Engineering a replication-incompetent viral vector for the delivery of therapeutic RNA in crustaceans

Abstract

Viral disease pandemics are a major cause of economic losses in crustacean farming worldwide. While RNA interference (RNAi)-based therapeutics have shown promise at a laboratory scale, without an effective oral delivery platform, RNA-based therapy will not reach its potential against controlling viral diseases in crustaceans. Using a reverse-engineered shrimp RNA virus, Macrobrachium rosenbergii nodavirus (MrNV), we have developed a shrimp viral vector for delivering an engineered RNA cargo. By replacing the RNA-dependent RNA polymerase (RdRp) protein-coding region of MrNV with a cargo RNA encoding green fluorescent protein (GFP) as a proof-of-concept, we generated a replication-incompetent mutant MrNV^(ΔRdRp) carrying the GFP RNA cargo resulting in MrNV^(ΔRdRp)-GFP. Upon incorporating MrNV^(ΔRdRp)-GFP in the diet of the marine Pacific white shrimp (Penaeus vannamei), MrNV^(ΔRdRp) particles were visualized in hemocytes demonstrating successful vector internalization. Fluorescence imaging of hemocytes showed the expression of GFP protein and the MrNV capsid RNA (RNA2) as well as the incorporated GFP RNA cargo. Detection of cargo RNA in hepatopancreas and pleopods indicated the systemic spread of the viral vector. The quantitative load of both the MrNV RNA2 and GFP RNA progressively diminished within 8 days postadministration of the viral vector, which indicated a lack of MrNV^(ΔRdRp)-GFP replication in shrimp. In addition, no pathological hallmarks of the wild-type MrNV infection were detected using histopathology in the target tissue of treated shrimp. The data unequivocally demonstrated the successful engineering of a replication-incompetent viral vector for RNA delivery, paving the way for the oral delivery of antiviral therapeutics in farmed crustaceans.

Significance Statement.

RNA-based therapy, such as RNA interference (RNAi) by double-stranded RNA (dsRNA) injection, is a promising antiviral strategy, especially for invertebrate organisms where conventional vaccination is not possible. However, therapeutic delivery by injection is not feasible for industrial-scale applications in aquatic invertebrates such as crustaceans. Our study presents an RNA oral delivery platform utilizing a shrimp viral vector. This viral vector was generated from a reverse-engineered shrimp RNA virus Macrobrachium rosenbergii nodavirus (MrNV), resulting in a replication-incompetent MrNV^(ΔRdRp) vector that can be simultaneously produced and self-assembled with any desired RNA cargo. This study demonstrated this vector's efficient production and successful oral delivery, which provides a bridge to allow RNAi to finally reach its potential as an effective treatment strategy against diseases in crustacean aquaculture.

Introduction

The development of gene editing technologies, namely RNA interference (RNAi) and clustered regularly interspaced short palindromic repeats (CRISPR)/Cas, has revolutionized the field of crustacean research, both for basic research and commercial development (1–4). Facilitating the loss-of-function assays, these technologies revealed the function of essential novel genes that improved our knowledge on host defense response and host–pathogen interactions in crustaceans and invertebrates in general. Furthermore, these gene editing tools opened doors for developing promising therapeutics for agriculturally important species such as farmed shrimp (5). Translating these gene editing discoveries into commercial applications is now desperately needed to enhance the sustainability of crustacean aquaculture since viral disease pandemics are widespread and conventional vaccines are not applicable due to the lack of an antibody-based immunity in invertebrates (6).

The development of nucleic acid–based therapeutics is gaining momentum in shrimp aquaculture and is geared toward addressing the ongoing problems posed by viral diseases (7). Results from studies involving injection of double-stranded RNA (dsRNA) to target viral transcripts have identified RNAi as a promising antiviral therapeutics in shrimp (8). Targeting key genes involved in viral pathogenesis using RNAi has been shown to confer protection against many economically important and the World Organization for Animal Health (WOAH, Paris, France)-listed diseases (e.g. white spot disease caused by the white spot syndrome virus and yellow head disease caused by the yellow head virus) (8–10).

The development of therapeutic agents using genomic tools faces major roadblocks, which include the lack of immortalized cell lines in crustaceans and practical yet effective production and oral delivery platforms (7). The unavailability of immortalized cell lines impedes both in vitro culture of viruses and the testing of therapeutic agents (11), while the lack of large-scale dsRNA production and effective oral delivery methods prevents RNAi therapeutics’ advancement to farm-scale application (7). Circumventing the lack of an immortal cell line, Dhar and colleagues engineered an infectious cDNA clone of an RNA virus infecting freshwater prawn, MrNV, by cloning full-length cDNAs representing MrNV RNA-1 and RNA-2 into the pFastBacDUAL vector and expressing them using a baculovirus expression vector system (BEVS) in the Sf9 cell line (12).

Classified as a member of the family Nodaviridae, MrNV is a nonenveloped virus with a bipartite, positive-sense, single-stranded RNA genome enclosed in T = 3 icosahedral capsids measuring about 40 nm in diameter (13). The viral genome consists of two short-genomic RNA fragments coding for three genes: (i) RNA-1 (3.2 kb) encodes an RNA-dependent RNA polymerase (RdRp) and a B2-like protein, and (ii) RNA-2 (1.2 kb) encodes the viral capsid protein (13, 14). The viral profile of MrNV, particularly as an RNA virus with a small genome, makes it an ideal nanocarrier for the delivery of RNA molecules. Additionally, the success in generating infectious MrNV clones provided evidence supporting the utility of BEVS in a production platform where the nanocarrier is simultaneously expressed and assembled with its cargo RNA.

In this study, we hypothesized that the production of infectious MrNV clones using BEVS can be reengineered to express a replication-incompetent MrNV shrimp viral vector where the viral RdRp gene is replaced with any desired RNA cargo (e.g. a protein-coding RNA or any therapeutic noncoding RNA). In this study, we explored the oral delivery of a vector–model RNA cargo complex by incorporating them into a commercial diet, feeding shrimp, and then confirming the detection of the delivered RNA model cargo and expression of the corresponding GFP protein in the shrimp tissues.

Results

Engineering and production of nonreplicating MrNV viral vector

Both cDNA sequences, the viral vector MrNV chimeric RNA1 MrNV^(ΔRdRp) and the GFP-coding region, were successfully cloned into the pFastBac Dual vector (Supporting information; Fig. S1). With this, coexpression through Bac-to-Bac baculovirus expression system facilitates vector–cargo self-assembly of the RNA2 capsid protein and the model gene cargo GFP RNA. The resulting construct was the donor plasmid to transpose the target genes to the bacmid to generate the recombinant baculovirus expressing MrNV^(ΔRdRp)-GFP in Sf9 cells.

The recombinant baculovirus expressing MrNV^(ΔRdRp)-GFP was successfully purified, and the multiplicity of infection (MOI) for GFP expression was optimized in inoculated Sf9 cells. Results show GFP-expressing Sf9 cells were observed after inoculation with the MOI at 1:1, 1:10, and 1:100 from 24 h postinfection. Irrespective of the MOI, GFP expression reached equivalent levels at 92 h postinfection (Fig. S2). By flow cytometry analysis, the population of GFP-expressing cells was 98% for the MOI of 1:1 at 92 h postinfection (Fig. 1A). Based on these results, further inoculations were done with an MOI at 1:1 and cells were incubated for up to 92 h postinfection.

Fig. 1.

Production of MrNV^(ΔRdRp)-GFP in SF9 cells by baculovirus expression system. A) Population of GFP-expressing SF9 cells in both infected (left) and uninfected (right) 92 hpi with 1:1 MOI of baculovirus vector expressing MrNV^(ΔRdRp)-GFP. Number of events (#Events) and percentage of GFP-expressing cells (%Parent) are graphed below the graph plots. B) TEM images of ultrathin sections of SF9 cells producing baculovirus and MrNV^(ΔRdRp)-GFP. Micrographs of lysed SF9 cells showing occlusion bodies (arrow) with scale bar at 2 µm (i). Higher resolution focusing on baculovirus (arrow) and MrNV^(ΔRdRp)-GFP-containing vesicles (arrowheads) with scale bar at 500 nm (ii). MrNV^(ΔRdRp)-GFP -GFP viral particles measured ∼30–40 nm with a scale bar at 200 nm (iii). C) TEM images of negatively stained baculovirus (arrows) and MrNV^(ΔRdRp)-GFP (triangle:▴) viral particles from transfected Sf9 cell supernatant (left panel) and MrNV^(ΔRdRp) particles separated by gradient centrifugation purification (right panel). Scale bars are set to 200 nm.

Assembly of MrNV^(ΔRdRp)-GFP was confirmed by transmission electron microscopy (TEM) imaging of ultrathin sections of Sf9 cells. Micrograph images revealed lysed Sf9 cells with occlusion bodies containing rod-shaped baculovirus nucleocapsids (Fig. 1B-ii). The MrNV^(ΔRdRp) virions appear as icosahedral viral particles (∼35–45 nm in diameter) packed in microvesicles (Fig. 1B-iii). The particle size of MrNV^(ΔRdRp)-GFP was consistent with the purified MrNV^(ΔRdRp)-GFP confirmed through the analyses of negatively stained viral particles from the cell supernatant of Sf9 cells infected with BEVS carrying MrNV^(ΔRdRp)-GFP genes (Fig. 1C). Micrographs of the mixture of baculovirus and MrNV^(ΔRdRp)-GFP particles as well as separated MrNV^(ΔRdRp)-GFP particles (purified) are shown; coimaging provided a better contrast for the analysis of particle sizes. Both TEM analysis of ultrathin Sf9 sections and purified MrNV^(ΔRdRp)-GFP virions showed a consistent average size of 40 nm.

Confirmation of the delivery of RNA cargoes by MrNV^(ΔRdRp)

To demonstrate the potential for large-scale applicability of MrNV^(ΔRdRp) in the oral delivery of therapeutic RNA molecules, we present evidence that the chimeric virions MrNV^(ΔRdRp)-GFP can successfully deliver its GFP payload in shrimp tissue when mixed with a commercial diet and orally administered to specific pathogen-free (SPF) shrimp.

We confirmed the delivery of MrNV^(ΔRdRp)-RNA2 and GFP RNA cargo by detecting their cognate RNA by RT-PCR (Fig. 2A and B). A cell suspension containing Sf9 cells expressing MrNV^(ΔRdRp)-GFP was incorporated in the diet and administered to healthy shrimp for 5 consecutive days. The MrNV^(ΔRdRp)-RNA2 gene was detected in hepatopancreas and pleopod tissues by RT-PCR at day 1 postfeeding. As control treatments, RNA from the same tissues was collected from shrimp at 5 days postinjection with a purified inoculum containing MrNV^(ΔRdRp)-GFP and from naïve shrimp maintained on the control diet. The MrNV^(ΔRdRp)-RNA2 was detected in both tissues of shrimp maintained on the treatment diet as well as in injected shrimp (Fig. 2A). This confirmed successful delivery of the marker gene using the MrNV-GFP viral vector via the oral route, as well as the enteric and systemic distribution of the viral vector MrNV^(ΔRdRp)-GFP. The injection method was also effective in delivering the marker gene in both hepatopancreas and pleopod tissues, and the intensity of the amplicons was higher compared with the corresponding amplicons when delivered via the oral route. Comparing the band intensity of the detected MrNV^(ΔRdRp)-RNA2 between the two delivery methods, slightly lower levels of RNA2 were detected in the hepatopancreas than pleopods were observed in the injected animals. On the other hand, delivery through the oral route showed a lower RNA2 level in the pleopods and higher in the digestive organ hepatopancreas (Fig. 2A). The orally delivered GFP RNA was also detected in the pleopods with a decreasing intensity from day 0 to day 5 postfeeding (Fig. 2B).

Fig. 2.

Oral delivery and detection of MrNV^(ΔRdRp) and GFP RNA. Oral delivery was conducted by soaking commercial feeds with cell supernatant of Sf9 cells infected with baculovirus vector carrying MrNV capsid protein (RNA2) and GFP RNAs. Penaeus vannamei shrimp hepatopancreas and pleopod tissues (n = 3) were sampled at days 0 (fifth day of feeding), 1, 3, 5, and 8 postfeeding for RNA extraction. Naïve and injected shrimp were used as negative and positive control, respectively. Extracted RNA were used for the RT-PCR detection of (A) MrNV RNA2 and (B) GFP cargo RNAs. Detection of MrNV RNA2 in the hepatopancreas and pleopods sampled at day 1 postfeeding with MrNV^(ΔRdRp)-GFP. On the other hand, time-course detection of GFP RNA cargo in the pleopods was performed in samples taken on days 0, 1, 3, 5, and 8. Both detections were done together with the detection of internal control gene EF-1α for all samples.

MrNV^(ΔRdRp)-GFP internalization and GFP expression in the hemocytes

The ability of the viral vector MrNV^(ΔRdRp) to be internalized by the hemocytes facilitates the delivery of the therapeutic molecule to all other organs. TEM micrographs of ultrathin-sectioned hemocytes drawn from shrimp fed MrNV^(ΔRdRp)-GFP-containing diet revealed the presence of icosahedral particles within the hemocytes (Fig. 3A–C). The size of the MrNV^(ΔRdRp) in the hemocytes matches the particle size (∼35–45 nm) observed in the Sf9 cells expressing MrNV^(ΔRdRp)-GFP (Fig. 1B). This confirms that MrNV-GFP produced in Sf9 cells can be taken up by the hemocytes in the same manner as a wild-type MrNV.

Fig. 3.

MrNV^(ΔRdRp)-GFP uptake by hemocytes and detection of delivered GFP cargo. Following the same experimental set-up on Fig. 2A, hemocytes were drawn from P. vannamei fed with a diet containing MrNV^(ΔRdRp)-GFP and used for (A) ultrathin section imaging by TEM (B), detection of GFP RNA by RT-PCR, and (C) GFP protein detection by fluorescence microscopy. Micrographs in (A) show MrNV particles in areas of hemocytes cells marked in a yellow box with a scale bar = 1 µm (left). Higher resolution focusing on MrNVcp particles (right) measuring ∼30–40 nm which are indicated by a triangle (▴) with scale bar = 200 nm. Gel image shown in (B) of the detection of GFP RNA cargo delivered into hemocytes cells drawn from shrimp (n = 3) together with the detection of internal control EF-1α. C) Bright field and fluorescent (GFP) microscope images of hemocytes drawn from MrNV^(ΔRdRp)-GFP-treated shrimp. Hemocytes from injected and naïve/SPF shrimp were used as the positive and negative control, respectively. Scale bar = 10 µm.

Aside from the internalization of MrNV^(ΔRdRp), we looked into the delivery of the GFP RNA cargo using both GFP RNA detection and detecting the translated cognate protein through the emitted green fluorescence in the primary culture of hemocytes collected from the treated shrimp. The GFP RNA was detected in total RNA from hemocytes of both MrNV^(ΔRdRp)-GFP injected and fed shrimp, while no detection was observed in the naïve shrimp (Fig. 3B). Comparing expression levels between delivery methods, higher GFP RNA levels were detected in the samples from injected shrimp. Corroborating the RNA detection, the expression of GFP protein was observed in the primary culture of hemocytes as seen through fluorescence microscopy, confirming the successful delivery and subsequent translation of a functional RNA by MrNV^(ΔRdRp)-GFP (Fig. 3C).

Time-course quantification of MrNV^(ΔRdRp)-RNA2 and histopathology analysis post-MrNV^(ΔRdRp)-GFP treatment

Through absolute quantification, we determined how much RNA was delivered and the duration of MrNV^(ΔRdRp) shedding after 5 consecutive days of feeding MrNV^(ΔRdRp)-GFP to healthy shrimp. At 1 day postfeeding, 1.0 × 10⁴ RNA copies of MrNV RNA2 per 1.0 ng of total RNA were detected in hepatopancreas tissue (Fig. 4A). This decreased by 50% at almost every 24 h until 8 days postfeeding, where ∼14 copies were detected. On the other hand, 1.0 × 10² copies were detected per 1.0 ng pleopod RNA until 3 days postfeeding (Fig. 4B). The RNA copy number then decreased at day 5 and was undetected at 8 days postfeeding.

Fig. 4.

Post-MrNV^(ΔRdRp)-GFP feeding biosecurity assessment. To rule out any biosecurity risks resulting from the use of MrNV^(ΔRdRp) as the viral vector, (A and B) time-course quantification of MrNV^(ΔRdRp), and (C) histopathology analysis following 5 days consecutive feeding with MrNV^(ΔRdRp)-GFP. Absolute quantification of MrNV^(ΔRdRp) -RNA2 using RT-qPCR analyzed with the standard curve. Penaeus vannamei shrimp were fed with a diet containing MrNV^(ΔRdRp)-GFP, and animals (n = 3) were sacrificed at days 0 (fifth feeding day), 1, 3, 5, and 8 postfeeding. Copy number of MrNV^(ΔRdRp)-RNA2 was measured and quantified in the RNA from hepatopancreas and pleopods. RNA extracted from shrimp injected with MrNV^(ΔRdRp)-GFP 5 and naïve SPF were used as the positive and negative control, respectively. As described above, MrNV^(ΔRdRp)-GFP treatment was fed to shrimp, and tissues were processed for staining with H&E and subjected to whole tissue imaging to assess tissues sampled on days 1, 3, 5, and 8 postfeeding; positive and negative control tissues are also shown from injected and naïve shrimp groups, respectively. MrNV wild-type target tissues (skeletal muscles) were chosen as representative tissues. Indicated scale bar = 30 µm.

Histopathology analysis did not detect any sign of pathological changes in the skeletal muscle tissue of shrimp in both MrNV^(ΔRdRp)-GFP injected and orally administered animals as compared with naïve shrimp tissue. Lesions indicative of muscle necrosis, a typical sign of wild-type MrNV infection, were not observed in any of the samples screened (Fig. 4C).

Real-time PCR was performed to detect the baculovirus gp64 gene in shrimp that were fed Sf9 cell homogenate containing recombinant baculovirus and MrNV^(ΔRdRp)-GFP virions. The gp64 gene could not be detected in shrimp hepatopancreas and pleopods tissue by real-time PCR (Table S2). This indicates baculovirus does not replicate in shrimp cells.

Discussion

Only with a successful oral delivery platform can RNA-based targeted therapy reach its full potential for protection against diseases of aquatic invertebrates (7). In this study, we demonstrated the utility of our MrNV^(ΔRdRp) shrimp viral vector to be used as an oral delivery platform that fulfills three key criteria: oral delivery, successful detection of delivered RNA cargo, and a pathway for economical large-scale production of the viral vector using an insect cell culture system.

The delivery of dsRNA by injection has been proven at a laboratory scale to be an effective treatment against major viral diseases in shrimp (5, 15, 16). From these results, several studies explored different methods for delivering dsRNA to address the challenges associated with large-scale commercial applications (7). As reviewed by Itsathitphaisarn et al. (2017), the two main challenges in adopting RNAi for therapy at a commercial scale in shrimp are the development of inexpensive large-scale dsRNA production and an oral delivery method (7). Several promising results were seen with the use of dsRNA expressed in Escherichia coli which were then coated onto aquatic feeds (17–19). While bacterial expression is a promising platform, it poses many challenges including regulatory and environmental issues related to the use of antibiotic-resistant E. coli cells at a farm level and immune exhaustion in treated animals due to the presence of bacterial pattern recognition molecules, like lipopolysaccharide (LPS) and β-glucan (20). The use of nonviral nanocarriers, such as cationic liposomes, chitosan, and virus-like particles (VLPs), is also a promising platform being explored for dsRNA delivery in shrimp (17, 21, 22). Aside from providing RNA stability, these nanocarriers show great promise because of their biocompatibility, biodegradability, and low toxicity (23). Recently, a promising nanocarrier, a VLP of MrNV expressed in E. coli BL21 (DE3) then chemically assembled to encapsulate dsRNA, has been reported (22). Using MrNV capsid as a nanocarrier is practical for RNAi therapeutics given the natural ability of RNA viruses to enter and release genomic content into shrimp host cells (24). For oral delivery, however, it is critical for the capsid protein to maintain its structural integrity to withstand incorporation in feed, dispersion in water, and digestive enzymes until the dsRNA can be taken up by shrimp cells. A previous study comparing MrNV VLP production by E. coli and BEVS in Sf9 cells reported the production of more structurally uniform VLP by BEVS in Sf9 cells that remained stable in a wide range of pH and temperature (25). We present here data to unequivocally demonstrate that the MrNV genome can be engineered to make a replication-incompetent virus that is useful for the delivery of a marker RNA in crustacean cells by using the GFP gene and a commercially farmed shrimp species as a proof-of-concept study.

The present finding was built on our earlier report that using the BEVS and Sf9 cells, when MrNV RNA1 and RNA2 are expressed simultaneously, mature infectious recombinant MrNV virions can be produced in Sf9 cells (12). The previous study paved the way for the genetic manipulation of RNA viruses in shrimp without the need of an immortal cell line in crustaceans, since no such immortal shrimp cell line exists. In the current study, we have taken the previous approach one step further to develop a viral vector that can be used to deliver a coding or a noncoding RNA with therapeutic potential. Most importantly, since the recombinant viral vector mimics the wild-type virus, such a viral vector could be used to orally deliver RNAi-based therapeutics based on the method of MrNV transmission in nature using the oral route.

The development of this crustacean viral vector platform involved engineering an infectious MrNV clone into a nonreplicating viral vector by replacing the viral RdRp gene with an RNA cargo. In this proof-of-concept study, the RNA cargo was represented by a GFP RNA (Fig. S1). In the previous infectious MrNV clone, recombinant RNA1 encoding RdRp and RNA2 encoding viral capsid were successfully packaged into mature virion and delivered into shrimp cells. From this result, we hypothesized that recombinant RNA1 containing the 5′ and 3′ untranslated region (UTR) but a different coding region instead of RdRp and RNA2 expressed simultaneously will be packaged into a nonreplicating mature virion. We used the protein-coding region of GFP RNA to serve as a reporter gene in the production of the MrNV^(ΔRdRp)-GFP vector–cargo complex in Sf9 cells (Fig. 1, Fig. S1). Besides the previous success in generating infectious MrNV in insect cells, the use of BEVS in insect cells has been widely utilized for large-scale production of the viral vaccines due to adequate posttranslational modifications; when compared with other cell lines, it provides higher versatility and yield (26, 27). Our data also show efficient production of a viral vector with a transfection efficiency of 98% of Sf9 cell population expressing MrNV^(ΔRdRp)-GFP (Fig. 1A, Fig. S1 and S2). This is corroborated by the images of uniform-size MrNV^(ΔRdRp) particles in ultrathin sections of Sf9 cells and negatively stained purified viral particles (Fig. 1B and C). In addition, this production does not require antibiotics for maintenance nor selection of transfected cells as in BEVS. This eliminates the risk of incorporating antibiotic-resistant microorganisms or antibiotic resistance–related genes into the feeds that may pose some health risk (28). In addition to the high yield, the use of serum-free media (SFM) with insect cells reduces costs for vector production in comparison with other eukaryotic cell lines (29, 30).

In marine shrimp, intramuscular injection remains the most effective method for dsRNA delivery, as it allows direct dispensing of the therapeutic RNA molecules into targeted tissues with a more controlled dosage (7). However, the intense labor, costs, and invasiveness of the method pose an overwhelming economic disadvantage, especially for a large-scale application to farmed aquatic organisms like shrimp. It is for these reasons that the delivery by the oral route is more practical, albeit not without challenges. In contrast to the injection method, the oral delivery needs the therapeutic RNA to withstand several steps in the delivery process: from diet incorporation to dispersion into the water, until the cellular uptake where digestive enzymes are present. To address these hurdles, the characteristics of the viral vector are critical to providing stability during the delivery of RNA cargo. As a capsid protein of a shrimp RNA virus, MrNV^(ΔRdRp)-GFP is well suited for oral delivery to encase and protect the RNA until it is taken up by the cells. Detection of MrNV^(ΔRdRp) and GFP RNAs in the hepatopancreas and pleopod (Fig. 2A and B) demonstrates the successful uptake of intact MrNV^(ΔRdRp)-GFP through the digestive system. Prior to cellular uptake, MrNV^(ΔRdRp)-GFP is exposed to mechanical and chemical breakdown by the mouth and the stomach and then into the hepatopancreas where it is further exposed to enzymatic digestion, absorption, and interfacing with the hemolymph that reaches the rest of the organs (31). The positive delivery of GFP RNA (and expression) also signifies that the MrNV^(ΔRdRp) capsid protein was successful in providing protection for the RNA cargoes against RNases that are ubiquitously present in the gut environment as well as protection from digestive enzymes during food uptake. Also, as the wild-type MrNV virus is known to infect the shrimp systemically (24, 32), we hypothesized MrNV^(ΔRdRp)-GFP would also have the same systemic reach and be distributed to the different shrimp tissues. In addition, supporting the possibility of oral delivery of MrNV^(ΔRdRp)-GFP is the previous infection study using the wild-type MrNV experimentally achieved with both oral and intramuscular inoculation (24). Given that the viral vector described in this study contains the identical capsid protein found in the wild-type MrNV, we hypothesized that MrNV^(ΔRdRp)-GFP will be able to spread systematically from the digestive system to different organs and tissues via hemolymph and through the open circulatory system in shrimp. The data presented here support this notion since GFP RNA was detected in hepatopancreas and pleopods (swimming legs) upon feeding animals with MrNV^(ΔRdRp)-GFP (Fig. 2). As expected, higher detection of the MrNV capsid gene, as well as the GFP gene, was observed when the MrNV^(ΔRdRp)-GFP viral vector was delivered via injection. Interestingly, the detected RNA levels corresponded accurately to the delivery method having a higher level at the site of administration. This is evident in the higher RNA2 levels in the hepatopancreas, which is expected since oral delivery's point of administration is the digestive system before it spreads to the entire body. This same logic explains the lower RNA2 levels detected in the pleopods (Fig. 2A). The opposite was observed with the injection method where higher levels of RNA2 were detected in the pleopods.

The successful transport of MrNV^(ΔRdRp)-GFP from the digestive tract to the pleopods signifies that the vector has a systemic reach, which is advantageous for therapeutic molecules. This distribution to the different tissues can be attributed to the structure of viral capsid and the open circulatory system of crustaceans. The circulating hemocytes play a key role in molecular transport with their ability to infiltrate all tissues. Considering this, we explored the ability of MrNV^(ΔRdRp) to be internalized by the hemocytes. As evident in the observed MrNV^(ΔRdRp) particles in the micrographs of hemocyte cells (Fig. 3A), as well as the detection of GFP RNA and the protein expression of GFP protein in hemocyte primary culture (Fig. 3B and C), we confirm one possibility of how MrNV^(ΔRdRp)-GFP is distributed throughout the organs which are through the hemocyte circulation. This also demonstrates the ability of MrNV^(ΔRdRp) to be internalized and deliver functional RNA cargo into the shrimp cells.

Owing to their versatility, efficiency, and biosafety profile, these replication-incompetent viral vectors are likewise gaining recognition in the development of vaccine delivery platforms (33). These biologically derived nanoparticles mimic the structure and natural function of live pathogens but are noninfectious and unable to replicate (33). However, as we utilized an expression system which was also used in a similar way to produce infectious recombinant MrNV clones (12), it prompted us to determine if there is any pathological changes caused by the nonreplicating MrNV^(ΔRdRp)-GFP such as seen when infected by the wild-type MrNV. When using viral vectors, risk assessment is done by considering factors such as replication ability, shedding, and presence of tissue damage or lesions (34). To produce replication-incompetent MrNV viral vector, we designed our vector plasmid construct to coexpress only the MrNV capsid protein (RNA2) and the cargo GFP excluding the RdRp (RNA1) gene (Fig. S1). This is based on prior studies showing RdRp is required for viral propagation among RNA viruses, where knocking out the RdRp in shrimp RNA viruses abolishes infectivity and viral replication (19, 35). This is seen in the gradual decrease in the copies of MrNV^(ΔRdRp)-RNA2 from the last day of feeding (day 0) until the eighth day postfeeding (Fig. 4). Such a result indicates that the RNA2 genome is not propagating as with the infectious wild-type MrNV. In the same way, the levels of the delivered RNA-GFP cargo decreased, which corroborated that, without the RdRp, there was no exogenous RNA replication in the cell (Fig. 4A and B). These results confirm that the viral vector is nonreplicating and is cleared within a week. If a longer clearing time is desired, further optimization is needed with the dosage and feeding routine. Likewise, the effectivity and stability of the RNA cargo, when applied as therapeutic molecule, will depend on the structural design of the RNA molecule (dsRNA and hairpin) and the nature of its target molecules. Supporting the findings that MrNV^(ΔRdRp) is nonreplicating and noninfectious, histopathology assessment of the treated shrimp (both fed and injected) showed no signs of tissue damage and were equivalent to that of the naïve (SPF) healthy shrimp (Fig. 4C). Moreover, the absence of baculovirus genes in tissues of MrNV^(ΔRdRp)-GFP fed shrimp denotes clearance of the expression vector and not being able to cross the gut barrier as opposed to the MrNV^(ΔRdRp)-GFP, which is distributed to other tissues and absorbed into the cells. Taken together, while having high biocompatibility as the wild-type MrNV, using MrNV^(ΔRdRp) as a nanocarrier poses no health risks to the shrimp host and are shed out of the shrimp's system.

The successful application of MrNV^(ΔRdRp) to P. vannamei shrimp allows us to speculate on the applicability of our delivery platform across different species, given that the freshwater M. rosenbergii is the known natural host for wild-type MrNV to which it causes the white tail disease (WTD). This disease also causes high mortality in other penaeid and nonpenaeid species (e.g. P. vannamei, Penaeus monodon, Penaeus indicus, and Marsupenaeus japonicus) (5, 36–40). The applicability of this viral vector to a broad range of host species is important when applied against a viral pathogen, such as the white spot syndrome virus (WSSV), which has a broad host range across crustaceans (37).

To summarize, our current study provided evidence supporting the utility of a promising oral delivery platform for therapeutic RNA molecules against shrimp diseases. We demonstrated the successful production, delivery, and detection of an engineering RNA cargo into shrimp tissues and cells. This oral delivery platform has the potential to provide a practical and safe application of RNAi therapeutics to shrimp farmed at an industrial scale. Considering that insect cell culture at a commercial scale is well established, the potential for developing an oral delivery of antiviral therapeutics for farmed crustaceans appears to be very feasible now. Due to the systemic nature and transovarial transmission of MrNV, the viral vector developed in this study could potentially open avenues in developing viral-resistant lines of farmed shrimp using precision breeding and in developing transgenics. Finally, since nodavirus infects insects, a similar approach could be adopted to develop a viral vector for delivering RNAi therapeutics for insect control.

Materials and methods

Construction of dual-expression donor plasmid

Full-length cDNA of MrNV RNA1 encoding viral RdRp and RNA2 encoding the capsid protein were custom synthesized and cloned downstream of polyhedrin and P10 promoters, respectively, in the pFastBac Dual vector (Gibco) by GenScript (Piscataway, New Jersey). The open reading frame (ORF) of RdRp gene was then replaced with the coding region of GFP resulting in a chimeric gene containing the 5′ and 3′ UTR of RNA1 flanking the GFP ORF. The resulting construct containing the mutant MrNV (referred to as MrNV^(ΔRdRp)-GFP) served as the donor plasmid for the transposition of the genes of interest into the baculovirus (bacmid) genome. MrNV^(ΔRdRp)-GFP was then expressed using a Bac-to-Bac baculovirus expression system in Sf9 cells adapted to Sf-900 II SFM (Gibco).

Production of MrNV^(ΔRdRp)-GFP in Sf9 cells

The baculovirus expressing MrNV^(ΔRdRp)-GFP was purified and resuspended in Sf900 II (Gibco, USA) SFM with a titer of 9.8 × 10⁸ pfu/mL. Successful expression of MrNV^cp-GFP was initially assessed through the expression of reporter gene GFP in Sf9 cells by fluorescence microscopy, which allowed the determination of the optimum inoculation condition to maximize the expression of the gene targets. For this optimization, Sf9 cells were cultured in T-25 flask containing SFM and were inoculated with baculovirus with the MOI of 1:1, 1:10, and 1:100 and GFP expression was analyzed and imaged using the EVOS fl microscope (AMG, USA) at 24, 48, and 92 h postinoculation (hpi).

Flow cytometric analysis of GFP-expressing Sf9 cells

The population of Sf9 cells expressing GFP was analyzed using BD FACS Canto II (BD Biosciences, USA). The 488 nm laser channel was used with filters 530/30 nm and 585/42 nm. Only single cells were analyzed, disregarding all doublets from the analysis. GFP (530 nm) was then gated against autofluorescence/PE (585 nm).

Purification of MrNV^(ΔRdRp) viral particles and TEM

To visualize the MrNV^(ΔRdRp)-GFP particles being generated in Sf9 cells through the baculovirus expression system, Sf9 cells were collected at 92 hpi, fixed in 5% glutaraldehyde, embedded in resin, and then used in an ultrathin section for TEM imaging. To validate the MrNV^(ΔRdRp)-GFP particle size, separation of MrNV^(ΔRdRp)-GFP from baculovirus particles was performed using density-gradient ultracentrifugation in OptiPrep^TM Density Gradient Medium (Sigma-Aldrich). Purified particles were embedded in a carbon grid and stained with uranyl acetate and visualized by TEM.

Delivery of MrNV^(ΔRdRp)-GFP by oral route

To confirm the ability of MrNV^(ΔRdRp) to deliver GFP RNA into shrimp cells and to explore its potential through oral administration, we conducted a feeding bioassay using SPF P. vannamei shrimp. The delivery of the recombinant MrNV^(ΔRdRp)-GFP via injection was used as the positive control treatment. For the bioassay, we prepared three groups comprising five individual shrimp (wt. ∼2.5–3 g size) per group. The first group was injected once with 50 μL of ∼1 × 10⁷ pfu/mL of the purified inoculum containing MrNV^(ΔRdRp)-GFP diluted in 1 × phosphate-buffered saline (PBS). The second group was fed with commercial shrimp feeds soaked with cell supernatant of SF9 containing MrNV^(ΔRdRp)-GFP 15 min prior to feeding at 5% of the biomass in the tank for 5 consecutive days. This oral inoculum contained 4 × 10¹² MrNV^(ΔRdRp) RNA copies/mL/1.5 g pellet, as determined by quantitative PCR (qPCR). The third group was fed with untreated commercial feed serving as the negative/naïve control.

Detection and distribution of MrNV^(ΔRdRp)-RNA2 and GFP cargo

To assess the successful delivery of the viral vector and its ability to cross the gut barrier and spread systemically to other tissues, hepatopancreas and pleopods samples were collected from three individual shrimp per treatment for RNA extraction at 24 h postfeeding. For the injected and naïve control groups, similar tissue samples were collected at 5 days from postinjected or untreated animals. Detection of target genes, MrNV^(ΔRdRp) capsid protein (RNA2), GFP, and elongation factor 1 alpha (EF-1α) (shrimp internal control gene) was performed by RT-PCR. The cDNA was synthesized using 1 μg of RNA and SuperScript IV Reverse Transcriptase (Invitrogen, USA), and detection of target genes was carried out using DreamTaq Green DNA Polymerase. The primer sequences of the target and reference genes are provided in Table S1. Thermal cycling conditions were as follows: predenaturation cycle at 95°C for 5 min, 30 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s and a final extension at 72°C for 5 min. The amplicons were electrophoresed using a 1% agarose gel in 1× TAE buffer, and the gel was imaged using Biorad Gel Doc Xr+ Imaging System.

Quantification and viability test of orally delivered RNA cargo, GFP gene

A second bioassay was performed following the same experimental set-up as described above except there were 15 shrimp per treatment. Both injection and oral feeding were performed as described above. Sampling of hepatopancreas and pleopod tissues from three individual shrimp was done for the naïve and injected group 5 days postinjection. For the feeding treatment group, the last day of feeding treatment diet (i.e. a commercial diet containing MrNV^(ΔRdRp)-GFP) was considered as day 0 sampling. This was followed by sampling at days 1, 3, 5, and 8 postfeeding.

For the absolute quantification of MrNV RNA2 copy numbers, RNA standards were generated by first cloning MrNV RNA2 into pCR 2.1-TOPO vector (ThermoFisher, USA), which attaches the T7 promoter sequence on the N-terminal of the inserted sequence. After linearizing T7-MrNV RNA2 by amplification using M13 primers, T7 transcription was used to generate RNA of MrNV RNA2. Together with the samples, the standards were run in triplicate by RT-qPCR using TaqMan Fast Virus 1-Step Master Mix and MrNV RNA2 primers and probe (Table S1).

Clearance of the BEVS from shrimp that were either injected or fed MrNV-GFP was assessed by TaqMan qPCR (S) using GP64 primers as described by Hitchman et al. (2007) (38).

Delivery into the shrimp hemocytes

To investigate cellular internalization of the viral vector and RNA cargo (i.e. GFP gene by the hemocytes), imaging of viral particles in the hemocytes and detection of cargo RNAs were performed. Hemolymph was drawn from three individual shrimp at 5 days postinjection and 24 h postfeeding. Hemocytes were separated by centrifugation and divided for primary culture and RNA extraction. For the primary culture, hemocytes were seeded in 24-well plate with L-15 medium (Sigma-Aldrich) and then viewed under a fluorescence microscope to assess GFP protein expression. The RT-PCR was also done to detect the RNA cargo of GFP and MrNV RNA2 following the same conditions as described above in Quantification and viability test of orally delivered RNA cargo, GFP gene. Hemolymph was collected and pooled from three individual shrimp at day 1 postfeeding. Hemocytes were separated by centrifugation and were fixed for ultrathin sectioning as described above in Purification of MrNV(ΔRdRp) viral particles and TEM.

Histopathology

After feeding with the MrNV^(ΔRdRp)-GFP-containing diet, shrimp were sampled at days 0, 1, 3, 5. and 8 postfeeding and were fixed with Davidson's alcohol–formalin–acetic acid (AFA). In the same way, samples from the injected group and naïve SPF shrimp were sampled and fixed. Fixed samples were then processed for paraffin embedding, microtome sectioned (5 μm thick), and then mounted on glass slides following established protocols (39). Slides containing tissue sections were stained with hematoxylin and eosin (H&E), subjected to whole slide imaging using Motic EasyScan Pro 6 using 80× magnification, and then viewed and analyzed using the Motic DS assistant software (Motic VM 3.0). Images of tissue samples were assessed and scanned for signs of infection and typical lesions of muscle necrosis (39, 40).

Animal experiments

SPF P. vannamei shrimp were procured from a United States Department of Agriculture (USDA)-certified hatchery in the United States. No ethical approval is required by the Institutional Animal Care and Use Committees (IACUC) for the use of invertebrates other than the animals belonging to the Class Cephalopoda. All procedures performed, however, adhere to the guidelines of the state government of New South Wales, Australia, for the Humane Harvesting of Fish and Crustaceans (www.dpi.nsw.gov.au). Sacrificing of live animals was done according to The Welfare of Crustaceans at Slaughter by The Humane Society of the United States (www.humanesociety.org).

Supplementary material

Supplementary material is available at PNAS Nexus online.

Funding

This study was supported in part by the Aquaculture Pathology Laboratory (APL), University of Arizona, and grant-in-aid by the U.S. Department of Agriculture—National Institute of Food and Agriculture (R.R.R.A. and A.K.D. USDA-NIFA Award Number 2022-67015-36332).

Author contributions

A.K.D. and R.R.R.A. designed the study and planned the experiments. All experiments and sample preparation were carried out by R.R.R.A. H.N.M. and A.K.D. mapped the plasmid construct. H.N.M. performed digital imaging of histological slides. The manuscript was drafted by R.R.R.A. and was reviewed and edited by A.K.D. and H.N.M.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary materials.