Plants are naturally colonized by many fungal species that produce effects ranging from beneficial to pathogenic. However, how many of these fungi are linked with a single host plant has not been determined. Furthermore, the composition of plant-associated fungal communities has not been rigorously determined. We investigated these essential issues by employing the perennial wetland reed Phragmites australis as a model. DNA extracted from roots, rhizomes, stems, and leaves was used for amplification and cloning of internal transcribed spacer rRNA gene fragments originating from reed-associated fungi. A total of 1,991 clones from 15 clone libraries were differentiated by restriction fragment length polymorphism analyses into 345 operational taxonomical units (OTUs). Nonparametric estimators for total richness (Chao1 and ACE) and also a parametric log normal model predicted a total of about 750 OTUs if the libraries were infinite. Sixty-two percent of the OTUs sequenced were novel at a threshold of 3%. Several of these OTUs represented undocumented fungal species, which also included higher taxonomic levels. In spite of the high diversity of the OTUs, the mycofloras of vegetative organs were dominated by just a few typical fungi, which suggested that competition and niche differentiation influence the composition of plant-associated fungal communities. This suggestion was independently supported by the results of nested PCR assays specifically monitoring two OTUs over 3 years, which revealed significant preferences for host habitat and host organ.

Ectomycorrhizal (ECM) communities were assessed on a 720 m(2) plot along a chronosequence of red oak (Quercus rubra) stands on a forest reclamation site with disturbed soil in the lignite mining area of Lower Lusatia (Brandenburg, Germany). Adjacent to the mining area, a red oak reference stand with undisturbed soil was investigated reflecting mycorrhiza diversity of the intact landscape. Aboveground, sporocarp surveys were carried out during the fruiting season in a 2-week interval in the years 2002 and 2003. Belowground, ECM morphotypes were identified by comparing sequences of the internal transcribed spacer regions from nuclear rDNA with sequences from the GenBank database. Fifteen ECM fungal species were identified as sporocarps and 61 belowground as determined by morphological/anatomical and molecular analysis of their ectomycorrhizas. The number of ECM morphotypes increased with stand age along the chronosequence. However, the number of morphotypes was lower in stands with disturbed soil than with undisturbed soil. All stands showed site-specific ECM communities with low similarity between the chronosequence stands. The dominant ECM species in nearly all stands was Cenococcum geophilum, which reached an abundance approaching 80% in the 21-year-old chronosequence stand. Colonization rate of red oak was high (>95%) at all stands besides the youngest chronosequence stand where colonization rate was only 15%. This supports our idea that artificial inoculation with site-adapted mycorrhizal fungi would enhance colonization rate of red oak and thus plant growth and survival in the first years after outplanting.

BackgroundFungal endophyte communities are often comprised of many species colonizing the same host. However, little is known about the causes of this diversity. On the one hand, the apparent coexistence of closely related species may be explained by the traditional niche differentiation hypothesis, which suggests that abiotic and/or biotic factors mediate partitioning. For endophytes, such factors are difficult to identify, and are therefore in most cases unknown. On the other hand, there is the neutral hypothesis, which suggests that stochastic factors may explain high species diversity. There is a need to investigate to what extent each of these hypotheses may apply to endophytes.ResultsThe niche partitioning of two closely related fungal endophytes, Microdochium bolleyi and M. phragmitis, colonizing Phragmites australis, was investigated. The occurrences of each species were assessed using specific nested-PCR assays for 251 field samples of common reed from Lake Constance, Germany. These analyses revealed niche preferences for both fungi. From three niche factors assessed, i.e. host habitat, host organ and season, host habitat significantly differentiated the two species. M. bolleyi preferred dry habitats, whereas M. phragmitis prevailed in flooded habitats. In contrast, both species exhibited a significant preference for the same host organ, i.e. roots. Likewise the third factor, season, did not significantly distinguish the two species. Differences in carbon utilization and growth temperature could not conclusively explain the niches. The inclusion of three unrelated species of Ascomycota, which also colonize P. australis at the same locations, indicated spatio-temporal niche partitioning between all fungi. None of the species exhibited the same preferences for all three factors, i.e. host habitat, host organ, and time of the season.ConclusionsThe fungal species colonizing common reed investigated in this study seem to exploit niche differences leading to a separation in space and time, which may allow for their coexistence on the same host. A purely neutral model is unlikely to explain the coexistence of closely related endophytes on common reed.