Insecticidal Activity of Photorhabdus luminescens against Drosophila suzukii

Drosophila suzukii causes considerable economic damage to small and thin-skinned fruits including cherry, blueberry, raspberry, grape and strawberry. Since it attacks fruits at the ripening stage, the use of chemical pesticides is limited due to the high risk of residues on fruit. Biological control is thus expected to play an essential role in managing this pest. The Gram-negative bacterium, Photorhabdus luminescens and its symbiotic Heterorhabditis spp. nematode have been shown to be highly pathogenic to insects, with a potential for replacing pesticides to suppress several pests. Insecticidal activity of P. luminescens at different bacterial cell concentrations and its cell-free supernatant were assessed against third-instar larvae and pupae of D. suzukii under laboratory conditions. P. luminescens suspensions had a significant oral and contact toxicity on D. suzukii larvae and pupae, with mortalities up to of 70–100% 10 days after treatment. Cell-free supernatant in the diet also doubled mortality rates of feeding larvae. Our results suggest that P. luminescens may be a promising candidate for biological control of D. suzukii, and its use in integrated pest management (IPM) programs is discussed.


Introduction
Drosophila suzukii (Matsumura) (Diptera Drosophilidae) is an invasive species that threatens soft fruit industries in America and Europe [1,2] through feeding on unripe and undamaged cherry, blueberry, raspberry, grape and strawberry [2][3][4][5][6], causing extensive economic losses [7]. Chemical pesticides are the main D. suzukii control methods [8], but their use has to be limited due to the high risk of residues on the fruit, insect resistance development and their negative impact on beneficial insects [9][10][11]. Alternative and more sustainable control strategies are therefore constantly being sought [12]. Biological control agents are expected to play an essential role, being a cost-effective and environmentally safe approach for the management of this pest [8]. Control approaches based on biological agents are thus highly recommended to establish effective and sustainable integrated pest management (IPM) programs.
Several commercially available biological agents including parasitoids [13,14], predators [15,16], nematodes [16,17], entomopathogenic fungi [16,18,19] and bacteria [20] have been evaluated against D. suzukii under laboratory conditions. Currently, formulations based on the entomopathogenic fungus Beauveria bassiana are allowed for treatment against D. suzukii on cherry, grapevine and strawberry in Europe [21]. Entomopathogenic bacteria, such as Bacillus thuringiensis, Serratia spp., Pseudomonas Insects used in the experiments originated from adult D. suzukii collected from field infested cherry and grape fruits in Verona Province, North-Eastern Italy and maintained in the laboratory of the Department of Agronomy, Food, Natural Resources, Animals and the Environment, Padova University under controlled conditions. Male and female adults (mixed ages) were placed in plastic vials (Falcon type with 50 mL capacity, diameter 30 mm, length 115 mm) with a specific medium for D. suzukii rearing [33,34].
Diet components (75 g raw cornmeal, 17 g dry-yeast, 15 g sucrose, 12 g soybean meal, 5.6 g agar and water adjusted to 1000 mL) were thoroughly mixed and cooked for 20 min at about 100 • C, 5 mL propionic acid were then added at a temperature of less than 50 • C just before pouring 15 mL of medium into the plastic vials. Cultures were maintained in climate chambers held at 23 ± 1 • C, 70 ± 10% relative humidity and 16L:8D photoperiod. Wild D. suzukii were introduced into the colony on multiple occasions to ensure that the genetic make-up of the individuals screened in the laboratory was representative of the field population.

Bacteria Isolation and Culture Conditions
P. luminescens was isolated from the soil of the experimental farm of Bologna University (Cadriano, Lat: 44.548985, Long: 11.386292). For the isolation, G. mellonella larvae, purchased at a local fishing gear retailer, were placed in 15 mL plastic tubes closed with 1 mm-mesh nets [35]. Twenty tubes were buried at 10 cm depth in a set-aside plot of an orchard where chemical pesticide had not been directly sprayed in the last 12 months, to attract entomopathogenic nematodes naturally carrying P. luminescens. The tubes were left underground for 3 days. Successively, to isolate the bacterium, each G. mellonella larva was homogenized in 2 mL of sterile MgSO 4 (10 mM). The homogenate was serially 10-fold diluted in the same buffer and 200 µL of each dilution was plated on a modified nutrient agar medium supplemented with 0.004% (w/v) triphenyltetrazolium chloride and 0.0025% (w/v) bromothymol blue (NBTA) [36] amended with cychloheximide (0.01% w/v) to prevent fungal contamination. The plates were incubated at 30 • C for 48 h. Colonies that absorbed bromothymol blue were selected and singularized onto MacConkey agar (Biolab) and tryptic soy broth (TSB) agar medium. The different isolates were identified as described by Liu et al. [37]. P. luminescens subsp. akhurstii (strain W14, DSM15138) was used as positive control from molecular identification. The strain was purchased from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures. Each identified isolate was stored at −80 • C in a cryotube containing TSB amended with 40% sterile glycerol.
The P. luminescens isolates were tested against D. suzukii. The bacterial suspension for the bioassay experiments on D. suzukii was prepared from P. luminescens culture grown in 200 mL of TSB incubated at 28 • C under moderate shaking (150 rpm) for 3 days. Bacterial concentration was estimated by optical density at 600 nm and adjusted to 3.5 × 10 8 cells mL −1 , which was higher than the average concentration (4 × 10 7 cells mL −1 ) shown to be effective against larvae of Aedes aegypti [38] and G. mellonella [12]. The final concentration was also confirmed by 10-fold serial dilutions and plating on TSB agar medium. To obtain a bacterial supernatant containing toxic metabolites, the bacterial suspension at 3.5 × 10 8 cells mL −1 concentration was centrifuged at 10,000× g for 10 min at room temperature. The obtained supernatant was filtrated with a 0.2 µm Millipore filter. The resulting filtrate supernatant was checked for sterility by placing a 100 µL aliquot on Luria-Bertani (LB) agar plates [34]. The supernatant was diluted 10-fold with TSB. The undiluted supernatant and 10-fold dilution were tested against larvae and pupae of D. suzukii.

Larval Oral Bioassay
For the oral bioassay [26,27] six bacterial concentrations (3.5 × 10 8 , 3.5 × 10 7 , 3.5 × 10 6 , 3.5 × 10 5 , 3.5 × 10 4 , 3.5 × 10 3 cells mL −1 ) of P. luminescens and bacterial supernatants were evaluated against third-instar larvae of D. suzukii. Sterile D. suzukii diet (10 mL), amended with 1 mL of the P. luminescens suspensions or its supernatant was placed in 9 cm diameter Petri dishes and allowed to dry. Thus, each treatment corresponded to a 1:10 dilution of each bacterial suspension. Ten third-instar larvae were put in each Petri dish. All dishes were sealed with Parafilm before incubation at 25 • C for 4 days. TSB treatments acted as control. Each treatment had three replicates, and the experiments were independently repeated three times. Larvae mortality was recorded at 2 and 4 days after application (DAA). Larvae were considered dead if they did not move when lightly touched with a camelhair needle [38,39]. The development of surviving larvae into pupae and adults was then followed until 10 DAA given that the developmental periods from egg to adult is 9-10 days at a constant temperature of 25 • C [40,41]. Pupae were considered dead if they did not produce live adults. Percentage mortality was calculated for immature stages [larvae (4 DAA) and pupae (10 DAA)], and all life stages [larvae (4 DAA), pupae (10 DAA) and adults (10 DAA)] of D. suzukii.

Larval and Pupal Dipping Bioassay
Dipping bioassays were used for evaluating the insecticidal activity of P. luminescens at concentrations of 3.5 × 10 8 , 3.5 × 10 7 and 3.5 × 10 6 cells mL −1 against larvae and pupae of D. suzukii. The same procedures considered in the larvae-oral bioassay were followed in the dipping bioassays, except that larvae and pupae were fully immersed in bacterial cells for 30 s before being placed in Petri dishes, instead of adding P. luminescens to the diet. Controls were dipped in TSB. Survival was determined as development to subsequent stages. Larvae mortality was recorded 2 and 4 DAA. Mortality of emerged pupae and adults was assessed at 9 and 10 DAA, respectively. For pupal-dipping bioassay, mortality of both pupae and emerged adults at 9 DAA was recorded.

Adult Insect Bioassay
P. luminescens was allowed to grow in the LB broth, then it was precipitated by centrifugation (5000× g, 10 min) and resuspended in sterile 10 mM MgSO 4 solution to a final concentration of 2 × 10 6 , 2 × 10 4 , and 2 × 10 2 cells mL −1 . The supernatant was filter-sterilised with a 0.2 µm Millipore filter.
The bacterial suspensions and the culture supernatant were tested against adults of D. suzukii (2-3 weeks old). After mixing 1:1 (v/v) with a 20 mL L −1 agar solution maintained at 50 • C, the bacterial suspension or supernatant were spread on the walls of a 50 mL Falcon tube. Five to ten insect flies were introduced into the tube, together with about 1 g of freshly-cut cherry slices, and closed with cotton tissue to allow for air exchange. Controls were treated with sterile MgSO 4 solution (to evaluate the effect of live bacteria suspensions) and LB medium (to evaluate the effect of solutes dissolved in the supernatant). Each treatment was replicated 3 times. The tubes were incubated as previously described, and mortality was recorded every two days until 9 DAA.

Bacterial Persistence on Cherry Fruit
Cherry fruit were surface-sterilized by washing in 1.5% NaOCl for 10 min. Subsequently, the fruit was rinsed twice in sterile water, and inoculated by 1-min dipping in a bacterial suspension (1.1 × 10 8 cells mL −1 ) in 10 mM MgSO 4 . After inoculation, the fruit was allowed to dry under laminar air flow, and then maintained at room temperature in a sterile jar allowing gas exchange.
Three fruit samples were taken 0, 3, 24, 48 and 96 h post inoculation. Each cherry was singularly washed in 10 mL MgSO 4 solution (30 min, 120 rpm shaking), and serial tenfold dilutions of the wash were plated on LB-Agar plates amended with 100 mg L −1 cycloheximide, to determine the bacterial concentration. Before inoculation, a fruit sample was subjected to the same process to verify the sample sterilization.

Data Analysis
Bioassay data were analyzed by one-way analysis of variance (ANOVA) followed by means separation with Fisher's least significant difference (LSD) test, using the appropriate models for a completely randomized design (laboratory bioassays) (SAS Institute, Cary, NC, USA, 2010). Percentages of adult mortality were transformed to arcsine [sqrt (%mortality)] before analysis to stabilize variance and reported means were back-transformed to percentages for presentation. Data were expressed as the mean and confidence intervals (CI) of all the replicates from the three experimental repetitions (n = 9). Differences were considered significant at p < 0.05 level. Abbott's formula [43] was used to correct for control mortality (efficacy).

Larval and Pupal Dipping Bioassay
In these bioassays, the potential activity of P. luminescens was tested on D. suzukii larvae and pupae (Figure 1c,d), while the supernatants were not evaluated. For larval-dipping bioassay ( Table 3), none of the tested P. luminescens concentrations were able to significantly decrease the number of treated larvae at 2 (F = 0.25; df = 3; p = 0.8592) and 4 (F = 1.07; df = 3; p = 0.4158) DAA. However, higher mortality than the control was observed in the 3.5 × 10 7 and 3.5 × 10 8 cells mL −1 treatments at 4 DAA. A significant decrease in the number of emerged pupae that were previously treated as larvae was caused in the treatment 3.5 × 10 8 cells mL −1 at 9 DAA compared to the control (F = 3.09; df = 3; p = 0.0481), showing pupal efficacy of 54%. No significant differences appeared in the mortality of emerged adults at 10 DAA (F = 1.10; df = 3; p = 0.4033). By the tenth day, P. luminescens significantly (F = 4.25; df = 3; p = 0.0451) reduced the number of treated immature stages, with the highest activity (64%) observed at 3.5 × 10 8 cells mL −1 concentration. D. suzukii individuals at all life stages were significantly (F = 5.40; df = 3; p = 0.0252) affected by the tested concentrations. High efficacy (82%) was provided with a concentration of 3.5 × 10 8 cells mL −1 ; while fairly good efficacy (50%) was shown with 3.5 × 10 7 . Table 3. Mortality of Drosophila suzukii larvae, pupae and adults following dipping the larvae in Photorhabdus luminescens at different concentrations. For the pupal dipping bioassay (Table 4), no significant differences were observed in the number of pupae (F = 2.83; df = 3; p = 0.1062) or emerged adults (F = 0.64; df = 3; p = 0.6116) at 9 DAA, however, all P. luminescens treatments caused higher percentage mortality than the control. Generally, all treatments caused greater mortality of pupa and adult D. suzukii individuals than the control due to their pupae being dipped in different bacterial concentrations, with significant mortality (≥70%) at 3.5 × 10 8 and 3.5 × 10 7 cells mL −1 . Means and CI were obtained from n = 9 samples. Values followed by the same letter(s) within columns are not significantly different according to LSD test for p < 0.05.

Pupal Direct-Spray Bioassay
In the pupal direct-spray bioassay (Table 5), significant reductions in live pupae numbers at 9 DAA were caused by all concentrations compared to the control (F = 4.38; df = 6; p = 0.0107), while the number of emerged adults was not significantly affected (F = 2.01; df = 6; p = 0.1326) even if the treatments caused a higher percentage mortality than the control. Overall, significant differences (F = 32.96; df = 6; p < 0.0001) in the mortality of pupating and adult D. suzukii individuals were caused in all treatments as a result of pupae being sprayed with different bacterial concentrations, with high efficacy (100%) at 3.5 × 10 8 and 3.5 × 10 7 cells mL −1 .  Regarding supernatant applications, the treated pupae were not significantly affected by either undiluted supernatant or the 10-fold dilution (F = 1.50; df = 2; p = 0.2621), even if mortality was higher than the control (Table 2). A significant effect of P. luminescens supernatant (F = 10.29; df = 2; p = 0.0025) appeared against the emerged D. suzukii adults at 9 DAA. Overall, a 74% cumulative mortality in all individuals was significant (F = 22.56; df = 2; p < 0.0001) in the undiluted supernatant treatment; the supernatant 10-fold dilution treatment was less effective (Table 2).

Adult Insect Bioassay
The inclusion of D. suzukii adult insects in tubes colonized by P. luminescens did not reduce their survival rate. Spreading the bacterial supernatant in the tube significantly increased insect survival 9 DAA (F = 3.19; df = 12; p = 0.04618) in comparison to any other treatment ( Figure 2): at that time point, mortality in the supernatant-treated samples was 65%, compared to approximately 95% in the controls. with Photorhabdus luminescens (10 6 cells mL −1 ) or P. luminescens culture supernatant over 9 days after application (DAA). Survival rates were compared to 10 mM MgSO 4 (control) or axenic Luria-Bertani (LB) medium. Different letters show significant differences among different treatments for each day, according to LSD test; p < 0.05.

Bacterial Persistence on Cherry Skin
P. luminescens maintained similar population levels (approx. 10 5 bacterial cells per fruit) over four days after inoculation. The bacterium was undetectable one week after inoculation (Figure 3).

Discussion
All of the currently known Photorhabdus spp. display entomopathogenic properties, which are expressed upon colonization of the target insect by nematode species forming a symbiosis with the bacterium. Isolates from human clinical cases, initially recognized as P. luminescens, were subsequently reassigned to the species P. asymbiotica basing on molecular taxonomy and pathogenetic potential [44]. Thus, P. luminescens is safe for agricultural operators and consumers. This study provides new information for the possible use of P. luminescens to control D. suzukii larvae and pupae. Indeed, the bacterium caused a high mortality of pre-immaginal stages through both oral and contact toxicity.
Interestingly, a significant efficacy was found also with very low bacteria concentration. In the oral bioassay, the bacterium increased the mortality of larvae and reduced the viability of the emerging adults even at a concentration of 3.5 × 10 2 cells mL −1 .
The larval bioassays showed significant insecticidal activities of P. luminescens against all D. suzukii development stages with mortality ranging between 86.7% and 100%. Parallel findings were obtained by Rahoo et al. [12] on Galleria mellonella larvae using P. luminescens at a concentration of 4 × 10 7 cells mL −1 and by da Silva et al. [38] on Aedes aegypti L. (Diptera: Culicidae) larvae. Within 24 h of being given the diet the treated larvae of D. suzukii stopped feeding and produced less faeces than usual. After cessation of feeding, larvae body colour changed from green to yellow or black (Figure 1a). However, larvae can survive for several days without feeding. This explains why the significant effect of P. luminescens bacteria began to appear 4 DAA. The bacteria seem to be more effective if they are swallowed; oral toxicity was about 37% vs. 10% contact toxicity. In our study, the bacterial supernatant caused 24% larval mortality, and 74% cumulative mortality of all D. suzukii development stages, confirming that toxins secreted by these bacteria have the potential to kill insects. The results agree with those previously obtained by Blackburn et al. [26] on larvae of Manduca sexta after they were fed a diet treated with 1 µg of Tca, a toxin complex secreted by P. luminescens. Moreover, the injection of 100 cells of P. luminescens (K −1 ) isolate was found to be lethal for G. mellonella larvae in 48 h. Likewise, the bacterial toxin secreted into the growth medium as the culture supernatant killed the insect in 48 h [29].
For the pupal bioassay, all concentrations were able to reduce pupae individuals compared to the control, achieving mortality ranges from 43.3-63.3%. A total mortality of 100% and 73.33% was caused by 3.5 × 10 8 cells mL −1 following pupal direct-spray and dipping bioassays, respectively. According to our visual observations, the treatment induced a high frequency of deformed adults (Figure 1c) emerging from pupae treated with P. luminescens cells (data not shown). Most of those deformed adults were unable to fully emerge and pursue their normal life stages. No deformed adults were noted in the control. Treatments (either with bacterial cells, or with toxin-containing supernatant) applied to adult insects had no insecticidal effect. This work did not investigate whether the offspring of treated adults is affected by P. luminescens. Neither the axenic bacterial medium (LB), nor the bacterial supernatant showed any toxic effects on D. suzukii adults. Since the bacteria growing medium share several components with the media used to rear insects, it may actually represent a nutrient source for the insects.
So far, the commercial use of these bacterial endosymbionts is related to the employment of the nematode species with which they are associated [20]. These results show the possibility of the use of P. luminescens in new IPM strategies to control D. suzukii. Bacterial suspension of P. luminescens has been used as a topical pesticide against a wide range of insects like aphids and mealybugs [45,46]. Considering that D. suzukii eggs are laid in the ripe fruits, the application should be done preventively at the beginning of oviposition (against eggs near surface of fruit). To apply the P. luminescens against D. suzukii larvae (first instar), further investigation is required to evaluate the ability of bacterial suspensions to penetrate through fruit skin. Topical application of biological agents on fruit is advantageous in terms of safety, potency and specificity compared to chemicals. These aspects, together with bio-degradability, may suggest a large and competitive market for P. luminescens-based biocontrol products. The high efficacy against larvae may also help in reducing the environmental population of the pest. In fact, bacteria can be sprayed also on wild host fruit plants that may act as pest reservoirs. Bacteria application could also be an economic alternative to clean harvest since the treatments of waste or non-harvested fruit may prevent the emergence of new pest generations. These kinds of application on fruit are particularly promising since, according to our data, a single bacterial application may maintain a sufficiently high population on fruit for at least 5 days. Moreover, P. luminescens acts against larvae at very low concentrations ranging between 10 2 and 10 3 cells mL −1 . Therefore, a natural spread of P. luminescens, up to an effective protective population, is likely to occur from treated plants to neighbouring ones.
Entomopathogenic bacteria can be used as stand-alone products for pest management in organic farming, their use in rotation or combination with chemicals is strongly encouraged to achieve full efficacy and eco-sustainability. Many studies have highlighted compatibility and synergistic effects of entomopathogenic bacteria and chemical substances [47,48]. For example, the conjugation of a pesticide that provides a rapid knock-down of adult population, followed by a bioinsecticide that targets larvae in fruit, may provide a successful strategy optimizing treatment efficacy and coverage. This work shows that P. luminescens is a promising tool for the containment of D. suzukii population. However, for its technological application in open field conditions, further studies are needed to assess the efficacy and formulation stability of products based on bacterial suspensions in different crops and environmental conditions.
In an alternative to topical applications of P. luminescens, other delivery methods should be investigated for the practical use in field conditions. For example, P. luminescens stabilized cells can be added to specific lures in order to obtain a horizontal spread of the pathogen from one individual to another. The use of stabilized, liophylized cells will also allow to recycle the purified cultural supernatant, which in our experiments showed a good degree of efficacy, for the application as a topical bioinsecticide.
To increase the efficacy of this bioinsecticide and make it comparable to chemical pesticides, it is essential to achieve a rapid control of the target pest. There is little information on conversion of bioagents or their by-products to enhance the toxicological properties. The efficacy of extracellular secretions of P. luminiscens could be improved by downsizing the particles through nanotechnology [49]. The results of Ramesh et al. [49] revealed that is possible to enhance the insecticidal property with a faster mortality of the target insect, attributed to higher penetration power of the nanoparticle, which acts as a small carrier of the toxin complex rapidly killing the insect.

Conclusions
Although agrochemicals are still the most convenient method for controlling pests, the use of alternative methods is increasing due to the negative side effects of pesticides. Even if a certain degree of prudence is recommended in the use of these biological pest control agents, entomopathogenic bacteria or their by-products could be a valid alternative or combined method to reduce the intensive use of xenobiotic chemicals, resulting in a significant environmental benefit.