Transgenic microalgae expressing double-stranded RNA as potential feed supplements for controlling white spot syndrome in shrimp aquaculture

Viral infection of farmed fish and shellfish represents a major issue within the aquaculture industry. One potential control strategy involves RNA interference of viral gene expression through the oral delivery of specific double-stranded RNA (dsRNA). In previous work we have shown that recombinant dsRNA can be produced in the chloroplast of the edible microalga, Chlamydomonas reinhardtii and used to control disease in shrimp. Here we report a significant improvement in antiviral dsRNA production and its use to protect shrimp against white spot syndrome virus (WSSV). A new strategy for dsRNA synthesis was developed that uses two convergent copies of the endogenous rrnS promoter to drive high level transcription of both strands of the WSSV gene element in the chloroplast. New vectors were designed that allow rapid Golden Gate-mediated assembly of a transformation plasmid in which the ‘dual promoter-WSSV DNA’ cassette is targeted into the chloroplast genome, with selection based on the restoration of photosynthesis. PCR analysis of transformant lines confirmed the integration of the cassette and homoplasmy of the polyploid genome. Transcribed sense and antisense VP28-RNA were hypothesised to form an RNA duplex in the chloroplast stroma, and quantitative RT-PCR indicated that ∼100 μg dsRNA is produced per litre of transgenic microalgae culture. This represents an ∼10,000-fold increase in dsRNA relative to previous reports using convergent psaA promoters. The engineered alga was assessed for its ability to prevent WSSV infection when fed to shrimp larvae prior to a challenge with the virus. Survival of shrimp fed with dsRNA-expressing C. reinhardtii was significantly enhanced (68.3%) relative to the negative control. The study suggests that this new dsRNA production platform is significantly more efficient than that reported previously, and merits further scale-up and downstream processing studies.


0
agar was melted and cooled to 42 °C prior to addition to the mixture followed by immediate pouring onto HSM agar 1 1 plates. The transformation plates were incubated overnight in the dark cultured under 50 µmol photon m -2 s -1 white 1 2 light at 25 °C. Successful transformation colonies where photosynthesis had been restored were obtained after 4-6 1 3 weeks.

4
Single colonies were individually picked and streaked on fresh HSM plates and continuously cultured for 3-4 1 5 weeks before next passaging. Homoplasmy was monitored by PCR following extraction via a modified CTAB  Table 1). The VP28 fragment was detected using specific primers, VP28_ESP_F and VP28_ESP_R, to confirm 1 9 insertion of the convergent rrnS promoter cassette. The primers TN72_F and TN72_R were used to confirm 2 0 homoplasmy of the transformed chloroplast genome. Once confirmed to be homoplasmic, the transformant line was 2 1 named TN72-dsVP28, and was further cultured for dsRNA quantification and viral protection studies. The pSS116 2 2 plasmid was also transformed using the same method to generate a photosynthesis-positive TN72 strain, called 2 3 TN72-SS, that was used as a negative control for further experiments. A full loop of TN72-dsVP28 was inoculated into 400 ml TAP medium and cultured by shaking at 100 rpm 2 7 under continuous light (50 µmol photon m -2 s -1 ) at 25 °C. The culture was harvested (centrifugation at 3,000 x g for 2 8 7 5 min), at late log phase (day 4) and the 1 st day of stationary phase (day 5) (growth analysis data not shown) and re-1 suspended with Trizol solution (Ambion, USA) for RNA extraction to obtain total RNA. Undesired DNA and 2 single-stranded RNA (ssRNA) were removed by DNase I and RNase A treatment as per the manufacturer's 3 instructions to obtain isolated dsRNA. VP28 specific dsRNA was detected by RT-PCR (RBC Bioscience, Taiwan) 4 using VP28_ESP_F and VP28_ESP_R primers. The formation of dsRNA was confirmed by observing degradation 5 of dsRNA by RNase III treatment.

6
Quantitative RT-PCR (qRT-PCR) was performed using the same set of primers as for RT-PCR (KAPA 7 Biosystem, USA). Isolated dsRNA was incubated at 95 °C for 5 min and immediately chilled on ice for 2 min prior 8 to addition into the qRT-PCR reaction to ensure denaturation of dsRNA. A series of ten-fold dilution of purified 9 dsRNA from the bacterial expression was used to generate a standard curve. The amount of standard dsRNA was 1 0 calculated according to a formula obtained from the standard curve (R 2 =0.958). shaking under continuous light at 25 O C for 3-4 days. Absorbance at 750 nm was measured and calculated for a 1 6 volume input into the polythene tubular bag containing 17 L TAP medium to make a starter absorbance at 0.1. The 1 7 hanging bag photobioreactor was set up according to Cui et al. (2021). Filter-sterilized air was provided from the 1 8 bottom of the bag and continuous illumination provided at 100 µE m -2 s -1 by Osram Lumilux Cool daylight 1 9 fluorescence tubes. The bag cultures were cultivated for 5 days at 25 O C with growth monitored by measuring 2 0 absorbance at 750 nm every 12 hours. At harvesting time, chemical flocculation by FeCl 3 at a final concentration 2 2 1 mM was performed with continuously air supplied for 10-15 min before leaving the bag to stand until the algae had 2 2 completely precipitated. Microalgal biomass were recovered by centrifugation of the concentrated culture at the 2 3 bottom of the bag at 4,500 rpm for 20 min and the paste was immediately subjected to freezing at -20 O C. This was 2 4 then lyophilized in a freeze-dryer over a period of 24 h, and the dried biomass was kept at -20 O C until further 2 5 analysis. Double-stranded RNA purification and detection were performed for confirmation of the accumulation of 2 6 specific dsRNA, as above.

Preparation of animal and viral inoculum
2 Specific Pathogen Free (SPF) Post larval (PL) (23 day old) Penaeus vannamei shrimp were purchased from a 3 major hatchery farm (supplier wishes to remain anonymous) and acclimatized in 10 parts per trillion (ppt) salinity 4 artificial seawater for 7 days prior to use as an animal model. Average animal mass and length were approximately 5 0.03 g and 1.5 cm respectively on the first day of experiment. To prepare the WSSV inoculum for the challenge 6 assay, a mixed gender group of 20 shrimps (10-15 g each) were fed with WSSV infected homogenized shrimp tissue 7 to a rate of 10% of total body mass. Visibly infected shrimps were euthanized and muscle and pleopod tissue 8 collected. WSSV infection was confirmed by the IQ2000 PCR-based detection kit (GeneReach Biotechnology 9 Corp., Taiwan). Infected shrimp muscle tissue was homogenized and kept at -20 °C for later use in the oral 1 0 challenge assay. 1 1 1 2

3
The experiment was dividing into four feeding groups of 100, representing four replicates of 25 animals. Each 1 4 replicate was cultured in its own glass tank containing 2 L of 15 ppt artificial seawater with continuous aeration.

5
Groups one and two represented negative and positive controls respectively and were fed with commercial feed 1 6 twice daily at 5% body weight throughout the experiment. Groups three and four were fed with commercial feed 1 7 supplemented at 1:1 ratio of TN72-SR (negative control transformant line from Charoonnart et al. (2019)) and 1 8 TN72-dsVP28, respectively, at the same amount as group one and two. The WSSV challenge was conducted on day 1 9 four of the experiment. Groups two, three, and four were exposed to oral administration of WSSV infected tissue at 2 0 a rate of 50% total body mass (1.25x10 8 WSSV copies/tank). Group one was not exposed. Animal survival was 2 1 observed daily until group 2 reached 100% mortality. The p2xTRBL vector is based on the principle of convergent promoter elements, each facing an in-sense 2 6 terminator located on the opposite side of a Golden Gate cloning region, and downstream of the opposite facing 2 7 promoter ( Figure 1A). For this vector, the promoter of the Chlamydomonas reinhardtii chloroplast gene rrnS, which 2 8 encodes the 16S ribosomal RNA, was chosen as it is the most transcriptionally active promoter in the C. reinhardtii 1 chloroplast (Blowers et al. 1990). Insertion of the target sequence is mediated by Golden Gate cloning using the type 2 IIS restriction enzyme Esp3I as discussed below. Once assembled, the expression cassette comprising the target 3 sequence flanked by the promoters and terminators can be excised using a second type IIS enzyme, BsaI and 4 transferred to a chloroplast integration vector such as pSS116 through a second Golden Gate cloning step, as 5 illustrated in Figure 1A. The presence of different antibiotic selection markers on the two vectors avoids the need to 6 purify the excised cassette prior to cloning into pSS116. 7 8

Golden Gate assembly allows for quick and highly efficient construction of plasmids 9
The insertion of the VP28 PCR product into the p2xTRBL vector, and of the resulting expression cassette into 1 0 the pSS116 vector, were both conducted by one-pot-one-step Golden Gate reactions. This method allows for the 1 1 directional cloning of one or more fragments into a vector without having to separate digestion and ligation steps: 1 2 the use of type IIS restriction endonuclease (RE) sites ensure that once the desired plasmid is assembled the enzyme 1 3 recognition sites are lost, precluding it from any further digestion. Both vectors also featured a colorimetric negative 1 4 screening device, such that successfully assembled plasmids lose the bacterial mRFP expression cassette located 1 5 between the pair of RE sites and thus yield white colonies, as opposed to the pink colonies generated by the parental 1 6 plasmids ( Figure 1B). Using a 1:8 molar ratio of p2xTRBL to VP28 DNA we achieved 90% successful assembly 1 7 based on colony colour. All selected white colonies (26) were shown by PCR analysis to be correctly assembled 1 8 p2xTRBL-VP28.

9
Downstream cloning of the expression cassette into pSS116 was conducted using a similar method, but with 2 0 BsaI used as the restriction endonuclease instead of Esp3I, and colonies selected on ampicillin as opposed to 2 1 tetracycline. Assembly as judged by pink/ white selection was at a similarly high level (98% white colonies), and 2 2 again all picked colonies gave the expected PCR products for the recombinant plasmid, pSS116_VP28. A diagnostic 2 3 digestion of the purified plasmids using XbaI and BstXI further confirmed this.

6
After assembly of the pSS116_VP28 plasmid in the standard E. coli strain DH5α, the plasmid was transformed containing either pSS116_VP28 or the recipient pSS116 were grown to mid-log phase (8 hours), and the purified 1 dsRNA analysed by agarose gel electrophoresis. As shown in Figure 2a, a dsRNA band of the expected size of ~0.4 2 kb is seen for the VP28 dsRNA from the pSS116_VP28 culture, whereas no such band was observed from the 3 pSS116 culture. Total dsVP28 accumulation was estimated from band intensity to be approximate 500 µg.

4
Furthermore, RT-PCR using specific primers to the VP28 sequence was conducted to confirm the presence of VP28 5 dsRNA. An amplicon of ~0.3 kb representing just the VP28 sequence without the upstream and downstream 6 transcribed regions was observed from pSS116-VP28, while this amplicon was not seen for the pSS116 line ( Fig.   7 2b). and pSS116 as a control, with the first colonies becoming visible on minimal medium after six weeks. A total of 1 5 seven colonies were observed for pSS116-VP28 and only two for pSS116 transformation.

6
After the second re-streak on minimal medium, DNA was extracted and PCR analysis conducted, with 1 7 amplification of the endogenous gene rbcL used as a positive control for extraction (Fig. 3a). The VP28 DNA was 1 8 successfully amplified from the pSS116_vp28 lines, and not seen from the pSS116 lines (Fig. 3b). As the copy   The yield of dsRNA in the TN72-dsVP28 strain was analysed following total RNA extraction, and DNase I and 1 RNase A treatment to digest all DNA and ssRNA leaving only dsRNA, as previously demonstrated in Charoonnart 2 et al. (2019). The specific presence of VP28 dsRNA in the TN72-dsVP28 RNA sample, but not in an equivalent 3 sample from the TN72-SS control strain, was investigated by RT-PCR. TN72-dsVP28 gave the expected amplicon 4 at 307 bp, whereas no band was seen for TN72-SS (Fig. 4a). Furthermore, this amplicon was lost when the dsRNA 5 template was treated with the dsRNA specific enzyme, RNase III (Fig. 4b), confirming that VP28 dsRNA was 6 indeed present in the TN72_VP28 line. To evaluate the efficiency of TN72-dsVP28 in protecting against WSSV, a feeding experiment following by a 1 9 WSSV challenge assay was performed. Prior to supplementing the feed with transgenic microalgae, cell biomass of 2 0 both TN72-SR and TN72-dsVP28 was analysed for the present of specific dsVP28 accumulation by RT-PCR 2 1 analysis and the result confirmed the present of dsRNA in TN72-dsVP28 but not TN72-SR (Supplementary data).

2
As shown in Figure 5, a lethal dose 50 was reached on days 4 or 5 post infection (dpi) in shrimp receiving the TN72-2 3 SR supplement and positive group (commercial feed without algae, full WSSV challenge), respectively whereas the 2 4 negative group (commercial feed without algae, no WSSV challenge) showed 91-94% survival. In contrast, the 2 5 shrimp that received the feed supplemented with TN72-dsVP28 showed a marked reduction in mortality compared 2 6 to the TN72-NR control with 69.49±6.25% at 6 dpi which then remained at this level ( Figure 5). On the last day of 2 7 the experiment (8 dpi), the survival of both the positive and TN72-SR supplemented groups was less than 2 8 1 2 10%,whereas shrimp receiving feed supplemented with TN72-dsVP28 showed survival of ~68%. Statistical analysis 1 of these numbers gives significant at 99% confidentiality.

6
Since chloroplast promoters are prokaryotic in origin, they are often active in E. coli, with the rrnS promoter 1 7 being no exception. This presented a useful opportunity to validate the dsRNA-expressing cassette in 1 8 pSS116_VP28, prior to algal transformation, and also to produce a stock of VP28 dsRNA to be used as a positive 1 9 control. Building on several reports where convergent bacteriophage T7 promoters have been used to produce 2 0 dsRNA in E. coli (García et al. 2015;Kim et al. 2015), we transformed the pSS116_VP28 vector into the RNase III 2 1 deficient E. coli strain HT115(DE3). The resulting system was shown to be able to produce reasonable amount of 2 2 dsRNA, but the amount was lower than that produced from hairpin expressing cassettes (Chen et al. 2018).

3
Nonetheless, hairpin cassettes are not stable in plastids due to the presence of a RecA recombinase that might loop 2 4 out the inverted repeat in hairpin cassette (Nakazato et al. 2003). A convergent cassette was therefore considered 2 5 preferable for dsRNA production in the C. reinhardtii chloroplast.

6
The use of traditional antibiotic resistance selection markers for engineering of microalgae to be used in 2 7 aquaculture is highly undesirable because of the risk of spreading such cassettes to other microorganisms by 2 8 1 3 horizontal gene transfer. The ability for chloroplast transformants to be selected without such markers is hence a 1 large advantage, as has be demonstrated numerous times (Changko et al. 2020;Charoonnart et al. 2019;Gangl et al. 2 2015;Zedler et al. 2015). Despite these benefits, the transformation with pSS116_VP28 suffered from long 3 incubation times and low transformation efficiency. The incubation period was 2 weeks longer than previous psaA 4 convergent promoter cassette (4 weeks), and those reported for protein production (Gangl et al. 2015). The delay of 5 transformation and low efficiency may have been due to the presence of the convergent rrnS promoters, with the 6 inverted repeat potentially interfering with homologous recombination; however, as colonies were ultimately 7 recovered and shown to all be correct. Moreover, the transformant strains rapidly reached homoplasmy, showing no 8 evidence of the recipient genotype after only two rounds of single-colony isolation. Similar results have been 9 achieved recently for two transformant lines where the rrnS convergent promoter system has been used to produce 1 0 dsRNA targeting the RdRp gene of YHV and VP9 and ORF366 of WSSV (unpublished results).

1
The resultant transplastomic line, TN72-VP28, was shown to express VP28-specific dsRNA by RT-PCR 1 2 following selective degradation of DNA and ssRNA. It was then not possible to amplify the VP28 sequence 1 3 following the use of RNAse III to specifically degrade dsRNA. Combined, these two results give strong evidence for 1 4 the production of VP28 dsRNA in the C. reinhardtii chloroplast. Subsequent RT-qPCR was used to quantify the 1 5 VP28 dsRNA, giving a yield of 119 µg of VP28 dsRNA from 1 L culture at an OD 750 of 3.9. Though the dsRNA 1 6 yield obtained is very much lower than a bacterial system, the value of the algal system is that it can be used as a 1 7 whole-cell feed. Furthermore, the marker-less transformation method means that the only transgenic DNA in the 1 8 feed is the ~0.3 kb section of the viral VP28 gene, and this DNA is endemic anyway in an infected shrimp pond.

9
This current study representing a ~10,000-fold increase in dsRNA compared to the earlier work using the psaA 2 0 promoter, although as this work expressed a different target sequence it is not possible to make a direct comparison.

1
It is of note that levels of dsRNA were observed to increase from day four to day five of the cultivation period.

2
Further work will be needed to confirm this trend, and give a more complete picture of dsRNA expression over a 2 3 full growth period. It would also be of interest to investigate whether other abiotic factors can influence dsRNA 2 4 accumulation, for example the effect of light levels or media composition.

5
The drying of the algal biomass represents an effective method of killing the microalgae thereby reducing the 2 6 chance of environmental contamination with a GM species when deployed in aquaculture. Furthermore, the drying 2 7 produces a powder suitable for formulation into shrimp feed where the dsRNA is potentially stable at room  The Penaeus vannamei shrimp is known to be highly susceptible to WSSV throughout its life cycle, but the post 1 0 larval (PL) stage of development is considered to be when it is must vulnerable to viral attack according to data 1 1 collection from hatchery farms and nursery farms throughout Thailand by the Department of Fisheries 1 2 (Kongkumnerd and Saleetid, 2019). It is therefore essential to ensure PL stocks are healthy and disease-free before 1 3 transferring them to grow-out ponds. Since WSSV can be transmitted horizontally, viral transmission in the TN72-1 4 SR group was potentially quicker than in the TN72-VP28 group resulting in 100% mortality being reached more 1 5 rapidly. Feeding PL shrimp with inactivated microalgae expressing dsRNA delayed the massive mortality of the 1 6 shrimp following viral infection. Together with good farm management, shrimp farmers may therefore be able to 1 7 rescue shrimp from mortality and gain some profits instead of completely discarding the whole crop..  Construction involved a two-step cloning strategy in which the VP28 DNA was first cloned into p2xTRBL using the 4 type IIS restriction enzyme Esp3I and then the dual transcription cassette was cloned into pSS116 using a second 5 type IIS enzyme, BsaI to flank the cassette with chloroplast DNA elements for recombination into the chloroplast 6 genome. (b) Cloning was aided at each stage by a simple 'pink/white' screening system where transformant colonies 7 and cultures carrying the correctly assembled plasmid gave rise to a 'white' phenotype.