Orally Delivered Scorpion Antimicrobial Peptides Exhibit Activity against Pea Aphid (Acyrthosiphon pisum) and Its Bacterial Symbionts

Aphids are severe agricultural pests that damage crops by feeding on phloem sap and vectoring plant pathogens. Chemical insecticides provide an important aphid control strategy, but alternative and sustainable control measures are required to avoid rapidly emerging resistance, environmental contamination, and the risk to humans and beneficial organisms. Aphids are dependent on bacterial symbionts, which enable them to survive on phloem sap lacking essential nutrients, as well as conferring environmental stress tolerance and resistance to parasites. The evolution of aphids has been accompanied by the loss of many immunity-related genes, such as those encoding antibacterial peptides, which are prevalent in other insects, probably because any harm to the bacterial symbionts would inevitably affect the aphids themselves. This suggests that antimicrobial peptides (AMPs) could replace or at least complement conventional insecticides for aphid control. We fed the pea aphids (Acyrthosiphon pisum) with AMPs from the venom glands of scorpions. The AMPs reduced aphid survival, delayed their reproduction, displayed in vitro activity against aphid bacterial symbionts, and reduced the number of symbionts in vivo. Remarkably, we found that some of the scorpion AMPs compromised the aphid bacteriome, a specialized organ that harbours bacterial symbionts. Our data suggest that scorpion AMPs holds the potential to be developed as bio-insecticides, and are promising candidates for the engineering of aphid-resistant crops.


Introduction
Aphids are among the most destructive agricultural pests, causing direct damage to crops by feeding on phloem, as well as indirect losses by transmitting viruses [1]. Aphids are also biological models for the investigation of insect-plant interactions and symbiosis [2]. Buchnera aphidicola is an obligate bacterial symbiont of aphids, and is exclusively localized in a specialized structure known as bacteriome, which consists of bacteriocytes. This species has coevolved with aphids to provide them with essential amino acids that are not supplied in sufficient quantities by the sugar-rich phloem sap on which aphids feed [3,4]. Aphids also frequently host one or more secondary bacterial symbionts, Uy234, D5, D10, and D11) were effective only at the medium (250 µg/mL) and high (500 µg/mL) concentrations, and UyCT1 and Um4 were only effective at the highest concentration.
We used the insecticide imidacloprid and the antibiotic rifampicin as controls to gauge the effectiveness of AMP treatments, although their modes of action differ from AMPs. Imidacloprid killed all the aphids in less than three days (survival rate 0%), whereas rifampicin did not significantly affect aphid survival as compared to the control AP3 diet (survival rate 92.5%). Table 1. Effect of antimicrobial peptides (AMPs) and control treatments after three days of feeding.

Treatment
Concentration (µg/mL) % Survival Significance Survival curves were constru AMPs in A. pisum ( Figure 1, Table   Insecticide Imidacloprid 5 0 **** We used the insecticide imidacloprid and the antibiotic rifampicin as controls to gauge the effectiveness of AMP treatments, although their modes of action differ from AMPs. Imidacloprid killed all the aphids in less than three days (survival rate 0%), whereas rifampicin did not significantly affect aphid survival as compared to the control AP3 diet (survival rate 92.5%). Survival curves were constructed to compare the insecticidal activity of scorpion and insect AMPs in A. pisum ( Figure 1, Table S1). Figure 1A shows the effect of the insect AMPs (500 µg/mL) Compared to control AP3 diet (survival = 90%); ns-not significant, p > 0.05; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. Survival curves were constructed to compare the insecticidal activity of scorpion and insect AMPs in A. pisum ( Figure 1, Table S1). Figure 1A shows the effect of the insect AMPs (500 µg/mL) compared with the control AP3 diet, imidacloprid (5 µg/mL) and scorpion AMP UyCT5 (500 µg/mL). The three insect AMPs (cecropin A, apidaecin and stomoxyn) had no significant effect (survival rate ≥75%) against A. pisum whereas UyCT5 was highly effective at the same concentration (survival rate <10%). Figure 1B compares three UyCT AMPs, which reduced aphid survival by 65-90%. UyCT5 was the most effective, killing 10% of the nymphs after the first day and more on the second and third days until only~10% of the aphids survived. Figure 1C compares the U. yaschenkoi AMPs Uy17, Uy192, Uy234 and the U. manicatus AMP Um4, revealing that Uy17 was the most potent. Figure 1D compares the D-peptides, indicating that D10 was the most effective, killing all aphids by the end of the third day.
These data show that some of the scorpion AMPs are comparable to imidacloprid in terms of potency, e.g., Uy17 and Uy234 at the highest concentration, and D10 at the medium and highest concentrations, resulting in 100% mortality ( Figure 1). Aphids that survived the three days of treatment were monitored for the following two weeks in order to detect any delayed effects of the AMP treatments. In most cases, there was no significant difference in survival as compared to the control AP3 diet. However, the survival rate continued to decline in the aphid groups fed on 50 µg/mL Um4, 50 and 500 µg/mL Uy234, and 500 µg/mL D5 (data not shown).
Toxins 2017, 9,261 4 of 16 compared with the control AP3 diet, imidacloprid (5 µg/mL) and scorpion AMP UyCT5 (500 µg/mL). The three insect AMPs (cecropin A, apidaecin and stomoxyn) had no significant effect (survival rate ≥75%) against A. pisum whereas UyCT5 was highly effective at the same concentration (survival rate <10%). Figure 1B compares three UyCT AMPs, which reduced aphid survival by 65-90%. UyCT5 was the most effective, killing 10% of the nymphs after the first day and more on the second and third days until only ~10% of the aphids survived. Figure 1C compares the U. yaschenkoi AMPs Uy17, Uy192, Uy234 and the U. manicatus AMP Um4, revealing that Uy17 was the most potent. Figure 1D compares the D-peptides, indicating that D10 was the most effective, killing all aphids by the end of the third day. These data show that some of the scorpion AMPs are comparable to imidacloprid in terms of potency, e.g., Uy17 and Uy234 at the highest concentration, and D10 at the medium and highest concentrations, resulting in 100% mortality ( Figure 1). Aphids that survived the three days of treatment were monitored for the following two weeks in order to detect any delayed effects of the AMP treatments. In most cases, there was no significant difference in survival as compared to the control AP3 diet. However, the survival rate continued to decline in the aphid groups fed on 50 µg/mL Um4, 50 and 500 µg/mL Uy234, and 500 µg/mL D5 (data not shown).  Table S1. The insecticide imidacloprid was used as a positive control (5 µg/mL). (A) Insect AMPs, 500 µg/mL. (B,C) Natural scorpion AMPs, 500 µg/mL. (D) Designed scorpion AMPs (D-peptides), 500 µg/mL. The most effective AMPs were UyCT5, Uy17, Uy192, and D10, causing ~90% mortality. Insect AMPs had no significant effect on aphid survival.

Effect of AMP Treatments on Aphid Reproduction
The effect of each AMP on A. pisum reproduction was determined by counting the number of offspring and recording the delay before reproduction in surviving aphids for two weeks after treatment ( Figure 2).
Most of the scorpion and insect AMP treatments affected the time to reproduction, in some cases even at the lowest tested concentration of 50 µg/mL, resulting in significant delays of several days as compared to the control AP3 diet (Figure 2A). The treatments that did not cause a significant reproductive delay were 500 µg/mL UyCT1, 500 µg/mL UyCT5, 500 µg/mL Uy192, 50 µg/mL Uy234, 500 µg/mL D3, and 500 µg/mL D5.
Most of the treatments did not affect the number of offspring regardless of the effect on survival ( Figure 2B). However, all three concentrations of cecropin A, the 500 µg/mL apidaecin treatment, and the 500 µg/mL Uy234 treatment had a significant impact on the number of offspring.
Rifampicin caused the most significant effect on reproduction, resulting in smaller and underdeveloped adults that produced hardly any offspring ( Figure 2B). However, rifampicin did not affect aphid survival after three days of feeding nor during the two weeks after treatment (data not shown).

Effect of AMP Treatments on Aphid Reproduction
The effect of each AMP on A. pisum reproduction was determined by counting the number of offspring and recording the delay before reproduction in surviving aphids for two weeks after treatment ( Figure 2).
Most of the scorpion and insect AMP treatments affected the time to reproduction, in some cases even at the lowest tested concentration of 50 µg/mL, resulting in significant delays of several days as compared to the control AP3 diet (Figure 2A). The treatments that did not cause a significant reproductive delay were 500 µg/mL UyCT1, 500 µg/mL UyCT5, 500 µg/mL Uy192, 50 µg/mL Uy234, 500 µg/mL D3, and 500 µg/mL D5.
Most of the treatments did not affect the number of offspring regardless of the effect on survival ( Figure 2B). However, all three concentrations of cecropin A, the 500 µg/mL apidaecin treatment, and the 500 µg/mL Uy234 treatment had a significant impact on the number of offspring.
Rifampicin caused the most significant effect on reproduction, resulting in smaller and underdeveloped adults that produced hardly any offspring ( Figure 2B). However, rifampicin did not affect aphid survival after three days of feeding nor during the two weeks after treatment (data not shown).

Effect of AMPs on Bacterial Growth In Vitro
The susceptibility of aphid symbionts to AMPs was tested using the only known cultivable strain for aphids: S. symbiotica CWBI-2.3 [29]. Most of the scorpion AMPs that affected aphids in the feeding experiments also showed in vitro activity against S. symbiotica CWBI-2.3, with MICs of 125-500 µg/mL corresponding to the in vivo range (Table 2). Interestingly, some scorpion AMPs that were active in the feeding assays showed no in vitro activity against S. symbiotica CWBI-2.3, even at the highest tested concentration of 500 µg/mL (Uy234, Um4, D10 and D11) suggesting that they affect the aphids without targeting these bacterial symbionts. Further research is required to determine the mode of action of such compounds. Insect AMPs were not active against S. symbiotica CWBI-2.3, even at 500 µg/mL.

qPCR-Based Quantification of Bacterial Symbionts in Treated Aphids
We used a quantitative PCR (qPCR) assay to determine the impact of scorpion AMPs on the population density of S. symbiotica and B. aphidicola in vivo. Two groups of samples were analysed: (i) after three days of AMP treatments in feeding chambers, and (ii) two weeks after treatment. These samples were compared to investigate whether S. symbiotica and B. aphidicola can recover from AMP exposure. As shown in Figure 3, many scorpion AMPs significantly reduced the density of S. symbiotica and B. aphidicola after three days as compared to the control AP3 diet: 50, and 250 µg/mL D3, 50, and 500 µg/mL D11, all three tested concentrations of UyCT3, and UyCT5, 250 µg/mL Uy192, and 500 µg/mL Um4. Furthermore, the density of symbionts was still significantly lower than the control level after two weeks in the groups treated with 50 µg/mL D3, 250, and 500 µg/mL D11, 250, and 500 µg/mL Um4, and all three concentrations of UyCT3 and UyCT5 (data not shown).
Rifampicin caused a significant reduction in the numbers of S. symbiotica and B. aphidicola after three days of exposure and two weeks after treatment, but none of the insect AMPs reduced the density of either symbiont. Indeed, both symbionts were slightly more abundant two weeks after the treatment with cecropin A. Negative control = control AP3 diet; positive control = 50 µg/mL rifampicin; AMP treatments = 250 µg/mL. Statistical significance indicated as follows: * p < 0.05, ** p < 0.01, **** p < 0.0001.

Localization of Bacterial Symbionts in A. pisum by Fluorescence In Situ Hybridization
Fluorescence in situ hybridization (FISH) was carried out with specific probes (Table S2) to establish the tissue distribution of S. symbiotica and B. aphidicola in aphids 24 and 48 h after exposure to the highest concentration of each AMP and the control treatments. In the negative control (AP3 diet) group, we found that B. aphidicola was exclusively localized in bacteriome of the nymphs, and its associated ovarioles ( Figure 4A,C), whereas S. symbiotica was detected in most tissues, including the gut, bacteriome, and ovarioles ( Figure 4B,C). The S. symbiotica signal remained visible in aphid tissues 24 h after the AMP treatments, and was prevalent in the gut ( Figure 4D). However, the signal could not be detected after 48 h, indicating that the S. symbiotica 16S rRNA had degraded by this point ( Figure 4E). We also found that treating aphids with the scorpion AMPs compromised the structure of the bacteriome ( Figure 4E). Rifampicin treatment also reduced the S. symbiotica signal after 48 h, whereas the signal remained strong 48 h after treatment with the three insect AMPs (data not shown).
In contrast to the results observed for S. symbiotica, neither the AMPs nor rifampicin reduced the intensity of the signal for B. aphidicola in aphid nymphs. However, the B. aphidicola signal was often detected in the siphunculi 48 h after treatment with D3, D5, D10, Um4, UyCT3, and UyCT5 ( Figure  4E), which might indicate that the bacteriome structure has been compromised. As expected, imidacloprid did not affect the localization of the bacterial symbionts, because it acts directly on the insect central nervous system.

Localization of Bacterial Symbionts in A. pisum by Fluorescence In Situ Hybridization
Fluorescence in situ hybridization (FISH) was carried out with specific probes (Table S2) to establish the tissue distribution of S. symbiotica and B. aphidicola in aphids 24 and 48 h after exposure to the highest concentration of each AMP and the control treatments. In the negative control (AP3 diet) group, we found that B. aphidicola was exclusively localized in bacteriome of the nymphs, and its associated ovarioles ( Figure 4A,C), whereas S. symbiotica was detected in most tissues, including the gut, bacteriome, and ovarioles ( Figure 4B,C). The S. symbiotica signal remained visible in aphid tissues 24 h after the AMP treatments, and was prevalent in the gut ( Figure 4D). However, the signal could not be detected after 48 h, indicating that the S. symbiotica 16S rRNA had degraded by this point ( Figure 4E). We also found that treating aphids with the scorpion AMPs compromised the structure of the bacteriome ( Figure 4E). Rifampicin treatment also reduced the S. symbiotica signal after 48 h, whereas the signal remained strong 48 h after treatment with the three insect AMPs (data not shown).
In contrast to the results observed for S. symbiotica, neither the AMPs nor rifampicin reduced the intensity of the signal for B. aphidicola in aphid nymphs. However, the B. aphidicola signal was often detected in the siphunculi 48 h after treatment with D3, D5, D10, Um4, UyCT3, and UyCT5 ( Figure 4E), which might indicate that the bacteriome structure has been compromised. As expected, imidacloprid did not affect the localization of the bacterial symbionts, because it acts directly on the insect central nervous system.

Discussion
Aphids are dependent on their association with bacterial symbionts, and antibiotics can therefore impair their fitness and fecundity [16,30]. The evolution of innate immunity in aphids has been accompanied by the loss of many genes encoding antibacterial peptides because their expression could damage bacterial symbionts [14]. This has led to a hypothesis in which engineered crops expressing AMPs could be used to target aphids via their bacterial symbionts [31,32]. Engineered pathogen-resistant crops already provide a sustainable strategy to counteract specific plant diseases. For example, the antifungal peptides gallerimycin from Galleria mellonella (Linnaeus) and metchnikowin from Drosophila melanogaster (Meigen) have been shown to confer fungal resistance in plants [33,34].
As previously stated, the efficacy of AMPs expressed in crops relies on the ability of orally ingested AMPs to function correctly even following exposure to digestive enzymes found in the aphid gut [18]. We therefore investigated whether feeding aphids with scorpion and insect AMPs can affect their survival and fecundity. We selected AMPs from two Australian scorpion species (U.

Discussion
Aphids are dependent on their association with bacterial symbionts, and antibiotics can therefore impair their fitness and fecundity [16,30]. The evolution of innate immunity in aphids has been accompanied by the loss of many genes encoding antibacterial peptides because their expression could damage bacterial symbionts [14]. This has led to a hypothesis in which engineered crops expressing AMPs could be used to target aphids via their bacterial symbionts [31,32]. Engineered pathogen-resistant crops already provide a sustainable strategy to counteract specific plant diseases. For example, the antifungal peptides gallerimycin from Galleria mellonella (Linnaeus) and metchnikowin from Drosophila melanogaster (Meigen) have been shown to confer fungal resistance in plants [33,34].
As previously stated, the efficacy of AMPs expressed in crops relies on the ability of orally ingested AMPs to function correctly even following exposure to digestive enzymes found in the aphid gut [18]. We therefore investigated whether feeding aphids with scorpion and insect AMPs can affect their survival and fecundity. We selected AMPs from two Australian scorpion species (U. yaschenkoi and U. manicatus) because the evolution of scorpions has involved the development of venom glands producing peptides and proteins that can efficiently kill insect prey [24,26]. Certain scorpion AMPs are also active against human pathogenic bacteria [21,28]. We used three insect AMPs (cecropin A, apidaecin and stomoxyn), as well as a synthetic insecticide (imidacloprid) and antibiotic (rifampicin) as controls to evaluate the scorpion AMPs.
Each of the scorpion AMPs we tested was active against A. pisum, affecting their survival and/or fecundity. UyCT3, UyCT5, and D3 were highly effective at all three tested concentrations, whereas UyCT1 and Um4 were effective only at the highest concentration (500 µg/mL). In contrast, the insect AMPs we tested had no effect on aphid survival, and only a minimal impact on reproduction (Figure 2). In addition, many of the tested scorpion and insect AMPs delayed reproduction, but only a few reduced the number of offspring (cecropin A, apidaecin and Uy234) (Figure 2). The impact of scorpion and insect AMPs on aphid reproduction is probably a non-specific consequence of AMP toxicity, which causes an overall decrease in the fitness of aphids, and thus impairs their reproductive ability.
One potential explanation for the differential activity of scorpion and insect AMPs against aphids and their bacterial symbionts is the origin and intrinsic characteristics of these peptides. Scorpion AMPs are short cationic amphipathic peptides that are produced in the venom gland [23]. They target cell membrane by a pore-forming mechanism resulting in the loss of electrolytes [35]. Their broad activity against bacteria, erythrocytes, and other mammalian cells has been attributed to their lack of selectivity. Their precise function in nature still remains unclear, but they may protect the telson (open end of the fifth metasomal segment) from bacterial infections and may also help neurotoxins reach their targets once the AMP has ruptured the cell membrane [19,36].
The insect AMPs used in this study are expressed in the haemolymph when the host insect is challenged by a pathogen [37][38][39][40]. These AMPs act selectively against the membranes of a wide range of human, animal, and plant bacterial pathogens, but they do not affect eukaryotic cells [41][42][43]. Insect AMPs usually disrupt bacterial membranes by forming pores, but the mechanism of apidaecin is different [35]. This proline-rich AMP not only breaches the bacterial membrane, but also binds intracellular targets. The ineffectiveness of insect AMPs in aphids may reflect their selective nature toward pathogens, whereas scorpion AMPs target different tissues, including the bacteriome, probably using the same lytic mode of action.
As well as assessing the impact of each AMP on aphid survival and fecundity, we evaluated their direct effect against both B. aphidicola and S. symbiotica. The CWBI-2.3 strain of S. symbiotica is the only aphid symbiont that can be cultivated under laboratory conditions [29]. This strain is a transitional form between a free-living bacterium and a host-dependent mutualistic symbiont, and is a close relative of the S. symbiotica strain found in the A. pisum population used in this study [44]. We were able to determine MICs for each AMP against S. symbiotica CWBI-2.3 in vitro. Most of the scorpion AMPs (UyCT1, UyCT3, UyCT5, Uy17, Uy192, D3, and D5) inhibited the growth of S. symbiotica CWBI-2.3 ( Table 2). We found that several of the scorpion AMPs that affected aphid performance in the feeding assays were also active against S. symbiotica in vitro and in vivo, whereas others (Uy234, Um4, D10, and D11) did not act directly against the symbiont but were nevertheless active against the aphids in feeding assays, suggesting an alternative mechanism of action or an alternative target.
We also investigated the effect of the AMPs by using qPCR and FISH to directly characterize the population density and localization of both B. aphidicola and S. symbiotica in aphid tissues. FISH analysis did not reveal any clear AMP-mediated effect on the abundance of intracellular B. aphidicola, but there was a remarkable reduction in the S. symbiotica population, which was more accessible to the AMPs due to its intracellular and extracellular localization ( Figure 4D,E). However, qPCR revealed a significant reduction in the density of both populations, confirming the antibacterial effect of the tested scorpion AMPs (Figure 3).
The compartmentalization of symbionts inside the bacteriome and specialized host-derived membranes is an evolutionary strategy to protect mutualistic symbionts from host innate immunity, including AMPs [45]. This special structure must be breached before AMPs can exert their antibacterial activity [46][47][48][49]. For these reasons, the selective insect AMPs were probably unable to reach the bacterial symbionts, whereas the non-selective scorpion AMPs were more likely to compromise the bacteriome, affecting both symbionts (Figures 3 and 4E).
The observed insecticidal and antibacterial activities of scorpion AMPs against A. pisum and its bacterial symbionts are supported by earlier research in which indolicidin, an AMP from bovine neutrophils, showed activity against the green peach aphid Myzus persicae (Sulzer) and also affected the bacteriome [50]. Furthermore, scorpion AMPs (UyCT3, UyCT5, Uy192, Um4, D11) and indolicin showed activity against Escherichia coli, which is closely related to B. aphidicola, providing further support for our observations [20,21,28,51,52].
In summary, we found that the scorpion AMPs UyCT3, UyCT5, and D3 were the most effective against aphids and their symbionts. These AMPs showed insecticidal activity at different concentrations and they clearly affected aphid survival and reproduction, but also significantly reduced the population size of both B. aphidicola and S. symbiotica. There is a growing interest in the development of bio-insecticides derived from the venom of arachnids that prey on insects [53][54][55][56]. The natural characteristics of scorpion AMPs make them attractive candidates for this purpose because they are short and linear, and therefore easy to synthesize at low costs. Scorpion AMPs are also suitable candidates for the engineering of aphid-resistant crops, although further research is required to determine whether there are any negative effects in the plants themselves and whether the scorpion AMPs confer a significant degree of protection against aphids when expressed in planta.

Aphids and the Detection of Bacterial Symbionts
A. pisum clone LL01 was reared on 2-3-week-old bean plants (Vicia faba var. minor) in a climate cabinet (KBWF 720, Binder GmbH, Tuttlingen, Germany) with a 16-h photoperiod and a day/night temperature of 24/18 • C, as previously described by [57]. Plants for experiments and aphid rearing were cultivated in a greenhouse at an average temperature of 20 • C under natural light, plus additional illumination (SONT Agro 400 W, Phillips, Eindhoven, The Netherlands) to maintain a 14-h photoperiod.
The A. pisum population was screened for the presence of bacterial symbionts, as previously described [58,59], with slight modifications. Total genomic DNA was isolated from individual aphids or pools of 10-20 aphids using the CTAB method [60]. Bacterial symbionts were detected by PCR using genus-specific primers to amplify 16S rRNA gene fragments (Table S2) [59,61]. The reaction volume was 25 µL, comprising of 4 µL DNA template (25 ng/µL), 10 µM of each primer (1 µL), 12.5 µL of GoTaq Green 2x Master Mix (Promega, Madison, WI, USA) and 6.5 µL nuclease-free water. PCR products were visualized by 1% agarose gel electrophoresis using SYBR Safe (Invitrogen, Darmstadt, Germany). Amplicons were eluted using the NucleoSpin ® Gel and PCR Clean-up kit (Macherey-Nagel, Düren, Germany), and sequenced for verification on a 3730xl DNA analyser (Macrogen Europe, Amsterdam, The Netherlands). Only B. aphidicola and S. symbiotica were detected in our aphid population and each individual harboured both bacterial symbionts (data not shown). The sequences were compared against NCBI databases using BLAST and deposited under accession numbers KX900450-KX900452 for S. symbiotica and KX910798-KX910801 for B. aphidicola [62].

Aphid Feeding with AMPs
A. pisum nymphs (48 h old) were fed for three days on an artificial AP3 diet in modified chambers [63,64]. The AP3 diet was mixed with the corresponding AMP or control treatment. Ten nymphs were placed in each chamber and five replicates were included per treatment. AMPs were tested at three different concentrations: 50, 250, and 500 µg/mL. Untreated aphids were fed on the control AP3 diet. Positive control treatments comprised aphids fed on the AP3 diet supplemented with the insecticide imidacloprid (5 µg/mL) or the antibiotic rifampicin (50 µg/mL) (Sigma-Aldrich, Taufkirchen, Germany) [30,64]. Imidacloprid is strongly hydrophobic, and was therefore prepared first as a highly concentrated stock (1000 µg/mL) in acetone and working solutions were diluted in the AP3 diet. The corresponding control (AP3 + acetone) was tested on the aphids, and survival was not affected when compared to AP3 diet alone or AP3 diet diluted with water (data not shown). Mortality was scored after 24, 48, and 72 h of feeding. Aphids that survived the three-day treatment were transferred to agar plates containing bean plant leaves and reared for another two weeks in order to determine the impact of the diets on survival and reproduction [65].

In Vitro Activity of Scorpion and Insect AMPs against Serratia Symbiotica CWBI-2.3
S. symbiotica strain CWBI-2.3, the only aphid symbiotic bacterium that can be cultivated in the laboratory, was purchased from the Leibniz Institute DSMZ (Braunschweig, Germany) and cultivated as recommended by the supplier. MICs were determined according to the CLSI guidelines using a broth microdilution assay in 96-well polypropylene microtiter plates. The bacteria were cultivated overnight at 28 • C using 535 medium (Tripticase soy broth) and diluted to 5 × 10 5 CFU/mL in broth. The AMPs were dissolved in water to a concentration of 4 mg/mL and a series of two-fold dilutions was prepared in 535 broths, ranging from 500 to 4 µg/mL. S. symbiotica CWBI-2.3 in an unmodified medium was used as a positive control, and blanks were prepared with medium only or with medium and water (the latter to exclude any possible negative effect of water on the bacteria). The bacteria were incubated for 18 h and the absorbance at 600 nm was recorded every 20 min. The MICs were defined as the lowest concentrations of AMP causing complete bacterial growth inhibition.

Relative Quantification of Bacterial Symbionts In Vivo
The density of the B. aphidicola and S. symbiotica populations in vivo was determined by qPCR. Genomic DNA was extracted from pools of five aphids, as previously described [60]. Three biological replicates were prepared per treatment. The primers used for the identification of bacterial symbionts and the reference genes are listed in Table S2 [66]. Amplifications were carried out using a StepOnePlus™ Real-Time PCR System (Applied Biosystems, Waltham, MA, USA). The reaction volume was 10 µL, comprising 2 µL of template DNA (25 ng/µL), 10 µM of each specific primer and 5 µL of SYBR Green PCR Master Mix (Applied Biosystems). Each reaction was heated to 95 • C for 10 min, followed by 40 cycles of 95 • C for 15 s and 60 • C for 60 s. Melting curve analysis was performed by increasing the temperature from 60 • C to 95 • C for 15 s, cooling to 60 • C for 60 s and heating to 95 • C for 15 s. The expression of each gene was tested in triplicate to ensure reproducibility. Relative abundance values for each symbiont were calculated by comparing the threshold cycle (Ct) of each target gene to that of the aphid ribosomal protein L32 gene [67] and efficiencies were calculated using LinReg PCR software.

Localization of Bacterial Symbionts In Vivo by FISH
FISH was carried out as previously described [68], with slight modifications. Treated A. pisum nymphs were fixed for three days in Carnoy's solution (6:3:1 chloroform:ethanol:glacial acetic acid) and then bleached in 6% H 2 O 2 in 96% ethanol for 1 week. After bleaching, samples were washed in 100% ethanol and then hybridized overnight in hybridization buffer (20 mM Tris-HCl pH 8.0, 0.9 M NaCl, 0.01% sodium dodecylsulfate, 30% formamide) containing 100 nM of each fluorescent probe and 500 nM DAPI. Different probes were used to label Buchnera (ApisP2a) and Serratia (SerratiaPA) as shown in Table S2 [15]. After hybridization, samples were rinsed three times with phosphate buffered saline containing 0.3% Triton X-100 and viewed under a Leica TCS SP8 confocal microscope (Leica Microsystems, Wetzlar, Germany). We analysed a minimum of 20 samples from each treatment. The specificity of detection was confirmed using controls with no probe and specimens were pre-treated with RNase.

Data Analysis
All data were analysed using SPSS v17.0 software (SPSS Inc., Chicago, IL, USA) and statistical significance was defined as p < 0.05. For mortality assessment, we used non-parametric survival analysis (Kaplan-Meier) and multiple pairwise comparisons among different groups were carried out using log-rank tests to assess efficiency. The total number of offspring, time to reproduction, and relative numbers of bacterial symbionts were analysed using the Wilcoxon ranked sum test for non-parametric data and a paired t-test for parametric data.