Throughout their lifetime, insects face multiple environmental challenges that influence their performance. Gregarines are prevalent endoparasites in most invertebrates that affect the fitness of their hosts, but are often overlooked in ecological studies. Next to such biotic factors, a current common challenge is anthropogenic pollution with pesticides, which causes a major threat to non-target organisms that are readily exposed to lethal or sublethal concentrations. In a laboratory study, we investigated whether the presence of gregarines modulates the food consumption and life-history traits of a (non-target) leaf beetle species, Phaedon cochleariae, in response to sublethal insecticide exposure. We show that the larval food consumption of the herbivore was neither affected by gregarine infection nor sublethal insecticide exposure. Nevertheless, infection with gregarines led to a delayed development, while insecticide exposure resulted in a lower body mass of adult males and a reduced reproduction of females. Individuals exposed to both challenges suffered most, as they had the lowest survival probability. This indicates detrimental effects on the population dynamics of non-target insects infected with naturally occurring gregarines that face additional stress from agrochemical pollution. Moreover, we found that the infection load with gregarines was higher in individuals exposed to sublethal insecticide concentrations compared to unexposed individuals. To counteract the global decline of insects, the potential of natural parasite infections in modulating insect responses to anthropogenic and non-anthropogenic environmental factors should be considered in ecological risk assessment.

Study organism and rearing

A rearing of P. cochleariae was maintained for several generations at Bielefeld University under constant climatic conditions (20°C, 16 h: 8 h light:dark, 70% r.h.). Adult beetles were mixed randomly for mating and groups of 100-200 individuals kept in boxes (20 × 20 × 6.5 cm) covered with gauze lids. Every year the rearing population was replenished with individuals collected in the field (51°51′21″ N, 8°41′37″ E). Larvae and beetles were supplied with leaves of non-flowering 8-10-week old cabbage plants (Brassica rapa L. spp. pekinensis), which were cultivated in pots in a greenhouse (20˚C, 16 h: 8 h light:dark, 70% r.h.). For the experiment, only middle-aged leaves were offered, and larvae were provided with greenhouse-grown cabbage, while adults received cabbage bought from an organic store due to plant shortage.

General experimental set-up and measurements of larval food consumption and life-history traits

A full-factorial design was set up to test the influences of gregarine infection and sublethal insecticide exposure on larval consumption and life-history of P. cochleariae. Three-week-old adults (about 200 individuals) were provided with cabbage leaves and their offspring was subsequently used for the experiment. Female beetles bite little cavities in the leaf surface, lay individual eggs in these cavities and cover the eggs with secretion. Infection of larvae of P. cochleariae with gregarines occurs via the uptake of spores excreted by infected conspecifics. Faecal remains can also cover the eggs. Thus, after 24 h of oviposition time, eggs were carefully removed from the leaves and the secretion and potential faeces removed with a paintbrush and tap water. Eggs were placed on fresh cabbage leaves and randomly distributed over two rearing boxes, one assigned to the uninfected gregarine treatment (G-: N = 325), the other to the gregarine infection treatment (G+: N = 320). Hatching larvae (G-: N = 195, G+: N = 204) were supplied with the respective food sources for four days as described below (experimental infection of P. cochleariae larvae with gregarines) to ensure a G- or G+ treatment.

At the fourth day after hatching, the larvae were divided into groups of 5-10 larvae in large Petri dishes (9 cm diameter) lined with filter paper and provided with cabbage leaf pieces (3 × 4 cm). From day five on, half of both G- and G+ larvae were assigned to one of the two insecticide treatment groups, either receiving no insecticide (I-) or receiving λ-cyhalothrin-treated leaf discs (I+) for 48 h (procedure see below), resulting in four treatment groups (G-I-: N = 50, G-I+: N = 113, G+I-: N = 60, G+I+: N = 130). Much higher numbers of I+ animals were set up to account for a lower survival of insecticide-treated individuals. From day seven after hatching, all individuals were provided with untreated cabbage leaves.

To investigate the effects of the different treatments on consumption, the amount of leaf mass consumed within 24 h by each larva was measured. At day 9 after hatching, two larvae were randomly selected from each Petri dish. Each larva was weighed (micro balance, ME36S, Sartorius AG, Göttingen, Germany), placed in a small Petri dish (5.5 cm diameter) lined with moistened filter paper and offered a leaf disc (24 mm diameter) of known mass (balance, LA120S-OCE, Sartorius AG, Göttingen, Germany). After 24 hours, the larvae were weighed again and the remaining leaf discs were scanned (Samsung ProXpress SL-M3375FD, Schwalbach/Taunus, Germany; resolution: 600 dpi). The remaining leaf area was determined with ImageJ (v 1.52a) and used to calculate the consumed leaf mass [initial leaf mass x (initial leaf disc area – remaining leaf area)/initial leaf disc area]. Afterwards, these larvae remained in their individual Petri dishes and were used for counting of gregarines (see below).

To test the effects of the different treatments on life-history traits, pupae were individually placed in small Petri dishes lined with filter paper and the time from larval hatching until adult eclosion was noted. The body mass of adult individuals was measured 24 hours after adult eclosion and the sex of the individuals determined. Adults were kept individually in small Petri dishes and supplied with cabbage leaves. Eight days after adult eclosion females were mated with males of the same treatment group, placing each male to one female for 24 hours. After mating, the pairs were separated again, and the number of eggs laid within the subsequent four days per female was counted. The hatching rate was determined by counting the number of hatched larvae relative to the number of eggs laid. Moreover, the survival of all individuals kept under the four treatment regimes was noted until day ten after adult eclosion. The experiment was ended when adults were 15 days old, while adults can live up to three months (Bogdanov-Katjkov 1923).

Experimental infection of P. cochleariae larvae with gregarines

To ensure larval infection with gregarines (G+ treatment), leaves covered with faeces were taken that had been placed in the boxes of the insect rearing stock for 24 hours. Random examinations of the rearing stock showed that all tested beetles were infected with gregarines and that infectious spores were present in their faeces. Larvae of the G- treatment were offered leaves that had been kept for 24 hours in a box without conspecifics and were damaged by regular cuttings to imitate feeding and thus provide leaves of comparable quality to those provided to the G+ group. Such leaves were offered to hatching larvae of the respective treatment groups and replaced every 2 days until larvae were 4 days old. Microscopy of larvae and adults (see below) confirmed that all dissected individuals of the G- treatment (N = 40 larvae, N = 46 adults) were not infected, while all individuals of the G+ treatment were successfully infected with gregarines (N = 40 larvae, N = 73 adults).

Preparation of sublethal λ-cyhalothrin concentration and insecticide exposure treatment

The pyrethroid λ-cyhalothrin was chosen as contact insecticide because it is widely used in agriculture, including fields with crops of Brassicaceae. A sublethal concentration was prepared from the insecticide LAMBDA WG, which contains 5% of the active toxin λ-cyhalothrin. The insecticide was dissolved in methanol (HPLC-grade, VWR International GmbH, Darmstadt, Germany), centrifuged and the supernatant used to prepare a 0.6 mg/L λ-cyhalothrin stock solution. This concentration was considered as sublethal, as less than 50% of the larvae died during the exposure period (following the definition by de França et al. 2017). In an agricultural system, this concentration is approximately less than half of the application recommended by the supplier and may thus occur in areas close to treated fields as residue. The experimental larvae were fed on day five and six after hatching with cabbage leaf discs (24 mm diameter) treated with 76.8 µl methanol (I-) or λ-cyhalothrin solution (I+). The surfaces of the leaf discs were evenly covered with the respective solutions using a pipette and leaf discs were kept under a fume hood until evaporation of the solvent (at least 20 min) before offering them to the larvae. Depending on the number of larvae, two to four leaf discs were offered per Petri dish to provide food ad libitum and after one day the leaf discs were exchanged by discs of the respective treatments for another day.

Counting of gregarines

Subgroups of both larvae and adults of P. cochleariae were examined to test whether the number of gregarines depends on the different treatments and developmental stage of the host. At day 12 after larval hatching, a subset of larvae of the different treatments (two larvae per Petri dish, used for consumption assay, N = 20 larvae per treatment group) was taken and individuals were frozen (-20 °C). Likewise, adult males and females of each treatment group (G-I-: N = 9 females, N = 12 males; G-I+: N = 12 females, N = 13 males; G+I-: N = 12 females, N = 15 males; G+I+: N = 26 females, N = 20 males) were randomly selected and frozen at day 15 after adult eclosion. Afterwards, individuals were thawed, the midgut dissected, spread on a microscope slide in 25 µl sodium phosphate buffer (0.1 M, pH = 7.2), and the number of gregarines (likely as trophozoites and gamonts) counted under a light microscope at 200 to 400 times magnification (ZEISS Axiophot, Carl Zeiss Microscopy GmbH, Jena, Germany) (ESM Fig. S2).

Statistical analyses

All statistical analyses were performed in R version 3.6.3 (R Core Team 2020) with RStudio version 1.2.5033 (RStudio Team 2019). Data were analysed using linear models (lm), generalised linear models (glm) (package MASS; Venables and Ripley 2002) and cox proportional hazards model (package survival; Therneau 2020). Model assumptions for lms and glms (normal distribution and homoscedasticity of residuals) and cox proportional hazard models (proportional hazard and influential observations) were checked using diagnostic plots. Stepwise backward model selection was applied to obtain the minimal adequate models. Based on F test or Chi-square test results (package MASS; Venables and Ripley 2002), non-significant (P-value > 0.05) interaction terms and/or predictors were excluded from the models.

The effects of the predictors gregarine treatment and insecticide treatment as well as their interaction on development time (from larval hatching to adult eclosion) were tested using a glm (poisson distribution, identity link function). The influences of the predictors larval body mass, gregarine treatment and insecticide treatment and their interaction on the food consumption of larvae were analysed using a lm. The effects of the predictors gregarine treatment, insecticide treatment as well as their interaction on the adult body mass were analysed separately for males and females using lm. Glms were performed to investigate the effects of gregarine treatment, insecticide treatment and their interaction on the number of eggs laid by adult females (poisson distribution, log link function) and the hatching rate of larvae [binomial family (success/failure of hatching), logit link function]. To investigate the effects of insecticide treatment on the gregarine load in the larval stage a glm (poisson distribution, identity link function) was calculated. For the gregarine load in adults a glm (poisson distribution, identity link function) with insecticide treatment and sex as predictors was performed. The influences of the predictors gregarine treatment, insecticide treatment and their interaction on survival probability were tested using a cox model, followed by pairwise log rank post hoc tests. Survival data were plotted using Kaplan-Meyer curves (package survival; Therneau 2020).

R Core Team (2020) R: a language and environment for statistical computing. R Foundation for statistical computing, Vienna, Austria. https://www.R-project.org/.

R Studio Team (2019) RStudio: integrated development for R. RStudio, Inc., Boston, MA. http://www.rstudio.com/.

Therneau T (2020) A package for survival analysis in R. R package version 3.2-7. https://CRAN.R-project.org/package=survival.

Venables WN, Ripley BD (2002) Modern applied statistics with S, 4. edn. Springer, New York.