Potentials and limitations of quantification of fungi in freshwater environments based on PLFA profiles

Highlights

• PLFA patterns of fungi can be applied to quantify them in mixed aquatic communities.

• Zoospores and sporangia of Chytridiomycota differ in their PLFA patterns.

• PLFA patterns of phytoplankton infected by fungal parasites change significantly.

• Quantifying aquatic fungi with a Bayesian mixed model based on PLFA is generally recommended.

• Quantification of Chytridiomycota based on PLFA patterns remains difficult.

Abstract

Aquatic fungi are increasingly recognized for their contribution to carbon cycling in aquatic ecosystems, both as saprotrophs and parasites. Their quantification in mixed communities is crucial to assess their ecological significance but remains challenging. We characterized the phospholipid-derived fatty acid (PLFA) composition of fifteen aquatic fungal isolates from Chytridiomycota (chytrids) and Dikarya. Additionally, we identified PLFA patterns of chytrids infecting phytoplankton and their zoospores. PLFA composition of zoospores was highly similar among different taxa, but were distinct from their respective sporangial life-stage. Finally, we applied a fatty acid-based Bayesian mixed model (FASTAR) and tested its potential to quantify fungi in complex mixtures with bacteria and phytoplankton using PLFA profiles. While the quantification of chytrid biomass in low quantities was rather imprecise, the model predicted the contribution of filamentous fungi and other components with fair accuracy, supporting the suitability of this approach to quantify fungal biomass in aquatic environments.

Introduction

Fungi have been described in freshwater environments since the 19th century (Bärlocher, 1992) and the question of their ecological function has been a scientific challenge ever since. Their role as decomposers of plant detritus in streams and marshes is traditionally acknowledged and has mostly been studied in terms of biomass and production (Gessner et al., 2007). Their importance relies in their capacity to produce a diverse suite of degrading enzymes and their ability to penetrate particulate substrates with their hyphae (Gulis and Suberkropp, 2003). Their growth on leaf litter improves, in most cases, the availability of nutrients to higher trophic levels (Krauss et al., 2011) and also provides resources that promote the activity of co-colonising bacteria (Romaní et al., 2004). Most species involved in plant litter decomposition are filamentous Ascomycota, whereas Basidiomycota, Chytridiomycota and yeasts are found in smaller and more variable quantities (Nikolcheva and Bärlocher, 2004; Sampaio et al., 2007; Marano et al., 2011; Bärlocher et al., 2012).

It has been repeatedly suggested that the ecological significance of saprotrophic aquatic fungi might go beyond plant litter decomposition and some proposed an overlooked significance as degraders of other substrates, particularly in lentic ecosystems (e.g. Bärlocher and Boddy, 2016; Chauvet et al., 2016; Grossart and Rojas-Jimenez, 2016). Until now, only a few studies have provided evidence for such a role (Czeczuga et al., 2000; Wurzbacher et al., 2014; Tant et al., 2015; Kagami et al., 2017). However, this perception is shifting due to molecular surveys revealing an unsuspected fungal diversity in both marine and freshwater ecosystems (Monchy et al., 2011; Jobard et al., 2012; Song et al., 2017; Wahl et al., 2018).

Parasitic chytrids can have strong impacts on phytoplankton populations by infecting >90% of the population (Ibelings et al., 2011; Rasconi et al., 2011; Gerphagnon et al., 2019). Beyond the direct effects on phytoplankton abundance, infection by chytrids can also drive successional community changes, delay or supress algal bloom formation, and modulate trophic relationships (e.g. Frenken et al., 2017). Chytrids produce single flagellated zoospores, which penetrate an algal host cell with rhizoids that extract nutrients from the host and develop a sporangium, which eventually releases new zoospores (Canter, 1967). Chytrid zoospores constitute a quantitatively and qualitatively important food source for zooplankton, especially when phytoplankton host species are inedible (due to their size and/or structure) for zooplankton (Rasconi et al., 2011; Kagami et al., 2014; Gerphagnon et al., 2019).

The existing gap in knowledge of fungal diversity in freshwaters, and their detailed ecological functions beyond the decomposition of leaf litter, is largely caused by methodological limitations. While molecular methods have great potential for investigating fungal phylogenetic diversity, it remains difficult to link this diversity with ecological functions, fungal impact, biomass or production. Studies focusing on leaf litter colonisation and decomposition have predominantly used ergosterol as a chemotaxonomic biomarker for fungi. However, ergosterol cannot be used to detect early diverging fungal lineages and can be highly persistent after fungal death (Mille-Lindblom et al., 2004) and is, therefore, not necessarily indicative of living fungal biomass. Chitin staining methods overcome taxonomic restrictions and allow the analysis of fungal colonisation on diverse substrates (Wurzbacher and Grossart, 2012). However, this approach is again not specific for living biomass and is generally applicable to qualitative studies only, rather than for biomass estimates and/or comparison with other fungal groups.

Phospholipid-derived fatty acids (PLFA) are essential structural components of all microbial cellular membranes (with the exception of some Archaea) and have been applied as chemotaxonomic markers to analyse microbial community composition for over 30 years (Frostegård and Bååth, 1996). This method has allowed countless research questions in microbial ecology, in particular in soil sciences, to be answered (reviewed among others by: Ruess and Chamberlain, 2010; Frostegård et al., 2011; Willers et al., 2015). The presence and quantity of certain fatty acids can reflect the presence and taxonomic composition of living biomass (White et al., 1979). The interpretation of fatty acids as markers for certain taxonomic groups is not always consistent between studies (Frostegård et al., 2011). But a few markers have proven to be consistent and universal within soil bacterial groups, i.e. i15:0 (14-methyl pentadecanoate for Gram negative bacteria), and C18:2n6 (linoleic acid (9,12-octadecadienoic acid) for fungi (Frostegård and Bååth, 1996). PLFA analyses have also been applied to characterize microbial communities in aquatic environments, both in field (Dijkman et al., 2009; Steger et al., 2011; de Kluijver et al., 2014) and laboratory studies (Fabian et al., 2016, 2018). The spectrum of analysed fatty acids in aquatic environments is broader due to the presence of polyunsaturated fatty acids (PUFAs) in phytoplankton (Taipale et al., 2013). The detection of fungi in natural aquatic communities based on single markers remains challenging, since fungal fatty acid composition partially overlaps with those of prokaryotic and eukaryotic photoautotrophs.

Strandberg et al. (2015) used a Bayesian mixing model to calculate the taxonomic composition of phytoplankton communities based on their PLFA profiles. Expanding this approach to other taxonomic groups is, therefore, promising and only limited by the availability of published PLFA patterns from the respective species. The simultaneous quantification of bacteria, fungi and phytoplankton, based on the apparent differences in PLFA composition would allow characterisation of microbial communities with unprecedented resolution. However, potential overlaps between fungal and phytoplankton PLFA patterns need to be evaluated. In a recent study on the quantification and distribution of benthic fungi in littoral systems using PLFA, a Bayesian mixed model was applied and validated in comparison to the ergosterol method (Taube et al., 2018). Two questions concerning the universal applicability of this approach remained open: first, no PLFA profiles of aquatic Dikarya were available at that time and could, therefore, not be evaluated and secondly, the quantification of chytrids could not be verified, since they do not contain ergosterol. Furthermore, suitable criteria for the delimitation of individual organismal groups remain unclear. In the case of fungi, different phyla show no categorical differences in PLFA patterns, suggesting that a delimitation of some organismal groups based on morphology might be suitable. As to phytoplankton, some groups are taxonomically distinct but similarities in PLFA profiles seem to justify grouping them together (Strandberg et al., 2015). PLFA profiles of terrestrial fungi have been characterized frequently (e.g. Dembitsky et al., 1992; Zelles, 1997; Klamer and Bååth, 2004). It is reasonable to assume that the habitat, aquatic or terrestrial, does not cause major differences in the PLFA profile, since many species can inhabit both habitats. However, there is currently no evidence for this assumption. The PLFA patterns of various saprotrophic chytrids were described by Akinwole et al. (2014), whereas PLFA patterns of parasitic chytrids are still lacking.

This study aims to explore the use of PLFA signatures to analyse fungal biomass and distribution in aquatic systems by applying Bayesian mixed models on environmental PLFA profiles. With this approach we aim to provide a methodological tool independent of microscopy and molecular methods for fungal quantification in complex samples to explore fungal potential and importance in aquatic ecosystems.

We present here PLFA profiles of saprotrophic species isolated from aquatic environments belonging to Ascomycota, Chytridiomycota and Basidiomycota. To identify potential differences between terrestrial or aquatic species respectively in PLFA patterns, we compared these with previously published patterns of terrestrial taxa. Furthermore, we investigated putative changes in PLFA patterns of phytoplankton when infected by chytrids. As proof of concept to quantify fungal biomass in complex samples, we: (1) applied individual PLFA patterns in a Bayesian mixed model to predict the composition of artificial mixed communities consisting of bacteria, phytoplankton and Ascomycota; (2) evaluated the ability of the model to detect smaller fractions of saprotrophic chytrid sporangia and zoospores by testing varying chytrid abundances within artificial mixes; and (3) tested the effect of fungal infections of phytoplankton on the stability of the Bayesian mixed models output.