Gregarines are ubiquitous endosymbionts in invertebrates, including terrestrial insects. However, the biodiversity of gregarines is probably vastly underestimated and the knowledge about their role in shaping fitness-related traits of their host in dependence of fluctuating environmental conditions is limited. Using morphological and molecular analyses, we identified a new gregarine species, Gregarina cochlearium sp. n., in the mustard leaf beetle, Phaedon cochleariae. Applying a full-factorial design, we investigated the effects of a gregarine infection in combination with fluctuating starvation conditions during the larval stage on the development time and fitness-related traits of adult beetles. Under benign environmental conditions, the relationship between gregarines and the host seemed neutral, as host development, body mass, reproduction and survival were not altered by a gregarine infection. However, when additionally exposed to starvation, the combination of gregarine infection and this stress resulted in the lowest reproduction and survival of the host, which points to a parasitic relationship. Furthermore, when the host experienced starvation, the development time was prolonged and the adult females were lighter compared to non-starved individuals, independent of the presence of gregarines. Counting of gregarines in the guts of larvae revealed a lower gregarine load with increasing host body mass under stable food conditions, which indicates a regulation of the gregarine burden in dependence of the host condition. Contrary, in starved individuals the number of gregarines was the highest, hence the already weakened host suffered additionally from a higher gregarine burden. This interactive effect between gregarine infection and fluctuating starvation conditions led to an overall reduced fitness of P. cochleariae. Our study emphasises the need to study endosymbionts as important components of the natural environment and to investigate the role of host-symbiont relationships under fluctuating environmental conditions in an evolutionary and ecological context.

Study Organism and Rearing

The leaf beetle P. cochleariae was reared for several generations at 20 °C, 16 h: 8 h light:dark, 70% r.h. in a climate cabinet and once a year crossed with field-collected beetles (51°51′21″ N, 8°41′37″ E). Groups of about 150 individuals were kept together in plastic rearing boxes (20 x 20 x 6.5 cm) with gauze lids. As food and for oviposition, leaves of Chinese cabbage (Brassica rapa spp. pekinensis) were provided. The plants were grown in a greenhouse (20˚C, 16 h: 8 h light:dark, 70% r.h.) and only 7-10 weeks old non-flowering plants were used.

Experimental Set-up to Test Effects of Gregarine and Fluctuating Starvation on Life-History Traits of Beetles

To test consequences of a gregarine on its host under different environmental conditions, a full-factorial set-up was used with gregarines (G-, G+) and fluctuating starvation (S-, S+) as treatments, resulting in four treatment groups (G-S-, G-S+, G+S-, G+S+) with 67 – 70 replicates per group. For the experiment, cabbage leaves were offered for 24 h to the beetles in one rearing box. Then, eggs were randomly collected from these leaves. For gregarine infection we followed a protocol developed previously in our group (Wolz et al., 2022). Each egg was carefully cleaned from faecal residues and female secretions with a moist brush. Eggs were then randomly distributed to one of two gregarine treatment groups (uninfected and infected), kept each in a rearing box with fresh cabbage leaves. Shortly before larval hatching (after 6 d), individuals of the uninfected (G-) treatment group received cabbage leaves that had been mechanically damaged by knife cuttings (to provide leaves of comparable quality between treatment groups) and left for 48 h in a rearing box without any insects. Individuals of the gregarine-infected treatment (G+) group received cabbage leaves, which had served as food in rearing boxes with beetles for 48 h and were contaminated with faeces from gregarine-infected conspecifics. Faecal residues contain oocysts with infectious sporozoites, which are ingested by the larvae and cause gregarine infection. Larvae of the two gregarine treatment groups received the respective food, which was replaced every 24 h, for three days.

On the fourth day after larval hatching, all individuals received untreated cabbage leaves, which were replaced every 48 h, until the end of the experiment (except during starvation periods). Within both gregarine treatment groups (G- and G+), larvae were subdivided into groups of five in Petri dishes (9 cm diameter) lined with filter paper. Half of the individuals of each of the gregarine treatment groups were fed ad libitum and assigned to the non-starvation treatment group (S-), the other half were starved three times (starvation group, S+), each time for a period of 24 h (d4, d7 and d11 after larval hatching), as similarly performed in an earlier study with sawfly larvae (Paul et al., 2019). In the field, larvae may experience repeated bouts of starvation when their host plants are over-exploited by high population densities. To prevent any potential cannibalism and to keep up humidity, small moistened paper balls were added to the Petri dishes during the starvation periods, providing hiding places. However, cannibalism never occurred in this or former experiments with this species. To investigate the impact of this fluctuating starvation treatment and larval body mass on the number of gregarines, one larva was taken from each Petri dish on d13 after larval hatching (n = 13 per group; for gregarine counting see below).

Remaining larvae that pupated were placed individually into Petri dishes (5 cm diameter) lined with filter paper. The day of adult eclosion was noted to determine the development time from larval hatch to reaching adulthood in dependence of the treatments. Two days after adult hatching, the beetles were sexed, weighed (micro balance, ME36S, Sartorius AG, Göttingen, Germany) and adult biomass used as further life-history parameter. Pairs of one male and one female were set up for mating within each treatment group (mating pairs: G-S-: n = 31, G-S+: n = 20, G+S-: n = 20, G+S+: n = 10). The pairs remained together until the seventh day after adult hatching, after which the males were removed. The females were weighed again and the number of eggs laid was counted for each female for four consecutive days (from day 7-10 after adult hatching) as measure of fecundity. From larval hatching until the seventh day of adulthood, the number of individuals that had died were monitored daily to calculate the probability of survival. As the sex of larvae cannot be determined, these data were not separated by sex. The adults can usually live up to three months under laboratory conditions.

Counting Total Number of Gregarines in Hosts of the Different Treatments

Larvae taken for gregarine counting (see above) were weighed and frozen at -20 °C (14 replicates per treatment group). To determine the total number of gregarines (only trophozoite stage) in the gut, the larvae were dissected, and their guts spread in sodium phosphate buffer (0.1 M, pH = 7.2) on microscope slides. The trophozoites were counted at 200 to 400 times magnification (ZEISS Axiophot).

Statistical Analyses

The statistical processing and visualisation of the data were performed with R (version 3.6.3, R Core Team, 2020) in RStudio (version 1.2.5033, RStudio Team, 2019). Model residuals were tested for normality and variance homogeneity and stepwise backwards deletion of non-significant interaction terms and predictors (F test or Chisq test) was computed to obtain the minimum adequate models (package: MASS; Venables and Ripley, 2002). In the results section, only the statistical values of the predictors that remained in the final models are reported. The effects of gregarine treatment, starvation treatment and their interaction on development time and number of eggs laid by females were tested using generalised linear models [GLMs: poisson distribution and identity link function (development time) or link log function (egg number)]. Treatment effects on the body mass of adult beetles were analysed separately for males and females by linear models (LMs: Gaussian distribution, identity link function). The treatment effects on survival data were analysed by a stratified cox model to control for proportional hazard assumption (package: survival, Therneau, 2020). The effects of starvation treatment, insect body mass and their interaction on the number of gregarines were tested using a GLM (poisson distribution, identity link function).