to play a crucial role in determining nuclear transport by interacting with the importin β family specifically bound to the GTP-bound form of Ran (33). Next, to elucidate whether crustacean CqImportin α1 indeed binds to CqImportin β1 to form a heterodimer for a cargo-importing function as proposed in vertebrates (32), recombinant CqImportin α1 fused with a Flag tag and CqImportin β1 fused with a Myc tag were constructed and overexpressed in HEK293T cells. As expected, CqImportin α1 showed clear binding affinity for CqImportin β1, as identified by the co-IP assay in HEK293T cells (Fig. 7B). Furthermore, a short basic N-terminal IBB domain in CqImportin α1 was also found to bind to CqImportin β1 (Fig. 7B), while the CqImportin α1^ΔIBB construct lacking the IBB domain exhibited no binding affinity with CqImportin β1 (Fig. 7B), strongly suggesting that the IBB domain is crucial for the binding of CqImportin α1 to CqImportin β1. Therefore, we proposed that this binding might occur to promote WSSV infection in Hpt cells because the dissociation of the IBB domain from the internal NLS binding site, via binding to importin β, contributes to the binding of cargo proteins, such as the nucleocapsid of WSSV, by vacating the NLS binding site in the ARM repeat domain as reported in vertebrates (34).

Additionally, to elucidate whether crustacean Ran has binding affinity for the shuttling receptor importin β1, recombinant CqRan was used in a co-IP assay with CqImportin β1 in vitro. As speculated, CqRan was found to bind with CqImportin β1 by binding to the IBN-N domain of CqImportin β1, as verified by co-IP assays (Fig. 7C). Furthermore, to determine whether the proposed crustacean CqImportin α1/β1-Ran complex formed similarly to the complex in vertebrates (35), the binding affinity of the rHis-CqImportin β1 protein, which was directly captured on anti-penta-HIS biosensors immobilized with an anti-His tag antibody, for both GST-tagged rCqImportin α1 and rCqRan was detected by the BLI method. As shown in Fig. 7D, the binding affinity of rCqImportin β1 to rCqImportin α1 or rCqRan was increased with the rising concentrations of rCqImportin α1 or rCqRan, respectively, in which the strong binding was observed (Fig. 7D, a and b). These data showed that rCqImportin β1 indeed had affinity to rCqImportin α1 and rCqRan, with K_D values of 53.71 ± 1.73 nM (Fig. 7D, a) and 13.79 ± 1.94 nM (Fig. 7D, b), respectively, based on the fitting curves of all concentrations and their associated residuals. However, there was not obvious binding affinity for the negative-control protein rGST (Fig. 7D, c). Meanwhile, the CqImportin β1 protein was capable of binding to both CqImportin α1 and CqRan simultaneously, as demonstrated by a protein pulldown assay (Fig. 7D, d). These findings demonstrate that crustacean importin α1, importin β1, and Ran are capable of forming the conserved classical pathway involved in nuclear import that is recruited for WSSV infection.

Importin β1 usually interacts with a series of nucleoporins for the nuclear import of the cargo-importin α1/β1 ternary complex (36). To examine whether crustacean nucleoporins were recruited for docking of the proposed cargo-importin α1/β1 complex to the nuclear pores via the binding of importin β1 to nucleoporins located on the cytoplasmic surface of the nucleus, CqNup35, CqNup37, and CqNup62 were used to create recombinant proteins, after which the protein interactions with CqImportin β1 were examined using a co-IP assay. As indicated in Fig. 7D, CqImportin β1 showed a clear binding affinity for both CqNup35 and CqNup62 (Fig. 7D, e and f) but not for CqNup37 (data not shown), suggesting that the binding of CqImportin β1 to CqNup35 and CqNup62 is involved in the initial docking step of the cargo-CqImportin α1/β1 complex at the nuclear pore for subsequent nuclear import. Taken together, our findings suggest, for the first time, that crustacean CqImportin α1 binds to both CqImportin β1 and the major nucleocapsid protein VP664 to form the proposed viral nucleocapsid cargo-CqImportin α1/β1 ternary complex, which dock to NPC via the binding of CqImportin β1 to the nucleoporin CqNup35/62. This process could facilitate the nuclear import of the WSSV genome with the action of CqRan as it binds to CqImportin β1 (Fig. 7C) to achieve the proposed catalytic activity (13).

Nuclear import of the expressed viral structural proteins is remarkably blocked by suppression of CqImportin α1/β1 activity.

To evaluate whether the assembly of WSSV progeny was affected after inhibition of the CqImportin α1/β1 activity, we primarily examined the time intervals for virion assembly by ultrastructural analysis with TEM in Hpt cells. As shown in Fig. 8A, a, the precursor empty shell structures, i.e., the initial assembly morphogenesis of WSSV progeny, were present at a middle infection stage of 12 hpi in the infected Hpt cell nucleus (Fig. 8A, a, a1); few virions were packed with an electron-dense DNA core and assembled with an integrated envelope at a relatively late infection stage of 18 hpi, whereas a large number of incompletely assembled provirions, i.e., those without an electron-dense DNA core and an integrated envelope, were present in the Hpt cell nucleus (Fig. 8A, b, b1); in contrast, most of the completely assembled WSSV progeny virions, i.e., those packaged with an electron-dense DNA core and encapsulated with the integrated viral envelope, were observed in the Hpt cell nucleus at 24 hpi (Fig. 8A, c, c1), except for a few incompletely enclosed empty shells without packaging of the WSSV genome and an integrated envelope (Fig. 8A, c, c2). This finding indicates that the assembly of WSSV progeny is to primarily form an incompletely enclosed empty shell that is probably composed of both envelope and capsid, after which the viral genome is packaged into it to assemble the mature progeny virions in the Hpt cell nucleus, similar to the assembly model of WSSV as reported by Tsai et al. (37).

FIG 8.

Pharmacological suppression of CqImportin α1/β1 activity led to the remarkably reduced nuclear import of the expressed viral structural proteins. (A) TEM analysis of the assembly of WSSV progeny virion in Hpt cell nucleus. The precursor structures of WSSV progeny were present in nucleus, i.e., the empty shells of the initial assembly at 12 hpi (a); most progeny virions were not fully assembled, i.e., without an electron-dense DNA core and integrated envelope, at 18 hpi (b); most progeny virions were completely assembled, i.e., with an electron-dense DNA core as well as integrated envelope, at 24 hpi (c). The empty shells of unassembled virions are indicated by yellow arrowheads. Fully assembled virions are indicated by red arrowheads, while incompletely assembled virions are indicated by black arrowheads. (B) Nuclear import of the expressed viral structural proteins was significantly reduced by blockade of CqImportin α1/β1 or of CqImportin β1/Ran in Hpt cells. (a) Dysfunction of CqImportin α1/β1 by IVM or of CqImportin β1/Ran by IPZ led to cytoplasmic accumulation of expressed VP28 and VP664 proteins in Hpt cells. CqImportin α1/β1 and CqImportin β1/Ran were blocked by IVM and IPZ, respectively, at 6 h post-WSSV infection, while control cells were treated by DMSO. The fluorescence was recorded at 18 hpi. (b) The fluorescence in Hpt cell nucleus for integrated optical density of VP28 or VP664 signals in panel a was quantified and is shown in histograms. (C) Propagation of WSSV was significantly reduced by the decreased nuclear import of the expressed viral structural proteins caused by pharmacological blocking of CqImportin α1/β1. The Hpt cells were infected by WSSV for 6 h before the pharmacological blockade, in which the viral titers were determined at 24 hpi. The control cells were treated by DMSO. Cell nucleus DNA was stained by DAPI (blue). **, P < 0.01; ns, no significant difference.

To further confirm whether nuclear import of these expressed viral proteins (Fig. 4B) is indeed mediated by the nuclear transporters CqImportin α1/β1 in crustacean host cells, crayfish Hpt cells were infected with WSSV for 6 h to achieve efficient production of viral structural proteins (data not shown). Next, the cells were treated with IVM or IPZ to examine its effect on nuclear translocation of the expressed viral structural proteins at 18 hpi by confocal fluorescence microscopy. As shown in Fig. 8B, the nuclear entry of the expressed capsid protein VP664 and the envelope protein VP28 was significantly inhibited in Hpt cells by blockade of CqImportin α1/β1 with IVM or of CqImportin β1/Ran with IPZ from 6 hpi and detected at 18 h after WSSV infection. The VP664 protein and the VP28 protein strongly accumulated on the periphery of the Hpt cell nucleus treated either by IVM or by IPZ but were rarely present within the nucleus compared to the proteins in the negative-control groups mock treated with DMSO (Fig. 8B, a). Accordingly, the relative fluorescence intensity of VP28 and VP664 in nucleus was significantly reduced by 65.0% and 79.2% or 54.7% and 75.7% (Fig. 8B, b), respectively, in Hpt cells treated with IVM or IPZ compared to control cells treated by DMSO. Furthermore, the virus titers of progeny virions were significantly decreased by 49.3% with IVM and 34.1% with IPZ at 24 hpi in Hpt cells (Fig. 8C) if infected by WSSV for 6 h before the pharmacological blockade. These findings demonstrated that the nuclear translocation of the expressed viral proteins, and the assembly of progeny virions, was substantially reduced after blockade of the CqImportin α1/β1 or of CqImportin β1/Ran, since pharmacological blockade was initiated in Hpt cells from 6 h post-WSSV challenge. Notably, the nuclear translocation of envelope proteins lacking an NLS such as VP28 is also obviously associated with the activity of CqImportin α1/β1 and CqImportin β1/Ran, and the mechanism might be different from that of capsid proteins containing NLS as proposed above; this possibility is worthy of further investigation.

Nuclear translocation of the expressed WSSV structural proteins containing NLS is dominantly dependent on the binding of NLS to the ARM of CqImportin α1.

The NLSs were predicted in six proteins among a total of eight known capsid proteins, including VP15, VP60, VP95, VP136, VP190, and VP664, but absent from VP51 and VP76 (Table 1), by performing NLS Mapper analysis on 526 proteins deduced from the complete genome of the WSSV-CN strain (26). A key question was then raised: are the nuclear import mechanisms for these two types of capsid proteins (proteins with and proteins without NLS) the same? To further investigate whether the predicted NLS sequences indeed affected the nuclear import of WSSV capsid proteins, the NLS of these capsid proteins, including VP15, VP60, VP95, VP136, VP190, and VP664, were selected for assessment of protein translocation activity from the cytoplasm to the nucleus in HEK293T cells, as described in Materials and Methods. In this experiment, the recombinant constructs deleted with bipartite or monopartite NLSs were fused with EGFP tags, which were defined as VP15^ΔbNLS2, VP60^ΔmNLS, VP95^ΔmNLS, VP136^ΔmNLS, VP190^ΔmNLS, and VP664^ΔbNLS accordingly. In addition, overexpression of the capsid proteins VP51 and VP76 (without any predicted NLS) was also performed to establish the control groups in HEK293T cells. Subsequently, the nuclear translocation of the selected NLS-defective constructs of capsid proteins, including EGFP-VP15^ΔbNLS2, EGFP-VP60^ΔmNLS, EGFP-VP95^ΔmNLS, EGFP-VP136^ΔmNLS, EGFP-VP190^ΔmNLS, and EGFP-VP664^ΔbNLS, was determined by immunofluorescence assay. As shown in Fig. 9A, a, the expressed EGFP vector protein showed a wide distribution of fluorescence in both cytoplasm and nucleus in the transfected HEK293T cells. Intriguingly, as expected, compared to the wild-type VP190 protein, which showed complete nuclear distribution, the VP190^ΔmNLS protein was substantially distributed throughout the cytoplasm, but not in the nucleus, in the transfected HEK293T cells (Fig. 9A, b). This finding indicates that the NLS in VP190 is a prerequisite for its nuclear translocation. In contrast to the complete nuclear translocation observed for the wild-type EGFP-VP136 (Fig. 9A, c), EGFP-VP95 (Fig. 9A, d), EGFP-VP664 (Fig. 9A, e), EGFP-VP15 (Fig. 9A, f), and EGFP-VP60 (Fig. 9A, g) proteins, the distribution of the NLS-deleted EGFP-VP136^ΔmNLS (Fig. 9A, c), EGFP-VP95^ΔmNLS (Fig. 9A, d), EGFP-VP664^ΔbNLS (Fig. 9A, e), EGFP-VP15^ΔbNLS2 (Fig. 9A, f), and EGFP-VP60^ΔmNLS (Fig. 9A, g) proteins was both cytoplasmic and nuclear in transfected HEK293T cells. In particular, the cytoplasm/nucleus ratios of the EGFP-VP136^ΔmNLS, EGFP-VP95^ΔmNLS, EGFP-VP664^ΔbNLS, EGFP-VP15^ΔbNLS2, and EGFP-VP60^ΔmNLS proteins were significantly increased by approximately 95.2%, 41.4%, 36.2%, 30.1%, and 26.8%, respectively (Fig. 9A, c, d, e, f, and g), compared with those of controls. This finding indicated that the nuclear translocation of nucleocapsid proteins with NLSs was substantially reduced by deletion of the NLSs. However, the NLS deletion of EGFP-VP136^ΔmNLS, EGFP-VP95^ΔmNLS, EGFP-VP664^ΔbNLS, EGFP-VP15^ΔbNLS2, and EGFP-VP60^ΔmNLS did not completely abrogate the nuclear translocation of these proteins like it did for EGFP-VP190^ΔmNLS (Fig. 9A, b, c, d, e, f, and g, right); the remaining translocation may have been achieved or facilitated by another unknown mechanism in addition to the NLS-dependent nuclear translocation pathway that can guide capsid proteins into the cell nucleus. In contrast, wild-type EGFP-VP51 and EGFP-VP76, both lacking any predicted NLS, were solely distributed in the cytoplasm of transfected HEK293T cells (Fig. 9A, h). Together, these data clearly indicate that the nuclear import of the expressed WSSV capsid proteins with the predicted NLS is dominantly dependent on intact NLS, even if it is not completely determined by the selected NLS in most of the NLS-containing capsid proteins except VP190 (the nuclear import of which is completely determined by the NLS). Importantly, our findings also demonstrate, for the first time, that the nuclear import of WSSV capsid components is both complicated and diverse, as other capsid proteins without NLS, such as VP51 and VP76, are not imported into the cell nucleus by the proposed NLS-mediated pathway like other capsid proteins with NLS, as described above. How these expressed capsid proteins without NLS are translocated into the cell nucleus for WSSV progeny assembly is still an enigma.

FIG 9.

Binding of viral NLS to ARM repeat domain in CqImportin α1 determines nuclear translocation of the expressed WSSV structural proteins. (A) The nuclear translocation of capsid proteins with NLS was substantially reduced by loss of their NLS. (a) The EGFP vector protein was widely expressed in both cytoplasm and nucleus of transfected HEK293T cells. The EGFP-tagged capsid proteins were transiently expressed in HEK293T cells. (b) The nuclear translocation of VP190 was completely abrogated by deleting NLS. The cytoplasm/nucleus ratios of the other capsid proteins containing NLS, including VP136 (c), VP95 (d), VP664 (e), VP15 (f), and VP60 (g), were clearly increased, i.e., decreased nuclear translocation, by deleting NLS. (h) No nuclear translocation activity was found with the capsid proteins lacking NLS, including VP51 and VP76. (B) The nuclear translocation of viral envelope proteins with NLS was significantly reduced by loss of their NLS. Nuclear translocation of envelope proteins VP52A (a) and VP124 (b) was completely abrogated by deleting NLS, while the predicted NLS in VP52B did not have nuclear translocation activity, i.e., lacking the authentic NLS activity (c). (d) Envelope proteins without predicted NLS, such as VP26 and VP28, exhibited no nuclear translocation activity. (C) The NLS containing viral structural proteins bound to CqImportin α1 via the binding of viral NLS to ARM repeat domain. (a to c) The NLS of viral structural proteins, including the capsid proteins VP15, VP60, VP95, VP136, VP190, and VP664, and the envelope proteins VP52A and VP124 (but not VP52B lacking the authentic NLS), bound to CqImportin α1. The predicted monopartite or bipartite NLS in capsid proteins and envelope proteins were deleted accordingly, and were all tagged with EGFP and transfected into HEK293T cells. Both capsid proteins and envelope proteins without any predicted NLS were also tagged with EGFP and overexpressed in HEK293T cells as the controls accordingly. Cell nucleus was stained by Hoechst 33258. (Right) The cytoplasm/nucleus ratio of fluorescent capsid proteins was quantified and is shown in histograms. Significant differences were analyzed using Student's t test. **, P < 0.01; ns, no significance. Binding between the NLS of viral structural proteins and CqImportin α1, ARM repeat domain, or CqImportin α1^ΔARM was performed accordingly in HEK293T cells and then detected by co-IP assays as described in Materials and Methods. The results shown are representatives of three biological repeats.

In addition to nuclear translocation of capsid proteins, nuclear import of envelope proteins must also be achieved in crustacean cells for WSSV progeny assembly. In contrast to the WSSV capsid proteins with NLS described above, most of the typical envelope proteins among the 22 currently annotated envelope proteins, such as VP28, VP26, VP24, and VP19, do not contain any predicted NLS; only three of the envelope proteins, i.e., VP52A, VP52B, and VP124, have been shown to have the predicted NLS (Table 2). To verify whether these predicted NLSs function in the nuclear translocation of VP52A, VP52B, and VP124, wild-type EGFP-VP52A, EGFP-VP124, and EGFP-VP52B and the NLS-deleted constructs EGFP-VP52A^ΔmNLS4, EGFP-VP124^ΔbNLS2, and EGFP-VP52B^ΔbNLS were transiently overexpressed in HEK293T cells as described above. In contrast to wild-type EGFP-VP52A, which was dominantly distributed in the cell nucleus but scarce in the cytoplasm, and EGFP-VP124, which was widely distributed in both the cytoplasm and nucleus, EGFP-VP52A^ΔmNLS4 and EGFP-VP124^ΔbNLS2 with deletion of the predicted NLSs were distributed throughout the cytoplasm and absent from the nucleus of the transfected HEK293T cells (Fig. 9B, a and b). This result confirms that the nuclear translocation of the envelope proteins VP52A and VP124 is indeed strongly dependent on the NLSs of these proteins. However, both the wild-type EGFP-VP52B with a predicted NLS and EGFP-VP52B^ΔbNLS solely exhibited cytoplasmic distribution, not nuclear distribution, in the transfected HEK293T cells, suggesting that the predicted NLS of the VP52B protein lacks authentic nuclear translocation activity (Fig. 9B, c). Moreover, overexpression of the main envelope proteins VP26 and VP28 without any predicted NLS, as mentioned above, was also performed in the control groups of HEK293T cells. As expected, wild-type EGFP-VP26 and EGFP-VP28 were entirely distributed in the cytoplasm of transfected HEK293T cells (Fig. 9B, d); similar distributions were also found for other envelope proteins lacking NLS, such as VP19 and VP24 (data not shown). Importantly, most of the envelope proteins without NLS, such as VP19, VP24, VP26, and VP28, were entirely distributed in the cytoplasm of the transfected HEK293T cells used as cell models for identification of the nuclear translocation activity of WSSV structural proteins; however, these proteins have to be translocated into crustacean cell nuclei for progeny virion assembly. Hence, we speculate that most envelope proteins without NLSs are not directly imported into crustacean cell nuclei by classical paradigms such as viral NLS-guided nuclear import pathways for both capsid proteins and envelope proteins with NLS; rather, these proteins might also be transported through the nuclear pores (Fig. 4B) in a manner manipulated or cofacilitated by WSSV factors such as other NLS-containing viral proteins for import into the nucleus during infection. This possibility is definitely worthy of further investigation, as the complete assembly of WSSV with an envelope in the crustacean cell nucleus is unique, which warrants further study.

TABLE 2.

The predicted NLSs in WSSV envelope proteins^a

Envelope protein	NLS	Sequence
VP52A	Monopartite NLS1	34PFRKRRKRKRY44
	Monopartite NLS2	34PFRKRRKRKR44
	Monopartite NLS3	36RKRRKRKRYR45
	Monopartite NLS4	36RKRRKRKRYRT46
	Monopartite NLS5	38RRKRKRYRTS45
VP52B	Bipartite NLS	35KKSKKRKIEDENEEEPVKTLED56
	Monopartite NLS1	35KKSKKRKIED45
	Monopartite NLS2	36KSKKRKIEDE45
VP124	Bipartite NLS1	78RKDPKKKRNLKGLEPASKKLA99
	Bipartite NLS2	78RKDPKKKRNLKGLEPASKKLAKNI102
VP12	No	No
VP14	No	No
VP16	No	No
VP19	No	No
VP24	No	No
VP26	No	No
VP28	No	No
VP31	No	No
VP32	No	No
VP33	No	No
VP38	No	No
VP39	No	No
VP41A	No	No
VP41B	No	No
VP56	No	No
VP90	No	No
VP110	No	No
VP150	No	No
VP187	No	No

^aPredicted amino acid sequences of bipartite NLS, monopartite NLS, or both bipartite and monopartite NLS were found only in three viral envelope proteins, including VP52A, VP52B (but lacking the authentic nuclear translocation activity), and VP124, deduced from WSSV genome. In contrast, most of the other envelope proteins are lacking NLS in the 22 annotated envelope proteins. The predicted NLS amino acids are indicated with underlining in the selected envelope proteins.

To further reveal whether CqImportin α1 and a key element, such as the ARM repeat domain, could directly bind to NLS of viral structural proteins for the proposed nuclear translocation mediated by the CqImportin α1/β1 transporter, wild-type CqImportin α1, the ARM repeat domain, and CqImportin α1^ΔARM constructs lacking the ARM repeat domain (all fused with a Flag tag) were cotransfected with NLSs of structural proteins, capsid proteins, or envelope proteins (as indicated in Materials and Methods) into HEK293T cells for co-IP assays. As shown in Fig. 9C, significant binding affinity was present between the indicated NLS from viral structural proteins, including both capsid proteins (such as VP15, VP60, VP95, VP136, VP190, and VP664) and envelope proteins VP52A and VP124, and the wild-type CqImportin α1 (Fig. 9C, a) as well as the ARM repeat domain (Fig. 9C, b). However, no binding affinity was found for the envelope protein VP52B lacking the authentic NLS (Fig. 9C, a, b, and c), which is in agreement with the result described above showing no nuclear translocation activity (Fig. 9B, c). In contrast, CqImportin α1^ΔARM, which lacked the ARM repeat domain, showed no binding affinity for any NLS of the viral structural proteins, as described above (Fig. 9C, c). This finding suggests that the ARM repeat domain is critical for the binding of WSSV structural proteins with NLS to CqImportin α1. Taken together, in addition to the nuclear targeting of the incoming nucleocapsid enclosing the WSSV genome verified above (Fig. 7), these data demonstrate that CqImportin α1/β1 also contributes to nuclear translocation of the expressed NLS-containing structural proteins of WSSV via binding of the ARM repeat domain within CqImportin α1 to viral NLS.

Assembly of WSSV progeny is strongly inhibited by suppression of CqImportin α1/β1 in vitro in Hpt cells.

Next, to better validate the effect of CqImportin α1/β1 activity on WSSV assembly via both targeting of nucleocapsids to nuclear pores for delivery of the viral genome and import of expressed viral structural proteins into the cell nucleus, the assembly of WSSV progeny virions was then determined by immunofluorescence analysis at both 18 hpi and 24 hpi in Hpt cells, in which the CqImportin α1/β1 was preblocked by pharmacological inhibitors before viral infection. Notably, as indicated by the colocalization of nucleocapsid and envelope proteins in the Hpt cell nucleus according to immunostaining with an anti-VP664 antibody and an anti-VP28 antibody, respectively, the assembly of WSSV progeny was strongly reduced at the relatively late infection stage of 18 hpi in the Hpt cell nucleus after blockade of CqImportin α1/β1 with IVM or of CqImportin β1/Ran with IPZ (Fig. 10A, a). The relative fluorescence intensity analysis showed that the amounts of VP28 and VP664 proteins were significantly decreased by 91.6% and 94.3%, respectively, in WSSV-infected Hpt cells pretreated with IVM or by 65.6% and 77.7%, respectively, in cells pretreated with IPZ (Fig. 10A, b). Additionally, the relative fluorescence intensity of the colocalized VP28 and VP664 was substantially reduced, with an 83.9% decrease in the Hpt cell nucleus after pretreatment with IVM (Fig. 10A, c). The reduction caused by IPZ was smaller, at 72.3%, than that caused by IVM (Fig. 10A, c). This finding indicated that the CqImportin α1/β1 accounted for greater efficiency of WSSV infection than the CqImportin β1 monomer alone under our experimental conditions. Furthermore, at the very late infection stage of 24 hpi, the relative fluorescent signal of the VP664 protein from progeny nucleocapsids was clearly reduced in the Hpt cell nucleus (Fig. 10B, a and b) compared to the signal in cells at 18 hpi (Fig. 10A, a and b). We speculate that the abundant fluorescent signal of VP664 was caused by incomplete packaging of the nucleocapsids by envelope proteins at 18 hpi, which exposed the nucleocapsids for efficient recognition by the VP664 antibody; in contrast, at the very late infection stage of 24 hpi, they were completely packed by envelope proteins, with the formation of a dense viral structure, which prevented the antibody from detecting the enclosed nucleocapsid proteins, in agreement with observations made by TEM analysis (Fig. 8A). Importantly, the assembly of WSSV progeny, as indicated by the immunofluorescent signal of the viral envelope proteins immunostained with an anti-VP28 antibody in the Hpt cell nucleus at the very late infection stage of 24 hpi, was strongly reduced after CqImportin α1/β1 and CqImportin β1/Ran activity was blocked with IVM and IPZ, respectively (Fig. 10B, a). The relative fluorescence intensity analysis showed that the amounts of the VP28 and VP664 proteins were significantly decreased in WSSV-infected Hpt cells if the cells were pretreated with IVM (87.9% and 91.6% decreases, respectively) or pretreated with IPZ (76.6% and 86.7% decreases, respectively) (Fig. 10B, b). In addition, as a portion of progeny virions were not fully assembled (Fig. 10A), the amount of VP664 protein capable of being detected by the anti-VP664 antibody but with reduced efficiency at the very late infection stage of 24 hpi (Fig. 10B, b) also showed a trend similar to that of VP28, even though the VP664 fluorescent signal was much lower in the Hpt cell nucleus. Moreover, the relative fluorescence intensity of VP28 was significantly lower (87.9% decrease) in the nucleus of Hpt cells treated with IVM than in those of cells treated with IPZ (76.6% decrease) at 24 hpi compared to the nucleus of the mock-treated cells. This finding indicates again that CqImportin α1/β1 accounts for much greater efficiency for WSSV infection than CqImportin β1/Ran (Fig. 10B, b), in agreement with the inhibitory efficiency at 18 hpi shown above (Fig. 10A, b and c). Accordingly, no obvious difference in the fluorescent signal of VP664 in the cell nucleus was found between the IVM- and IPZ-treated groups, which exhibited 91.6% and 86.7% decreases, respectively (Fig. 10B, b), as most of the nucleocapsids had been completely enclosed in envelopes in the intact virions at 24 hpi (Fig. 8A, c); this enclosure prevented detection of the nucleocapsids by the anti-VP664 antibody, as proposed above. These data together strongly indicate that the CqImportin α1/β1 is critical for WSSV infection, at least by promoting both the targeting of the viral nucleocapsid toward the cell nuclear pores for the subsequent release of the viral genome into the host cell nucleus and the delivery of expressed viral structural proteins into the host cell nucleus.

FIG 10.

Pharmacological suppression of CqImportin α1/β1 activity led to the remarkably reduced assembly of WSSV progeny in vitro. (A) Assembly of WSSV progeny virion was significantly reduced by the decreased import of both viral genome and expressed NLS-containing proteins into Hpt cell nucleus via blockade of CqImportin α1/β1 or of CqImportin β1/Ran. (a) The incomplete assembly of WSSV progeny virions was significantly decreased by inhibition of CqImportin α1/β1 or CqImportin β1/Ran in Hpt cells. Most of the incompletely assembled WSSV virions were present in control cells treated by DMSO but rarely present in cells pretreated by IVM or IPZ at the relatively late infection stage of 18 hpi. (b) Relative quantification of the integrated optical density of fluorescent envelope protein VP28 or capsid protein VP664 in panel a by histogram analysis. (c) The assembly of WSSV progeny virion was significantly reduced by pharmacological blockade of CqImportin α1/β1 or CqImportin β1/Ran in Hpt cells. WSSV assembly was statistically analyzed according to the colocalized fluorescence intensity of both envelope protein VP28 and capsid protein VP664 in panel a. (B) Propagation of WSSV progeny virion was significantly reduced by the decreased import of both viral genome and expressed NLS-containing proteins into Hpt cell nucleus via blockade of CqImportin α1/β1 or of CqImportin β1/Ran. (a) The completely assembled WSSV progeny virions were indicated by localization of fluorescent envelope protein VP28 in Hpt cell nucleus. WSSV progeny virions were rarely present in Hpt cell nucleus, if pretreated by IVM or IPZ, but were mostly present in control cells treated with DMSO at a very late infection stage of 24 hpi. (b) Relative quantification on the integrated optical density of fluorescent capsid protein VP664 or envelope protein VP28 in panel a by histogram analysis. Viral components were immunostained by both anti-VP28 and VP664 antibodies. Cell nucleus DNA was stained by DAPI (blue). **, P < 0.01.

WSSV propagation is substantially inhibited by suppression of CqImportin α1/β1 activity in vivo, resulting in a significantly promoted survival rate in crayfish under viral challenge.

To further elucidate whether inhibition of CqImportin α1/β1 could protect the host from WSSV challenge, the survival rate of crayfish was determined after single gene silencing of CqImportin α1 or CqImportin β1 alone or double gene silencing of both CqImportin α1 and CqImportin β1 in the animals followed by WSSV challenge. The results showed that the transcription of both the CqImportin α1 and CqImportin β1 genes was significantly reduced in vivo in Hpt in crayfish at 36 h after the second round of RNAi via dsRNA injection. The transcript levels of CqImportin α1 (approximately 88.2% and 86.7% decreases; P < 0.01) or CqImportin β1 (approximately 86.5% and 86.8% decreases; P < 0.01) were significantly reduced by either single gene silencing or double gene silencing, respectively, compared with the levels in the control groups subjected to mock treatments with GFP dsRNA (Fig. 11A, a and b). Consequently, both gene expression and protein expression of WSSV proteins, such as VP28, were significantly reduced at 24 hpi in vivo in Hpt in crayfish after single gene silencing of CqImportin α1 or CqImportin β1 or after double gene silencing of CqImportin α1 and CqImportin β1 compared to the levels in mock-treated control animals (Fig. 11A, c and d; P < 0.01). Furthermore, the virus titers, as determined by absolute quantification of VP28 gene with PCR, also exhibited substantial reductions of 42.1%, 45.0%, and 74.1% in the animals with silencing of CqImportin α1, CqImportin β1, or both CqImportin α1 and CqImportin β1 compared to the animals subjected to mock treatments with GFP dsRNA (Fig. 11A, e; P < 0.01). Intriguingly, the survival rate was significantly higher, at approximately 29.1%, in crayfish after double gene silencing of both CqImportin α1 and CqImportin β1 than after mock treatment with GFP dsRNA, which resulted in 100% mortality (Fig. 11A, f; P < 0.01). However, there was an approximately 21.5% survival rate verified in CqImportin β1 silencing group than CqImportin α1 silencing, while the survival rate of CqImportin α1 silencing had no significant difference from the mock treatments with GFP dsRNA or with a double dose of GFP dsRNA (Fig. 11A, f; P < 0.01), as all the crayfish in these three groups died on the 8th day after viral challenge. Accordingly, this survival rate was in accordance with that of the WSSV gene replication assay, which showed significantly reduced viral replication in vivo in the Hpt of the animals after double gene silencing of both CqImportin α1 and CqImportin β1 (Fig. 11A, c and d) and with that of WSSV propagation, which showed substantially decreased viral titers in vivo in Hpt of the silenced animals (Fig. 11A, e). Furthermore, the survival rate of crayfish was also analyzed after the animals were fed a diet containing IVM and then subjected to WSSV challenge. As shown in Fig. 11B, a, the viral gene expression of VP28 in Hpt was significantly decreased by 78.8% at 24 hpi in vivo in the animals fed the diet containing IVM compared to the control animals without inhibitor feeding (P < 0.01). Accordingly, the expression of WSSV structural proteins, as detected by anti-VP28 antibody binding, was also significantly reduced by 86.7% at 24 hpi in vivo in Hpt from crayfish fed the IVM-containing diet (Fig. 11B, b; P < 0.01). Intriguingly, the WSSV titers exhibited a substantial reduction of 77.1% in the Hpt of the animals fed the IVM-containing diet compared with the control animals without inhibitor feeding (Fig. 11B, c; P < 0.01). Consequently, the crayfish survival rate was significantly increased in the animals fed the IVM-containing diet to block the activity of CqImportin α1/β1. Specifically, it was approximately 42.6% higher than the survival rate of the animals fed the diet without IVM (Fig. 11B, d), indicating that a possible protective strategy against WSSV disease in crustaceans could involve blocking CqImportin α1/β1 activity with IVM. IVM has already been widely used in livestock as an antiparasitic remedy but thus far has been less commonly used in aquaculture (38, 39). Collectively, these data demonstrate that the importin pathway is critically recruited for WSSV infection in a crustacean, the red claw crayfish, and provide promising insights for anti-WSSV target design in crustaceans.

FIG 11.

Blockade of CqImportin α1/β1 results in significant protection of crayfish against WSSV challenge. (A) Survival rate of crayfish under WSSV challenge was significantly increased after gene silencing of CqImportin α1/β1. Gene expression of CqImportin α1 (a) and CqImportin β1 (b) was significantly decreased by RNAi in vivo in Hpt. Transcription of VP28 gene (c), protein expression of VP28 (d), and viral titers of WSSV (e) were significantly decreased by gene silencing of CqImportin α1, CqImportin β1, or CqImportin α1/β1 in vivo in Hpt at 24 hpi. (f) Survival rate was significantly higher in crayfish after double gene silencing of both CqImportin α1 and CqImportin β1 or single gene silencing of CqImportin β1 if compared with those animals by single gene silencing of CqImportin α1 or by mock treatments with GFP dsRNA. (B) Survival rate of crayfish under WSSV challenge was significantly increased by oral administration with the diet containing IVM. Both WSSV replication, evaluated by presence of VP28 gene transcript (a) and expression of VP28 protein (b), and propagation of WSSV progeny determined by viral titers (c) were clearly inhibited by oral administration of crayfish with the diet containing IVM. (d) Survival rate of crayfish was significantly increased by oral administration with IVM-containing diet. Histograms were performed with at least three biological replicates. All data are presented as the means ± SD. Significant differences were analyzed using Student's t test: **, P < 0.01; *, P < 0.05. The differences are represented with P values for survival rate of crayfish between the two groups analyzed with log-rank test using the software of GraphPad Prism 5.0.

DISCUSSION

Knowledge of the host factors and regulation required for WSSV infection is largely limited. In this study, the crayfish actin cytoskeleton was shown to be recruited for the cellular uptake of WSSV mediated by protruding filopodia (Fig. 1A) in Hpt cells. We further proved that WSSV entry is facilitated by actin nucleation promoted by the CqArp2/3 complex to achieve actin polymerization to push forward the leading edge of motile cells for endocytosis, similar to the role of the Arp2/3 complex in vertebrates (40). As the polymerization of actin filaments may provide forces that contribute to deformation of the plasma membrane to bring the cargoes into the cytoplasm (41), our findings suggest that remodeling of the actin cytoskeleton is a prerequisite for the efficient endocytosis of WSSV, i.e., WSSV entry into host cells via the CME pathway at the molecular level. Thus, we conclude that WSSV, upon attachment to the cell surface, may activate the CqArp2/3 complex to nucleate actin filaments to support its entry into Hpt cells. Additionally, the frequently observed association of viruses with filopodia is not coincidental; rather, it represents an efficient infectious pathway in the case of endocytosed viruses such as herpes simplex virus and human papillomavirus type 16 that can result in the direct transport of viruses into clathrin-coated pits (42, 43). By localizing WSSV-containing endocytic vesicles, we found that these vesicles were colocalized with the crayfish actin cytoskeleton component F-actin, which may also participate in the detachment of WSSV-containing vesicles from the Hpt cell plasma membrane (Fig. 2A). Intriguingly, the formation of enlarged WSSV-containing CME vesicles was substantially induced by viral infection at an early infection stage in Hpt cells. However, how the induction of the formation of these endocytic vesicles is triggered at the molecular level by WSSV challenge remains to be further elucidated.

In the present study, we further demonstrated that incoming WSSV particles were first encapsulated in endocytic vesicles (Fig. 2A). With the fusion of endocytic vesicles and endosomes, the virions successfully entered the endosomes (Fig. 3A, a), as verified by the cytoplasmic localization of WSSV, endocytic vesicles, and endosomes. Endosomes have been well recognized as the main sorting stations in the endocytic pathway, the membrane, and contents of which are mainly derived from primary endocytic vesicles that fuse with each other. In particular, the formation of enlarged endosomes, in agreement with the formation of the WSSV-containing CME vesicles mentioned above, was also obviously induced after WSSV infection, which is consistent with our previous report that cytoplasmic WSSV virions tend to accumulate in endosomes in the early infection stage (9). Importantly, we also identified the direct fusion process between the WSSV envelope and endosomal vesicle membrane by TEM analysis at the early infection stage, which strongly supports our conclusion that WSSV envelope fusion indeed occurs in endosomes (in agreement with our previous report), as primarily indicated by the fluorescence colocalization between the WSSV envelope and endosomal vesicles (9). This finding is similar to a finding indicating that the viral core of a nucleocytoplasmic large DNA virus, the African swine fever virus, is released into the cytosol by membrane fusion (44). As envelope fusion is one of the determinant steps for WSSV infection, how this fusion is initiated, and particularly which key molecules from both the WSSV envelope and endosome membrane are recruited at the first trigger, is worthy of clarification and further investigation, even though we have already identified that the reduced pH in endosomes is a prerequisite for viral fusion (9).

In our present study, formation of enlarged endosomes induced by WSSV challenge was prevented when microtubules were depolymerized with ND in Hpt cells. This finding suggests that microtubules play a positive role in the formation of enlarged endosomes and even in further transport of endosomes after fusion with endocytic vesicles containing virions, especially considering that transport of Semliki Forest virus from early to late endosomes is assisted by microtubules (45). This finding is also similar to findings in rat kidney cells with ceased movement of endosomes and in Vero cells with dispersed late endosomes and lysosomes throughout the cytoplasm; delayed maturation of endosomes was observed if microtubules were depolymerized with ND (46). Therefore, the formation of enlarged endosomes containing WSSV virions was disturbed due to the depolymerization of microtubules, which prevented envelope fusion and nucleocapsid penetration. Unfortunately, we do not have commercial antibodies suitable for distinguishing late endosomes from early endosomes in more detail in crayfish at present. Thus, the mechanism needs further confirmation once our monoclonal antibodies (currently in preparation) are available. Furthermore, the cytoplasmic transport of incoming WSSV nucleocapsids (after exit from endosomes and movement toward the perinuclear membrane) was substantially blocked by depolymerization of microtubules with ND in Hpt cells, and this effect was accompanied by significantly reduced WSSV replication (Fig. 3D and E). This result demonstrates that transport of nucleocapsids, after shedding of the envelope, also depends on microtubules. The microtubules could act as bridges during both the transport of WSSV within endosomes and exit of the penetrated nucleocapsids from the endosomes. Together, these findings demonstrate that microtubules are hijacked for the cytoplasmic transport of nucleocapsids toward the perinuclear membrane in crustaceans. Further investigations are necessary to reveal the molecular regulation (for example, the dynamic mechanism of the intracellular transport of the penetrated nucleocapsids and possibly also the endosome transport before nucleocapsid penetration) mediated by microtubules and the associated molecules.

By using TEM analysis together with immunogold labeling, we showed for the first time, at the ultrastructure level, that nucleocapsids containing the WSSV genome were located in the vicinity of the nuclear pores in Hpt cells. The enclosed viral genome was believed to be released from these nucleocapsids into the host cell nucleus via the nuclear pores. After the release of the WSSV genome, empty capsids without an electron-dense viral DNA core were present within the nuclear basket (Fig. 4A, c) but not inside the nucleus, again indicating the importance of nucleocapsid targeting to nuclear pores for WSSV genome release into the host cell nucleus. Importantly, we found that targeting of the incoming nucleocapsids to nuclear pores is mediated by CqImportin α1/β1, as both pharmacological blocking and gene silencing of CqImportin α1/β1 resulted in the substantial accumulation of incoming WSSV nucleocapsids on the periphery of the Hpt cell nucleus and subsequently reduced viral replication (Fig. 5D and 6). The NLS contributes to the nuclear import activity of CqImportin α in red claw crayfish, similar to the nuclear import activity of importin α1 in vertebrates (47). Furthermore, CqImportin α1 binds to the NLS from capsid proteins of WSSV (Fig. 9C). As the capsid proteins of most DNA viruses contain one or more NLSs required for guiding viral capsid components into the host cell nucleus through the nuclear pores, we propose that the multiple NLSs at least within VP664 also facilitate the efficient targeting of incoming nucleocapsids toward the host cell nuclear pores for genome release. In addition to VP664, whether the other capsid proteins of WSSV with NLSs for which localization information within the capsid is lacking due to the unavailable antibodies at present, such as VP15, VP60, VP136, and VP190, are also involved in the nuclear targeting of incoming nucleocapsids during the early infection stage is worthy of further study.

Notably, the expressed structural proteins, such as envelope proteins, in addition to capsid proteins, must be imported into crustacean cell nucleus for WSSV progeny assembly. Despite the importance of this critical aspect of WSSV infection, very little is known about how WSSV succeeds in the management of its expressed capsid proteins to gain access to the cell nucleus for genome packaging and progeny virion assembly as well as for encapsulation of the nucleocapsids with envelope proteins within the crustacean cell nucleus. Importantly, we found that deletion of NLSs in the capsid proteins VP15, VP60, VP95, VP136, VP190, and VP664 or the envelope proteins VP52A and VP124 substantially reduced the nuclear translocation of these proteins (Fig. 9A and B), particularly that of VP190, VP52A, and VP124, as it completely abrogated the nuclear translocation of these proteins. This finding demonstrates that the nuclear translocation of WSSV structural proteins with NLS for progeny virion assembly is dependent on intact NLS. Intriguingly, after blockade of CqImportin α1/β1 with IVM or of CqImportin β1/Ran with IPZ, VP664 strongly accumulated on the outer membrane of the Hpt cell nucleus, and this accumulation was accompanied by a subsequent substantial reduction in the presence of VP664 in the nucleus (Fig. 8B). Hence, CqImportin α1/β1 is also recruited, via binding to viral NLS by ARM repeat domain within CqImportin α1, to form NLS-containing viral protein cargo-CqImportin α1/β1 ternary complex (Fig. 7 and 9C) to transport expressed WSSV structural proteins with NLS, including the indicated capsid proteins (Fig. 9A) and envelope proteins (Fig. 9B), into the Hpt cell nucleus for progeny virion assembly through nuclear pores mediated by the binding of CqImportin β1 to nucleoporin CqNup35/62 (Fig. 7D). Collectively, the nucleoporins in crustaceans, such as CqNup35 and CqNup62 at least, are recruited for the docking of WSSV cargo-CqImportin α1/β1 ternary complex to cell nuclear pores via the binding of CqImportin β1 to the evolutionarily conserved NPC proteins CqNup35/62 localized on the surface of the nucleoplasm, which is similar to that in vertebrates (14, 15); the cargo can be either the incoming nucleocapsids or the expressed structural proteins containing NLSs.

On the other hand, no NLS was present in the other two predicted nucleocapsid proteins, VP51 and VP76. Furthermore, except for VP52A and VP124, which have characterized NLS activity, most of the typical envelope proteins among the 22 currently annotated envelope proteins of WSSV, such as VP19, VP24, VP26, and VP28, do not contain any predicted NLS (26). Unfortunately, the mechanism by which these types of structural proteins (both envelope proteins and capsid proteins without NLS) are imported into the host cell nucleus for WSSV assembly is unclear. Intriguingly, the expressed VP28 protein without an NLS also accumulated on the outer membrane of the nucleus in Hpt cells after blockade of CqImportin α1/β1 with ivermectin or of CqImportin β1/Ran with importazole during WSSV infection (Fig. 8B). We speculate that the nuclear import pathways or key factors mediating the nuclear translocation of the VP28 protein were also blocked by inhibition of CqImportin α1/β1, resulting in retention of VP28 on the outer membrane of the nucleus in Hpt cells. Consequently, the assembly of progeny WSSV was significantly decreased at the late infection stage (Fig. 8B). In herpes simplex virus 1 (HSV-1), the major capsid protein VP5, the capsid protein VP23, and the small capsid protein VP26, all of which lack an NLS, are not capable of nuclear import on their own (48). However, the nuclear import of VP5 is facilitated by the capsid scaffolding protein VP22a for localization to the cell nucleus (49), an NLS of the triplex capsid protein VP19c is responsible for the nuclear import of the other triplex protein VP23, and the nuclear localization of VP26 is only observed when VP5 is present together with either VP19C or preVP22a (50). Given the above-described example of HSV-1 proteins imported into the cell nucleus, we had examined whether the WSSV capsid proteins VP15 and VP664 as well as the envelope proteins VP52A and VP124 (all of which have both bipartite NLS and monopartite NLS) could facilitate the nuclear import of structural proteins without NLS, including the capsid proteins VP51 and VP76 and the envelope proteins VP19, VP24, VP26, and VP28, by cotransfection of their recombinant plasmids in HEK293T cells followed by localization of the overexpressed proteins. Unfortunately, the nuclear import of these structural proteins without NLS was not clearly promoted by the presence of VP15, VP664^2160–2714, VP52A, or VP124^1–546 in HEK293T cells under our experimental conditions (data not shown). This lack of promotion might have been caused by the absence of some putative viral factors as well as host factors uniquely expressed in crustaceans during WSSV infection that need further investigation. Additionally, whether the expressed envelope proteins without NLS enter the Hpt cell nucleus via a mechanism that is currently unverified, such as by binding to nucleoporins or other putative transporting factors (51), certainly needs more experimental evidence. Taken together, the evidence indicates that the different types of WSSV structural proteins, with or without NLS, complicate the nuclear translocation of these expressed proteins. Notably, the detailed molecular regulation mechanism, particularly regarding envelope proteins in terms of their unique encapsulation within the host cell nucleus and their determinant role in WSSV invasion (52), is definitely worthy of more attention, as elucidating these key points can definitely enhance understanding of the molecular interaction and coevolution between hosts and viruses.

As proposed in Fig. 12, WSSV hijacks microfilaments via actin nucleation enhanced by the CqArp2/3 complex and the CME pathway to initiate invasion. After delivery into endosomes from endocytic vesicles mediated by microtubules, viral envelope fusion occurs within endosomes and is followed by exit of the nucleocapsids from endosomes into the cytosol. Subsequently, the penetrated nucleocapsids are transported along microtubules in a manner mediated by the NLS of capsid proteins such as VP664 to the peripheral nuclear area. In particular, both the nuclear pore targeting of the nucleocapsids for viral genome release into the cell nucleus and the nuclear import of expressed viral structural proteins containing NLS bound by the ARM repeat domain of CqImportin α1 are mediated by the nuclear transporter CqImportin α1/β1 with the action of CqRan. They formed viral cargo-CqImportin α1/β1 ternary complex docks to cell nuclear pores via binding of CqImportin β1 to the nucleoporin CqNup35/62. This is a smart strategy for hijacking of host molecules by WSSV to achieve successful infection. In summary, our findings shed new light on the pathogenesis of WSSV, which contributes to the design of novel targeting strategies against WSSV disease, such as inhibition of nuclear import factors like CqImportin α1/β1 by the IVM inhibitor.

FIG 12.

Schematic model for nuclear import of WSSV: the incoming genome for viral gene replication and the expressed structural proteins for progeny assembly. For successful infection, WSSV hijacks microfilaments via actin nucleation enhanced by the CqArp2/3 complex and the CME pathway to initiate invasion. The microtubules are hijacked by WSSV for cytoplasmic trafficking. The WSSV cargo-CqImportin α1/β1 ternary complex docks to cell nuclear pores via binding of CqImportin β1 to the nucleoporin CqNup35/62. Both the nuclear pore targeting of the nucleocapsids, for viral genome release into the cell nucleus, and the nuclear import of expressed viral structural proteins, containing NLS bound by the ARM repeat domain of CqImportin α1, are mediated by the nuclear importer CqImportin α1/β1 with the action of CqRan. In contrast, the WSSV infection could be significantly inhibited after dysfunction of the actin nucleation promoted by the CqArp2/3 complex, the microtubule activity, the CqNup35, and particularly the CqImportin α1/β1-Ran complex.

MATERIALS AND METHODS

Animals, Hpt cell culture, WSSV preparation, and infection.

Healthy red claw crayfish (30 ± 5 g for preparing Hpt cell culture, 15 ± 2 g for survival rate after viral challenge) were purchased from Yuansentai Agricultural Science and Technology Co., Ltd., of Zhangzhou, Fujian Province, China. Crayfish was acclimated in aerated tanks at 25°C for at least 2 weeks before experiments. Only intermolting male crayfish free of WSSV were used in this study. Hpt cells isolated from crayfish were cultured as described by Söderhäll et al. (52) and Liu et al. (21). WSSV inoculum was prepared and quantified by absolute quantification via PCR as described by Xie et al. (5). WSSV infection in vitro in Hpt cell cultures was performed at the multiplicity of infection (MOI; the ratio of virus to cell) we previously described (8, 9). For WSSV challenge in vivo in crayfish, each animal was injected with 100 μL of crayfish saline (CFS) buffer containing 1 × 10⁵ of virions via the base of the fourth walking leg.

For WSSV challenge in Hpt cells, the infection times chosen were based on the interpretation for different stages of WSSV invasion and replication, which were primarily determined with a time course of infection, including endocytosis of virion at 0.5 hpi, presence in the cytoplasm at 1 hpi, intracellular transport within cell cytoplasm at 1.5 hpi to 3 hpi, viral gene replication from an early time around 3 hpi, viral protein expression from 6 hpi, nuclear translocation of the expressed viral structural proteins at 12 hpi, initial assembly of progeny starting from 12 hpi, and complete assembly of progeny around 24 hpi in nucleus. These time points were in agreement with our previous description (9).

Reagents and antibodies.

Chemical inhibitors CK666, IVM, IPZ, CHX, and ND (all from Sigma-Aldrich, USA) were optimized by primary experiments to determine the concentrations without cytotoxicity for in vitro assays in Hpt cell cultures. Hpt cell nucleus was stained by 4′,6-diamidino-2-phenylindole (DAPI; Beyotime, China) in fixed cells and by Hoechst 33258 (Beyotime, China) in live HEK293T cells. Mouse monoclonal antibody against VP28 and rabbit polyclonal antibody against VP664 of WSSV were kindly provided by Feng Yang (Third Institute of Oceanography, Ministry of Natural Resources, Xiamen, Fujian, China) and Han-ching Wang (National Cheng Kung University, Tainan, Taiwan), respectively. Rabbit polyclonal antibody against RabGEF1 (ABclonal, China) and mouse monoclonal antibody against CqCLC (previously prepared in our lab) were used for the immunostaining assay. Tetramethyl rhodamine isocyanate-phalloidin and tubulin-tracker red were purchased from Yeasen (China) to stain F-actin and microtubules accordingly. FM1-43 was purchased from ThermoFisher to stain the endocytic vesicles.

Mouse monoclonal antibodies against β-actin, horseradish peroxidase (HRP)-conjugated secondary antibodies against mouse or rabbit IgG, Alexa Fluor 488-conjugated donkey anti-rabbit, and Alexa Fluor 633-conjugated goat anti-mouse were provided as we previously described (8, 9). Mouse monoclonal antibody against Flag tag, hemagglutinin (HA) tag, GST tag, and 6× His tag were purchased from Cell Signaling Technology. The secondary antibodies of 12-nm and 25-nm colloidal gold anti-rabbit or -mouse IgG were purchased from ThermoFisher.

Gene cloning and sequence analysis.

Full-length cDNA sequences of CqArpc1A, CqArpc3, CqArpc4, CqArpc5, CqImportin α1, CqImportin β1, CqRan, CqNup35, CqNup37, and CqNup62 were cloned from a transcriptomic library of Hpt from C. quadricarinatus (21). The open reading frame and amino acid sequences of genes mentioned above were analyzed by using the ExPASy translate tool (http://web.expasy.org/translate/). The NLSs of viral structural proteins and CqImportin α1 were predicted by the NLS mapper program (http://nls-mapper.iab.keio.ac.jp/cgi-bin/NLS_Mapper_form.cgi).

Recombinant expression and purification of crayfish proteins.

For recombinant protein preparations, protein coding sequence of CqArpc1A, CqImportin α1, CqImportin β1, or CqRan was expressed with pGEX-4T-1 or PET-28a vector followed by protein purification as we previously described (8, 9) and fused with GST or His tag accordingly at the N terminus of recombinant protein. The presence of recombinant protein was confirmed by Western blotting.

Gene silencing or pharmacological blockade in vitro in Hpt cell cultures and in vivo in crayfish.

For gene silencing by dsRNA, dsRNA preparation and RNAi of CqArpc1A, CqArpc3, CqArpc4, CqArpc5, CqImportin α1, CqImportin β1, CqRan, CqNup35, CqNup37, and CqNup62 were performed as we previously described (8, 9). The expression of viral VP28 was detected after gene silencing or pharmacological treatment in Hpt cells in vitro. For WSSV challenge in vivo after gene silencing of importins, the animals were divided into five groups: CqImportin α1, CqImportin β1, CqImportin α1 plus CqImportin β1, GFP, and GFP plus GFP dsRNA groups. RNAi assay was performed by injecting dsRNA (1 μg/g of body weight) of CqImportin α1, CqImportin β1, and CqImportin α1 plus CqImportin β1, respectively, and the GFP dsRNA was taken as the mock control. The second injection of dsRNA was given after 24 h of the first injection. RNAi efficiency at mRNA level in Hpt dissected from crayfish was detected by qRT-PCR. Crayfish was challenged with WSSV at 36 h after the second dsRNA injection. For WSSV challenge experiments in vivo after pharmacological blocking of importins, crayfish was fed with IVM-containing pellet diet (0.1 g/kg of body weight) once every 2 days for 2 weeks; crayfish were then challenged with WSSV and fed with the pellet containing IVM as described above during the rest of the postviral challenge. After viral challenge, the survival rate was calculated by counting the number of surviving animals every day. All experiments were repeated three times, in which 20 crayfish were used in each treatment per assay. The survival curves were presented as Kaplan-Meier plots, in which differences were analyzed with log-rank test using the software GraphPad Prism 5.0. To examine WSSV replication in crayfish after inhibition of CqImportin α1/β1, the transcript level of VP28 gene was examined by qRT-PCR at 24 hpi in Hpt tissue dissected from crayfish after gene silencing of CqImportin α1/β1 genes or pharmacological blockade by IVM, respectively, and the corresponding viral titers were determined using an absolute quantitative PCR (26). This experiment was biologically repeated three times, and the data were expressed as means ± standard deviations (SD) and statistically analyzed by Student's t test.

Proteins binding assays and chemical inhibition.

The rHis-CqImportin β1, rGST-CqImportin α1, and rGST-CqRan proteins were prepared for binding affinity examination by a ForteBio Octet R8 device (Sartorius, Germany). rCqImportin β1 protein was captured on anti-penta-HIS biosensors immobilized with an anti-His tag antibody and tested for binding with a gradient concentration of 0, 0.002, 0.02, 0.2, 2, and 20 mM for rCqImportin α1 protein or rCqRan protein. The control protein rGST was used with the same concentrations as those of the tested proteins described above. Binding affinity of IVM or IPZ to rCqImportin α1 or rCqImportin β1, respectively, was determined as described by Concepcion et al. (53). Purified rCqImportin β1 and rCqImportin α1 were labeled with biotin and immobilized on Super Streptavidin biosensors to test the binding affinity with 4, 8 16, 32, and 64 μM IVM or IPZ, and data were collected over time. To further detect inhibitory binding of rCqImportin β1 to rCqImportin α1 or rCqRan by IVM or IPZ, respectively, the affinities were determined as shown above. Briefly, the rCqImportin β1 was captured on the anti-penta-HIS biosensors and the affinities were detected in the presence of a gradient concentration of 0, 8, 32, and 128 μM for IVM or IPZ, respectively. The binding kinetics and the affinity constants were calculated using the Octet 1.1 analysis software.

Protein pulldown assay and co-IP assay.

To examine interaction between rCqArpc1A and Cqβ-Actin, crayfish Hpt was isolated and solubilized for GST pulldown assay. In addition, protein binding between rHis-CqImportin β1 and rGST-CqImportin α1 or rGST-CqRan was performed by His pulldown assays with 10 μg of purified recombinant proteins as we previously described (8, 9). rGST protein was taken for mock treatments accordingly. For co-IP assays, selected cDNAs were cloned into expression vectors pXJ40-Flag (Miaolingbio, China), pXJ40-Myc (Miaolingbio, China), and pCMV-HA (Beyotime, China), respectively. CqImportin α1, ARM repeat domain, IBB domain, CqImportin α1^ΔIBB lacking IBB domain, and CqImportin α1^ΔARM lacking ARM repeat domain were constructed with Flag tag. CqImportin β1 was fused with Flag or Myc tag, respectively. IBN domain and CqImportin β1^ΔIBN lacking IBN were tagged with Flag, and CqRan was tagged with Myc. The NLSs of WSSV structural proteins were designed with HA tag accordingly in capsid protein residues, including full-length VP15 (1 to 61) and truncated VP60 (304 to 411), VP95 (547 to 653), VP136 (60 to 166), VP190 (374 to 480) and VP664 (2405 to 2535), except that VP664 protein residues (2160 to 2714) were fused with Myc tag and in truncated envelope proteins VP52A (1 to 107), VP52B (1 to 107), and VP124 (34 to 143). HEK293T cells were transfected with various combinations of constructs as indicated; transfected cells were lysed at 48 h posttransfection and incubated with 10 μL of anti-FLAG M2 beads (Sigma-Aldrich, USA) with constant agitation overnight at 4°C. The immunoprecipitated samples were collected and subjected to Western blotting assays using anti-Flag, anti-Myc, and anti-HA antibodies accordingly.

Immunoblotting assays.

For WSSV internalization examination in Hpt cell cultures after gene silencing or pharmacological blockade, the cells were treated twice with dsRNA of CqArp2/3 complex (CqArpc1A, CqArpc3, CqArpc4, and CqArpc5) as we previously described (9) or pretreated with CK666 (100 μM) for 1 h followed by viral infection. For WSSV infection examination in Hpt cell cultures after gene silencing or pharmacological blockade, the cells were treated twice with dsRNA of CqImportin α1, CqImportin β1, CqRan, or CqNup35 as we previously described (9), or optimized concentrations of IVM (8 μM) and IPZ (8 μM) were used to pretreat Hpt cells for 1 h followed by WSSV infection for 12 hpi. The VP28 protein expression was examined by Western blotting as performed above. For Hpt cell culture, the Hpt dissected from three crayfish was defined as one biological replicate; all experiments were biologically repeated at least for three times.

Immunofluorescence assays.

To examine WSSV internalization and endocytic vesicle formation, Hpt cells were infected by WSSV for 0.5 and 1 h, respectively, and subjected to fixation and permeabilization followed by blocking and incubation with the corresponding antibodies as we previously described (8, 9); in addition, F-actin and vesicles were stained by phalloidin (100 nM) and FM 1-43 (1 μg/mL), respectively. Hpt cells were infected by WSSV for 0.5 hpi followed by ND (50 μM) and CHX (35 μM) treatment (9), with the presence of inhibitor until samples collection for detecting nucleocapsids by anti-VP664 antibody and microtubules by tubulin-tracker red (200 nM) staining, respectively. To examine the effect on targeting of incoming nucleocapsid to NPCs for genome releasing into nucleus by pharmacological blockade or gene silencing, Hpt cells were pretreated with IVM or IPZ for 1 h or dsRNA of CqImportin α1, CqImportin β1, and CqImportin α1 plus CqImportin β1 genes described above; Hpt cells were then challenged by WSSV in the presence of CHX and fixed at 3 hpi for analysis by confocal microscopy. To evaluate effect on importing expressed structural proteins of WSSV into Hpt cell nucleus for progeny virion assembly by blocking of importins, Hpt cells were challenged by WSSV for 6 h, followed by treatment with IVM or IPZ, in which the assembly of progeny virions was examined at 18 hpi accordingly and the corresponding viral titers were determined at 24 hpi as shown above. The cell nuclear DNA was stained with DAPI (1 μg/mL). Cell imaging was recorded with an LSM 780 confocal microscope (Carl Zeiss Microscopy, Germany). At least 120 Hpt cells, with WSSV infection or without viral infection for a control, were randomly selected for scoring with each group; at least three independent experiments were employed for statistical analysis.

Transient gene expression and fluorescence quantification in HEK293T cells.

All recombinant plasmids used for transient gene expression were prepared by standard molecular cloning protocols. The indicated genes, gene truncations, and genes with epitope tags were inserted into pEGFP-C1 vector (Clontech) in which the WSSV proteins containing NLSs were designed as the following: full-length amino acids of viral capsid proteins, including VP15, VP51, VP60, VP76, and VP95, and envelope proteins, including VP52A, VP52B, VP26, and VP28, that are suitable for recombinant expression; the truncated segment residues, with molecular sizes of more than 50 kDa, of capsid proteins, including VP136 (1 to 548), VP190 (1 to 555), VP664 (2160 to 2714), and envelope protein VP124 (1 to 546) were constructed. Due to the large molecular size, which is difficult to use for overexpression of the intact proteins at present, a molecular size over 50 kDa could hardly diffuse freely across the nuclear pore (11). CqImportin α1 and NLS-deleted CqImportin α1^ΔmNLS were also inserted into pEGFP-C1 vector (Clontech). The pEGFP-C1 vector protein was employed as a control accordingly. For subcellular localization analysis, HEK293T cells were transfected with the indicated plasmids for 48 h and stained with Hoechst 33258 (1 μg/mL) for 10 min, followed by confocal microscopy analysis. Fluorescence quantification data were obtained using Image J software, and Student's t test was performed to compare the differences between tested samples. At least 120 cells with indicated protein expression were scored for each group, and the data from at least three independent experiments were employed for statistical analysis.

TEM and IEM analysis.

Ultrastructural observation of WSSV infection was examined by TEM in Hpt cells, which were infected by WSSV for different time intervals, including 1.5, 2, 2.5, 3, 8, 12, 18, and 24 hpi, respectively, as indicated. Hpt cells were fixed and sectioned with a microtome (Leica), and sections were double stained with uranyl acetate and lead citrate before examination under a TEM (Hitachi HT-7800) as we previously described (8, 9). For immunogold (IgG) labeling of WSSV in Hpt cells, the cells were infected by WSSV for 3 and 8 h, respectively, followed by standard sample preparation for automatic ultramicrotome and blocking treatment with bovine serum albumin (BSA; Solarbio, China) (6, 54); subsequently, the samples were incubated with mouse-anti-VP28 antibody and rabbit-anti-VP664 antibody after primary optimization in blocking solution overnight at 4°C. Furthermore, the samples were washed and then incubated in goat-anti-mouse IgG conjugated to 25-nm gold particles and goat-anti-rabbit IgG conjugated to 12-nm gold particles, respectively, at room temperature for 3 h, which were further determined by standard procedures for TEM.