The Japanese beetle (Popillia japonica) is one of the most important invasive insects threatening the agricultural and horticultural sectors in its invasive range, and is listed as a priority quarantine pest in the European Union (1) and other countries (2, 3). It was accidentally introduced to the USA in the beginning of the 20^th century (4) and spread from New Jersey to the west coast, up to Canada, and to the south of the USA (5). Since 2014, the Japanese beetle is present on mainland Europe (6), where it has spread from northern Italy (Piedmont and Lombardy) into southern Switzerland (Ticino), expanding its range each year. Both countries have designated infested zones where phytosanitary measures are in place to limit its spread (7–9).

Unlike many other insect pests, Japanese beetle cause significant damage as both larvae and adults. Larvae are white grubs that feed preferentially on grass roots (4), causing damage to grasslands, football fields, golf courses and other areas with turf (10–12). After emerging in early summer, adults move from the larval habitats to feed on their host plants, where they live for four to six weeks. They are gregarious and polyphagous, feeding on leaves, flowers, and fruits of more than 300 plant species including grapevines, stone fruits, berries, maize, soybean, roses and forest trees (4). Currently, the control of Japanese beetle adults and larvae mainly depends on the use of synthetic insecticides (13, 14), with the cost of damage and control measures estimated at more than $400 million per year in the USA alone (15). Besides these monetary costs, large-scale insecticide applications to control this invasive pest pose a risk to the environment and human health (16–18).

To date, no efficient and environmentally friendly control strategy exists against Japanese beetles (14). However, several biological agents have the potential to control this species, including parasitic nematodes (Steinernema sp. and Heterorhabditis sp.), bacteria (Paenibacillus popilliae and Bacillus thuringiensis var. galleriae), and entomopathogenic fungi (Metarhizium spp. and Beauveria spp.; 13, 14, 19–21). Entomopathogenic fungi have proved effective against many insect pests in the Coleoptera, Lepidoptera and Diptera, and fungi of the genera Beauveria, Metarhizium, Isaria and Lecanicillium are commercially applied worldwide (22–26). These fungi infect their insect hosts by attaching to and breaching the cuticle (27). For successful infection, the number of adhering spores is crucial, as mortality is dose dependent (28, 29). Once the fungi have reached the hemolymph of the insect, they form blastospores and exploit the nutrients of the insect (30). Upon the death of the host the fungus grows out of the insect to form new conidiospores (31, 32). The duration of fungal incubation can vary from a few days to a couple of months, and depends on the insect species, the virulence of the fungus and environmental conditions (32, 33).

In Switzerland, adults and larvae of native scarab beetles such as the cockchafer (Melolontha spp.), the June beetle (Amphimallon spp.), and the garden chafer (Phyllopertha horticola) are highly susceptible to different strains of Beauveria brongniartii or Metarhizium sp., depending on the scarab species (33–36). In contrast to the Japanese beetle, these native scarabs almost exclusively cause damage in their larval stages. Thus, the state-of-the-art control strategy against native scarabs consists of culture of the appropriate fungal strain on sterilized barley kernels and application of fungus-colonized barley kernels (FCBK) with a no-till seeder to the larval habitats (meadows, pastures, or turf; 36–38). The Japanese beetle is closely related to these native scarabs and shares a similar ecological niche and life cycle. Under laboratory conditions, several fungal strains (Metarhizium spp. and Beauveria spp.) have proven successful in infecting and killing Japanese beetle larvae or adults (39–42). Furthermore, Behle et al. (21) achieved a moderate to good control of larvae when applying the fungal strain Metarhizium brunneum BIPESCO5/F52 (Bip5) to turf in well-controlled small scale-field experiments. However, there remains a lack of research to test the efficacy of large-scale field application of entomopathogenic fungi against Japanese beetle larvae and adults.

Here, we tested the application of Bip5 against Japanese beetle larvae and adults in the field. We hypothesized that the Japanese beetle can be controlled in a similar way to its native relatives in Europe. Thus, we tested the application of FCBK for control of larvae on the Japanese beetle. Furthermore, we also assessed the efficacy of a spray application of a Bip5 conidiospore suspension to control adult populations. The objective of the field experiments was to monitor the ability of Bip5 to establish and survive in both adult and larval habitats and to measure its effect on insect survival and the damage caused by adult beetles.

Additionally, we carried out two sets of laboratory experiments with the overall aim of assessing the susceptibility of Japanese beetle larvae and adults to three commercially available fungal strains, M. brunneum ART 212, M. brunneum BIPESCO 5, and B. brongniartii BIPESCO 2, under standard laboratory conditions. These experiments revealed host stage-related differences in susceptibility to fungal infection and we hypothesized that they may be explained by cuticular defense mechanisms. We anticipated that larvae have a stronger cuticular defense against soil-borne pathogens than adults, owing to the long subterranean development of the former. Moreover, we expected blastospores to induce faster speed of kill under laboratory conditions due to their fast germination and growth, while more robust conidiospores may be slower in infesting their host.

Our experiments help to assess the potential of entomopathogenic fungi as biocontrol agents against Japanese beetles in the field, and to better understand the mechanisms underlying host-stage-specific virulence of the applied fungal strains.

Japanese beetle adults and larvae cause major damage in their invasive range. To efficiently protect valuable crops and grassland, a control strategy targeting both life stages is crucial. We tested the impact on Japanese beetle larvae in infested meadows of an application method using M. brunneum Bip5 that is well established for the control of scarab larvae native to Europe. In addition, we targeted adults feeding on a soybean field with Bip5 conidiospore spray applications. We found that Bip5 established and persisted in the soil of larval habitats and on the leaves of soybeans. However, neither larval nor adult population sizes were reduced by these treatments at the study sites. Subsequent laboratory experiments revealed that young third instar larvae were not susceptible to superficially applied blasto- or conidiospores of two M. brunneum (Bip5 and ART 212) and one B. brongniartii (Bip2) strain. In contrast, all spray applications of blasto- or conidiospores of the same strains increased the mortality of adults. Both life stages were susceptible to Bip5 and Bip2 when blastospores were injected directly into the hemolymph of the insects, which suggests that the cuticle is an important factor in determining the difference in susceptibility. We did not detect any differences in the virulence of the different fungal strains, species or spore types which shows the robustness of our findings. Our results suggest that larvae are in general resistant to the three commercially available fungal strains tested here. Thus, we do not see great potential in attempting to adapt to Japanese beetle larvae the state-of-the-art control of native scarabs. However, adults are susceptible to all three fungal strains. Hence, the control of adults with entomopathogenic fungi appears to be more promising.

In our meadow field experiments, we found evidence for good establishment and persistence of Bip5 in the soil, but only a marginal effect of Bip5 FCBK treatment on the abundance of P. japonica larvae. This is in contrast with results of Ramoutar et al. (52) and Behle et al. (21), who found clear, albeit variable, control effects when applying Bip5 to turf in well-controlled small-scale field trials. In addition, FCBK application at even lower doses has proven effective in the control of native grubs, and the method is well established (34). Two main factors are probably crucial for the success of the studies cited above. First, clear control effects on white grubs of either Japanese beetle, cockchafer (Melolontha spp.), June beetle (Amphimallon spp.) or garden chafer (P. horticola) were usually obtained when experiments were conducted in moist environments. Additionally, high soil temperature in summer promotes fungal growth and larval infection (53). Furthermore, an intermediate level of moisture and high temperatures have been shown to be most favorable for the infection of P. japonica larvae by Metarhizium anisopliae (41), while infectiveness is reduced under dry conditions (54). Second, the effectiveness of the fungal treatments is affected by the larval stage targeted (55) with greatest success reported for measures targeting first or second instar larvae (21). We conclude that the optimal setting for the control of P. japonica larvae with entomopathogenic fungi of the genera Metarhizium are moist soils during summer months, when early larval instars are present, and larvae are actively feeding directly below the sod.

These conditions were not met in our studies. In our first field experiment, Bip5 FCBK were applied in autumn and the soil was moist during the winter season due to frequent rainfall. However, most larvae were in their third instar and the activity of both P. japonica larvae and Bip5, was limited by low soil temperatures. In our second field experiment, we applied Bip5 FCBK in May to target eggs and young larval instars. The experimental site received very limited summer precipitation, and the resulting dry soils diminished the efficacy of Bip5. We therefore argue that in the actual infested zone in continental Europe, optimal conditions for entomopathogenic fungi to infect P. japonica larvae are rarely met, since moist soils, high temperatures and susceptible larval stages do not coincide.

The control of Japanese beetle adults in the infested zone in northern Italy relies upon insecticide spraying, leaving organic farmers without effective control measures against the invasive pest (phytosanitary service Piedmont, pers. comm.). The spraying of entomopathogenic fungi would offer an environmentally friendly option for organic farmers to protect their crops. Our study shows that Bip5 conidiospores can be applied effectively to soybean plants and that infective propagules persist on leaves over the entire flight period of the Japanese beetle, despite exposure to high temperatures and strong solar radiation. However, Bip5 conidiospore treatment was not more effective than the spore-free surfactant control. The effectiveness of these two treatments averaged around 20–30% (compared to around 70% for the Karate Zeon insecticide treatment; effectiveness calculated according to Abbott; 56), and likely is due to the surfactant in the formulation (57). The lack of an effect of Bip5 conidiospore suspension on the abundance of beetles in the plots was in stark contrast to the difference in the mortality of beetles collected from Bip5 conidiospore treated and surfactant control plots ( Figure 2D ). One explanation may be that a stronger or more rapid effect of the surfactant masked the effect Bip5. The relatively slow action of Bip5 also comes as a disadvantage in comparison to the insecticide Karate Zeon, which caused an immediate knock down of the beetles, with effects that lasted throughout the entire flight period. To accelerate the speed of kill of Bip5 conidiospore treatments, more spores need to come in contact with the beetles, since time-to-death is directly correlated with spore dose (unpublished data; 28). One solution might therefore be to increase the spore concentration of the applied suspension, or to spray repeatedly at shorter time intervals.

Soybean infestation levels in our experimental field were generally low, with an average of less than three beetles per plant. Leaf damage did not exceed one fully skeletonized leaf per plant, even in control plots. This damage level would not have justified the application of an insecticide against Japanese beetles (14), and it may be that effects of the Bip5 conidiospore treatment would have been more readily detectable under a higher insect infestation level.

Our laboratory experiments showed that spray applications of blasto- or conidiospores caused high mortality in adults but not in larvae, independent of the fungal strain. This contrasts previous studies (40, 41, 58); for example, Giroux et al. (40) found no difference in mortality between adults and larvae of Japanese beetle. However, that work used extremely high spore doses (40), which were 8000 times higher than those used here (10⁷ spores mL^-1, which is commonly used in bioassays). In other laboratory studies, mycelial particles were added to soil as source of infection, resulting in elevated mortality of larvae caused by the applied fungi (41, 58), which contrasts with our findings. To the best of our knowledge, our study is thus the first that reports striking differences in the susceptibility of Japanese beetle adults and larvae to M. brunneum and B. brongniartii. There is, however, evidence from studies on other insect species that susceptibility to fungal infection differs between developmental stages, especially when larvae and adults do not share the same habitat and are consequently not equally exposed to entomopathogenic fungi (59, 60). Based on these findings, we hypothesize that developmental stages with long periods of exposure to soil-born fungal pathogens show higher resistance to infection than stages that are exposed for only a short time. In Japanese beetles, adults are exposed to attack by soil-borne pathogens during only a very short time at emergence, and in females during oviposition. In contrast, larval stages are exposed for more than half a year, between egg hatch and pupation. It follows that defense mechanisms in the latter should be stronger than in short-lived adults with limited exposure to soil-borne pathogens.

Results from our injection experiments provide evidence that this may be true at least for cuticular defense mechanisms in P. japonica. When we injected blastospores of Bip5 and Bip2 directly into the insects to override cuticular defense, we were able to infect both adults and larvae. This indicates that larvae possess a cuticle that protects them efficiently against fungal attack, while cuticular defense mechanisms appear to be negligible in adults. This may seem unexpected, since the heavily sclerotized cuticle of adult beetles appears very robust when compared to the soft-bodied larvae. Several studies have shown that M. anisopliae and other entomopathogenic fungi have difficulties penetrating thick and highly sclerotized areas of the integument of other beetle species and cicadas (22, 59, 61). However, spores of M. anisopliae preferably attach to the intersegmental membranes and around the setae of locusts (62), and Beauveria bassiana presumably penetrates Tribolium castaneum larvae through intersegmental membranes (59). Based on these findings, we hypothesize that the intersegmental membranes of adult Japanese beetles may be more prone to fungal attack than the less sclerotized but generally more robust cuticle of their larvae.

However, our results ( Figure 3 ) suggest that cuticular defense mechanisms only partially explain the difference in susceptibility of adults and larvae. Larvae were still more resistant to Bip5 and Bip2 than adults, even when blastospores were injected. Adults reached more than 90% mortality after 7 days. In contrast, only around 30% of the larvae were dead after this period, and mortality averaged around 90% after 10 weeks only. This suggests that larvae are not only better protected by their cuticle but also possess more efficient internal defense mechanisms than adults.

Overall, we were able to successfully establish high numbers of Metarhizium sp. CFU in the soil by applying Bip5 FCBK. However, these did not effectively control Japanese beetle larvae in the field. Our results were similar in two independent field sites at two different dates, thus encompassing potential differences in susceptibility of P. japonica according to the larval stage. We are therefore confident that this finding is robust. We thus cannot recommend Bip5 FCBK application against Japanese beetle larvae in the infested zone in northern Italy as it is used against its native relatives in Europe. While there clearly is room for improvement by adopting more virulent and better-adapted fungal strains (63), and by optimizing their field application, our laboratory experiments indicate that Japanese beetle larvae are generally resistant to entomopathogenic fungi. This conclusion is in line with a literature review of laboratory and field experiments that found that the control of Japanese beetle larvae with entomopathogenic fungi is erratic and thus not recommended (13). Other environmentally friendly alternatives such as the application of entomopathogenic nematodes may be more promising for the control of P. japonica larvae (20).

In contrast, the control of adult Japanese beetles with entomopathogenic fungi appears more promising. We found high susceptibility of adult beetles in laboratory settings to Bip2, Bip5 and ART 212. Furthermore, we found that Bip5 can be used in foliar sprays on crops with conidiospores persisting under high UV radiation and heat. Although we did not find clear effects of the Bip5 conidiospore suspension on Japanese beetle abundance or crop damage in a soybean field, we were able to prove Bip5 infections in field-collected adults after application. More efficient spraying techniques or alternative spore dissemination strategies, such as attract and infest approaches (64, 65), may lead to a greater impact on adult P. japonica populations. In field experiments in which M. anisopliae treated and untreated adult Japanese beetles were released, a significantly lower number of treated individuals were recaptured compared to the control group (66). Furthermore, adults that are attracted to traps equipped with Metarhizium spp. conidiospores as inoculum are contaminated with a sufficient spore dose to cause and increase mortality (64, 65). Those contaminated beetles spread the spores to non-infected conspecifics (64, 65), further increasing fungal disease within the population. We suggest that it would be valuable to further test these attract and infest methods in regions with low Japanese beetle population densities to assess whether it may serve as a tool to reduce the flying population and, consequently, further spread of this invasive pest. We conclude that the application of entomopathogenic fungi can be an important tool in an integrated pest management strategy targeted against P. japonica adults but not against larvae.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

TG: conceptualization, methodology, investigation, statistical analysis and writing original draft. FS: investigation and formal analysis. PN: statistical analysis and writing. GG: funding acquisition, project administration, conceptualization, and writing. All authors contributed to the article and approved the submitted version.