Fungal community assembly in the primary successional glacier foreland soil
Fungal PCR-amplicons were successfully obtained from environmental soil samples collected at the forefront of a receding glacier. The two controls, one that was included in the DNA extraction and another from which the template was omitted in the PCR amplification, remained free from visible PCR products indicating that the clone libraries indeed represented fungi obtained from soil samples. A large proportion of the sequences was determined to be chimeric by the Chimera Check software of the RDP. Analyses conducted after exclusion of the sequence data potentially obtained from another target organism confirmed that the placement of most cloned sequences was insensitive to exclusion of the potentially chimeric data. In other words, the placement of a majority of the cloned sequences was similar whether or not the data identified as possibly chimeric were included in the analyses.
Thirty-seven unique RFLP phenotypes were identified and sequenced from 16 environmental samples. A majority of these clones was present in only one soil sample. The observed low similarity among samples emphasizes the high heterogeneity in soil-inhabiting fungal communities and underlines the difficulty of making general conclusions about factors governing fungal occurrence in an ecosystem scale. Samples from the young substrate tended to represent greater diversity than the older substrates as indicated by exclusive detection of basal asco- and basidiomycetes. Clones obtained from young substrates contained taxa from three of the four main phyla in the Kingdom Fungi: Ascomycota, Basidiomycota and Chytridiomycota; no zygomycetous taxa were observed. The samples collected adjacent to the terminal moraine and therefore representing older substrate contained exclusively higher fungi, filamentous ascomycetes and hymenomycetous basidiomycetes. These results are likely due to the dormant spore bank in the young substrate, whereas the fungal communities in the older substrate were mainly comprised of active biotrophic and saptrotrophic mycelium.
Early community assembly by airborne heterotrophs has been proposed to precede assembly of autotrophic communities in primary successional environments (Hodkinson et al., 2002). It was therefore hypothesized that diverse saprotrophic fungal communities would dominate in the young substrates. Unexpectedly, the clone libraries contained mycorrhizal, parasitic or pathogenic ascomycetes and basidiomycetes unlikely to be able to sustain considerable growth in the absence of autotrophic hosts. Most surprisingly, three biotrophic clades – the Taphrinomycetes (Ascomycota), Urediniomycetes (Basidiomycota) and Ustilaginomycetes (Basidiomycota) – were detected adjacent to the present terminus of Lyman Glacier. It is unlikely that the detected biotrophic fungi were present in soil as active mycelium or unicellular, yeast-like colonies. Although Ustilaginomycetes may have brief saprobic periods after basidiospore germination and Taphrinomycetes may have a budding unicellular stage after ascospore release, the three taxa depend largely on their host plants and have limited growth in soil substrate in the absence of their hosts (Alexopoulos et al., 1996). Similarly, potentially mycorrhizal fungi were detected although no susceptible hosts were in the early successional areas. Taken together, data presented here suggest that early community assembly has a significant component of aerially deposited, dormant spore bank in soil. This conclusion corroborates earlier experiments conducted at the Lyman Glacier study site. The fungal biomass in soil was estimated to be very low in the early successional communities but significantly increased with time since deglaciation (Ohtonen et al., 1999). Focusing on ectomycorrhizal fungi, Jumpponen et al. (2002) concluded that fungi and their propagules were sparse in the young soil substrates but their numbers increased with time since deglaciation. The few resident spores in the soil spore bank detected in the present study may not be adequate for establishment of a functional mycorrhizal symbiosis due to unpredictable germination of the spores in the spore bank. It is important to note that the areas near the glacier terminus are completely void of any mycorrhizal plants.
The fourth surprising group was identified as chytridiomycete Neocallimastix frontalis (95% similarity) in BLAST analyses and placed among Zygomycota and Chytridiomycota in the NJ analyses without a well-supported affinity to either. The true affinity of these clones remains unclear. The two obtained clones may represent a new higher level taxon (class or order) as has been proposed in a study that assessed fungal communities in plant roots by random sequencing of PCR amplicons (Vandenkoornhuyse et al., 2002). More likely, the two clones emphasize the limited taxon sampling of lower fungi in GenBank.
The unexpected, biotrophic groups (Taphrinomycetes, Urediniomycetes and Ustilaginomycetes) observed in the young substrate were absent in the clone libraries from older soil substrate. By contrast to the dormant soil fungi in the spore bank in young soil substrates, many of the fungi observed in the older soil substrates likely represent active soil mycelium. For instance, mycorrhizal fungi (clones with well-supported affinities to Russulaceae and Thelephoraceae) likely match macrofungi collected in earlier surveys of ectomycorrhizal fungi in the Lyman Glacier study site (Jumpponen et al., 1999a; Jumpponen et al., 2002). The inability to detect the biotrophic Ascomycetes and Basidiomycetes in the older substrates does not necessarily indicate their absence in the soil. Ohtonen et al. (1999) observed increasing fungal biomass based on estimated hyphal length and measured fungus-specific phopholipid fatty acids with increasing age of the soil substrate. Accordingly, it is most likely that the members of Taphrinomycetes, Urediniomycetes and Ustilaginomycetes are not exclusive to young substrate on this glacier forefront but, rather, that the resident, dormant sporebank is masked by more abundant, actively growing fungal mycelium.
Additional basidiomycetes likely representing ectomycorrhizal taxa were detected in both young and old substrates. Clones with high similarities to Thelephora sp. and Inocybe geophylla were detected in BLAST, the affinity to Thelephora sp. was also supported in NJ analyses. One clone originating from a sample collected adjacent to the glacier terminus was placed in Tricholomataceae and found 99% similar to Laccaria pumila. Laccaria pumila is closely related to L. cf. montana, which has been frequently observed in previous sporocarp surveys on this site (Jumpponen et al., 1999a; Jumpponen et al., 2002). Nonetheless, it is likely that the observed fungi in the older substrates are active, whereas the fungi detected in the young substrate, void of any adjacent vegetation, represent a resident dormant spore bank.
Other clones from both young and old substrates were placed among basidiomycetous Hymenomycetes and filamentous ascomycetes. The nutritional modes of these taxa remain unclear as they may be saproptrophic or symbiotically associated with plant host mycorrhizas. For example, several taxa belonging to Pezizomycetes were detected in both old and young substrates. These likely form symbiotic associations with mycorrhizal plants as well as occur as free-living saprotrophs (Danielson, 1984; Egger & Paden, 1986a; Egger & Paden, 1986b).
The effect of chimeric sequences on placement of environmental clones
The majority of the cloned sequences determined to be possibly or likely chimeric were congruently placed, whether or not the upstream or downstream chimeric regions were excluded. However, several clones were placed on different phyla in the analyses after exclusion of the chimeric data indicating their likely chimeric origin. For example, dramatic rearrangements such as those of clone Young_08–11 (placed basal to Hymenomycetes in the analyses with full data set, but showed a marginally supported affinity to L. pumila when upstream chimeric data were excluded) and Old_21–14 (grouped with pezizalean P. archeri in the analysis of the full data set, but was nested within Agaricales when upstream chimeric data were excluded) clearly show sequence data originating from more than just one organism. Other changes among the analyses of the different data sets were rather minor and involved loss of bootstrap support but similar placement of the environmental clones. Determining whether these clones represent truly chimeric sequences is difficult at best. For instance, clones Young_02–04 and Young_03–06 (affinity to Lecanoromycetes was not supported in the bootstrap analyses when the upstream or downstream data were excluded) or clones Old_26–05, Old_33–04 and Old_33–19 (placement as sister group to Ophiostomatales was not supported in the bootstrap analyses when the upstream or downstream data were excluded) are unlikely true chimeras for two reasons. First, closely similar sequences were obtained from more than just one sample. It is unlikely that similar sequences would be obtained by random chance of pooling DNA from different sources. Second, the placement of these clones was insensitive to the exclusion of the possibly chimeric data although the support for their placement was reduced because of lesser available data. To avoid errors due to inclusion of chimeric data or false exclusion of nonchimeric data requires specific effort in order to confirm the correct assessment of the community structure.
It was fortunate that these analyses detected obligate biotrophs whose biological activity is unlikely in the absence of their autotrophic hosts and thus indicated the presence of a soil-borne, dormant spore bank. In other words, the studies presented here did not confirm the original hypotheses but provided other valuable insights into the initial fungal community assembly in the absence of established vegetation. The resident spore bank and its biotrophic fungi underline the importance of aerial deposition and stochastic events in early community assembly. Although providing a unique view into the early fungal communities, these observations also outline the shortcomings of the environmental DNA analyses. First, it can be difficult to infer the function of the detected organisms because gross taxonomic resolution on the level of family or order is often inadequate for that task. Second, the conclusions may be shadowed by the uncertainties of the data quality. Chimeric sequences are difficult to detect and present detection tools only indicate need for reanalyses of the data. In this study, several sequences were confirmed to be chimeric based on the inconsistent placement of the cloned sequences when some (potentially chimeric) data were omitted. In light of these limitations environmental rDNA analyses will be most valuable when combined with traditional microscopic and culture-based techniques.