Recruitment of entomopathogenic nematodes by insect-damaged maize roots Sergio Rasmann 1 , Tobias G. Ko¨ llner 2 , Jo¨ rg Degenhardt 2 , Ivan Hiltpold 1 , Stefan Toepfer 3 , Ulrich Kuhlmann 3 , Jonathan Gershenzon 2 & Ted C. J. Turlings 1 1 University of Neuchaˆtel, Institute of Zoology, Laboratory of Animal Ecology and Entomology, CP 2, CH-2007 Neuchaˆtel, Switzerland 2 Max Planck Institute for Chemical Ecology, Hans-Kno¨ll-Strasse 8, D-07745 Jena, Germany 3 CABI Bioscience Switzerland Centre, Rue des Grillons 1, 2800 Dele´mont, Switzerland ........................................................................................................................................................................................................................... Plants under attack by arthropod herbivores often emit volatile compounds from their leaves that attract natural enemies of the herbivores. Here we report the first identification of an insect-induced belowground plant signal, (E )-b-caryophyllene, which strongly attracts an entomopathogenic nematode. Maize roots release this sesquiterpene in response to feeding by larvae of the beetle Diabrotica virgifera virgifera, a maize pest that is currently invading Europe. Most North American maize lines do not release (E )-b-caryophyllene, whereas European lines and the wild maize ancestor, teosinte, readily do so in response to D. v. virgifera attack. This difference was consistent with striking differences in the attractiveness of representative lines in the laboratory. Field experiments showed a fivefold higher nematode infection rate of D. v. virgifera larvae on a maize variety that produces the signal than on a variety that does not, whereas spiking the soil near the latter variety with authentic (E)-b-caryophyllene decreased the emergence of adult D. v. virgifera to less than half. North American maize lines must have lost the signal during the breeding process. Development of new varieties that release the attractant in adequate amounts should help enhance the efficacy of nematodes as biological control agents against root pests like D. v. virgifera. Plants are not simply passive victims of attacking herbivores; they have evolved an arsenal of physical and chemical defences to protect themselves. Often these defences are mobilized only in response to herbivory 1,2 . Among the proposed inducible defences is the production and release of volatile chemicals that could serve as signals to attract natural enemies of the herbivores 3–5 . Manipulating these signals can help increase the effectiveness of these natural enemies as control agents 6–8 . The induced emission of chemical signals is not limited solely to aboveground plant parts. The entomopathogenic nematode Heterorhabditis megidis was found to be attracted to exudates emitted by plant roots after damage by weevil larvae 9,10 , but the nature of the attractants involved is unknown. Here we show that maize roots damaged by larvae of the economically important coleopteran pest Diabrotica virgifera virgifera LeConte are attractive to entomopathogenic nematodes, and we identify the chemical compound responsible for the attrac- tion. D. v. virgifera or Western corn rootworm (WCR) is a voracious pest of maize that is responsible for the use of the bulk of pesticides applied in the cultivation of this crop in the USA 11 . The recent introduction and rapid spread of WCR into Europe has caused major concern for maize production on this continent and has stimulated the search for new methods of maize protection 12,13 . The use of nematodes to control WCR is an ecologically sound option 14,15 , especially if researchers can optimize their efficacy at finding and killing WCR. Attraction of nematodes by WCR-damaged roots To determine whether or not WCR-infested maize plants would attract nematodes, three glass pots each containing one 10-day-old maize plant (var. Delprim) were attached to the arms of a custom- made six-arm olfactometer filled with moist (10% water) sand (Fig. 1a). The plants had been grown on clean sand in the pots, starting 5 days after seed germination. Three additional pots, containing only sand, were attached to the remaining three arms of the olfactometer. Four such olfactometers, each containing three plants plus three sand controls, were prepared on a given day. One plant of each set of three received four second-instar or third-instar WCR larvae, the roots of a second plant were damaged daily by stabbing them five times with a metal corkborer 7 mm in diameter, and the third plant was left unharmed. On day 3 after initial damage, about 2,000 Heterorhabditis megidis nematodes were released in the centre of each olfactometer, where they were free to enter the arms until their passage was blocked by an ultra-fine metal screen (see Fig. 1a and Methods). One day after release, the number of nematodes in each arm was recorded. Significantly more nematodes were recovered from arms connected to the pots with the WCR- damaged plants than from the arms connected to the other treat- ments or controls (Fig. 1b), indicating that damage by WCR induces maize roots to release a nematode attractant. Identification of the attractant Maize leaves had previously been shown to emit a mixture of volatile compounds in response to damage by caterpillars 4 . To determine whether WCR damage induces similar changes in plant volatiles, the leaves and roots from WCR-damaged (3 days) and healthy maize plants were ground and volatiles collected by solid-phase micro- extraction (SPME) were analysed by gas chromatography–mass spectrometry (GC–MS). A marked difference between the treat- ments was that the sesquiterpene (E)-b-caryophyllene was present in roots damaged by WCR but was completely absent from undamaged roots (Fig. 1c). The damaged roots contained small amounts of a-humulene and caryophyllene oxide as well. To a smaller extent, the WCR-induced increase in (E)-b-caryophyllene content was also apparent in the leaves (Fig. 1d). To test whether (E)-b-caryophyllene was indeed attractive to H. megidis, an authentic standard (Sigma-Aldrich, more than 98% pure) was tested in the olfactometer. For this purpose the system was entirely filled with clean moist sand and a 0.2-ml dose of (E)-b-caryophyllene was articles NATURE | VOL 434 | 7 APRIL 2005 | www.nature.com/nature 732 © 2005 Nature Publishing Group