Fungal and bacterial communities and their associations in snow-free and snow covered (sub-)alpine Pinus cembra forest soils

Abstract

Background: In Europe, Pinus cembra forests cover subalpine and alpine areas and they are of high conservational and ecological relevance. These forests experience strong seasonality with alternating snow-free and snow covered periods. Although P. cembra is known for mycorrhization and mycorrhizae usually involve fungi, plants and bacteria, the community compositions of fungi and bacteria and their associations in (sub-)alpine P. cembra forests remain vastly understudied. Here, we studied the fungal and bacterial community compositions in three independent (sub-)alpine P. cembra forests and inferred their microbial associations using marker gene sequencing and network analysis. We asked about the effect of snow cover on microbial compositions and associations. In addition, we propose inferring microbial associations across a range of filtering criteria, based on which we infer well justified, concrete microbial associations with high potential for ecological relevance that are typical for P. cembra forests and depending on snow cover.

Results: The overall fungal and bacterial community structure was comparable with regards to both forest locations and snow cover. However, occurrence, abundance, and diversity patterns of several microbial taxa typical for P. cembra forests differed among snow-free and snow covered soils, e.g. Russula, Tetracladium and Phenoliphera. Moreover, network properties and microbial associations were influenced by snow cover. Here, we present concrete microbial associations on genus and species level that were repeatedly found across microbial networks, thereby confirming their ecological relevance. Most importantly, ectomycorrhizal fungi, such as Basidioascus, Pseudotomentella and Rhizopogon, as well as saprobic Mortierella changed their bacterial association partners depending on snow cover.

Conclusion: This is the first study researching fungal-bacterial associations across several (sub-)alpine P. cembra forests. The poorly investigated influence of snow cover on soil fungi and bacteria, especially those mycorrhizing P. cembra roots, but also saprobic soil organisms, underlines the relevance of forest seasonality. Our findings highlight that the seasonal impact of snow cover has significant consequences for the ecology of the ecosystem, particularly in relation to mycorrhization and nutrient cycling. It is imperative to consider such effects for a comprehensive understanding of the functioning resilience and responsiveness of an ecosystem.

Keywords: Microbial association networks; Microbial interaction; Mountain forest soil; Mycorrhiza; Seasonality.

Fig. 1

Schematic overview of the network inference procedure and the statistical comparison of the networks inferred

Fig. 2

Overview of the microbial composition in all Pinus cembra forests investigated. A, B The taxonomic composition of the A fungal and B bacterial community at the different sampling sites is visualized separate for snow covered and snow-free soils. C, D The relative abundances of C fungal and D bacterial core ASVs (detected in at least one sample from all three locations) at the different sampling sites is visualized separate for snow covered and snow-free soils. Colour shades highlight the diversity of orders within phyla. The relative abundance was calculated as read proportion of microbial units referring to all reads within a sample group (ranging from 0 to 1, i.e. 0–100% of reads)

Fig. 3

A, B Top abundant A fungal and B bacterial ASVs across all sample groups. Those ASVs with highest average relative read abundance within sample groups were selected. For fungi and bacteria, these ASVs accounted for > 7% and within 1.0–2.5% of the total read abundance, respectively. C, D Heatmaps of the core C fungal and D bacterial ASVs. E, F ASV richness within E fungal and F bacterial phyla. Only phyla with significant differences among snow covered and snow-free soils were illustrated; for bacteria, only phyla with ASV richness > 10 ASVs were displayed

Fig. 4

Comparison of snow-free and snow covered association networks. Points connected by dashed lines refer to network pairs, i.e. snow-free and snow covered networks inferred after applying different filtering criteria. A The number of nodes in the networks. B The number of edges = associations in the networks. C The density of the networks as the number of inferred associations relative to the number of possible associations. D, E The percentages of D fungal and E bacterial ASVs detected in both the snow-free and snow covered network of each network pair. F, G The percentage of associations detected also in the paired network: F all associations considered (fungal-fungal, fungal-bacterial, and bacterial-bacterial associations), G only fungal-bacterial (fb) associations considered. Boxes depict the first to third quartile, whiskers indicate the maximum value to the third quartile, the black line indicates the median

Fig. 5

Fungal-bacterial (fb) associations frequently detected in snow-free and snow covered networks. A The influence of snow cover on the fb association frequency of fungal phyla was modelled using generalized linear model assuming Poisson distribution. Each point represents the estimate predicted by generalized linear modelling assuming Poisson distribution. B, C Frequent fb associations in B snow-free and C snow covered networks. For panels B–E, associations observed < 5 times across networks were omitted from display. D, E The Sankey plots visualize frequent fb associations summarized on genus level. Here, only those frequent fb genera associations are displayed that (i) received a taxonomic annotation on genus level and (ii) that were observed at least twice. The total number of associations displayed in the D snow-free and the E snow covered network (bars) is 20 and 25, respectively. Cand. = Candidatus, BCP = Burkholderia-Caballeronia-Paraburkholderia, CX = Cand. Xiphinematobacter