Potential of entomopathogenic nematodes against the pupal stage of the apple maggot Rhagoletis pomonella (Walsh) (Diptera: Tephritidae)

Abstract The apple maggot, Rhagoletis pomonella (Walsh) (Diptera: Tephritidae), is considered a key pest of apples and is native to the eastern United States. The virulence of seven different species of entomopathogenic nematodes (EPN) was assessed against pupae of R. pomonella under laboratory conditions. Nematode species and strains included Steinernema carpocapsae (ALL strain), Steinernema feltiae (SN strain), Steinernema riobrave (355 strain), Steinernema glaseri (VS strain), Heterorhabditis bacteriophora (VS strain), Heterorhabditis indica (HOM1 strain), and Heterorhabditis megidis (UK211 strain). We conducted three bioassays: (i) short-term exposure cup bioassay (7 d), (ii) long-term cup bioassay (30 d), and (iii) pot bioassay (30 d). In the short-term exposure bioassay, all nematode strains (applied at 54 infective juvenile nematodes (IJs) cm−2) significantly reduced (range: 42.9-73.8%) insect survival relative to the control, but no differences were observed among the treatments. For the long-term exposure bioassay, using the same EPN application rate as the short exposure assay, all treatments reduced adult R. pomonella emergence compared with the control. Steinernema riobrave was the most virulent (28.3% survival), and S. glaseri and H. megidis were the least virulent (53.3% survival). In the pot experiment, S. riobrave and S. carpocapsae (applied at 27 IJs cm−2) had the highest virulence (23.3 and 31.7% survival of R. pomonella, respectively), while H. bacteriophora was the least effective (68.33% survival). Our results indicate that S. riobrave, S. carpocapsae, and S. feltiae have substantial potential to attack R. pomonella pupae, and their field application under the tree canopy (prior to adult emergence) in the spring when temperatures are conducive might be a good option for successful IPM of apple maggot fly.

Rhagoletis pomonella has one generation per year throughout its range (Dean and Chapman, 1973). Mature larvae exit the fruit, drop to the ground, and overwinter as pupae in the soil at 2 to 5 cm depth under the infested tree. The pupae overwinter in a stage of facultative diapause that is regulated by environmental factors (Prokopy, 1968). Adults emerge from puparia beneath infested abandoned or insufficiently managed host trees and immigrate into nearby commercial apple orchards, where they oviposit into fruit flesh (Boller and Prokopy, 1976;Reissig, 1979).
To achieve commercially acceptable levels of control, apple growers typically apply up to three broad-spectrum insecticide sprays to the entire orchard, beginning in early July and ending in late August (Prokopy et al., 1990;Bostanian et al., 1999). One alternative management tactic that reduces reliance on the widespread application of broadspectrum insecticides was developed in the form of an attract-and-kill system involving red spheres coated with Tangletrap in association with an attractive lure (Prokopy, 1968;Bostanian et al., 1999;Wright et al., 2012). Low-maintenance spheres combined with attractive visual cues, a toxicant, and a phagostimulant such as sucrose have provided effective control of R. pomonella in commercial apple orchards (Green and Wright, 2009;Wright et al., 2012;Morrison et al., 2016). All these management strategies target the aboveground adult stage of R. pomonella rather than the soil dwelling stages of R. pomonella.
Chemical insecticides may be effective but there are a variety of problems associated with the use of broad-spectrum chemical insecticides such as the development of pesticide resistance, resurgence, hazards to humans and the environment (Coppel and Mertins, 1997). Therefore, the development of biologically based alternatives to control fruit flies is warranted (Spinner et al., 2011). In terms of natural enemies, several species of parasitoids have been found to attack the R. pomonella but they are not effective under commercial orchard situations (Feder, 1995). AliNiazee (1985) evaluated two parasitoids Opius lectoides Gahan and O. downesi against R. pomonella in apple and found only 2% parasitism rate compared to Rhagoletis zephyria Snow for which 60% parasitism was observed in hawthorn fruit. Entomopathogenic nematodes may be another alternative approach for the biological control of R. pomonella.
Entomopathogenic nematodes (EPNs) from genera Steinernema and Heterorhabditis have the ability to infect and kill insect pests and they are naturally found in all types of agricultural and natural soils (Grewal et al., 2005). Entomopathogenic nematodes are asso ciated with symbiotic bacteria, i.e. Xenorhabdus spp. bacteria are associated with Steinernema spp. and Photorhabdus spp. are associated with Heterorhabditis spp. Entomopathogenic nematodes enter the host via the mouth and anus (Salame and Glazer, 2000), spiracle (Triggiani and Poinar, 1976), and directly through the softer portion of insect integument (Koppenhofer et al., 2000). Upon entrance into the host, infective juvenile nematodes (IJs) release their associated bacteria; host death occurs within 24 to 48 hr due to septicemia or toxemia (Shapiro-Ilan et al., 2018). EPNs are extensively used to combat root feeding insects (Johnson and Murray, 2008;Shapiro-Ilan et al., 2017).

Entomopathogenic nematodes
The nematodes used in this study were from the USDA International Culture Collection held in Byron, Georgia, USA. The seven EPNs species tested were Steinernema carpocapsae (ALL strain), Steinernema riobrave (355 strain), Steinernema feltiae (SN strain), Steinernema glaseri (VS strain), Heterorhabditis bacteriophora (VS strain), Heterorhabditis indica (HOM1 strain), and Heterorhabditis megidis (UK211 strain). The nematodes were cultured in vivo on the last instar of Galleria mellonella L. (Lepidoptera: Pyralidae) and IJs were collected using the White trap method (Shapiro-Ilan et al., 2016). G. mellonella larvae were obtained from Vanderhorst Wholesale, Inc. (St. Mary's, Ohio). The EPNs were stored in aqueous suspensions in 250 ml tissue culture flasks at 14°C. Nematodes were stored for less than two weeks before using them in experiments.

Rhagoletis pomonella colony
Pupae of R. pomonella from a non-diapausing strain were supplied by United States Department of Agriculture, Agriculture Research Service (USDA-ARS), Kearneysville, West Virginia. Briefly, mated females kept in a colony room at 16:8 L:D and 25°C were provided organic 'Red Delicious' apples as an oviposition substrate. Apples were exposed to adults for 3 to 4 days, then removed and suspended over trays of moistened sand for four weeks at 16:8 L:D, 24°C and 45% RH. Pupae were removed from sand using water and agitation to float them to the surface of trays. Removed pupae were placed in small plastic cups with tissue to absorb excess moisture and shipped overnight within two days of removal from sand.

Short-term exposure bioassay
This experiment as well as the others described below were organized as completely randomized designs (CRD) and conducted at USDA-ARS research station in Byron, GA. Bioassay procedures were based on prior EPN virulence assays (Shapiro-Ilan et al., 2002). The experimental arena consisted of a lidded 30 ml plastic cup filled with 10 g of autoclaved sand with 0% soil moisture content. Each cup contained one pupa (recently pupated) at the bottom of the cup underneath the sand. In total, 1 ml of each nematode suspension containing 500 IJs (54 IJs cm −2 ) was applied to the top of sand via pipette, resulting in a final moisture content of 10% in each cup. The control group received the same amount of distilled water without IJs. Thus, the experiment had eight treatments consisting of seven nematode species and one negative control treatment. Each treatment was replicated three times. Lidded cups were placed on trays and bagged with damp paper towels to help retain moisture, and incubated at 25°C. Seven days after treatment application, pupae were examined under a stereomicroscope for infection by the presence of nematode inside the pupae. The experiment was conducted twice, resulting in a total of six replicates per treatment and each replicate had seven cups with (so a total of 42 pupae per treatment).

Long-term exposure bioassay
The experimental approach was identical to the short exposure bioassay except for its duration. After inoculating nematodes to the cups and placing them in the incubator, we observed the cups daily for adult emergence until 30 days post treatment application. Adults that successfully emerged were considered to have survived the nematode treatment. All other experimental parameters were the same as described above except we used 10 pupae per replicate (based on the availability of insects).

Pot bioassay
Five nematode species, S. riobrave, S. carpocapsae, S. feltiae, H. indica, and H. bacteriophora were selected for the pot assay based on their virulence performance in the prior experiments (S. glaseri and H. megidis were eliminated). Plastic pots (10.16 cm diameter) were filled with non-sterile oven-dried soil (200 g) (0% moisture). The soil was a loamy sand (84% sand, 10% silt, 6% clay; 2.8% organic matter; pH 6.1). Approximately, 27 ml of distilled water was added and mixed manually for equal distribution of water throughout the soil. After mixing, 10 pupae were loosely buried inside the soil (3 cm deep). Each nematode species was applied at 1,377 IJs ml −1 , which is equivalent to 27 IJs cm −2 to the top of soil throughout the pot (with final moisture content 14%). The negative control received the same amount of distilled water (1 ml) without IJs. A 100-mm Petri dish cover lined with yellow sticky trap was placed on the top of each pot. Pots were put onto a plastic tray and bagged with damp paper towel to help retain moisture and incubated at 25°C. We observed the pots daily and monitored adult emergence until 30 days post treatment. Successfully emerging adults were considered as survived individuals. There were three replicate pots per treatment and control. The experiment was conducted twice in time.

Statistical analysis
Treatment effects were analyzed with analysis of variance (ANOVA). If the ANOVA detected a significant difference (P ≤ 0.05), then treatment differences were elucidated through Tukey's test (SAS Version 9.4, 2002). Data from repeated experiments (trials) were combined for analysis when the treatment×trial interaction was not significant. Based on the inspection of residual plots, percentage data were arcsine transformed prior to analysis (Southwood, 1978;Steel and Torrie, 1980;SAS, 2002). Nontransformed means are presented in the Results section and associated figures.

Short-term exposure bioassay
In the short exposure bioassay, no significant interaction (P = 0.8427) was detected between the treatment and trial effects, and so data from the two trials were combined. All seven nematode species significantly reduced pupal survival compared with the control group (F = 3.87; df = 15.47; P = 0.0006, Fig. 1). The nematode treatments reduced pupal survival at a similar level, with no significant differences detected among them. Pupal survival among the treatments ranged from 42.85% (for S. carpocapsae) to 73.8% (for H. bacteriophora) (Fig. 1).

Long-term exposure bioassay
In the long exposure bioassay, no significant interaction (P = 0.9618) was detected between the treatment and trial effects and so data from the two trials were combined. Similar to the short exposure assays, all nematode species induced lower adult emergence compared with the control (F = 4.68; df = 15.47; P = 0.0001, Fig. 2). Steinernema riobrave was the most virulent EPN, resulting in the lowest adult emergence (28.3%), which was significantly lower than those treated with S. glaseri and H. megidis (53.3%). Steinernema carpocapsae, S. feltiae, H. indica, and H. bacteriophora had intermediate effects, which were not significantly different from any treatment other than control (Fig. 2).

Pot experiment
No significant interaction (P = 0.6224) was detected between the treatment and trial effects and so data from the two trials were combined. All five EPN species significantly reduced R. pomonella adult  Fig. 3). Numerically, the lowest R. pomonella adult emergence was observed in S. riobrave (23.3%) and S. carpocapsae. In contrast, the maximum adult emergence was recorded in H. bacteriophora (68.3%) and H. indica (56.7%) (Fig. 3). Steinernema riobrave and S. carpocapsae had similarly high levels of virulence, which were significantly stronger (with lower adult emergence) than H. indica and H. bacteriophora but not different from S. feltiae. S. feltiae caused lower adult emergence than H. bacteriophora only (Fig. 3).

Discussion
In the current study, we tested seven species of entomopathogenic nematodes against pupae of R. pomonella in a short-term and long-term exposure assays and also in a soil pot experiment. Although no differences in nematode virulence were observed in the short-term assay, the experiment demonstrated that all EPN species are virulent to the pupae even when the duration of exposure is relatively short. When pupae were exposed to the nematodes for a longer period, all nematodes were separated from the control but S. riobrave was the only one that separated from other species showing higher virulence than two of the heterorhabditids. In the pot test, S. riobrave and S. carpocapsae were more effective than other treatments (though not statistically separated from S. feltiae). Thus, we conclude that S. riobrave, S. carpocapsae, and S. feltiae show the greatest promise for control of R. pomonella.
In the pot experiment we used the non-sterile soil, which is similar to soil that would occur under field conditions. Under these conditions, we observed substantial virulence of EPNs, especially S. riobrave and S. carpocapsae, which caused the lowest survival of R. pomonella. During a preliminary pot bioassay, we compared R. pomonella emergence among three EPN species and a control in sterile vs. non-sterile soil (four replicates each) and we did not detect any differences in EPN virulence (Usman et al., unpublished data). Moreover, in the pot experiment we placed pupae at a depth of 3 cm, which mimics the depth of pupae found under field conditions (https://apples.extension.org/apple-maggot/; Hulthen and Clarke, 2006). Therefore, we speculate that the virulence observed in our bioassays may be predictive of what we may observe in the field. Nonetheless, we realize some other biotic and abiotic factors that were not present in our study would be present in an agricultural setting (Shapiro-Ilan et al., 2017, 2018. Similar to our study, there are various reports of pupal susceptibility among fruit flies exposed to EPN species (Stark and Lacey, 1999;Hernández, 2003;  Barbosa-Negrisoli et al., 2009;Cuthbertson and Audsley, 2016;Heve et al., 2017). For example, Stark and Lacey (1999) reported that S. carpocapsae (72.5 and 83.3%), S. feltiae (70.3 and 73.3%), H. bacteriophora (62.5 and 60.0%), and S. riobrave (40.0 and 40.3%) caused pupal mortality in two experiments when R. indifferens was exposed during the larval stage. Also, in agreement with our study, S. riobrave was reported to cause high levels of pupal mortality in another fruit fly A. fraterculus (Barbosa-Negrisoli et al., 2009). In contrast, other studies did not observe pupal infection among different species of fruit flies when exposed to various EPNs strains (Yee and Lacey, 2003;Malan and Manrakhan, 2009;Langford et al., 2014;Garriga et al., 2020). For example, Yee and Lacey (2003) did not observe any pupal mortality in R. indifferens when treated with S. carpocapsae, S. interdium, and S. feltiae. The lack of infection in Yee and Lacey's (2003) study may have been due to age of pupae used; they were later in development compared to Stark and Lacey (1999) study. We conducted our experiments on pupae from a continuous, non-diapausing colony. Thus, additional research is needed to verify that similar results would be obtained in wild populations.
In this study, we selected pupae of R. pomonella as the target stage, because the pupae undergo longer time in soil compared to the larval stage and thus are a more realistic option in practical applications. In this study we used a non-diapausing strain so additional research may be required to determine if any differences in EPN virulence exist with other wild strains. Late third-instar tephritid larvae only last for a few hours depending upon the species before pupating (Özdem and Kılınçer, 2008;Sirjani et al., 2009;Kamali et al., 2013). Under field conditions, to infect the larvae, EPNs must be present in the soil before the larvae drop from fruit or at least close to that timing (thus there is a narrow window of exposure). A number of studies indicated that when targeting the larval stage fly mortality actually occurred during pupal stage due to very short duration of late third instars (Lindegren and Vail, 1986;Karagoz et al., 2009;Sirjani et al., 2009;Kamali et al., 2013;Heve et al., 2017). Also, another benefit of using pupae as the target stage is that R. pomonella overwinters in the soil providing more exposure time to EPNs. EPNs can potentially be applied to the soil in the spring when pupae are present and soil temperatures are conducive to infection (Shapiro-Ilan et al., 2017).
Entomopathogenic nematodes enter the host via mouth and anal opening (Salame and Glazer, 2000), spiracle (Triggiani and Poinar, 1976), and directly through the softer portion of insect integument (Koppenhofer et al., 2000). Several studies have described possible mechanisms of infection in the pupal stage. Fraenkel and Bhaskaran, (1973) observed EPNs to be pathogenic to pupae due to their entrance via intersegmental membranes prior to completion of cuticle development. In addition, according to Barbosa-Negrisoli et al. (2009), soft portion of integument might be responsible for A. fraterculus pupal susceptibility. Along these lines, we exposed early stage pupae to EPNs in our study, which probably had a relatively soft integument. In another study, the mouth, anus, and spiracles were found to be still open during the pupal stage and hence offer several options for EPN entry routes (Minas et al., 2016). Heve et al. (2017) stated that larger and longer steinernematids such as Steinernema diaprepesi, S. glaseri and others in the S. glaseri group are more virulent to pupae of A. suspensa due to the presence of more fat reserves in the nematode that allow for prolonged exposure to the pupae. However, this premise is not supported in our results because the larger nematodes (S. glaseri, H. megidis, and S. feltiae) were not the most virulent.
Entomopathogenic nematodes exhibit different foraging behaviors that lie on a continuum between ambushers and cruisers (Grewal et al., 1994;Lewis et al., 2006). Ambush foraging nematodes tend to remain near the soil surface, nictate, and attack passing insects. Cruiser behavior consists of active searching for the host in response to various volatile cues. A number of EPNs display foraging strategies that are intermediate between ambush and cruiser types. The nematodes that we observed to be the most virulent to R. pomonella can be classified as ambusher (S. carpocapsae) or intermediate foragers (S. riobrave and S. feltiae) (Lewis et al., 2006). The cruiser-type EPNs that we tested (the heterorhabditids and S. glaseri) did not perform as well. Nonetheless, it is not clear that foraging strategy played a major role in causing differential virulence in our study. In our relatively small arenas, foraging behavior may not have been critical to successful infection; field testing will be needed to confirm any trends in EPN foraging behavior as a factor in biocontrol efficacy.
Our results indicate that EPNs such as S. riobrave, S. carpocapsae, and S. feltiae have significant potential to suppress R. pomonella pupal populations. A possible approach would be to apply the nematodes under the tree canopy in spring when the soil temperature is conducive, before pupae emerge into adults in the summer. Rhagoletis pomonella normally pupate in the fall and remain inside the soil until late June or July, thus providing sufficient time for EPNs to infect. Expanding on the soil application concept, it may be possible develop a trap-tree approach using EPNs against R. pomonella similar to what has been developed for the plum curculio, Conotrachelus nenuphar (Herbst.) (Coleoptera: Curculionidae). This novel attract-andkill approach calls for baiting the branches of several perimeter-row trees with a synergistic lure, which results in aggregations of adult C. nenuphar on those trap trees, and then confining insecticide applications to those trees only (Leskey et al., 2008). EPNs are then applied to the soil underneath the canopies of trap trees, which are known to concentrate injury by C. nenuphar, to control resulting ground-dwelling stages (Leskey et al., 2008;Shapiro-Ilan et al., 2013). Before attempting such field applications, the impact of climate conditions such as temperature and soil moisture on EPN efficacy and post-application persistence should be evaluated so that the optimum conditions for infection are identified (Shapiro et al., 1999(Shapiro et al., , 2000. Additional research will be required to assess the efficacy of EPNs under semi-field pot experiments and field cages so that a successful IPM program for R. pomonella can be developed.