Mycologia, 95(6), 2003, pp. 1021–1031. q 2003 by The Mycological Society of America, Lawrence, KS 66044-8897 Population genetics and spatial structure of the fairy ring fungus Marasmius oreades in a Norwegian sand dune ecosystem Emnet Abesha menting dikaryotic, vegetative mycelium, as previous- Gustavo Caetano-Anolle´s1 ly proposed. Division of Molecular Biology, Department of Biology, Key words: basidiomycetes, basidiocarps, clado- University of Oslo, P.O. Box 1066, Blindern, grams, DNA amplification fingerprinting, genetic dis- 0316 Oslo, Norway similarity Klaus Høiland2 Division of Botany and Plant Physiology, Department of Biology, University of Oslo, P.O. Box 1066, INTRODUCTION Blindern, 0316 Oslo, Norway Organismal units must be identified when studying populations at the genetic level to establish patterns Abstract: The population genetics and spatial struc- of propagation, inheritance and evolution. In fungi, genetic clones resulting from asexual reproduction ture of the fairy ring fungus Marasmius oreades (Bolt. : Fr.) Fr. was studied by DNA amplification fin- can be characterized by recurrent multilocus geno- gerprinting (DAF). Basidiocarp samples were collect- types (Milgroom 1996, Anderson and Kohn 1998). ed from fairy rings from two separate sand dune sys- Fungal clones generally are of recent origin (Guidot tems of about 560 m2 and 1750 m2, respectively, on et al 1999, Gryta et al 2000). However, in some cases, the Lista Peninsula in southwestern Norway in 1996. such as those that establish symbiosis with fungus- Samples were collected after a careful mapping of farming ants (Mueller et al 1998), they can be an- fairy rings and a vegetation survey of the composition cient. Clones can separate from their origin for dis- and spatial structure of vascular plants, bryophytes persal, while in other cases they can be physically con- and lichens. DAF with standard arbitrary oligonucle- nected and can be highly territorial. Territoriality can otide primers was used to examine the genetic rela- occur at different spatial scales, ranging from centi- tionship between basidiocarp samples. The study meters in Coriolus versicolor growing on birch stumps showed that the fungal population contained a high (Rayner and Todd 1978) to kilometers in Phellinus number of genotypes and that about 90% of the fairy weirii growing in a mountain hemlock forest (Dick- rings represented a separate genet. Both cluster and man and Cook 1989), and local populations can be phylogenetic analyses of DAF amplification products almost as diverse as the entire meta-population (Gui- established relationships between fairy rings and dot et al 1999). Spatial connections between mycelia showed that genetically similar basidiocarps were can be favored or broken by somatic incompatibility found close to each other. Overall results showed a reactions between neighbors, sometimes resulting in weak correspondence between genotype and spatial few and large (e.g., Armillaria bulbosa, see Smith et distribution and no correspondence between geno- al 1996) or many and small individuals, respectively. type and composition of the surrounding vegetation. The study of these spatially connected fungal clones Furthermore, the occurrence of the four dominant offers a unique opportunity to address important sand dune grass species was randomly distributed phenomena in biology, such as the existence of bar- among the localities housing the various fungal ge- riers to gene flow, the role of somatic incompatibility, notypes, indicating that the fungus did not exhibit the evaluation of the cost and benefits of sex in pop- genotypic specialization to the various grass species ulations of varying size, and the role of deleterious that could host it as a pathogen. Results show that mutations in evolution (e.g., Carbone et al 1999). establishment of new individuals generally was me- Marasmius oreades (Bolt. : Fr.) Fr., commonly diated by basidiospore dispersal and not by frag- known as the fairy ring mushroom, belongs to the family Tricholomataceae, order Agaricales, division Basidiomycota. It can form fairy rings, distinct ring Accepted for publication April 15, 2003. or arch-like structures generally recognized by the 1 Present address: Department of Crop Sciences, University of Illi- nois, 332 NSRC, 1101 West Peabody Drive, Urbana, IL 61801. stimulation or suppression of the surrounding plant E-mail: [email protected] growth and the seasonal production of basidiocarps. 2 Corresponding author. E-mail: [email protected] Marasmius oreades can be destructive to lawns, parks, 1021 1022 MYCOLOGIA golf courses and pastures (Lebeau and Hawn 1961, dunes in southern Norway. Experimental objectives Couch 1973) and has been reported to be a patho- included mapping the distribution pattern of each gen of grasses such as Poa pratensis and Festuca rubra fairy ring in two separate dune systems, determining (Blenis et al 1997). The mycelium, which is found in genetic variation among these rings and comparing the soil beneath the ring, interferes with plant-water this variation with the spatial structure and compo- relationships and produces hydrogen cyanide, poly- sition of the surrounding vegetation. acetylene and sesquiterpene metabolites capable of damaging grass roots (Traquair and McKeen 1986, Ayer and Craw 1989). The mycelium from each fairy MATERIALS AND METHODS ring is a genetically homogenous entity that can be Study area. The study area was in the sand dune systems considered a discrete fungal individual (Burnett and on the Lista Peninsula in the Farsund community, Vest-Ag- Evans 1966). Furthermore, these fairy ring individu- der County, southwestern Norway. Dune vegetation has als can be as old as 100–150 yr and possibly 500 yr been described in detail (Høiland 1978). The dominating (Bayliss-Elliott 1911, Schantz and Piemeisel 1917, grasses in the dune pastures on Lista were Festuca arenaria Burnett and Evans 1966). The fungus is heterothallic (a segregate from F. rubra) and Corynephorus canescens and has a unifactorial mating system controlled by a (Høiland 1978). In less established pastures, Marram Grass multiallelic locus (Mallett and Harrison 1988). In Ammophila arenaria was frequent. Poa pratensis ssp. subca- southern Norwegian sand dune landscapes, M. orea- erulea was common in vegetation under weak cultural influ- ence. Outside Poaceae, numerous lowland or coastal plant des is common in the dry dune pastures behind outer species occurred in the pastures, some more frequent on dune ridges (Høiland 1977). A significant positive pioneering sites, other more dominant on the more estab- correlation was found between the diameter of the lished areas (Høiland 1978). Bryophytes and lichens also rings and their distance from the outermost dune were common. The species of grasses, herbs, mosses and ridge in this dynamic ecosystem (Høiland 1993). The lichens usually occurred in mosaic patterns (Høiland 1978). age of the oldest rings was estimated to be less than In the dune ecosystem, M. oreades appeared to be confined 15 yr (Høiland 1993). However, the fungus has not to the dry dune pastures (Høiland 1977). The fungus avoid- been studied with contemporary molecular tools that ed the outer dunes and was absent in the wetter dune slacks would better define its genetics and population struc- and Salix repens-dominated vegetation types (Høiland and ture. Elven 1980). DNA amplification fingerprinting (DAF) is a nu- Field work. Two separate dune systems, areas A and B, of cleic acid scanning technique (Caetano-Anolle´s et al about 560 and 1750 m2, respectively, were chosen on the 1991) capable of resolving taxa efficiently at the sub- small Einarsneset Peninsula (Universal Transverse Mercator species level (Caetano-Anolle´s 1996). Short oligonu- Grid: UTMED50 LK 69 38), SE on Lista. Area A was a sand cleotides of arbitrary sequence, generally 5–8 nucle- plain sheltered from the prevailing western winds by rocks otides in length, are used to amplify a collection of (on the west and north) and otherwise delimited by dune anonymous nucleic acid segments in a genome. ridges and blowouts with naked sand. Area B was part of a These oligonucleotide primers bind to naturally oc- moving sand dune system containing building Ammophila curring sets of short, complementary and closely dunes in the west. Going eastward, there was a succession gradient of nonestablished to fixed Ammophila dunes, dune spaced inverted repeats, driving the DNA polymer- pastures and dune heaths dominated by Empetrum nigrum, ase-based amplification of the spanning sequences. in that order. The delimitation of Area B was set arbitrarily The very high primer-to-template mass ratios provide to cover a representative part of the vegetation types where both a highly stringent and stable amplification re- M. oreades could be found. Area B was 175 m N and 150 m action and result in relatively complex and highly re- E of Area A. producible DNA profiles. These profiles are visual- All data presented in this paper were collected the first ized by silver staining and prove far more reliable and week of Sep 1996, with some additional replicates from the robust than those generated using other techniques same rings in Sep 1997 and 1998. All fairy rings in areas A (e.g., random amplified polymorphic DNA [RAPD] and B were identified, mapped and permanently marked analysis). DAF can characterize closely related organ- for future investigation. Size and form of the rings were isms and has been used successfully to study fungal depicted on the map. In the part of the ring with the high- est density of basidiocarps, a plot of 1 m2 was laid in the S- populations that genetically were highly homoge- N direction. A full analysis of percentage cover of vascular neous (Trigiano et al 1995, Caetano-Anolle´s et al plants, bryophytes and lichens were performed for each 1996, 2001, Bentley et al 1998). plot. A basidiocarp representing the actual fairy ring was In this study, we explored the population structure collected for DNA extraction from the middle of the 1 m2 of M. oreades at the molecular level. DAF was used to plot.