A novel formulation technology for baculoviruses protects biopesticide from degradation by ultraviolet radiation

Biopesticides are biological pest control agents that are viewed as safer alternatives to the synthetic chemicals that dominate the global insecticide market. A major constraint on the wider adoption of biopesticides is their susceptibility to the ultraviolet (UV: 290–400 nm) radiation in sunlight, which limits their persistence and efficacy. Here, we describe a novel formulation technology for biopesticides in which the active ingredient (baculovirus) is micro-encapsulated in an ENTOSTAT wax combined with a UV absorbant (titanium dioxide, TiO2). Importantly, this capsule protects the sensitive viral DNA from degrading in sunlight, but dissolves in the alkaline insect gut to release the virus, which then infects and kills the pest. We show, using simulated sunlight, in both laboratory bioassays and trials on cabbage and tomato plants, that this can extend the efficacy of the biopesticide well beyond the few hours of existing virus formulations, potentially increasing the spray interval and/or reducing the need for high application rates. The new formulation has a shelf-life at 30 °C of at least 6 months, which is comparable to standard commercial biopesticides and has no phytotoxic effect on the host plants. Taken together, these findings suggest that the new formulation technology could reduce the costs and increase the efficacy of baculovirus biopesticides, with the potential to make them commercially competitive alternatives to synthetic chemicals.

www.nature.com/scientificreports/ when the maximum continuous exposure to simulated sunlight was 16 h, the non-encapsulated virus and the commercial virus formulation showed a non-linear decline in efficacy with increasing simulated sunlight dose, with minimum efficacy at 16 h of around 35% (Fig. 1a). In contrast, the ENTOSTAT-TiO 2 -encapsulated virus showed an initial small decline in activity but thereafter activity remained at a relatively constant level at around 87%. This observation is reflected in the statistical analysis, with a significant interaction between virus formulation and the number of hours simulated sunlight exposure (Generalised linear model, GLM: Hours exposure: χ 2 1 = 104.77, P < 0.0001; [Hours exposure] 2 : χ 2 1 = 20.96, P < 0.0001; Formulation: χ 2 2 = 158.19, P < 0.0001; Formulation*Hours exposure: χ 2 2 = 9.59, P = 0.0083). When considered alone, there was a significant decline overall in the performance of the ENTOSTAT-TiO 2 -encapsulated virus (GLM: χ 2 1 = 6.00, P = 0.014), but when the 0 h simulated sunlight-exposure time point was excluded, there was no significant decline in performance thereafter (χ 2 1 = 1.73, P = 0. 19), indicating that beyond 1 h exposure to simulated sunlight there was no further degradation of virus efficacy.
96 h simulated sunlight exposure. To explore this further, the duration of exposure to simulated sunlight in the ATLAS SUNTEST XLS + cabinet was increased up to a maximum of 96 h and the results again showed that TiO 2 protects the NPV from degradation from simulated sunlight. Whilst the ENTOSTAT-formulated virus showed only a limited decline in efficacy after the first exposure to simulated sunlight and maintained > 80% efficacy throughout, the non-formulated NPV and the commercial virus showed a non-linear decline in efficacy to a minimum of around 30% (Fig. 1b). This is again reflected in the statistical analysis (GLM: Hours exposure: χ 2 1 = 107.29, P < 0.0001; [Hours exposure] 2 : χ 2 1 = 55.84, P < 0.0001; Formulation: χ 2 2 = 190.50, P < 0.0001; Formulation*Hours exposure: χ 2 2 = 9.34, P = 0.0094). When the ENTOSTAT-TiO 2 -encapsulated virus was considered alone, there was a marginally non-significant decline in performance over the 96 h exposure period (GLM: χ 2 1 = 3.69, P = 0.055), and when the 0 h simulated sunlight-exposure time point was excluded, there was no significant decline in performance (χ 2 1 = 0.31, P = 0.57), suggesting again that the virus is not subject to any further loss of efficacy after the first hour exposure to simulated sunlight. phytotoxicity. There was no effect of virus formulation on any aspect of the growth of the cabbage plants (Table 1). titanium dioxide toxicity. Titanium dioxide had no effect on larval mortality either when NPV-free ENTOSTAT-TiO 2 was compared to controls that were fed dH 2 0 only (z value = 0.326, P = 0.744), or when TiO 2 -formulated NPV ENTOSTAT was compared to ENTOSTAT NPV alone (z value = 0.174, P = 0.862). Overall, however, there was a significant difference between the four treatments in the levels of larval mortality they generated (Treatment: χ 2 3 = 525, P < 0.0001) because the two NPV treatments caused higher mortality rates than treatments that did not include NPV (Fig. 3). www.nature.com/scientificreports/ Storage stability. After six months' storage, virus-induced mortality was significantly lower for viruses stored at 30 °C than at 4 °C (GLM: χ 2 1 = 24.70, P < 0.0001) and differed across the three virus formulations (χ 2 2 = 29.54, P < 0.0001), with non-formulated NPV causing lower mortality in S. littoralis larvae than the two formulated NPVs (z value = -4.414, P < 0.0001) ( Table 2). There was no significant difference between the mortality rates caused by the commercial standard and the wax-encapsulated NPV (z value = − 1.287, P = 0.198); the interaction between temperature and formulation was also non-significant (χ 2 2 = 0.86, P = 0.65), indicating additive effects of temperature and formulation.   www.nature.com/scientificreports/

Discussion
Overall, the results of the work reported here using simulated sunlight show that ENTOSTAT-TiO 2 formulated NPV is considerably more resistant to UV degradation than either non-formulated NPV or an existing commercial NPV product, and that it has storage and crop safety characteristics that are at least equal to those of existing formulations.
In any formulation of an infectious biological agent, it is important that the formulation processes result in little or no significant loss of activity 19,45 . The bioassay results confirmed that the ENTOSTAT encapsulation process, which involves heating and milling the mixture, does not reduce the infectivity of the baculovirus. Thus, the encapsulation process does not appear to damage the virus OB and, upon ingestion, the protective ENTOSTAT-TiO 2 wax coat successfully dissolves in the insect midgut to liberate the infectious virions.
A key finding from the current study is that by combining ENTOSTAT wax with a UV protectant (TiO 2 ), the efficacy of the wax-encapsulated baculovirus could be dramatically improved, even under intense UV radiation, from just a few hours to at least several days. This ENTOSTAT-TiO 2 -NPV formulation is superior to both nonformulated NPV and a commercial NPV formulation (LITTOVIR). The non-formulated virus and commercial NPV formulation had half-lives on glass slides of < 17 h continuous simulated sunlight exposure (Fig. 1b), while the ENTOSTAT-TiO 2 formulated virus retained > 80% activity even after 96 h continuous exposure. The initial drop in the efficacy of ENTOSTAT-TiO 2 formulated virus during the first hour of UV exposure, is most likely due to denaturing of virus that was on not fully encapsulated, whilst subsequent long-term persistence was probably due to virus that was completely encapsulated with the wax. The results from both the slide and plant systems show that the half-lives for non-formulated virus and commercial formulation fall within the reported persistence values of about 24 h in continuous simulated sunlight, and within the half-lives of 1-7 days reported in the field, depending on sunlight intensity and the degree of crop shading 2,6,20,46 . The results from the ENTO-STAT-formulated virus are very promising as no previously reported studies using a baculovirus formulation have demonstrated anything approaching this degree of UV-stability. The slide exposure system used here (ATLAS SUNTEST XLS + cabinet), differs from many other published formulation studies in that it was calibrated to an actual sunlight exposure level, and the intensity and duration was monitored throughout using a system already validated for UV-stability studies with baculoviruses 30,35,47 . It has been pointed out that laboratory studies of baculovirus sunlight stability using ad hoc arrangements of UV or germicidal lamps not calibrated to real sunlight can be very hard to relate to field conditions 30 . The slide simulated sunlight exposure system used here also involved controlling the slide temperature, avoiding the confounding effects of the heating of glass slides, sometimes seen when slides are exposed to artificial or natural sunlight without adequate temperature control 20 . Thus, the ATLAS SUNTEST system is seen as giving a realistic model of natural sunlight under temperate conditions and has previously been used with both the bacterium Bacillus thuringiensis and the baculovirus Anagrapha falcifera multiple nucleopolyhedrovirus, MNPV 35,48 . It is widely used as an industry standard equipment for testing products for solar stability and temperature (Atlas 2018), conforming to American Society for Testing and Materials (ASTM) and International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) standards for coatings, cosmetics and pharmaceutical products.
Formulations showing improved UV stability in the laboratory, however, do not always show improved persistence or efficacy in crop trials 6,35 . For example, although the UV protectant BLANKOPHOR BBH increased baculovirus persistence in lab trials, it did not increase effectiveness in field trials 49 . The crucial test of UV formulations must remain the use of plant studies, as this alone can most adequately approximate to the oncrop situation. The plant-based system used here draws on approaches that are standard in plant UV research to ensure radiation treatments were as close to the field as possible, for example filtering the UVB sources to remove wavelengths below those present in sunlight 50,51 . The crop results on tomato and cabbage show that the ENTOSTAT-TiO 2 formulated baculovirus is substantially better than non-formulated virus or an existing commercial formulation, showing > 80% activity after 16 d diurnal simulated sunlight exposure, while the natural and commercial formulations both retained < 20% activity under these conditions. It remains to be established whether these findings are replicated in field trials under a range of natural sunlight conditions.
The bioassays reported here indicate that TiO 2 does not, in itself, have any insecticidal activity, as the mortality of insects fed on ENTOSTAT-TiO 2 encapsulated virus, was no greater than that of the virus alone (but possible sub-lethal effects of TiO 2 nanoparticles have been reported in larvae of the waxworm moth, Galleria mellonella 52 ). Table 2. Virus-induced mortality of S. littoralis larvae inoculated with one of three SpliNPV formulations stored for 0, 3 or 6 months at 4 °C or 30 °C. Values are means (range), n = 5 replicate bioassays. Comparison with LITTOVIR commercial standard at 4 °C (LM: *P < 0.05, ***P < 0.001). www.nature.com/scientificreports/ Photostabilized Titanium dioxide is widely used in sunscreens and cosmetic products because it has been identified as safe, a factor in its favour as a potential formulation ingredient for biopesticides 30,53 . An issue for some formulation additives is the potential for phytotoxicity. It has been reported that when stilbene-derived UV blockers were sprayed onto plants, they had significant negative effects on the growth of several crop species tested 54 . Trials here showed that ENTOSTAT-TiO 2 had no adverse effect on cabbage plant growth. This could be ascribed to the very low application rate of TiO 2 when delivered as part of the ENTOSTAT formulation, which are substantially lower than in previous studies. Titanium dioxide used in suspension when sprayed at 200 L per ha would require 2 kg of TiO 2 per ha 30 , whereas in an ENTOSTAT formulation, because the TiO 2 is intimately bound to the wax-encapsulated OB rather than in general suspension, ≤ 20 g per ha only would be applied.
The research reported here is the first use of ENTOSTAT technology for encapsulating a biopesticide, as previous work has involved biopesticides being formulated to adhere to the outside of ENTOSTAT wax particles 40,55 . ENTOSTAT-encapsulation of baculovirus is a novel technology developed and patented during this research collaboration (see Competing Interests below for patent information). There have been previous studies using encapsulation of baculoviruses using polymers or lignin 20,35,37,39,56 . Some showed enhanced storage or UV-protection but none for the duration reported here, and only one of these studies with lignin formulation progressed to successful plant trials 56 .
An absolute requirement for a biopesticide formulation is that it can be applied through existing commercial spray systems 6 . Preliminary results, using a spray boom attached to a Hardi ilemo 1,000 L orchard sprayer using a lilac Albuz ATR80 nozzle, show that the ENTOSTAT formulation is sprayable (data not shown).
Another important issue for any biological pesticide is storage stability, as agents that cannot be formulated to remain stable for long periods have low viability as commercial biopesticides 33,45 . Baculoviruses are relatively stable robust agents 23 , but as formulated in water or glycerol have struggled to match the 2-4 years storage stability of chemicals and require freezer or cool chain storage 5,20,25 . The results presented here show that ENTOSTAT-TiO 2 -NPV has storage properties at least equal to that of a commercial glycerol-based formulation at 30 °C over 6 months. However, longer-term trials are needed to see if the ENTOSTAT-formulated NPV can reach the 18-24 month target proposed for biopesticides 20 .
A key issue with any commercial formulation is to ensure that the financial benefits outweigh the costs 19 . Whilst many studies have highlighted formulation additives that improve the performance of biological agents 20,57 , very few discuss the cost-benefit ratio of the additives/formulations. The use of UV protectants in biopesticides has been constrained because the typical inclusion rates are too high and too costly for routine field use 5,13,30 . With ENTOSTAT-TiO 2 , the UV protectant is tightly bound to the baculovirus OB, so only small quantities (g per ha) are required compared to when blockers are used in suspension (Kg per ha). For example, TiO 2 applied as a suspension at an effective 187 L per ha was costed at $18.06 per ha, a significant issue that has constrained its adoption so far in commercial formulations 30 . In an ENTOSTAT-TiO 2 formulation the equivalent cost for the TiO 2 would be around $0.08 per ha (in 2003 prices). Thus ENTOSTAT-TiO 2 can be said to address this issue as the key ingredients, wax and photo-stabilizing TiO 2 , are readily and cheaply available, and at the rates that would be used in the field would cost < $1 per ha.
A major cost associated with any commercial biopesticide is that associated with producing the active ingredient, and this is particularly true of baculoviruses, which commercially are produced in vivo. If the UV persistence of baculoviruses can be increased by using the ENTOSTAT-TiO 2 formulation, then the amount of baculovirus applied might be drastically reduced. The current high application rates of around 1-5 × 10 12 OB/ha are needed to overcome the short persistence issues. It has been suggested that with more effective UV protection, application rates could be reduced by a factor of × 10 30 . It is thus conceivable that adopting ENTOSTAT-TiO 2 formulation could enable producers to reduce active ingredient rates and that this would have a drastic effect on the cost of baculovirus products. It may even enable baculovirus products to reach the goal proposed by Reid et al. 23 of bringing biopesticide product costs below US$ 20/ha; making them, for the first time, competitive in broad acre crops. This could vastly increase their potential market share. Greater UV persistence could also allow biopesticides to be sprayed prophylactically in response to cues of imminent pest attack (e.g. large numbers of reproductive adults in traps) rather than timed precisely to coincide with pest appearance on crop. Moreover, it may be hypothesised that baculoviruses could be successfully micro-encapsulated in ENTOSTAT waxes with a broad range of different additives to give additional desirable characteristics, such as phagostimulation, broader host ranges, enhanced kill rates, etc., to deliver much 'smarter' biopesticides that enhance the beneficial properties of these biological entities 24 .
The issue of limited UV persistence is not restricted to baculovirus biopesticides, but also affects all of the major groups of pathogens used in crop protection 20 . Thus, it is possible that the ENTOSTAT-encapsulation system described here could be used with other pathogens, such as the widely-used bacterium Bacillus thuringiensis, whose on-crop persistence while longer than that of baculoviruses, is similarly limited to a few days 20,48 . conclusion We present here a novel formulation technology that effectively safeguards the efficacy of a baculovirus biopesticide by protecting the sensitive viral DNA from damaging UV radiation in sunlight on the crop by encapsulating it in an ENTOSTAT-TiO 2 waxy coat. This technology is a highly promising candidate formulation, whose adoption in baculovirus and other biopesticide formulations could greatly increase the persistence and effectiveness of biopesticides while reducing costs. This novel formulation could be a means of greatly expanding the use of biopesticides to move them from the role of niche products for high-value and protected crops into large-scale field crop use. This would meet the need for a safer, more ecologically-acceptable, pest control approach to replace www.nature.com/scientificreports/ those chemical pesticides that have been, or are currently being, removed from the market-place in response to public pressure for more environmentally-sustainable crop production.

Materials and methods
insects. Egyptian cotton leafworm (Spodoptera littoralis) used were originally collected in Egypt and had been maintained at Lancaster University since 2011. Except where stated, larvae were reared in isolation in 25 mL plastic pots containing a wheatgerm-based semi-artificial diet 58,59 . Virus. The S. littoralis nucleopolyhedrovirus (SpliNPV) used here was NRI strain #0,084 produced in the laboratories of the Natural Resources Institute, University of Greenwich, and purified for formulation using a standard purification technique for NPV 45,58 . The virus was enumerated using a Neubauer improved haemocytometer and viewed under phase contrast microscopy at × 400 using a Leica DMR microscope 45 . The virus was freeze-dried using the published protocol 58 in a SUPERMODULOYO 20 (Edwards).
Surface dose bioassays. The laboratory bioassays reported here employed a modified surface dose bioassay method 58,60 that utilised 96-well flat-bottomed cell culture plates (CORNING COSTAR). The bioassay used a smooth semi-synthetic diet using ground wheatgerm 58 and was dispensed into each well before being stored in a fridge (4 °C) for later use. For bioassays, virus formulations were prepared in dH 2 0 containing 4% food dye (Dr. Oetker). After vortexing, 10 µl of test suspension was pipetted onto the surface of the diet in the 96-well plates in a structured randomised pattern. After drying, a single starved L2 larva was added to each well and each plate was wrapped in PARAFILM. These were then left for 24 h in an incubator at 27 °C after which, each larva was transferred to an individual 25 mL plastic pot containing fresh diet. After two days, handling deaths and missing larvae were recorded and mortality assessed at 8 and 15 days post-inoculation.
formulation process. The formulation process is proprietary information and so only brief details are provided here. Pilot studies tested a range of different waxes, chosen based on their suitability for formulation with NPV. The NPV was incorporated into the selected ENTOSTAT wax in a melt phase with a virus loading of 2% NPV/wax (w/w) and stirred to form a uniform liquid using a high sheer blender. The UV protectant was then added, and the mixture cooled before being milled using a kibbling mill and micronized in a jet mill to a fine powder (X50 ~ 5-15 um). Photo-stabilised Titanium dioxide (TiO 2 ) was selected as the formulation UV protectant. The suspended concentrate (SC) formulation was made by suspending the NPV/wax micro-powder with a blend of proprietary combinations of wetting agents, dispersers, rheology modifiers and other co-formulants with TiO 2 , to create a uniform SC. Unlike non-formulated NPV, it is not possible to visually determine the concentration of OBs in wax-formulated virus, therefore the virus loading for the ENTOSTAT-TiO2-encapsulated NPV is an estimate based on how much NPV was added into the formulation.
Simulated sunlight slide exposure bioassays. Slide exposure. Test formulations were applied to blank ground glass slides. 200 µL of formulation was applied to each slide as ten 20 µL droplets then air dried at room temperature prior to exposure. The slides were exposed to UV using an ATLAS SUNTEST XLS + cabinet with the SunCool attachment comparable to other UV studies 35,47 . The chamber was set to apply 65 W m −2 using the daylight filter and was calibrated to match standard sunlight exposure at solar noon on the vernal equinox at Miami, Florida 61 ; the chamber temperature was standardised to 20 °C using the SunCool attachment. Five slides of each formulation were exposed for each time point tested over 0-96 h. Slides were arranged within the cabinet in a randomised three by five formation, to minimise the effect of any variation in UV exposure across the cabinet. After exposure, slides were immersed in 0.02% sodium dodecyl sulphate (SDS) and brushed to resuspend and recover the exposed formulation that was then placed into a universal vial for storage at 4 °C until bioassaying. Seven replicate bioassays were completed across eighteen blocks in a structured randomised design. Three formulations were tested: non-formulated virus, a commercial standard (LITTOVIR), and NPV formulated in ENTOSTAT wax with TiO 2 additive. The first four replicates included UV doses up to 16  Simulated sunlight plant exposure bioassays. Virus formulations were applied to plants that were then exposed to artificial lighting that included wavelengths in the UV range of the spectrum (see below). Two plant species were used: tomato (variety Ailsa Craig) and cabbage (variety Greyhound). These were grown from seed (Moles Seeds, Colchester, U.K.) in a glasshouse in John Innes No. 2 compost. When the plants were at an appropriate stage of growth (8-12 weeks old), the plants were sprayed with either dH 2 O (control), LITTOVIR (6 × 10 8 OB/mL) or ENTOSTAT-TiO 2 -encapsulated NPV (~ 7 × 10 8 OB/mL). Between one and three plants were used per UV dose per treatment group. The suspension was applied evenly to the plants using 1.25 L pressure sprayers (Hoselock, Birmingham, U.K.) immediately following preparation. Plants were then left to dry overnight in a glasshouse before being moved to the controlled environment (CE) room the following morning (25 ± 2 °C).
Plant exposure to simulated sunlight used facilities and approaches that we have described previously 50 . Plants were arranged in a completely randomised design in a CE room under artificial illumination provided by LED arrays (Valoya BX180, Valoya Oy, Helsinki, Finland) for photosynthetically active radiation (400-700 nm, mean irradiance over the growing area 300 micromole quanta m -2 s −1 ), and fluorescent tubes for UVA and UVB www.nature.com/scientificreports/ radiation (UVA340 and UV313 respectively, both from Q-Panel Laboratory Products, Bolton, UK). The UVB tubes were filtered with 0.13 mm thick cellulose diacetate (Clarifoil, Courtaulds Ltd, Derby, UK) to remove wavelengths below ~ 290 nm. The light environment was measured using a double monochromator scanning spectroradiometer (model SR991-v7; Macam Photometrics, Livingston, UK). The UV treatment provided a mean total UV irradiance (290-400 nm) of 11.55 W m −2 . The LED lights were switched on for 12 h per day, which gave a daily UV dose in the plant system roughly equivalent to 2 h in the ATLAS SUNTEST system (i.e. 499 kJ m −2 day −1 ). The control treatment (no UV) was provided on the same bench as the UV treatments, but used UV313 UVB tubes wrapped in clear UV-opaque polyester (Lee Filters, Andover, UK) which absorbed 97% of the UV less than 400 nm (unweighted UV irradiance 0.36 W m −2 ). Once exposed, the plants were removed from the CE room and 5-15 leaf discs were cut from each plant using a 5 mm corkborer. These were each immediately placed in an individual cell of a 25-well square plate (10 cm × 10 cm). Individual L2 S. littoralis larvae that had been starved for 4 h were then added to each cell of the plates and bioassayed as detailed above. Each treatment group comprised 10-15 larvae per UV dose, making a total of 120-270 larvae per bioassay.

phytotoxicity.
To determine any effects of the ENTOSTAT-TiO2-encapsulated NPV on the growth of the plants, the following attributes were measured in the 36 cabbage plants used in the final plant bioassay prior to cutting out the leaf discs at day 16 post-application: total number of leaves, number of healthy-looking leaves, plant height, fresh weight and dry weight. Two control treatments were used for comparison: dH 2 O (negative control) and LITTOVIR. titanium dioxide toxicity. To determine if there was any mortality associated with the TiO 2 additive, L2 S. littoralis larvae were inoculated with one of four treatments: dH 2 O (control), ENTOSTAT-TiO 2 -encapsulated NPV, ENTOSTAT-TiO 2 wax blank (lacking NPV), and non-encapsulated SpliNPV using the standard bioassay. Each bioassay comprised 60 larvae per treatment (12 in the controls) and was repeated four times.

Storage stability.
To compare the storage properties of three formulations of SpliNPV, they were stored at two temperatures (4 °C and 30 °C) and sampled at three time-points (0, 3 and 6 months). The three formulations were: freeze-dried SpliNPV in aqueous suspension; LITTOVIR (commercial standard SpliNPV, Andermatt Biocontrol); and ENTOSTAT-encapsulated SpliNPV. Replicate samples of the two non-commercial virus formulations were bottled in air and sealed, as per the commercial standard LITTOVIR. Five replicate bottles of each virus sample were bioassayed at each time point and storage temperature (i.e. a total of 30 bottles per bioassay), conducted over a single 5-day period, thus providing five independent replicates of each formulation at each storage temperature and time-point. Bioassays used the standard 96-well plate surface dose assay with 30 larvae per formulation. A single dose of each virus was used (5 × 10 6 OB/mL), chosen to achieve close to 100% mortality at time point 0 (i.e. pre-storage). Larvae in the control group were exposed to an aqueous suspension of the wax blank formulation. Statistical analysis. Data analyses were performed using the R statistics package (R Statistical Software, version 3.3.3 2017-03-06 62 ). Mortality data were analysed using logistic regression (generalized linear models, GLMs, with binomial errors and logit link function) using a stepwise deletion approach. All other analyses used linear models (LMs) and data were tested for normality and transformed if required.

Data availability
All data will be made available on Dryad upon acceptance.