Behavioral and Developmental Responses of Habrobracon hebetor (Hymenoptera: Braconidae) to Larvae of Helicoverpa armigera (Lepidoptera: Noctuidae) Inoculated With Various Concentrations of Bacillus thuringiensis var. kurstaki (Bacillales: Bacillacae)

Abstract Bacillus thuringiensis Berliner subsp. kurstaki (Btk) and Habrobracon hebetor Say are both biological control agents of Helicoverpa armigera Hubner. The present study evaluated their compatibility for combined application against this pest by examining the acceptability of Btk-inoculated hosts for H. hebetor females and testing for negative life-history impacts on developing progeny. Second-instar H. armigera larvae fed for 72 h on potted chickpea plants treated with three concentrations of Btk (LC15, LC35, and LC70) and were then used in bioassays of parasitoid development and parasitism behavior. Survival of parasitoids was significantly reduced, and immature development prolonged, on hosts fed chickpea plants treated with LC35 and LC70 Btk, but not on plants treated with LC15 Btk. Parasitoids failed to discriminate against hosts treated with LC15 or LC35 Btk in choice tests, but attacked fewer hosts treated with LC70 Btk, paralyzing and parasitizing more healthy hosts, and laying more eggs on them. In contrast, a no-choice test revealed that more hosts treated with LC35 and LC70 Btk were paralyzed compared with control or LC15-treated hosts, but the numbers of hosts parasitized and eggs laid did not vary among Btk treatments. Thus, females required an experience with healthy hosts, as they had in the choice test, to discriminate against diseased ones. We conclude that H. hebetor and Btk are compatible for joint application against H. armigera, which could potentially improve biological control of this pest.

Bt toxins have the potential to affect natural enemy development by diminishing the growth and vigor of their hosts, but can also make the host more vulnerable to attack by other natural enemies (Mohan et al. 2008). However, there exists variance among parasitoid species in susceptibility to Bt toxins; immature survival of Meteorus pulchricornis Wesmael (Hymenoptera: Braconidae) reared on larvae of Spodoptera litura F. (Lepidoptera: Noctuidae) fed diet amended with the field rate of Btk was not significantly different from controls (Walker et al. 2007). Similarly, Chilcutt and Tabashnik (1999) reported that Btk treatment of Plutella xylostella (Lepidoptera: Plutellidae) larvae had no effect on oviposition preference of the endoparasitoid Cotesia plutellae Kurdjumov (Hymenoptera: Braconidae), and direct feeding of the pathogen to wasps in honey had no detrimental effects. However, prolonged feeding of host larvae on Btk-contaminated diet was detrimental to parasitoid survival because of premature host mortality. Therefore, the ability of parasitoids to discriminate against infected hosts can be an important factor affecting parasitoid fitness, biocontrol efficacy, and persistence in the field when these agents are used in combination (Wan et al. 2019).
Parasitoid fitness is intrinsically linked to host quality because the latter affects juvenile survivorship, fecundity, and the body size of emerging adults (Vinson andIwantsch 1980, Ghimire andPhillips 2014). Natural selection should favor parasitoids that maximize their fitness by distinguishing and avoiding unsuitable hosts (Jiang et al. 2014). The present study was conducted to quantify the negative effects of Btk on development and survival of H. hebetor when they parasitize larvae of H. armigera inoculated with various concentrations of Btk and to determine whether H. hebetor females are able to discriminate inoculated hosts from healthy ones. The results were expected to clarify the compatibility of these two biocontrol agents for joint application in integrated pest management programs for H. armigera.

Plant Culture
Three to five seeds of chickpea, Cicer arietinum L. (cv. Bivanij), were sown in each plastic pot (10 cm ht × 7 cm diam) filled with a mixture of soil (30%), coco peat (30%), peat moss (30%), and perlite (10%). The pots were held at a daytime temperature of 23 ± 3°C, lowered to 20 ± 3°C at night in a greenhouse at the Department of Horticulture, Ilam University. Plants were watered every 2-3 d, as required, and used in rearing and bioassays when they reached the six to eight leaf stages (8-12 d).

Insect Colonies
Insect rearing and all experiments were carried out in climatecontrolled chambers set to 28 ± 1°C, 60 ± 5% RH, and a 16:8 (L:D) photoperiod. An H. armigera colony was established by collecting larvae (ca. 250) from chickpea fields in Ilam Province, Iran, in the spring of 2018. Each larva was transferred to a ventilated plastic container (4 × 6 × 8 cm) and fed an artificial diet based on that of Twine (1971) with some modification: bean flour (205 g), yeast (35 g), wheat germ (30 g), sunflower oil (5 ml), formaldehyde 37% (2.5 ml), agar (14 g), ascorbic acid (3.5 g), methyl-parahydroxyl-benzoate (2.2 g), sorbic acid (1.1 g), and distilled water (700 ml). Larvae were provided with a piece of diet (ca. 1 cm 3 ) that was refreshed daily until they pupated. Emerged moths were transferred to a transparent Plexiglas oviposition chamber (20 × 30 × 30 cm), 30-40 per container, and provisioned daily with 10% honey solution provided on cotton balls in open Petri dishes (6 cm diameter) and allowed 4-5 d to mate. The top and walls of the chamber were covered with a fine mesh that served as an oviposition substrate. The mesh nets bearing eggs were collected daily and transferred to plastic bags held under the same physical conditions. After hatching, neonate larvae were transferred to potted chickpea plants (three or four per plant) and allowed to feed 4 d until they reached the second instar, whereupon they were either used in experiments or transferred to larval rearing containers (as above) to maintain the colony.
A laboratory culture of H. hebetor was maintained on late instars of Ephestia kuehniella (Zeller) obtained from the Plant Protection Bureau of Ilam Province. The parasitoid colony was established by collecting parasitized H. armigera larvae (ca. 60) or pupae (ca. 30) from chickpea fields in Ilam Province, Iran, in the spring of 2018. These were reared out under the same physical conditions as described above for H. armigera. Emerging wasps were transferred to plastic Petri dishes (9 cm diameter), fed with a 10% honey solution on strips of paper (1 × 3 cm). Generations were started by placing two male and two female wasps in each of a series of Petri dishes (as above) and providing each dish with 10-12 fifth-instar larvae of E. kuehniella daily. Adult wasps were introduced to clean Petri dishes with new hosts daily until both female wasps died. Parasitized host larvae were kept until emergence of adult wasps and the colony was reared for eight generations before use in the experiments. All experiments employed 3-d-old mated females that had each been provided with three late-instar larvae of E. kuehniella on each of two successive days to verify successful mating and oviposition prior to use.

Helicoverpa armigera Bioassay
A wettable powder formulation of Bacillus thuringiensis subsp. kurstaki (Belthirul, 32,000 IU/mg) was obtained from Probelte Pharma, SL (Madrid, Spain). Five concentrations of Btk were prepared (250, 625, 1,250, 2,000, and 2,500 ppm), and in each treatment, the foliage of potted chickpea plants was immersed for 10 s in one of the Btk suspensions (n = 90 plants for each concentration), then air-dried on a laboratory bench for 10 min; control plants were immersed in distilled water. Each treated chickpea plant was then placed in a ventilated plastic container (12 cm diameter × 12 cm height). Second-instar H. armigera larvae were transferred to Btktreated chickpea plants, one larva per plant. The basal container was covered with a second container (15 cm diameter × 13 cm height) ventilated by means of a fine mesh glued to an aperture cut in the bottom (7 cm diameter). Both containers had sloping sides, so that a larva-proof seal was obtained when the larger container was placed upside down over the smaller one (Allahyari et al. 2020b). The second-instar H. armigera larvae were then left to feed on the Btk-treated chickpea plants for 3 d under the same physical conditions as the insect colony. Larval mortality was recorded on day 3 post-inoculation, and lethal concentrations were determined using Probit analysis.

Parasitoid Bioassay
After determining the lethal concentrations of Btk in the first bioassay, second-instar larvae of H. armigera were starved for ca. 8 h and then transferred to chickpea plants treated with one of three concentrations of Btk (LC 15 , LC 35 , and LC 70 ), or distilled water as controls, one larva per plant, and allowed to feed for 72 h. In total, 85, 120, 250, and 60 larvae were fed plants treated with LC 15 , LC 35 , LC 70 , and water, respectively, in consideration of lower expected host survival at higher concentrations. Surviving larvae were isolated in plastic Petri dishes (6 cm diameter) and provisioned with artificial diet, as described above, for the remainder of their development. Experimental wasps were held as couples in Petri dishes for 2 d after emergence and provided E. kuehniella larvae to ensure successful mating. Next, two female wasps, each 3 d old, were transferred to each Petri dish containing a single Btk-treated or control larva of H. armigera for 24 h. After the wasps were removed, a single parasitoid egg was left on each larva and parasitized larvae were held under the standard experimental conditions until emergence of adult wasps. The duration of immature parasitoid development, numbers of pupae formed, and numbers of adults emerging were all recorded. After emergence, adults were transferred to Petri dishes (6 cm diameter) and provisioned with diluted honey (10%) on strips of paper (1 × 3 cm) every other day until death of the last wasp to estimate adult longevity.

Choice Test
Second-instar larvae of H. armigera were fed for 3 d on chickpea plants treated with one of three concentrations of Btk (LC 15 , LC 35 , and LC 70 ), or distilled water as a control, as described above. Treated larvae that survived each of three Btk concentrations, and healthy larvae (controls), were isolated in plastic Petri dishes (9 cm diameter), which had been divided into eight equal radial sections by means of a wire mesh to isolate a single larva in each section and prevent cannibalism among them (Fig. 1). The wire mesh was coarse enough to permit wasps to move easily through the partitions. Four treated and four healthy larvae were arranged alternately among the eight sections of each dish, and each was provisioned with a piece of artificial diet (ca. 1 cm 3 ). A single, 3-d-old female wasp was introduced to each Petri dish and allowed forage for 24 h, whereupon the wasp was removed and all paralyzed and parasitized hosts, as well total eggs laid, were recorded. If a larvae was still alive but showed no movement when stimulated by a fine brush, it was considered to be paralyzed and parasitized if it received at least one egg (all parasitized larvae were paralyzed, but not all paralyzed larvae were parasitized). Altogether, 60 female wasps were tested at each concentration, one female per replicate. Data for replicates in which the wasp was killed by the host larvae were discarded.

No-Choice Test
A no-choice test was conducted using the same procedures as the choice test, except that eight Btk-treated larvae from each of three concentrations (LC 15 , LC 35 , and LC 70 ) or eight healthy larvae (as control) were placed in each Petri dish (9 cm diameter) and a 3-d-old female wasp was introduced to each (n = 60) as described above. Data for replicates in which the wasp were killed by the host larvae during the experiment were discarded.

Statistical Analyses
Lethal concentrations of Btk were calculated by SPSS software package (SPSS 1998) using mortality data of the bioassay subjected to Probit analysis (Finney 1971). Parasitoid survival was analyzed by Kaplan-Meier (α = 0.05). Developmental times were compared by one-way analysis of variance (ANOVA; SPSS 1998), and Fisher's least significant difference (LSD) test was used to separate means (α = 0.05). Mean numbers of paralyzed and parasitized larvae and numbers of eggs laid in the choice test were compared by paired t-test (two tailed). In the nonchoice test, the data were analyzed by one-way ANOVA (SPSS 1998) followed by Fisher's LSD test to separate means (α = 0.05).

Helicoverpa armigera Bioassay
The lethality of Btk against second-instar larvae of H. armigera was measured after 72 h of continuous feeding on treated chickpea plants and increased significantly with the concentration applied (Table 1).

Choice Test
Female parasitoids failed to discriminate between healthy hosts and those treated with the LC 15 of Btk, paralyzing and parasitizing similar numbers of them, and laying similar numbers of eggs on them (Table 3). When LC 35 -treated hosts were available, females paralyzed more treated hosts than healthy ones, but parasitized equal numbers of them and laid similar numbers of eggs on them. When LC 70treated hosts were available, parasitoids paralyzed and parasitized more healthy hosts than treated ones and laid more eggs on them.

No-Choice Test
In a no-choice situation, similar numbers of hosts were paralyzed in the LC 15 treatment as in controls, but more hosts were paralyzed in the LC 35 and LC 70 treatments (Table 4). However, paralyzed hosts in LC 35 were not significantly different from those in LC 70 . There were no differences among treatments in the numbers of hosts parasitized or the numbers of eggs laid.

Discussion
Female parasitoids did not discriminate against second-instar H. armigera treated with the LC 15 concentration of Btk in either choice or no-choice tests, and similar numbers were paralyzed and parasitized as healthy ones. This would suggest that parasitoids did not detect any symptoms of Btk infection in these hosts, which is ostensibly characterized by sluggish movement and weight loss (Blumberg et al. 1997).
However, hosts treated with LC 35 Btk were paralyzed more often than healthy hosts in both choice and no-choice bioassays, suggesting that their infections rendered them more vulnerable to initial attacks by the parasitoid, without rendering their behavior sufficiently abnormal to cue wasp avoidance, even when they were presented with healthy alternatives. Host defensive responses represent a hazard to wasps, as these large host larvae are sometimes able to injure or kill attacking parasitoids (Mohan et al. 2008). Hosts treated with LC 70 were also paralyzed more often than controls in the no-choice test, but the reverse was true in the choice test, suggesting that disease symptoms in this treatment were pronounced enough that wasps were more likely to avoid them when given a chance to compare them to healthy hosts. Allahyari et al. (2020a) showed that H. armigera larvae inoculated with sublethal concentrations of a nucleopolyhedrovirus, HearNPV, were more often paralyzed by H. hebetor females than controls and inferred reduced defensive responses in infected hosts. Sometimes, Bt-infected hosts can be more susceptible to parasitism than untreated hosts, due to their slower development and smaller size (Mascarenhas andLuttrell 1997, Erb et al. 2001). However, H. hebetor females in the no-choice test parasitized similar numbers of LC 35 -treated and LC 70 -treated hosts as they did healthy ones, and laid equal numbers of eggs on each. Therefore, whatever infection-related factors increased their susceptibility to attack did not alter their acceptability for oviposition for these females.
Parasitoid females discriminated against hosts treated with the LC 70 Btk concentration in the choice test and paralyzed and parasitized fewer of them compared with healthy hosts. We conclude that these hosts inoculated with higher doses of Btk displayed sufficiently abnormal symptoms that their acceptability to parasitoids for oviposition was reduced. However, they were parasitized at similar rates and received similar numbers of eggs in the no-choice test. The different results of these two tests suggest that females require experiences with healthy hosts, as they had in the choice test, to discriminate disease symptoms in infected ones. Similarly, Akinkurolere et al. (2009) reported that female H. hebetor parasitized more less-preferred, early-instar larvae of Plodia interpunctella Hubner (Lepidoptera: Pyralidae) when they were offered in no-choice situations, having had no experience with larger ones, and Wan et al. (2019) found that Microplitis pallidipes Szepligeti (Hymenoptera: Braconidae) required prior experience with healthy larvae of Spodoptera exigua (Lepidoptera: Noctuidae) before they could discriminate against those with nucleopolyhedrosis virus (NPV) infections. Development and survival of H. hebetor larvae were both negatively affected when they developed on hosts subjected to the LC 70 and LC 35 Btk treatments. Although immature development was delayed by less than 1 d, the adult longevity of wasps was also reduced significantly. It has been argued that natural selection should favor the avoidance of entomopathogen-infected hosts, which should select for parasitoids that can distinguish them from healthy ones (Flexner et al. 1986, Jiang et al. 2014. For example, both the growth and immature survival of the solitary endoparasitoid C. chlorideae were reduced in a dosage-dependent manner, relative to controls, when their H. armigera larval hosts fed continuously on a diet contaminated with sublethal  concentrations of Btk (Mohan et al. 2008). Similarly, Rolim et al. (2020) reported negative impacts on P. elaeisis when this endoparasitoid developed in host larvae of Spodoptera frugiperda (Smith) (Lepidoptera: Noctuidae), many of which extended to the subsequent generation. These impacts included impaired foraging behavior and reduced reproduction and longevity that the authors inferred would reduce compatibility between these biocontrol agents. These various studies suggest that infection with Bt will reduce the suitability of lepidopteran hosts for parasitism, but provided the hosts survive long enough for completion of wasp development, the parasitoid can survive. Sedaratian et al. (2014) examined H. hebetor larvae developing on H. armigera larvae treated with sublethal concentrations of Btk and found negative effects on their development, survival, and reproductive rate, which was reflected in diminished life table parameters such as finite and intrinsic rates of increase. However, they did not assess wasp behavioral responses to inoculated hosts. Bt-infected hosts may be smaller and provide poorer nutrition for parasitoid growth and development (Romeis et al. 2006), so a failure to discriminate against infected hosts under field conditions could potentially impair parasitoid reproductive success and reduce fitness. Our behavioral observations indicate that H. hebetor has some ability to discriminate against infected hosts, but it depends on both the degree of infection and whether the wasp has an opportunity to compare them to healthy hosts. Given that foliar applications of Btk have relatively short persistence in the field, many pest larvae will probably consume sublethal doses and remain as suitable hosts for parasitism by H. hebetor. Any ability of wasps to discriminate against heavily infected hosts could aid biological control outcomes by shifting more parasitism toward uninfected and sublethally infected host larvae. Means within rows followed by the same letter were not significantly different among treatments (ANOVA followed by Fisher's LSD, α = 0.05). Treated host larvae were second instars that fed for 72 h on chickpea plants treated with Btk; healthy larvae were fed on plants treated with distilled water (paired t-test, two tailed). Table 4. Mean (± SE) numbers of Helicoverpa armigera larvae (out of n = 8 per female) paralyzed and parasitized by Habrobracon hebetor females (n = 60 per treatment) and numbers of total eggs laid on the healthy (control) or treated hosts in a no-choice test Furthermore, laboratory studies have shown that H. armigera has the ability to evolve resistance to Btk (Lu et al. 2004), and resistant populations have been detected in the field (Liu et al. 2010). This eventuality might be delayed or precluded if releases of H. hebetor reduce the survival of resistant genotypes. Allahyari et al. (2020b) found evidence of additivity between Btk and H. hebetor in controling H. armigera on chickpea in an earlier field study, which would support their effectiveness in joint application. Thus, even though Btk had some negative effects on H. hebetor under laboratory conditions, and the parasitoid failed to discriminate against hosts treated with low concentrations of Btk, we conclude that these two biocontrol agents are potentially compatible for combined application against H. armigera under field conditions.