The second instar P. japonica grubs were obtained by caging the adult beetles in PVC cylinders (20 cm diam. × 15 cm tall) dug into a turfgrass lawn following procedures described by Klein et al. [4]. The beetles were caught early in the summer using trécé catch can traps (Trécé Inc., Adair, OK, USA) containing a feeding lure. When the grubs had developed to approximately 50% first and 50% second instars, the area within each cylinder was dug and second instar grubs were collected from the rhizospere of turfgrass. All the collected grubs were then stored in plastic trays containing field soil at 10 °C until used in the assays. The third instar C. borealis grubs used in this study were collected from the naturally infested turfgrass areas on different golf courses in Wayne County, Ohio and stored as described above until used.

Inbred lines used in this study were established from the GPS11 strain of H. bacteriophora, which was originally isolated from a single larva of C. borealis in 1998 at Atwood Lake Golf course, Dellroy, Ohio, USA [5]. Each last instar Galleria mellonella larva was exposed to a single IJ in the one-on-one sand-well method as previously described [12]. As H. bacteriophora IJs always develop into self-reproducing hermaphrodites [13], it is possible to establish inbred lines from single IJs. Each infected insect cadaver was then placed on a separate White trap [14] and next generation IJs were collected from the cadavers as individual inbred lines. This procedure was repeated 10 times (≈20 to 30 generations), resulting in a total of 35 inbred lines [11]. Out of these, seven inbred lines designated as A2, A6, A7, A8, A12, A18 and A21 were selected for this study as they exhibited superior IJ longevity, heat and UV tolerance, and virulence against the wax moth G. mellonella larvae as determined previously [11]. All the lines were stored in tissue culture flasks at 10 °C until used in the tests. Before the tests, all the lines were recycled once through the wax moth G. mellonella larvae at 25 °C [15]. Similarly, cryopreserved IJs of GPS11 (parent strain) were removed from liquid nitrogen storage and recycled once through G. mellonella larvae at 25 °C. Only IJs emerging from host cadavers within the first 2–5 days were used in all tests.

Separate bioassays were conducted for P. japonica and C. borealis to compare the virulence of inbred lines and the parental strain GPS11 of H. bacteriophora. The virulence of nematodes was tested in 30 mL plastic cups containing 29 g autoclaved Wooster silt loam soil. The soil was first passed through a 4 mm mesh sieve, and then transferred into each cup, and soil moisture was adjusted to 85% field capacity by adding the required amount of water. One P. japonica or one C. borealis was placed on the surface of the soil in each cup, lids were replaced, and the grubs were allowed to acclimate overnight at room temperature 25 °C. The grubs that did not burrow into the soil within 12 h were replaced. Then 100 µL nematode suspension containing 100 IJs was added to the surface of the soil in each cup. As a control treatment, each cup received 100 µL of water without nematodes. All the cups were arranged in a randomized block design with four replications and each replication contained 10 cups (n = 10 grubs). All cups were incubated at 25 °C. Both bioassays were repeated using new batches of grubs and freshly cultured IJs. Grub mortality was recorded twice at 1 and 2 weeks after treatment (WAT). Percentage mortality data were corrected using Abbott’s formula [16]. The corrected percent mortality data from both bioassays were pooled, as there were no differences in the two tests based on one-way ANOVA. The corrected percentage mortality data were transformed to arcsine and nematode penetration data were normalized by transforming to log+1 prior to ANOVA. Significant differences between treatments were determined using Tukey’s test at P < 0.05.

Two bioassays, one on P. japonica and one on C. borealis, were conducted to compare the penetration, encapsulation escape and survival capacities of inbred lines that showed the highest and lowest virulence against both grub species. The parental strain GPS11 was also included in the tests. Tests were conducted following the previously described method [5]. Briefly, 30 g autoclaved fine sand was transferred into 30 mL plastic cups and the moisture was adjusted to 8% by adding the required amount of water. The sand was used for its ease in being washing off the grub body as described below. Then, a 100 µL suspension, containing 200 IJs of A2, A12 or GPS11 was added to the surface of sand in each cup and the lid was replaced. As a control treatment, each cup received 100 µL of water without nematodes. A total of 8 such cups were prepared for each treatment and each cup was considered as a separate replication. One actively moving P. japonica or C. borealis grub was introduced in the center of each cup by making a hole in the sand with a glass rod, and lids were replaced. All the cups were then arranged in a randomized block design at 25 °C. After 24 h of incubation, each grub was removed from the cup and washed in a 50 mL beaker containing 10 mL tap water to remove any sand particles and nematodes adhering to the body. Only 24 h incubation period was used to capture initial host response against the leading nematode penetrants. Each grub was dissected in a 5 cm diameter petri dish containing 10 mL of tap water. The grub was decapitated, the ventral tip of the abdomen was partially excised and a ventral cut was made along the body. The digestive system was then removed. The number of nematodes penetrated was recorded as numbers of dead, alive and encapsulated nematodes using a stereomicroscope. The dark-colored and blood cell-attached nematodes were recorded as encapsulated. Only the P. japonica bioassay was repeated using a new batch of grubs and freshly cultured nematodes because of the lack of availability of C. borealis grubs. The penetration data for P. japonica from both bioassays were pooled, as there were no differences in the two tests based on one-way ANOVA. The percentage data were transformed to arcsine and normalized by transforming to log+1 prior to ANOVA. Significant differences between treatments were determined using Tukey’s test at P < 0.05.

We compared the interactions of two inbred lines A2, A12 and the parent strain GPS11 with immune response of P. japonica grubs over time. The IJs of lines A2, A12 and GPS11 were directly injected into the hemocoel of P. japonica grub from the foreleg. In order to obtain reproducible injection of nematodes, we used disposable sterile pipet tips with microcapillary (1–200 µL). The pipet tip was cut off using a sterile scissor to obtain a sharp pointed tip, and the uptake and release of nematodes was verified under the dissecting microscope by observing nematodes in the plastic capillary. Prior to injection, the grubs were washed with sterile tap water and cleaned with 70% ethanol at the injection site. About 20 IJs with 10 µL sterile water were injected laterally into the base of the first leg. Grubs in the control treatment were injected with hot water killed IJs. If the digestive system was damaged during injection, the grub was discarded. For each nematode line, there were three time intervals (6, 12, and 24 h post injection) at which injected grubs were dissected to observe the fate of the nematodes. The injected grubs were incubated at 25 °C. At each time interval, 10 P. japonica grubs for each treatment were decapitated, the ventral tip of the abdomen was partially excised and a ventral cut was made along the body. The digestive system was then removed. Water was flushed through the open hemocoel and examined for nematodes under a dissecting microscope. Other tissues were then carefully dissected and examined. The number of encapsulated living, non-encapsulated living and dead nematodes was recorded. The experiment was repeated once, and data indicated no differences in two experiments by one-way ANOVA and were pooled. The percentages of nematodes in three categories, relative to the total number injected, were then arcsine transformed and subjected to one-way ANOVA.

Overall there were significant differences in virulence of inbred lines against P. japonica (F = 3.04; df = 3, 15; P < 0.05) and C. borealis (F = 13.30; df = 7, 15; P < 0.05). At 1 week after treatment (WAT), line A2 caused more mortality of P. japonica than most of the other lines (A7, A8, A12, A18, A21) and even the parent strain GPS11 (Figure 1). Similarly, at 2 WAT, line A2 caused significantly higher P. japonica mortality than A7, A8, A12, A21 and the parent strain GPS11 (Figure 1). Comparatively, line A12 caused lowest P. japonica mortality (Figure 1). In case of C. borealis, the inbred lines A2, A8 and A21, and the parent strain caused significantly higher grub mortality than A6, A7 and A12 at 1 WAT (Figure 1). At 2 WAT, the inbred lines A2, A8, A18 and A21 caused significantly higher C. borealis mortality than A6 and A12 but it was not different from the parent strain (Figure 1). Taken together, line A2 caused significantly higher mortality than A12 in both P. japonica and C. borealis (Figure 1).

Overall penetration of H. bacteriophora inbred lines differed with white grub species (F = 1.85; df = 7, 5; P < 0.05). Total penetration ranged from 3 to 6% in P. japonica and 2.18 to 2.25% in C. borealis during the first 24 h, and specifically line A12 showed significantly higher penetration in P. japonica than C. borealis. While the total number of nematodes penetrated in P. japonica (F = 1.92; df = 15, 2; P < 0.05) and C. borealis (F = 0.37; df = 7, 5; P < 0.05) grubs did not differ significantly between the inbred lines, there were significant differences in the survival of nematodes between inbred lines in both grub species (Figure 2). Of the total number of nematodes that penetrated, significantly higher percentages were found alive in case of line A2 compared to A12 in both P. japonica and C. borealis (F = 2.20; df = 15, 5; P < 0.05 for P. japonica and F = 1.13; df = 7, 5; P < 0.05 for C. borealis) (Figure 2). The percentage of dead nematodes was significantly higher (F = 2.20; df = 15, 5; P < 0.05) in line A12 compared to A2 in case of P. japonica only. None of the lines differed from the parent strain GPS11 in either survival or encapsulation in either grub species.

Almost 100% of the nematodes in each treatment were encapsulated at 6 h post hemocoel-injection, and no non-encapsulated and dead nematodes were observed at this time. Nematode treatment influenced the percentages of nematodes encapsulated at 24 h (F = 10.83; df = 2, 57; P = 0.000), non-encapsulated living nematodes at both 12 h (F = 5.42; df = 2, 57; P = 0.007) and 24 h (F = 38.17; df = 2, 57; P = 0.000), and dead nematodes at 24 h (F = 12.89; df = 2, 57; P = 0.000) post infection. Compared to the A12 line, the A2 line had a higher percentage of non-encapsulated living, a lower percentage of dead at both 12 and 24 h post nematode injection, and lower percentage of encapsulated nematodes (including both alive and dead nematodes) at 24 h post injection. In contrast to line A2, line A12 had the least non-encapsulated living (P < 0.05) at both 12 and 24 h post injection, and the most dead (P < 0.05) at 24 h post injection IJs.