Skip to main content

Bioturbation by endogeic earthworms facilitates entomopathogenic nematode movement toward herbivore-damaged maize roots


Rasmann, Sergio et al. (2022), Bioturbation by endogeic earthworms facilitates entomopathogenic nematode movement toward herbivore-damaged maize roots, Dryad, Dataset,


Entomopathogenic nematodes (EPNs) have been extensively studied as potential biological control agents against root-feeding crop pests. Maize roots under rootworm attack have been shown to release volatile organic compounds, such as (E)-β-caryophyllene (Eβc) that guide EPNs toward the damaging larvae. As yet, it is unknown how belowground ecosystems engineers, such as earthworms, affect the biological control capacity of EPNs by altering the root Eβc-mediated tritrophic interactions. We here asked whether and how, the presence of endogeic earthworms affects the ability of EPNs to find root-feeding larvae of the beetle Diabrotica balteata. First, we performed a field mesocosm experiment with two diverse cropping systems, and revealed that the presence of earthworms increased the EPN infection potential of larvae near maize roots. Subsequently, using climate-controlled, olfactometer-based bioassays, we confirmed that EPNs response to Eβc alone (released from dispensers) was two-fold higher in earthworm-worked soil than in earthworm-free soil. Together our results indicate that endogeic earthworms, through burrowing and casting activities, not only change soil properties in a way that improves soil fertility but may also enhance the biocontrol potential of EPNs against root feeding pests. For an ecologically-sound pest reduction in crop fields, we advocate agricultural practices that favour earthworm community structure and diversity.


Mesocosm experiment - To study the effect of earthworms’ bioturbation activities on EPNs recruitment by maize roots, we performed an outdoor mesocosm experiment at the Botanical Garden of Neuchâtel, Switzerland. The mesocosms consisted of 50 × 50 × 30 cm wooden raised garden beds layered with a 1 mm diameter plastic mesh to prevent earthworm escape (Figure S1 in Supplementary material). Each mesocosm was filled with approximately 60 L of the A-layer of an Anthrosol (organo-mineral horizon) enriched with 10% compost and 10% sand. The initial soil was sieved once at 2 cm, and subsequently hand-sieved twice to remove all potential indigenous earthworms present. Before adding compost and sand, natural soil samples were collected, homogenized, dried at 40 °C for 48 hours, sieved at 2 mm, and ground using agate mortars for subsequent physicochemical analyses. Specifically, we measured the particle size distribution (modified Robinson pipette method), the organic matter content through loss on ignition by weighing before and after burning 10 gr of soil at 450°C for 2 hours, carbon to nitrogen ratio (CN) using an elemental analyser (FLASH2000, Thermo Fisher Scientific, Waltham, Massachusetts, United States); the pH in 1:2.5 soil to water ratio; the cation exchange capacity (CEC) following the cobaltihexamine chloride method; and the total phosphorous (using the Kjeldahl digestion method) 1. The initial A-layer of the Anthrosol so was thus characterized as a silty-loamy soil (23% sand, 65% silt, 11% clay), with 7.06% organic matter content, 2.95% organic carbon, with a CN of 11.36, and with 19 ppm of total phosphorus content. The soil pH was 7.65 and the CEC was 5.0 cmolc/kg.

Four experimental treatments were tested (Figure S1): monocultures with and without the endogeic earthworm species Allolobophora icterica; and polycultures with and without the same earthworms. To reduce risks of interspecific variation in earthworms, we used a commercial strain of A. icterica supplied by the Ecotoxicology Department of National Institute for Agricultural Research (INRA Versailles, France).

The simulated monoculture consisted of three maize plants (Zea mays var. Delprim, UFA Delley Semences et Plantes, Delley-Portalban, Switzerland) per mesocosm, whereas one squash plant (Curcubita pepo, var. Rondini, Sativa Rheinau AG, Switzerland) and two bean plants (Phaseolus vulgaris, var. Neckargold, Sativa Rheinau AG) were growing with three maize plants in the simulated polycultures (Figure S1). In total we built 10 mesocosms per treatment (N = 40 mesocosms). All plants were sown early July and grown until mid-October 2018 before the onset of the experiment (Figure S2). Seven weeks after sowing, half of the mesocosms were inoculated with 15 earthworms each. The earthworms were standardized to 6 g of total fresh weight biomass per mesocosm.

In mid-October, the roots of one maize plant per mesocosm were mechanically damaged with a cork borer (punched three times near root area) and watered with 25 ml of a solution containing 500 μg (2.4 µmoles) of jasmonic acid (JA; (±)-Jasmonic acid, CAS Number: 77026-92-7, Sigma, St Louis, IL, USA) per plant to induce emission of volatile defence compounds 2. JA is a growth phytohormone also called the “wound hormone” as it plays a central role in plant defence and has been shown to induce the release of Eβc in herbivore-attacked maize plants 3. Mechanical damage was preferred over direct herbivory on roots to ensure damage reproducibility and to standardize the production of defence volatile compounds. Two days later, four Galleria mellonella (Lepidoptera: Pyralidae) larvae per mesocosm were placed in the soil as sentinel hosts to quantify EPN infection success. Specifically, in each mesocosm, the first two G. mellonella larvae were buried 5 cm deep in the soil and 5 cm away from the stem of a root-broken maize plant (damaged roots), while the second pair of larvae was placed in the same conditions, but close to the roots of an undamaged plant (control roots). Because late-instars G. mellonella larvae are immobile and highly susceptible to EPN infection, they have been extensively used for monitoring EPNs’ presence in soil 4,5, as was done here. One day after G. mellonella addition, a solution of less than two-week old 3000 infective juveniles H. megidis EPNs was inoculated at the centre of all mesocosms. The used H. megidis EPNs (Nematoda: Heterorhabditidae) were supplied by Andermatt Biocontrol AG, Switzerland, and reared on late-instar G. mellonella larvae in accordance with an in vivo rearing protocol described step by step 6. Five days after EPN inoculation, all G. mellonella larvae were collected. Dead larvae were directly transferred into White traps to confirm infestation by EPNs, while living larvae were kept in soil-filled 5×6×4 cm plastic boxes for measuring potential EPN infection. Next, we collected plant traits related to biomass accumulation, including: total aboveground biomass, total vegetative height, and fitness, as the total biomass of all corncobs on each plant. Finally, a fraction of the soil was sampled in each mesocosm for fertility-related analyses; which included CEC and CN measures.

Olfactometer-based bioassays

To dissect the interactive effect of root herbivory and earthworm presence near the roots of maize plants on EPN recruitment, a first (four-arm) olfactometer bioassay was conducted in controlled conditions of temperature, light and humidity (22±2 °C day / 16±2 °C night, 55% RH, daytime 08:00am – 06:00pm, 230 μmol/m2.s). The belowground olfactometer device (Figure 2A), modified from Rasmann et al. 7, consisted of a central glass chamber filled with white sand (Spielsand classic, Hamann Mercatus GmbH, Germany) extending in side arms connected to terminal glass pots (10 cm high, 15 cm diameter). Pots were filled with 1.2 L of soil (1/4 sand, and 3/4 standard potting soil; Ricoter, Aarberg, Switzerland, 10% relative humidity) as a standard for growing maize in the non-soil substrate when tested in olfactometers 8. Three A. icterica earthworms were inoculated in two terminal pots (Figure 2A). Simultaneously, one maize seedling was sown in each of the four terminal pots, and left to grow for 20 days (two-leaf to three-leaf stage). After 20 days of growth, which is considered as a minimum period for significant bioturbation of the soils 9, three second instars of the banded cucumber beetle Diabrotica balteata (Coleoptera: Chrysomelidae) were added to two opposite glass pots containing the plants for the herbivore treatment setup (Figure 3A). We used D. balteata instead of D. v. virgifera because of quarantine restrictions impeding the use of this species in the climate chambers in Switzerland. D. balteata is a generalist beetle that can feed on maize roots, and has been previously shown to induce maize plants to produce Eβc and attract EPNs 10. Therefore, while slight differences might exist in terms of defence induction between the two Diabrotica species, they should be negligible and the results generalizable across Diabroticine beetles. Eggs of D. balteata were supplied by Syngenta Crop Protection (Stein, Switzerland) and larvae reared on a corn-based diet. Overall, the treatments followed a two-by-two factorial design experiment with the presence or absence of earthworms and root herbivores in each four-arm olfactometer (Figure 2A). After three days, a solution of 2000 H. megidis infective juveniles was inoculated 5 mm below the sand surface in the middle of the central arena. After an additional 24 hours, EPNs were retrieved from each side arm using the Baermann decantation funnel method. Next, roots were harvested, carefully washed and flash frozen in liquid nitrogen. Roots were ground in liquid nitrogen and Eβc production from each root system was measured using solid-phase microextraction (SPME) coupled to gas chromatography-mass spectrometry (GCMS) as described in Rasmann et al. 7. Finally, for each plant, we scored root and shoot fresh biomass and plant height, measured as the longest leaf length from the ground. The same experiment consisting of 5 olfactometers each time was repeated four times for a total of N = 20 replicates.

A second (two-arm) olfactometer bioassay was performed in order to explore whether earthworm-emitted exudates via the epidermal mucus or faeces would directly interfere with the production of Eβc from maize roots, and the subsequent EPN movement. For this, maize seeds were sown in olfactometer glass pots (10 cm high, 5 cm diameter) with commercial substrate in similar conditions as described above (see Figure S3). Half of the plants were watered with tap water (control) while the other plants were watered with mucus solution. The mucus solution was obtained by caging 10 earthworms (adults and juveniles) in two 1mm plastic mesh sieves in contact with each other’s open edge. The cage was submerged in 3 mm of tap water and left in darkness at room temperature (25 °C) for one hour, allowing earthworms to move into water and rub their skin against the sieve. The mucus of 20 earthworms (two cages) was collected to water 10 plants (equivalent of two earthworms per plant). The mucus solution (200 mL) was freshly prepared an hour before direct application onto the plants. All plants received the same volume of liquid at the same interval in order to keep substrate between plants as homogenously moist as possible. After 20 days, three second-instar D. balteata larvae were placed in every pot (controls and treatments) and left to feed on maize roots for three days. Connection with the olfactometer system was made 1 day before EPN inoculation, after which, a solution of 2000 infective juvenile EPNs was inoculated in the olfactometer central arena. After 24 hours, EPNs were extracted from the two side arms of each olfactometer with the Baermann funnel method and counted under the microscope. Finally, root biomass and Eβc emissions were recorded as described above. The experiment was replicated 10 times.

A third (two-arm) olfactometer bioassay was performed to the test whether earthworm bioturbation activity in bulk soil affects nematode mobility alone, independently of earthworms being in contact with the maize root system. For this, soil-filled side arms and central arena and sand-filled terminal pots (10 cm high, 5 cm diameter) were assembled into a two-arm olfactometer (Figure 4A). The same natural soil that was used for the mesocosm experiment at the Botanical Garden was used and sieved to 2 mm to ensure effective bioturbation and burrowing by earthworms. Ten soil samples were randomly taken from the soil stock, put on filter paper, emerged in water and left for decantation for 48 hours to test the presence of any indigenous nematodes. None were observed and the soil was consequently not sterilized. Based on the study of Chiriboga et al. 11 on diffusion of Eβc in different soil textures, soil and sand moisture were set respectively at 20% and 10% to maximise diffusion of volatiles. On the first day, three A. icterica earthworms were added to one half of olfactometer, and none in the other half. Earthworm bioturbation was restricted to half of the central arena by a 0.5 mm-mesh screen dividing it and by an anti-EPN mesh screen at the end of the side arms. Earthworms were left to work the soil for 4 days in climatic chamber (18 ± 2 °C, continuous darkness). Four days were considered enough time for three earthworms to properly burrow 0.5 L of soil. On day 5, the terminal pots were connected to olfactometer central system, and custom-made dispensers containing CO2 generating material (300 mg and sodium hydrogencarbonate and citric acid 3:1) and synthetic Eβc (300 μl, β-Caryophyllene, CAS Number 87-44-5, Sigma, St Louis, IL, USA) were prepared as described in Turlings et al. 3, and inserted into the terminal pots filled with sand to ensure that EPN attraction was equally stimulated by both sides of the olfactometer (Figure 4A). Five hours after inserting the dispensers, a suspension of 2000 H. megidis infective juveniles was inoculated in the central arena and left for 24 hours, after which EPN presence in each arm was retrieved using Baermann funnels. The experiment was replicated eight times.

Statistical analysis – All statistical analyses were performed on R 12.

Mesocosm outdoor experiment – We scored the probability of infection by dividing the number of larvae infected by EPNs around each plant by two. We then assessed the full interactive effect of culture type (two levels), earthworms (two levels), and root induction (two levels) on the probability of infection with generalized linear model analysis (GLM) with quasi-binomial distribution. We next performed the same GLM, but by comparing the interactive effect of earthworms and root induction, by splitting the data into monoculture and polyculture systems. Probabilities of infection scores were visualized using the library popbio 13. The interactive effect of culture type and earthworms on CEC, CN, plant biomass and corn earcobs biomass was assessed using two-ways ANOVAs, followed by TukeyHSD post-hoc tests. For plant traits, we included mesocosm as a blocking effect in the model.

Four-arm olfactometer bioassay - Analysis of variation in nematode recruitment across earthworms by root herbivory treatments was performed using generalized linear model analysis (GLM) with quasi-Poisson distribution to take data overdispersion into consideration, and by including the experiment date as a blocking factor in the model. Differences among treatments were assessed using analyses of deviance and F statistics. The analysis of the effect of herbivores, earthworms and their interactions on log+1-transformed Eβc emissions were performed using two-way ANOVA, and by including experiment as blocking factor, and root biomass as covariate in the model. Differences among treatments were assessed using Tukey HSD tests. The effect of root herbivory and earthworms on plant biomass accumulation was assessed on the first principal component analysis (PCA) axis that included plant height, root and shoot fresh biomass (Figure S4). 

Mucus experiment - Analysis of variation in nematode recruitment across treatments was performed using GLM with quasi-Poisson distribution. Analysis of the effect of earthworm mucus on log+1-transformed Eβc emissions and root biomass were performed using one-way ANOVAs.

Bioturbation and synthetic Eβc olfactometer bioassay - The analysis of variation in nematode recruitment across earthworm presence/absence treatment was performed using GLM with quasi-Poisson distribution, and followed by analysis of deviance.

1          Allen, S. E., Grimshaw, H. M., Parkinson, J. A. & Quarmby, C. Chemical analysis of ecological materials.  (Blackwell Scientific Publications., 1974).

2          van Dam, N. M., Witjes, L. & Svatos, A. Interactions between aboveground and belowground induction of glucosinolates in two wild Brassica species. New Phytol. 161, 801-810, doi:10.1111/j.1469-8137.2004.00984.x (2004).

3          Turlings, T., Hiltpold, I. & Rasmann, S. The importance of root-produced volatiles as foraging cues for entomopathogenic nematodes. Plant Soil 358, 51-60, doi:10.1007/s11104-012-1295-3 (2012).

4          Woodring, J. L. & Kaya, H. K. Steinernematid and Heterorhabditid Nematodes: a Handbook of Techniques.,  (Arkansas Agricultural Experiement Station, 1988).

5          Rasmann, S., Erwin, A. C., Halitschke, R. & Agrawal, A. A. Direct and indirect root defences of milkweed (Asclepias syriaca): trophic cascades, trade-offs and novel methods for studying subterranean herbivory. J. Ecol. 99, 16-25 (2011).

6          McMullen, J. G. & Stock, S. P. In vivo and in vitro rearing of entomopathogenic nematodes (Steinernematidae and Heterorhabditidae). Journal of Visualized Experiments, doi:10.3791/52096 (2014).

7          Rasmann, S. et al. Recruitment of entomopathogenic nematodes by insect-damaged maize roots. Nature 434, 732-737, doi:10.1038/Nature03451 (2005).

8          Degenhardt, J. et al. Restoring a maize root signal that attracts insect-killing nematodes to control a major pest. Proceedings of the National Academy of Sciences of the United States of America 106, 13213-13218, doi:10.1073/pnas.0906365106 (2009).

9          Le Couteulx, A., Wolf, C., Hallaire, V. & Pérès, G. Burrowing and casting activities of three endogeic earthworm species affected by organic matter location. Pedobiologia 58, 97-103, doi:10.1016/j.pedobi.2015.04.004 (2015).

10        Rasmann, S. & Turlings, T. C. J. First insights into specificity of belowground tritrophic interactions. Oikos 117, 362-369, doi:doi:10.1111/j.2007.0030-1299.16204.x (2008).

11        Chiriboga, M. X., Campos-Herrera, R., Jaffuel, G., Röder, G. & Turlings, T. C. J. Diffusion of the maize root signal (E)-β-caryophyllene in soils of different textures and the effects on the migration of the entomopathogenic nematode Heterorhabditis megidis. Rhizosphere 3, 53-59, doi:10.1016/j.rhisph.2016.12.006 (2017).

12        R Development Core Team. R: A language and environment for statistical computing.  (2019). <>.

13        Stubben, C. & Milligan, B. Estimating and analyzing demographic models using the popbio package in r. 2007 22, 23, doi:10.18637/jss.v022.i11 (2007).


Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Award: 31003A_159869

Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung

Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Award: 31003A_163334