SEARCH

SEARCH BY CITATION

Keywords:

  • environmental RNA sampling;
  • Myxogastria;
  • Myxomycetes;
  • protists;
  • soil biodiversity;
  • soil microbiology

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

In spite of the ecological importance of protists, very little data is available on their distribution in soil. This investigation is the first of its kind on what could be the major components of the soil protistan community, the Myxomycetes, or plasmodial slime-moulds, a monophyletic class in the phylum Amoebozoa. Myxomycetes have a complex life cycle culminating in the formation of mainly macroscopic fruiting bodies, highly variable in shape and colour, which can be found in every terrestrial biome. Despite their prevalence, they are paradoxically absent from environmental DNA sampling studies. We obtained myxomycete SSU rRNA gene sequences from soil-extracted RNAs using specific primers. Soil samples were collected in three mountain ranges (France, Scotland and Japan). Our study revealed an unexpectedly high diversity of dark-spored Myxomycetes, with the recovery of 74 phylotypes. Of these, 74% had < 98% identity with known sequences, showing a hidden diversity; there was little overlap between localities, implying biogeographical patterns. Few phylotypes were dominant and many were unique, consistent with the ‘rare biosphere’ phenomenon. Our study provides the first detailed insight into the community composition of this ecologically important group of protists, establishing means for future studies of their distribution, abundance and ecology.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

New insights into protist diversity have been achieved in the last decade thanks to the increasing use of environmental DNA sampling (Countway et al., 2007; Brown et al., 2009; Sauvadet et al., 2010; Scheckenbach et al., 2010). Most studies have been conducted in marine and fresh-water environments, and only a few have focused on soils (Willerslev et al., 1999; Moon-van der Staay et al., 2006; Bailly et al., 2007; Lesaulnier et al., 2008; Buée et al., 2009; Damon et al., 2012). As a consequence, the species richness of most natural soil protist assemblages is still an uncharted territory (Mora et al., 2011). In the largest dataset for the soil protist community generated to date, members of the Mycetozoa were found to represent c. 25% of total protist biodiversity in a single soil sample (Urich et al., 2008). As the Myxomycetes (also called Myxogastria) are the most species-rich taxon in the Mycetozoa (c. 10 times more species than Dictyostelida and 30 times more species than Protostelida) (Lado, 2001), it is likely that Myxomycetes were the most abundant protist group in the soil studied. Urich et al. (2008) applied an RNA-centered meta-transcriptomic approach, without an amplification step dependent on the use of primers. By contrast, in a metagenomic study in which universal primers were used to amplify the whole SSU rRNA gene from soil DNA (Lesaulnier et al., 2008), Myxomycetes were absent. The most likely explanation of the failure to find Myxomycetes in such environmental DNA sampling is their unusually long SSU rRNA and relatively divergent sequences (Nikolaev et al., 2006; Fiore-Donno et al., 2012). Consequently, universal primers usually do not match even conserved regions (Stephenson et al., 2011).

The diversity and ecology of Myxomycetes has been studied so far only from inventories based on the occurrence of fruiting bodies. From such worldwide data it appears that many species are cosmopolitan; however, some show a geographically limited distribution (Stephenson et al., 2008). As Myxomycetes may spend most of their life-cycle as amoebae (for a description of the life-cycle, see Stephenson et al., 2008), such inventories are likely to strongly underestimate the real diversity. What is known about the distribution of Myxomycetes in soil specifically comes mainly from studies conducted in the 1980s in the UK based on direct cultivation of soil extracts. Although this method (counting of plasmodium-forming units: Feest & Madelin, 1985) underestimates actual diversity, it nevertheless shows that Myxomycetes are abundant and widespread in virtually all soil types (Feest, 1987; Madelin, 1990). Dark-spored Myxomycetes (the newly erected order Fuscisporida: Cavalier-Smith, 2012) are the subgroup best represented in public sequence databases, but few attempts have been made using environmental DNA or RNA sampling to detect them directly from soil (Kamono & Fukui, 2006; Kamono et al., 2009b), air (Kamono et al., 2009a) or decaying wood and forest litter (Win Ko Ko et al., 2009). Although limited by the choice of denaturing gradient gel electrophoresis, these initial studies demonstrated that PCR with specific primers reliably detects myxomycete DNAs and RNAs from the environment.

The present study focused on an ecologically well-defined group of nivicolous Myxomycetes with very narrow ecological requirements. They are mainly found in high-latitude or mountainous grasslands and forests, forming fruiting bodies next to old patches of snow during spring. They have been observed only when snow coverage has lasted for at least 3 months (Meylan, 1931). Under proper ecological conditions, they can cover several square meters with their highly decorated fruiting bodies. The amoebae probably live in the water film under the snow, which hosts a wide diversity of bacteria on which they can feed (Lipson & Schmidt, 2004). Nivicolous Myxomycetes have been studied worldwide and comprise 73, mainly dark-spored, species (Ronikier & Ronikier, 2009 and references therein).

This study aimed to reveal the diversity in soil during the fruiting time of the nivicolous Myxomycetes, using environmental RNA sampling. Soil samples were collected from three sites in Europe and Japan to get a worldwide first impression of their diversity and how they spread. For each soil sample the taxonomic composition of the dark-spored myxomycete community contributing to the soil RNA pool was estimated using SSU rRNA gene fragments PCR-amplified from reverse-transcribed soil RNA. To estimate their soil biodiversity we combined these new environmental sequences with a reference database of SSU sequences, using the largest taxon sample of Fuscisporida for SSU to date.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Soil sampling

Soil samples were collected in spring (2008) under and beside patches of snow in the French Alps (Méribel, Savoy), in Scotland (Cairngorms) and in Japan (Uryu experimental forest of Hokkaido University, Hokkaido) (Table 1). At each site, three soil samples were collected at different positions, each 2 m apart in a straight line: one below the patch of snow, one at its edge, and the third in a drier spot beyond the snow. At each position, soil samples were collected from a small plot (30 × 30 cm) with a sterile spoon and stored in a 100-mL Falcon tube at 4 °C until processed within 1–3 days.

Table 1. Place of collection of the soil samples
Place of collectionDateCoordinatesAltitude (m)Snow depth (cm)SubsamplesLabelpHTemperature
SoilAir
France, Savoy, Les Allues, Méribel Altiport30/4/2008N 45.409°E 06.577°170025Under the snowMS5.60.07.5
At the edgeMM7.50.07.5
Two m apartMD6.51.67.5
Scotland, Cairngorms, Coire an Lochain6/5/2008N 57.116°W 03.570°83833Under the snowCS4.90.414.0
At the edgeCM5.59.014.0
Two m apartCD5.07.614.0
Japan, Hokkaido, Uryu experimental forest29/5/2008N 44.433°E 142.148°59530Under the snowUS5.380.718.2
At the edgeUM5.144.618.2
Two m apartUD5.059.018.2

RNA extraction, amplification and sequencing

We targeted a fragment of c. 600 bases at the beginning of the SSU rRNA gene, using primers specific for Fuscisporida. Primers were designed to target Fuscisporida unique signatures, using our database of Myxomycetes and selected eukaryotic sequences, and tested for specificity on DNAs from both Fuscisporida specimens and soil samples. RNA extraction and reverse-transcription of RNA into cDNA were carried out as described elsewhere (Kamono et al., 2009b). Contaminating DNA was removed from RNA extracts by treatment with RNase-free DNase. Reverse transcription of RNA into cDNA used the SuperScript_III First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA) and primer 718RL: CGTATGCTATTAGAGCTGGAATTAC (all primer sequences are given in the 5′–3′ direction). For the subsequent amplification step, 3 μL of cDNA was used as template with primers S2 TGGTTGATCCTGCCAGTAGTGT (Fiore-Donno et al., 2008) and SP03r TCCTCTAATTGTTACTCGAG. To check for the absence of DNA contamination, the same reactions without the reverse transcriptase (RT) enzyme were also performed with all purified RNA samples. Amplicons were purified using the ExoSAP-IT kit (USB Corporation, Cleveland, OH) and ligated into pCR2.1-TOPO vector using the TOPO TA Cloning® kit (Invitrogen), and then transformed into competent TOP10 cells (Invitrogen). A PCR using the M13 primer was conducted and amplicons were purified using SureClean (Bioline, London, UK) and sequenced at the Zoology Department, University of Oxford (UK) and at the Institute of Low Temperature Science, Hokkaido University (Sapporo, Japan).

DNA extraction from fruiting bodies, amplification and sequencing

To improve the resolution of the reference database of the SSU rRNA, 17 partial (5′) sequences of dark-spored myxomycete fruiting bodies were obtained (Table 2). Specimens were field-collected, often near the soil-sampling site, kept in herbaria and identified by the third author. DNA was extracted, amplified and sequenced as previously described (Fiore-Donno et al., 2012).

Table 2. List of the new myxomycete specimens used in this study and collection information
SpeciesAuthorsVoucherDatePlace of collectionAltitude (m)SubstrateLat.Long.GenBank access. #
  1. Herbaria: MM, Marianne Meyer; AMFD, Anna Maria Fiore-Donno; AK, Akiko Kamono.

Badhamia utricularis (Bull.) Berk.MM 2937108/11/04FR, 73, Rognaix400Sambucus nigra living trunkN 45.590°E 06.445° JQ898087
Diderma alpinum MeylanMM 3745827/05/08FR, 73, Bonneval Les Monts1553Rubus sp.N 45.533°E 06.444° JQ898088
Diderma fallax (Rostaf.) LadoMM 3698123/04/07FR, 73 Méribel, Mottaret1700Dry grassN 45.367°E 06.585° JQ898089
Diderma meyerae H.Singer, G.Moreno, Illana & A. SanchezMM 3062528/05/08FR, 73, Méribel, les Allues, altiport1670Dry grassN 45.404°E 06.574° JQ898090
Diderma microcarpum (Scotland 2)MeylanMM 3052206/05/08GB, Scotland, Cairngorm, Coire an Sneachda825Living and dry plantsN 57.119°W 03.669° JQ898091
Diderma microcarpum (France)MeylanMM 3741819/05/08FR, 73, Esserts-Blay1611 Vaccinium myrtillus N 45.620°E 06.396° JQ898092
Diderma microcarpum (Scotland 1)MeylanMM 3052506/05/08GB, Scotland, Cairngorm, Coire an Lochain838Living and dry plantsN 57.116°W 03.570° JQ898093
Diderma niveum (Rostaf.) T.Macbr.MM 3052305/05/08GB, Scotland, Cairngorm, Coire an Sneachda825Living and dry plantsN 57.119°W 03.669° JQ898094
Lamproderma echinosporum MeylanAK U1129/05/08JP, Hokkaido, Uryu Forest, site 1595Leaves of Sasa kurilensisN 44.433°E 142.148° JQ898095
Lamproderma maculatum var. ‘macrosporum’Mar. Mey. & PoulainMM 3818421/05/09FR, 73, Arvillard, Val Pelouse1696Living Rhododendrum stemsN 45.420°E 06.169° JQ898096
Lamproderma splendidissimum ad int.MM 3851624/05/01FR, 73, Bonneval Les Monts1607 Vaccinium myrtillus N 45.532°E 06.443° JQ898097
Lepidoderma chailletii Rostaf.AK U0629/05/08JP, Hokkaido, Uryu Forest, site 1595Leaves of Sasa kurilensisN 44.433°E 142.148° JQ898098
Lepidoderma peyerimhoffii (Maire & Pinoy) H.Neubert, Nowotny & K.BaumannAMFD 35607/05/06IT, Piemonte, CN, Demonte, Vallone dell'Arma2013Dead twigsN 44.376°E 07.139° JQ898099
Physarum albescens D.Ellis ex T.Macbr.AMFD 21707/05/05FR, 73, Les Saisies, Bisanne1700Dead twigsN 45.743°E 06.509° JQ898100
Physarum alpestre Mitchel, S.W.Chapm. & M.L. FarrMM 3699322/04/07FR, 73, Beaufort, Col du Méraillet1600Dry grassN 45.694°E 06.632° JQ898101
Physarum compressum Alb. & Schwein.AMFD 15117/07/04IT, Puglia, LE, Ciccorusso20Opuntia leavesN 40.382°E 18.260° JQ898102
Physarum vernum Sommerf. ex Fr.AMFD 26006/05/05IT, Piemonte, CN, Vallone dell'Ischiator1920Dry grassN 44.293°E 07.057° JQ898103

Identification of operational taxonomic units (OTUs)

Unique sequences were compared with the GenBank database with entries from all traditional divisions including EMBL and DDBJ, excluding bulk divisions and whole-genome shotgun entries, using the Basic Local Alignment Search Tool (blast). geneious pro version 5.5.6 (available from http://www.geneious.com/ ) was used to search the GenBank database for the best hit for each clone, with a maximum expectation value of 1e−5, and all default options, except that the low complexity filter was not used. Because myxomycete SSU sequences are highly variable, blastn was preferred to megablast, the latter being specifically designed for very similar sequences. These sequences were then blasted as explained above against our own database of Fuscisporida, which included 66 unpublished sequences at the time (Fiore-Donno et al., 2012). The MESA program (Caron et al., 2009) was then used for grouping OTUs, with two percentages of similarity, 98% and 96%, and to produce the rank-abundance curve.

Statistical analyses

Individual-based rarefaction curves were used to determine to what extent the number of clones sampled reflected the biodiversity. These curves were constructed using the software r (version 2-13.2) (R Development Core Team, 2011). Sørensen similarity indexes between soil samples and between subsamples were computed using estimates (version 7.5, R. K. Colwell, http://viceroy.eeb.uconn.edu/estimates/, last accessed June 2012). Twelve indexes were obtained: nine by comparison of the subsamples (three per site) and three by comparison of the sites. They were used as input for an analysis of variance (one-way anova), which was performed on spss 11.0 software (SPSS Inc., Chicago, IL); patterns between subgroups were determined by the post-hoc Fisher's least significant difference (LSD) test. Beforehand, the homogeneity of the variances was determined with Levene's test (= 0.166). For all tests, the critical value of 0.05 was used.

Alignments and phylogenetic analyses

The first sequence of each of the 73 genuine Fuscisporida OTUs (98% similarity, one OTU excluded, see Results below) was aligned in our existing dark-spored alignment of 114 taxa (Fiore-Donno et al., 2012) using bioedit (Hall, 1999), augmented with the 17 sequences obtained from sporophores for this work and newly available ones in public databases (430 positions, 204 taxa). First, to check for taxonomic discrepancies with the blast results, a maximum likelihood (ML) tree was obtained (results not shown). As no discrepancies were found at the level of the order, the alignment was split into Stemonitida (467 positions, 108 taxa) and Physarida (477 positions, 95 taxa) (Supporting Information, Data S1 and 2). The best evolutionary models were selected using jmodeltest 0.1.1 (Posada, 2008) and were GTR+I+G and, using the AIC correction, the TIM2ef+G, reflecting the high dissimilarities in substitution rates (Table 3). ML analyses were run using treefinder (version October 2008) (Jobb, 2008) with the TIM2ef+G model of substitution, the gamma distribution being approximated by four categories. The best tree was searched from a starting neighbour joining (NJ) tree, with edge supports calculated by 1000 replicates and search level two. Using mrbayes version 3.2 (Ronquist & Huelsenbeck, 2003) a Bayesian search of tree space was conducted with the GTR+I+G model of substitution, the gamma distribution being approximated by eight categories (Table 3). Resulting trees were rooted according to current phylogeny (Fiore-Donno et al., 2012).

Table 3. (A) Characteristics of the alignments of the myxomycete sequences. (B) Detailed results of the Bayesian inference
(A)
 Number of sequencesNucleotide positionsConstant sitesAmbiguous charactersGC contentBest evolutionary model
MyxomycetesSoil samplesTotaljModelTestAIC criterion
Physarida67289547743.60%1.48%52.68%GTR+I+GTIM2ef+Ga
Stemonitida634410846735.50%0.70%53.13%GTR+I+GTIM2ef+Ga
(B)
 Substitution rates (G <->T = 1, fixed)P-invGamma shapeNumber of generationsLog likelihoodGenerations before convergence
A<->CA<->GA<->TC<->GC<->T
  1. a

    The model TIM2ef has been used in maximum likelihood analyses and the GTR in Bayesian analyses, as Bayesian inference is relatively robust to over-parameterization.

Physarida1.973423.475432.239080.782997.44290.10.2851.5 millions−6229.1752650000
Stemonitida1.918743.248072.27491.121936.57170.1030.3563 millions−7584.55251645000

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

OTU identification, distribution and generic diversity

In total we obtained 546 sequences: 152 from Scotland, 206 from France and 188 from Japan (Table 4). Of these, 348 were unique. Their length varied from 537 to 641 nucleotides, most variation being restricted to the variable helices E8_1, 10, E10_1 and 11 (the latter the most variable). The blast of the 348 unique sequences resulted in a list of 41 distinct myxomycete taxa (best hit for each sequence) and six non-Myxomycetes. As the high-divergent Fuscisporida SSU sequences possess many signatures, non-myxomycete sequences can be detected readily. The six non-myxomycete sequences, as identified by blast (e-value = 0.0; pairwise identities given in brackets) included two uncultured fungi (99.3% and 97.3%), one ciliate (97.1%), one angiosperm (98.3%), one uncultured bacterium (99%) and one uncultured eukaryote (88.6%) (Fig. 1a). Four sequences with < 90% of query coverage were suspected to be chimeras and were checked carefully: if the blast gave a match for only a portion of a sequence, the other part was blasted separatedly: an incongruence in the two blast results indicated a chimera, which was then aligned manually to look for variable regions containing specific signatures for different lineages in one sequence. Only two of these sequences were found to be chimeric and therefore excluded (CM-01, CS-22). Three nearly identical sequences from Japan had a reasonably good match with Fuscisporida but lacked the dark-spored clade signature (Fiore-Donno et al., 2012); they were not chimeras. Because their inclusion in the phylogenetic analyses resulted in a very long branch between the outgroup and the Fuscisporida (result not shown), they were excluded from final analyses. A consensus sequence for these three has been submitted to GenBank as ‘Uncultured eukaryote clone UD_67_ JQ900843’. In the remaining 537 genuine myxomycete sequences, 74 and 66 OTUs with the 98 and the 96% pairwise similarities cut-off, respectively, were identified. At 98% similarity, the classical Stemonitida was the dominant group (60% of the OTUs), of which 52% belong to the paraphyletic genus Lamproderma, whereas the most common genera in the classical Physarida were Diderma and Didymium (both 35%) (Fig. 1b). The geographical distribution of these 348 unique sequences shows the largest number at the French site (41% of all sequences), which also provides the richest diversity (Fig. 1c). In the French Alps site (Méribel), in a range of 1 km from the soil collection point, 389 collections of dark-spored Myxomycetes were done from 1992 to 2012, comprising 55 taxa, of which 47 were nivicolous (Data S3). Stemonitida accounted for 63% of the total, Lamproderma being the most represented taxon (72%), whereas Diderma was the most common genus in Physarida (46%) (Fig. 1d).

Table 4. Summary of the myxomycete sequences obtained for each site (divided in three subsites)
 Méribel, FranceCairngorms, ScotlandUryu Forest, JapanTotal
Under the snow766561202
At the edge611965145
Two m apart676657190
Contaminants (chimeras)2(2)59
Total206152188546
Total genuine204150183537
image

Figure 1. Relative contribution to taxonomic diversity of Myxomycetes of the soil environmental partial SSU rRNA sequences, according to blast results. (a) 348 unique sequences, all sites together. (b) 98% sequence similarity, all sites together. (c) Unique sequences, three sites separately. (d) Relative contribution to taxonomic diversity of the 20-year inventory of fruiting bodies, in a circle of 1 km diameter centred on the spot where the soil was sampled in Méribel.

Download figure to PowerPoint

Statistical analyses

Rarefaction curves suggested that the datasets represented a small proportion of the total sequence diversity in the studied soils, so that more sampling would reveal more diversity (Fig. 2a). Estimation of the total number of taxa in all soils based on 537 genuine myxomycete sequences by the Chao 1 richness estimator predicted 109.8 taxa (95% lower bound = 86.62, upper bound = 169.71, standard deviation = 19.45). According to this estimate, we probably recovered only 44–67% of the extant diversity, but this figure may be biased by underestimation. The abundance of the OTUs shows a dominance of few species, only 11 being found 10 or more times (15%) and 33 only once (45%) (Fig. 2b).

image

Figure 2. (a) Rarefaction curves calculated for the entire dataset (All) and for each site (France, Scotland and Japan). (b) Rank abundance curve of the 74 OTUs (< 98% similarities), with the taxonomic identification of the four most abundant OTUs. (c) OTUs classified according to their percentage of similarity to the next kin by blast. Vertical bars show the threshold for taxonomic identification.

Download figure to PowerPoint

Taxonomic assignment

OTUs were named according to the percentage of sequence similarity to the next kin by blast: for OTUs with between 100% and 98% sequence similarity, the species name was assigned (19 OTUs); from < 98 to ≥ 85%, a genus name was assigned (except for one ‘Hyperamoeba’, see below) (47 OTUs); < 85 to ≥ 74.7%, a family name was assigned (seven OTUs) (Fig. 2c). The presence in GenBank of sequences of the now invalidated genus ‘Hyperamoeba’ is confusing, as its representatives are non-fruiting morphotypes that can belong to both classical orders of Fuscisporida. The sequence of ‘Hyperamoeba’ sp. EJC AF411290 is included in the sister-group of ‘Stemonitis’. OTUs belonging to this clade were named ‘Undetermined Stemonitidae’. The seven OTUs with only < 85 to ≥ 74.7% similarity to the next kin belonged mostly to Stemonitida.

Inter-site comparison

Only one OTU, the second most abundant (Lamproderma arcyrioides CD_02/MS_79/UD_51), was common to all three sites (Fig. 3a, Data S4A). Only four OTUs were shared between France and Japan (Lepidoderma sp. MD_42/US_64, Lepidoderma chailletii MD_10/UD_02, Didymium dachnayum MM_19/UD_03, Lamproderma sp. MD_22/UM_07) (Fig. 3a, Data S4). Sørensen similarity indexes between each pair of sites were lower than between the subsamples of each site (Fig. 3b). The variance of the similarity indexes (one-way anova) of the three sites was not statistically significant (= 0.264); the variance of the same indexes plus the inter-site indexes gave less clear results (= 0.067), indicating that perhaps at least one sample had a significantly higher variation. Fisher's LSD test showed contrasting results for the variance of all three sites when compared with the subsamples of each site: for France and Japan, a statistically significant difference was obtained (= 0.027 and 0.028, respectively), but not for Scotland (= 0.369). The Scotland site was also the most undersampled and least diverse (Fig. 1). Despite the undersampling, these results suggest that species assemblages probably differed somewhat between the studied sites.

image

Figure 3. (a) Number of OTUs per site (represented by ellipses) and in the intersections, shared between sites. S, Sørensen similarity index between two sites. Standard error, in all cases = 0. (b) Sørensen similarity index between subsamples of each site (as a table and in grey in the histogram) and between sites (in black in the histogram).

Download figure to PowerPoint

Phylogenetic analyses

In total, 74 sequences were submitted to GenBank (73 Fuscisporida, all possessing the dark-spored clade signature TCTCTCT at the end of the helix 6 and in the three following positions, and one ‘undetermined eukaryote’ (Data S4); GenBank accession numbers JQ900770JQ900843). Both Bayesian and ML tree topologies are broadly consistent with previous results (Fiore-Donno et al., 2012), although our dataset consists of short sequences with large evolutionary distance and a strong rate heterogeneity (L-shaped gamma distribution, Table 3). This could account for the low resolution of basal branches, particularly in Stemonitida (Data S4). This classical order is now separated into two distinct suborders: Lamprodermina (Lamproderma and allied genera) have fruiting bodies distinctly different from those in the derived suborder Stemonitina (Cavalier-Smith, 2012). The classical order Physarida, now the third suborder in Fuscisporida (Cavalier-Smith, 2012) is formed by four main clusters, the basal Lepidoderma (including Diderma fallax and excluding Lepidoderma tigrinum), the genus Didymium (including Diachea subsessilis, Hyperamoeba ATCC PRA39, Mucilago crustacea, Protophysarum phloiogenum and environmental sequences), the genus Diderma (including Ltigrinum) and the family Physaridae. In the latter, the genera Badhamia, Fuligo, Physarella and Physarum appear intermingled (Data S4A). Stemonitida has been tentatively rooted with the genus Meriderma (including Comatricha rubens). The paraphyletic assemblage of Lamproderma and allied genera Colloderma, Diacheopsis and Elaeomyxa appears here as ancestral to the new suborder Stemonitina, but in Fuscisporida trees it is ancestral to Physarina (Fiore-Donno et al., 2012). The basal branches are not well-resolved, yet the main clades are retrieved as in Fiore-Donno et al. (2012). The clade of the remaining Stemonitida is well-supported (97.2/1) as is each of the three clades included therein: ‘Comatricha’ (including Amaurochaete comata, Brefeldia maxima, Enerthenema melanospermum and Paradiacheopsis solitaria); undetermined Stemonitidae (Hyperamoeba EJC and environmental sequences); and ‘Stemonitis’ (including Symphytocarpus impexus) (Data S4B). For the latest classification of the clade Fuscisporida, which embraces both the classical Stemonitida and Physarida and is divided into three suborders, see Cavalier-Smith 2012).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The rare biosphere

In this study, we retrieved environmental RNA of soil-inhabiting dark-spored Myxomycetes (Fuscisporida) nearly simultaneously in three sites, two in Europe and one in Japan. The results follow the general trend of the rare biosphere: the diversity of protists is composed of a majority of ‘rare’ taxa and few common ones (Caron & Countway, 2009). There is much speculation about the forces sustaining this astonishing diversity. The very existence of the rare biosphere has been questioned (Caron, 2009; Stoeck & Epstein, 2009), as a significant fraction of eukaryotic DNA sequences could belong to randomly scattered remains of dead organisms. This suspicion has been recently fuelled by the finding of DNA sequences of metazoan and photosynthetic taxa from deep sea-floor samples, probably coming from free, extracellular DNA or remains and resting stages of planktonic species (Pawlowski et al., 2011). By extracting RNA, we avoided the bias of free DNA or remains of dead cells; however, the distinction between resting stages such as cysts or spores and active stages such as amoebae and plasmodia cannot be made, because spores contain ribosomal RNA that can successfully be reverse-transcribed (Fiore-Donno et al., 2012). Therefore, both active and easily dispersed dormant stages may have contributed to the diversity we found. It is conceivable that the more abundant sequences represent actively growing species that have bloomed in abundance temporarily because of conditions prevailing at the time of sampling and that rarer ones are resting spores awaiting better conditions when they can germinate and feed actively. Adequate technical means for differentiating between resting and feeding cells have not yet been established. They will be important in the future to determine whether some species are much rarer than others all the time or whether relative abundance primarily reflects temporal fluctuations in ecological conditions.

How far do Myxomycetes disperse?

Another mechanism invoked to explain the rare biosphere is to assume that the transport of protists over great distances is extremely common and extensive. In particular, protists with a life-cycle comprising a stage favouring wind-dispersal, such as Myxomycetes, which build mostly macroscopic fruiting bodies on aerial substrates, should be expected to be found in any suitable habitat. Accordingly, a population genetic study could not find any geographical pattern in the occurrence of distinct genotypes (identical for three genetic markers) of two species of Lamproderma in a transect of c. 50 km, indicating that spores were able to disperse freely in that range (Fiore-Donno et al., 2011). Although examples are few and the phenomenon may be rare, intercontinental transportation does occur. For example, L. chailletii MD_10/UD_02 from France and Japan (present study), Lamproderma aeneum from France and Japan (JQ031969, JQ031970) (Fiore-Donno et al., 2012), Stemonitis sp. UD_53 from Japan and Stemonitis sp. AH1 AY321108 and BuP AY321110, the two latter identified as ‘Hyperamoeba’ from tree bark in Germany (Walochnik et al., 2004), have identical sequences, supposedly indicating that they share a recent common origin. In this study, based on the Sørensen similarity index and the one-way anova test, the species assemblages show more similarity between subsamples of the same locality than between distant localities, although the effect of undersampling should be considered. The Sørensen index is an incidence-based index, known to underestimate compositional dissimilarity when data are undersampled and with a high proportion of few common species and many rare ones (Chao et al., 2006), as in our dataset. On the other hand, abundance-based non-parametric estimators for compositional dissimilarity are also sensitive to undersampling, and would not cover the true value unless at least 66% of the taxa have been sampled (Mao & Colwell, 2005), and were therefore not used in this study. However, the marked differences in taxon composition between our three geographically distant but ecologically similar sampling sites once again call into question the idea that dispersal homogenizes protist populations globally (Finlay, 2002) and is consistent with other evidence that several Myxomycetes have geographically limited distributions (Stephenson et al., 2008).

Hidden diversity revealed by environmental sampling

We arbitrarily applied the 98% cutoff for defining OTUs, as it was suggested that this threshold would prevent confusion with intra-specific polymorphism among multiple rRNA cistrons copies in ciliates (Nebel et al., 2010), and that it might be more appropriate for protists (Caron et al., 2009) than the 95% criterion fashionable for ‘species’-level distinctions among bacteria. The great majority of our data (74%) (Fig. 2c) did not match any known sequence at the 98% similarity cutoff. The classical order Stemonitida was more genetically diverse than Physarida and not surprisingly provided the majority of non-precisely identified OTUs. Since the SSU database only covers c. 10% of the Fuscisporida, a more comprehensive database would certainly increase the percentage of OTUs with a more precise attribution. Nevertheless, the dominance of the genus Lamproderma, already noticed in the first studies on nivicolous species (Meylan, 1931), is confirmed in this study by both the soil and the above-ground inventories (Fig. 1). One soil sample in spring revealed 33 OTUs, while it took 20 years of fruiting body inventories to summarize 47 taxa, with only 10 taxa (all nivicolous) being common to both (Data S3). Strikingly, our environmental soil sampling has revealed the presence of non-nivicolous taxa never recorded by the above-ground inventory, i.e. Diachea subsessilis and Comatricha rubens. Although we cannot be sure that those sequences do not come from spores, it is known that myxamoebae may be present in many environments without forming fruiting bodies: they have been isolated from the coelomic cavity of sea urchins in the Adriatic (Dyková et al., 2007) and several times from fresh water in both natural and artificial habitats (Fiore-Donno et al., 2010 and references therein). Also, a few environmental sequences revealed the presence of Myxomycetes in unusual habitats, such as two Didymiidae from a study aimed at characterizing algae in sloth fur (Suutari et al., 2010). Representatives from this family have also been found twice in marine environments: in sea urchins (Dyková et al., 2007) and in sequence GU320584 of a Didymium sp. from a coastal marine environment (Mt Hope Bay, MA). Thus, Myxomycetes are probably substantially more widespread than inventories of fruiting bodies suggest. Our study has probably recovered only a small sample of a much larger assemblage, suggesting a vast array of unexplored myxomycete diversity, which the combination of methods used here could do much more to unravel. They could also be used to tackle numerous ecological problems concerning Myxomycetes, which may be the quantitatively most important organisms feeding on microbes in soil.

Conflicts in identification

Two examples of sequences referring to the same nominal species that do not cluster in the same clade deserve comment: Diderma niveum and Lepidoderma carestianum (Data S4). Two sequences of D. niveum, AM231291 (Wikmark et al., 2007) and HE614617, are nearly identical but are very different from our sequence JQ898094. This apparent conflict is unsurprising given the very confused taxonomic history of D. niveum and its closest relatives, referred to as the ‘niveum-alpinum’ group in Poulain et al. (2011) (Diderma meyerae, Diderma globosum var. europaeum, D. niveum s. str., Diderma microcarpum and Diderma alpinum) but these species limits are not recognized by all authors (Lado, 2001). Our identification of JQ898094 followed the criteria of Poulain et al. (2011), whereas the specimens that yielded sequences AM231291 and HE614617 must have been identified as D. niveum with different criteria; we called our environmental sequence MS-07 JQ900785 D. niveum because of its similarity to these two sequences, even though this clade of three sequences almost certainly belongs to a different species than D. niveum JQ898094. There are also two very different phylotypes of L. carestianum, a taxon now split into L. carestianum sensu stricto and L. chailletii (Poulain et al., 2011): L. carestianum AM231296 (Wikmark et al., 2007) identified before that split has 99.82% similarities with our new environmental sequence of L. chailletii JQ900744 (named thus because of its near identity to sequence JQ898098 from a fruiting body identified after that split), and so may actually be L. chailletii. On the other hand, the three sequences L. carestianum HE614618, Lepidoderma crustaceum HE614619 and Lepidoderma peyerimhoffii JQ898099 are so closely related that they may not deserve the species status. Critical examination of all sequenced specimens by a single expert in myxomycete taxonomy is clearly needed to verify this interpretation. Further sampling and combined morphological and sequence analysis may be necessary to delimit species more accurately in these two clades.

Methodology

Our method allows for identification of dark-spored Myxomycetes from soil with high fidelity (≤ 1.7% contaminant sequences) and seems to encompass all taxonomic groups, although ‘Comatricha’ and Physaridae could be underrepresented. ‘Comatricha’ is mainly lignicolous, so its absence in soil may not be artefactual. In contrast, there is a bias against some Physaridae due to the first base of the reverse primer: it only matches some of the published sequences of Physarum and Badhamia, and none of the sequences of Fuligo spp. A combination of reverse primers should be designed to overcome this flaw. On the whole, this flaw seemed not to affect our results, as the relative contribution of the dominant genera are the same in the above-ground inventory and in our study in the French site. Both inventories are dominated by Lamproderma and Diderma (Fig. 1), a similarity that lends support to the accuracy of our methodology.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

This work was supported by a Royal Society International Joint Project between the Zoology Department of the University of Oxford (UK) and the Institute of Low Temperature Science, Hokkaido University (JP). We thank M. Fujii, M. Tsutsumi and Y. Yoshikawa for assistance with field surveys in Japan. We are grateful for the support of the staff of the Uryu Experimental Forest of Hokkaido University, Japan. We thank B. Ing, Scotland, for indicating possible places of collection in the Cairngorms, Dr E. Pellegrino, Italy, for help with the statistical analyses, and two anonymous reviewers for constructive suggestions.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
fem12042-sup-0001-DataS1.docWord document30KData S1. Physarida alignment in fasta format.
fem12042-sup-0002-DataS2.fastaplain text document78KData S2. Stemonitida alignment in fasta format.
fem12042-sup-0003-DataS3.fastaplain text document104KData S3. Myxomycete species collected in the site of Méribel (French Alps, Savoy) from 1992 to 2012.
fem12042-sup-0004-DataS4.xlsapplication/msexcel25KData S4. Partial SSU trees of each main taxonomic group of Myxomycetes showing the phylogenetic position of the soil environmental sequences.
fem12042-sup-0005-SupportingInformation4.pdfapplication/PDF179K 

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.