Attack and Success of Native and Exotic Parasitoids on Eggs of Halyomorpha halys in Three Maryland Habitats

Egg parasitoids of the exotic invasive brown marmorated stink bug, Halyomorpha halys (Stål), were investigated using lab-reared fresh (live) and frozen (killed) lab-reared sentinel egg masses deployed for 72h on foliage in three habitats—woods, orchard, and soybean field—in Maryland, USA, in summer 2014. Four native hymenopteran species, Telenomus podisi Ashmead (Scelionidae), Trissolcus euschisti (Ashmead) and Tr. brochymenae Ashmead (Scelionidae), and Anastatus reduvii (Howard) (Eupelmidae), developed and emerged from H. halys eggs. One exotic parasitoid, Trissolcus japonicus (Ashmead), emerged, providing the first known occurrence of this species in North America. Native parasitoids emerged from frozen eggs significantly more often than from fresh eggs (89.3% of egg masses and 98.1% of individual eggs), whereas the exotic Tr. japonicus did not show a similar difference, strongly suggesting adaptation to H. halys as a host by Tr. japonicus but not by the native species. Parasitoids were habitat-specific: all three Trissolcus species were significantly more likely to occur in the woods habitat, whereas Te. podisi was found exclusively in the soybean field. Further investigations are required to elucidate evolving host-parasitoid relationships, habitat specificity, and non-target effects of Tr. japonicus over the expanded range of H. halys in North America.


Introduction
The brown marmorated stink bug, Halyomorpha halys (Hemiptera: Pentatomidae), is native to Asia and was first discovered in the United States in 2001 in Allentown, Pennsylvania [1]. This polyphagous pest has since spread to 41 states causing economic damage to a wide variety of both fruit and vegetable crops and posing a nuisance in the fall and winter as the adults overwinter in residential and commercial buildings. The range of H. halys is expanding in Europe and North America, and is a threat to crops world-wide [2][3][4][5][6]. Chemical pesticides have met with mixed success in controlling H. halys, and deployment of pheromone baited traps has so far served to monitor widely their numbers and spread [7,8], but not to manage populations.
A potential long-term and low-cost strategy for control of H. halys in North American environments is biological control using egg parasitoid wasps. Although several egg parasitoid species in the genus Trissolcus (Hymenoptera: Scelionidae) attack and successfully parasitize H. halys in Asia, currently egg parasitoids native to the United States parasitize H. halys with very limited success [9]. Instead, native parasitoids typically fail to develop and/or emerge from H. halys eggs which thereby act as a sink for native parasitoids in both North America [10] and Europe [11]. Beginning in 2007, the USDA-ARS Beneficial Insect Introduction Research Unit (ARS/BIIRU) brought species of Trissolcus from China, Korea and Japan to quarantine laboratories in the U.S. for evaluation as potential biological control agents of H. halys in North America. Trissolcus japonicus (Ashmead) and Trissolcus cultratus (Mayr) predominated in the Asian collections obtained from H. halys [12]. Prior to possible implementation of classical biological control (the introduction of exotic natural enemies to control exotic pest species), several researchers had begun investigating the ability of native parasitoids to successfully parasitize H. halys eggs by placing sentinel egg masses in the field (e.g., [9,10]). Unexpectedly, we found that the exotic Tr. japonicus was present in the United States and attacking H. halys eggs in the field, presumably as a result of an accidental introduction [13].
Sentinel egg masses of exotic hosts such as H. halys may be subject to low rates of successful parasitism not only because native parasitoids may be maladapted to exotic hosts, but also because of a lack of host finding cues available to female parasitoids. Parasitoids use host insect and plant volatiles as long-range host finding cues [14]. These cues are not available to the female parasitoids responding to sentinel egg masses laid by colony insects and then placed out into the habitat. Jones et al. [9] showed that naturally laid H. halys egg masses in orchards experienced higher rates of parasitism than sentinel egg masses placed in the same habitat, likely due to the increase in host finding cues available to the wasps.
The primary goal of our study was to quantify field rates of attack and successful parasitism of H. halys eggs, using frozen eggs [11] as described below, fresh eggs from a lab colony, and eggs naturally laid by field-caged H. halys females. Previous studies in the U.S. have deployed sentinel egg masses exclusively in managed crops such as orchards, vineyards, tree nursery and row crops [5,6]. We decided to include a wooded unmanaged habitat, along with two agroecosystems, to expand our knowledge of field parasitism of H. halys.
We tested the following hypotheses: (1) Frozen sentinel egg masses yield higher rates of successful parasitism than fresh sentinel or cage laid egg masses at all sites; (2) Egg masses laid inside of mesh cages by H. halys in the field should result in higher parasitism than comparable colony laid sentinel egg masses; (3) Parasitoid species are habitat specific; and (4) Native parasitoids only successfully parasitize, develop, and emerge from sentinel eggs that were dead prior to deployment. In addition, upon discovery of the exotic parasitoid Trissolcus japonicus, we further hypothesized that its successful emergence from its Asian host H. halys would exceed that of North American native parasitoids.

Materials and Methods
Field permits were not required because all research was conducted at sites on land within our USDA ARS facility in Beltsville, MD. Our studies did not involve endangered or protected species.
We conducted a sentinel egg mass study to assess parasitoids of the invasive pest, H. halys, in three field habitats during the summer of 2014 on the Beltsville Agricultural Research Center North Farm in Beltsville, MD. All H. halys adults and eggs used in this experiment came from our laboratory colony, which was established by Jeffrey Aldrich in 2007 with adults collected in Allentown, PA, and was reportedly supplemented annually by~20 adults field-collected in the Beltsville area until 2010. Insects in this colony were reared in growth chambers at 25°C, 40-60% humidity and a 16L:8D cycle. They were provided with certified organic green beans (MOM's organic market College Park, MD), hulled raw sunflower seed (Meyer Seed Co. Baltimore, MD), buckwheat seed (Meyer Seed Co. Baltimore, MD) and water ad libitum. Three different egg mass treatments were deployed weekly in three different habitat types.

Sentinel egg mass treatments
H. halys masses were deployed in three different ways: fresh from rearing, frozen, and cage-laid.
Fresh from rearing. Fresh egg masses ( 24-hours-old) laid by colony insects on paper towels were collected and pinned to the underside of leaves of various vegetation using a sewing pin at each of the three sites. An additional set of fresh egg masses were collected from our colony and reared in a growth chamber at 25°C, 40-60% humidity and a 16L:8D cycle as a control treatment.
Frozen. Fresh egg masses ( 24-hours-old) laid by colony insects on paper towels were collected and immediately frozen at -80°C for 2 minutes. The egg masses were then pinned to the underside of leaves of various vegetation using a sewing pin at each of the three sites.
Cage-laid. Two mesh cages (30.5 cm x 30.5 cm x 61 cm with 20x20 mesh aluminum screening; BioQuip, Rancho Dominguez, CA) were secured on the ground at each site with twine and stakes where they remained for the course of the experiment. Each cage contained 15 female and 5 male H. halys adults. Adults in cages were replenished as needed and were provided with a diet of green beans, hulled sunflower seed, and buckwheat seed (as in the laboratory rearing), and one potted bush bean plant (E-Z Pick Bush Bean, Johnny's Selected Seeds, Winslow, ME) all replaced weekly. At each site, one cage was placed in a shaded area and the other was placed in direct sunlight. Throughout the course of the experiment, the female H. halys adults laid eggs on the bush bean plants inside the cages. Parasitoids could enter though the mesh to find and parasitize the fresh cage-laid eggs.

Habitat characteristics
To gain insight into the effect of habitat on parasitoids of H. halys, we stationed each type of egg mass in each of three contrasting habitats: soybean field, apple orchard, and second-growth woods, on the North Farm of Beltsville Agricultural Research Station, Beltsville, Maryland.

Egg deployment, collection, and evaluation
We deployed each egg mass treatment on 8 dates in 2014 (July 21 and 28, August 4, 18, and 25, and September 2, 8, and 15) in each of the three habitats, all within~20m of one another at each site. All egg masses were exposed for 72 hours in the field, at which time they were returned to the laboratory where they were held for emergence to determine survival and patterns of parasitism. Cage-laid eggs were collected from the bush bean plants after field exposure by cutting a section of leaf surrounding the egg mass. All egg masses were reared in plastic zip top bags in a growth chamber at 25°C, 40-60% humidity and a 16L:8D cycle, and checked circa daily for emergence of H. halys or parasitoids. Any parasitoids that emerged were placed in 95% ethanol for identification. Parasitoids that were either fully or partially developed, but unable to emerge from the host eggs, were extracted for identification. If this dissection was required, the egg mass was placed in a Petri dish, covered with gel hand sanitizer (62% ethanol, HDX, Atlanta, GA) to prevent live wasps from flying or the eggs from bouncing away during dissection. Dissected wasps were then placed in 95% ethanol to await identification.

Statistical analysis
Amongst habitats and egg mass types, numbers of egg masses from which adults successfully emerged, and numbers predated, were tested as to treatment effects by Fisher's exact test, with Freeman-Halton extension for 2xn tests where n>2 [15]. We used marginal rate analysis as described in Elkinton et. al, [16] to determine underlying parasitism rates of all sentinel eggs, including those eggs not recovered or in which we could not determine the cause of mortality.

Results
Frozen egg masses yielded higher rates of successful parasitism than fresh egg masses. Of the 117 egg masses each frozen and fresh, 22 frozen egg masses, but only 7 fresh egg masses, yielded parasitoid adults, a significant difference (p = 0.0015, Fisher exact 1-tail test). On the basis of individual host eggs, 121 of a total of 2966 fresh sentinel eggs yielded adult parasitoids, 97 of which successfully emerged from the host eggs, whereas 468of a total of 2898 frozen sentinel eggs yielded adult parasitoids, 292 of which successfully emerged. Both the numbers of parasitoids maturing to adulthood, and the lesser numbers that successfully emerged from host eggs, were significantly higher in frozen eggs (χ 2 (Pearson) = 235.3 and 101.2 respectively, p<0.0001 for both).
Four species of scelionid wasps (Trissolcus japonicus, Trissolcus brochymenae, Trissolcus euschisti, and Telenomus podisi) and one species of eupelmid (Anastatus reduvii) emerged from sentinel BMSB eggs, (Table 1). Seven egg masses yielded the exotic Trissolcus japonicus (Ashmead), a new record for North America [13]. By egg mass, native parasitoid adults were much more likely to develop from frozen egg masses, developing in 23 egg masses, 20 of which were frozen, but only 3 of which were fresh, a significant difference between egg mass types (Fisher exact test P<0.0001). However, the exotic Tr. japonicus was just as likely to emerge from fresh egg masses (4) as from frozen egg masses (2)(Fisher exact test P = 0.360, not significant). Overall, over 98% of native adult parasitoids emerging from sentinel H. halys eggs emerged from the frozen eggs, whereas over 70% of the exotic Tr. japonicus emerged from fresh eggs, reflecting its known ability to reproduce successfully on H. halys eggs in their native Asia. This difference between natives and the exotic Tr. japonicus was also highly significant (Fisher 2x2 exact test P<0.0001).
Parasitoid occurrence was very different by species among different habitats ( Table 2). All three Trissolcus species were more likely to occur in the woods habitat (P<0.00001), whereas Telenomus podisi was recovered exclusively from soybeans (P0.00001 habitat difference), and Anastatus reduvii was recovered exclusively from the apple orchard (a non-significant habitat effect due to low number of egg masses parasitized). As a group, native parasitoids did not exhibit a difference in proportion of egg masses attacked by habitat, but with the inclusion of the exotic Tr. japonicus, did show a significant habitat difference (P = 0.00162), with a majority (20 of 35) of parasitized egg masses in the woods habitat.
Eggs laid in cages on green beans by captive H. halys females resulted in 10 of 61 egg masses laid being parasitized in soybean habitat, zero of 35 egg masses parasitized in the orchard habitat, and 2 of 57 egg masses parasitized in the woods habitat, all by Telenomus podisi (F.) (Hymenoptera: Scelionidae). Of these, 169 individual parasitoids developed in soy, and 29 in woods habitat, but only 29 and 2 adult wasps, respectively, emerged successfully from these eggs. Within the soy habitat, we may compare the fresh sentinel egg masses (0 of 39 parasitized) with the cage-laid egg masses (10 of 61 parasitized), showing that Te. podisi was recovered significantly more frequently from the cage-laid egg masses on common bean, than from the fresh sentinel egg masses on the surrounding soybean (Fisher 2x2 Exact test, P 1-tail = 0.00521).
Although there is a plant species difference between common bean and soybean, these results may suggest a parasitoid preference for the caged (immediately laid) versus fresh sentinel eggs (laid over the past 24 hours and without the adult females being present in the field). Predation did not differ by egg mass treatment (Table 1) nor by habitat type (Table 2); a mean of 17.4% of egg masses, and 13.0% of individual eggs, showed evidence of predation. Because of these eggs predated, lost, or unhatched due to undetermined causes, we used marginal rate of parasitism analysis to determine the level of parasitism in those eggs. According to marginal rate analysis (Tables 1 and 2), observed rates of parasitism ranged from 7.6% in the orchard to 22.8% in the woods habitat; the marginal rate of parasitism ranged respectively from 9.4% to 26.5%.
Of the H. halys control eggs kept in the laboratory, 81% successfully hatched (M. Cornelius, pers. comm.). It is unclear why the remaining 19% of eggs were not viable.

Discussion
Inclusion of our wooded habitat site allowed us to be the first to detect the exotic Trissolcus japonicus, previously thought to only be found in quarantine in the United States [13]. Previous H. halys sentinel egg mass studies focused solely on agricultural habitats including both vegetable crops and orchards, and generally reported very low rates of parasitism. By using fresh laid eggs in screen cages we were able to detect higher levels of parasitism than fresh sentinel eggs. However, this study also suggests that H. halys egg parasitoids are habitat specific and because we only used bean plants in our screen cages, we only recovered the parasitoid wasp specific to the open soybean field habitat, Te. podisi.
Although we did not include multiple sites of each habitat type, our results are consistent with those of Okuda and Yeargan [19], who found Telenomus species to prefer herbaceous hosts and Trissolcus species to prefer woody host plants. Each parasitoid species seems to have a distinct habitat and/or host plant preference. We found only Te. podisi in the soy bean field and on the bean plants in the cage treatments, while Trissolcus species were found only on woody arboreal host plants at both the orchard and woods sites. Anastatus reduvii were found only on woody hosts at the orchard site. Further investigation into host plant and habitat preferences of these parasitoids should be conducted by replicating each habitat type.
Overall, frozen egg masses were more likely to yield adult parasitoids versus fresh. This is likely due to suppression of the immune response of the host egg due to freezing the eggs [11]. Parasitoids were also more likely to develop to adulthood in frozen egg masses versus fresh even if they were not successful at emerging from the host egg. This suggests that despite being a suitable host for completion of development to adulthood, native scelionid parasitoids are not well adapted to the invasive H. halys as a host, since they cannot exit the egg. Native parasitoids overall were less successful at both developing to adulthood and emerging from H. halys fresh and frozen egg masses than Tr. japonicus likely because the parasitoids native to North America did not co-evolve with H. halys.
Because this study relied on morphological features for identification of parasitism, we cannot exclude the possibility of multiple parasitism in egg masses in which parasitoids did not completely develop and did not emerge. Marginal rates of parasitism help to determine a more realistic level of parasitism, but molecular identification techniques would be even more helpful in illustrating an even more accurate picture of parasitism rates in the field.
It is unclear why 19% of the control eggs did not hatch in the laboratory. It is likely that a similar percentage of eggs placed in the field were also not viable. Given the low success rate of parasitism in fresh H. halys eggs (7%), this suggests the possibility that native parasitoids are only successfully parasitizing and emerging from the non-viable eggs within the fresh H. halys egg masses. Molecular techniques have the potential for determining these scelionid parasitoidpentatomid host relationships, including where the parasitism is unsuccessful, as shown by Gariepy et al. [20].
Detection of Tr. japonicus in wooded habitat in the United States is concerning for two reasons and warrants further investigation. First, although Tr. japonicus could decrease the population of H. halys overall, this parasitoid was not found in the soybean field habitat in which H. halys can be abundant and damaging during the late summer. Second, it has been shown in the laboratory that Tr. japonicus readily and successfully parasitizes eggs of native beneficial stink bugs including the spined soldier bug, Podisus maculiventris (Say), which is an important predator of both native and invasive pests such as Mexican bean beetle (Epilachna varivestis Mulsant (Coleoptera: Coccinellidae), Colorado potato beetle (Leptinotarsa decemlineata (Say) (Coleoptera: Chrysomelidae)), cabbage looper (Trichoplusia ni (Hübner)(Lepidoptera: Noctuidae), and gypsy moth (Lymantria dispar (L.)(Lepidoptera: Lymantriidae)) [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24]. Given that woody habitats are preferred for overwintering and in the early season by P. maculiventris [25], and that our wooded site is the only one where Tr. japonicus was detected, this poses non-target concerns and will be explored in more detail in future field studies.
Supporting Information S1 Data. All raw data is contained in the supporting information file, "Herlihy 2014 data. xlsx." (XLSX)