Research
What determines Alnus-associated ectomycorrhizal community diversity and specificity? A comparison of host and habitat effects at a regional scale Melanie Roy1, Juliette Rochet1,2, Sophie Manzi1, Patricia Jargeat1, Herve Gryta1, Pierre-Arthur Moreau3 and Monique Gardes1 1
Laboratoire Evolution et Diversite Biologique, Universite Toulouse 3 Paul Sabatier, UMR 5174 UPS, ENFA, CNRS, 118 route de Narbonne, 31062, Toulouse Cedex, France; 2UMR
BioEMCo, equipe Ibios, Faculte des Sciences et Technologie, Universite Paris Est Creteil, 61 avenue du General de Gaulle, 94010, Creteil Cedex, France; 3Laboratoire des Sciences Vegetales et Fongiques, UFR Pharmacie, Universite Lille Nord de France, EA GRIIOT 4481, BP83, 59006, Lille Cedex, France
Summary Author for correspondence: M elanie Roy Tel: +33 5 61 55 64 33 Email:
[email protected] Received: 3 January 2013 Accepted: 29 January 2013
New Phytologist (2013) 198: 1228–1238 doi: 10.1111/nph.12212
Key words: alder, Alnus, coevolution, community, ectomycorrhiza, mutualism, mycorrhizal network, specificity.
Global-scale analyses of ectomycorrhizal (ECM) fungi communities emphasize host plant families as the main drivers of diversity. This study aims to test, on Alnus–ECM communities, which fungi are said to be ‘host-specific’, to what extent host species, habitat and distance explain their alpha and beta diversity variations, and their specificity. In France, ECM communities associated with two subgenera and five species of Alnus, were sampled on 165 trees from 39 lowland to subalpine sites. In all, 1178 internal transcribed spacer (ITS) sequences of ECM fungi clustered in 86 molecular operational taxonomic units (MOTUs). The species richness was low but still variable, and the evenness of communities was lower on organic soils and in Corsica. Similarity between communities was influenced both by host, soil parameters, altitude and longitude, but not by climate and distance. A large majority of ‘specific’ fungi were shared between host species within a subgenus, and showed habitat preferences within the subgenus distribution range. Our study confirms that Alnus ECM communities are low in diversity, highly conserved at a regional scale, and partly shared between congeneric host species. A large part of alpha and beta diversity variations remained unexplained, and other processes may shape these communities.
Introduction Ectomycorrhizal (ECM) fungi are root symbionts, mutualistic and horizontally transmitted, associated with most genera of temperate and some tropical tree families (Smith & Read, 2008). An ECM fungus may associate with multiple host plants from distinct lineages (generalist strategy) or with hosts from a single lineage at various levels, for example family, genus or species (specialist strategy). Generalist ECM fungi dominate most ECM communities, and are thus ‘said to be the rule’ (Horton & Bruns, 1998). Generalist fungi might be more successful than specialists for long-distance dispersal in a spatially variable environment. By contrast, cases of host-specificity have been reported, such as the Boletales genus Suillus, specifically associated with Pinaceae (Bruns et al., 2002). The ratio between generalist and specific fungi varies between communities, according to host plant and possibly nutrient uptake (Bruns et al., 2002). However cases of host species-specificity are quite rare. Host specificity of ECM fungi may have evolved from local adaptation to a host without reciprocal selection on this host 1228 New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
(Hoeksema, 2010). As a consequence, different hosts may associate with distinct ECM communities, and ECM fungi may acquire more specificity. On another hand, the specificity of ECM fungi may also arise from coincident geographic isolation of host and symbionts, followed by speciation of the symbionts (Timms & Read, 1999), or from adaptation to the local host environment. Assuming that the host represents a relatively constant selection pressure over space and time, adaptation to a host is likely to predominate over isolation or local adaptation to abiotic parameters (Hoeksema, 1999). At the global scale, meta-analyses have suggested that different plant families (such as Fagaceae or Betulaceae) associate with distinct ECM communities (Tedersoo et al., 2012). However, at this large geographical scale, it is still difficult to assert if specificity results from adaptation to the host, or to its habitat, or from geographic isolation within the host distribution range. At a local scale, ECM fungi often associate with different hosts (Bruns et al., 2002; Ishida et al., 2007), and show preferences toward a host family or genus rather than a strict host-species specificity (Dickie, 2007). As an example, the couple Pinaceae–Suillus, a Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
New Phytologist
Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
communities) were measured to test to what extent host identity, habitat parameters and distance may shape these communities at a regional scale, and explain their specificity.
Materials and Methods Sampling sites
42
43
44
45
46
47
A total of 39 sampling sites were selected in three regions of southern France (Alps, southwest France, and Corsica), as representative of lowland to subalpine habitats (Fig. 1). In each site, one or two species of alder were present, either dominant or scattered in mixed forests or linear stands. The most widespread alder in western Europe, Alnus glutinosa (L.) Gaertn. (black alder), was sampled on 23 sites in the Alps, southwest France and Corsica. Four other alder species were also sampled: A. incana (L.) Moench (white alder, five sites) and A. alnobetula (Ehrh.) K.Koch ssp. alnobetula (green alder, five sites) in the Alps, and Alnus cordata (Loisel.) Duby (Italian alder, four sites) and A. alnobetula ssp. suaveolens (Req.) Lambinon & Kerguelen (fragrant green alder, two sites) in Corsica (Supporting Information, Table S1). One sampling was processed for each site, either in late spring or early autumn, in 2007, 2008 and 2009. The sampling sites covered the wide range of natural habitats of each species (from mineral (e.g. river banks) to highly organic substrates, e.g. acidic peat bogs). All sites were described as in Rameau (1994) based on vegetation survey and pedology, from which categories were drawn (Table S1), allowing better comparison among habitats (defined as ‘site category’). A larger category was drawn from soil observation around roots, distinguishing mineral (for trees growing close to rivers or on rocky slopes) from organic microenvironments (for trees growing close to peat bogs or in forests). Climatic data (temperature and precipitation, annual mean, maximum, minimum and annual range, per month and per quarter, isothermality and precipitation
0
200
400
600 km
41
well-studied tree–ECM specific system, suggests that host specificity occurs mainly at plant genus to subgenus levels (Kretzer et al., 1996). The genus Alnus (alders) is a famous example of host-genus specificity (Molina, 1981). Alders are associated with specific fungal lineages belonging to monophyletic clades, such as Alnicola and Alpova, along with a few other Alnus-specific fungal taxa (Moreau et al., 2006, 2013; Rochet et al., 2011). This observation, initially based on coinoculation experiments (Molina, 1981), has been confirmed by several studies of ECM communities (Becerra et al., 2005; Tedersoo et al., 2009; Kennedy & Hill, 2010; Kennedy et al., 2011). This high degree of specificity has been interpreted partly as a selection by coevolution by Rochet et al. (2011). However, many host shift events of ECM fungi coupled with speciation are assumed, even in specific genera such as Alnicola or Alpova (Rochet et al., 2011). The main event in Alnus speciation, possibly correlated with a speciation in ECM phylogeny, seems to be the divergence of the subgenus Alnobetula from the subgenus Alnus, dated at c. 48.6 million yr ago (Rochet et al., 2011). As an illustration of this pattern, ECM fungal lineages, such as Alnicola or Alpova, strictly associate with a subgenus, but do not distinguish between closely related host species (Rochet et al., 2011). Moreover, the two European alder subgenera colonize distinct habitats and rarely overlap: the subgenus Alnobetula is a pioneer shrub of subalpine habitats, while the subgenus Alnus occurs from lowland to subalpine, and from riparian to peaty habitats. Therefore, distinct ECM lineages or communities may have been selected by abiotic parameters or isolation rather than by the host subgenus itself. Abiotic soil parameters, such as pH, phosphate or organic matter content (Pritsch et al., 1997; Dilly et al., 2000; Becerra et al., 2005; Tedersoo et al., 2009), as well as vegetation factors, such as the successional stage and stand age (Helm et al., 1996; Kennedy & Hill, 2010) or the season (Becerra et al., 2005), affect Alnus ECM communities. Abiotic effects have been studied on a single subgenus mostly, and they may be difficult to compare among Alnus subgenera. For instance, Alnus glutinosa and Alnus incana (both from the Alnus subgenus) grow in wet environments and their ECM communities are influenced by water content (Baar et al., 2000; Dilly et al., 2000; Tedersoo et al., 2009), while the ice cover is known to shape the communities of the subalpine Alnus alnobetula (subgenus Alnobetula; Helm et al., 1996). Linking the phylogenies of Alnus–ECM fungi that show a conserved specificity with ecological studies (that emphasize how variable these communities can be) requires an understanding of what shapes Alnus–ECM communities, and whether the ‘specific’ ECM fungi are affected by host identity, habitat or isolation by distance. To address these issues, ECM communities of five Alnus spp. belonging to the two Alnus subgenera present in Europe were studied in various habitats in the south of France, including some locations with two co-occurring species of the subgenus Alnus. Specific ECM fungi were identified by comparing our sequences with those of Rochet et al. (2011) and previously published sequences from Alnus–ECM fungi. The alpha (local diversity) and beta diversities of ECM communities (described as the turnover of species between sites, and thus as a dissimilarity between
Research 1229
0
2
4
6
8
10
Fig. 1 Sampling locations in France (red, Alnus alnobetula ssp. alnobetula; dark blue, A. alnobetula ssp. suaveolens; cyan, Alnus cordata; red, Alnus glutinosa; black, Alnus incana). Axes indicate latitude (x) and longitude (y). New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
New Phytologist
1230 Research
seasonality, see Table S1) were retrieved from Worldclim database (http://www.worldclim.org) by raster package in R (Hijmans & van Etten, 2012). Sampling strategy At each site, a minimum of five alder trees, 10 m apart from each other, were selected per host species. Soil and roots were collected from two sampling points per tree, located up to 1 m distance from the trunk and defined after tracing alder roots (identified by the occurrence of Frankia nodules). Samples were pooled per tree, and kept at 4°C until processing. Back in the laboratory, roots were gently washed under tap water over a 500 lm mesh grid, and finally examined under a binocular microscope in distilled water. At least 20 ectomycorrhizae per sample were picked out with tweezers, and kept separately in 2% CTAB (100 mM Tris HCl, pH 8; 1.4 M NaCl; 20 mM Na2EDTA; 2% N-AcetylNNN-trimethyl ammonium bromide) buffer at 20°C, until DNA extraction. Soil was removed from samples, and dried at 60°C. Soil analyses were carried out following the protocol of Tedersoo et al. (2009), at the Instituto de Recursos Naturales y Agrobiologıa de Sevilla (IRNAS) – CSIC, on a subset of 19 sites and for three host species (A. glutinosa, A. incana and A. alnobetula ssp. alnobetula in continental France, see Table S2).
analysis handled through the CIPRES website (Miller et al., 2010). Molecular operational taxonomic units were assimilated to phylogenetic species, that is, groups of samples forming monophyletic terminal clades with identical or highly similar ITS sequences (> 97% similar, which is a commonly accepted threshold; Nilsson et al., 2008). Species were recognized as distinct with 97–99.7% of similarity when morphological and ecological data were congruent with phylogenetic distinction (not shown; for the genera Alnicola, Alpova and Lactarius; see Rochet et al., 2011). Morphologically recognized species with identical ITS sequences are here assimilated to the same MOTU (e.g. Alnicola umbrina and Alnicola badiofusca; Cortinarius americanus, Cortinarius atropusillus and Cortinarius badiovestitus). MOTUs not documented by sporocarp sequence were kept unnamed (cited as ‘Genus sp.’, see Table S3). For MOTUs not yet recorded on Alnus in databases or literature, host identity was checked by amplification of the same extract using the ITS1 plant and ITS4 universal primers (White et al., 1990), then sequencing. Ectomycorrhizas finally identified as nonAlnus by sequencing were removed from the dataset. A representative sequence of each MOTU per site was deposited in GenBank (353 sequences: JQ890228–JQ890314 and JX989834 –JX990108 and JX999824–JX999829). Statistical analysis on community matrix
Molecular biology DNA extraction was performed on each ectomycorrhiza individually, using the Wizard genomic DNA purification kit (Promega) as described in Rochet et al. (2011). The fungal ITS region was amplified using fungus-specific primers ITS-1F/ITS-4 (White et al., 1990; Gardes & Bruns, 1993), and the Basidiomycetesspecific primer pair ITS-1F and ITS-4B (Gardes & Bruns, 1993). PCR conditions were the same as in Rochet et al. (2011). The sequencing of the PCR product was done by the MilleGen company (Labege, France). Sequences were checked and cleaned using FourPeaks 1.7.1 (A. Griekspoor & T. Groothuis, www. mekentosj.com). Attribution of sequences to molecular operational taxonomic units (MOTUs) Generic attribution of fungal sequences was first assessed by using the best result from the BLAST algorithm (Altschul et al., 1990), performed on the GenBank database (http://www.ncbi.nlm.nih. gov/). Secondly, for each genus, and subgenus when opportune, a sequence dataset was built including a selection of sequences from BLAST results, downloaded from GenBank, and sequences obtained from sporocarps identified by morphology (all preserved in the herbarium at Lille Institute of Pharmacy (LIP); see Rochet et al., 2011). Alignments for each dataset were processed on MUSCLE (http://www.drive5.com/muscle/) and manually checked. Phylogenies were built with RAxML (Stamatakis et al., 2008), by maximum-likelihood analysis following a GTR + GAMMA model of evolution, and tested through a fast-bootstrap New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
All statistical analyses were performed with R 2.3.4.4 packages (R Development Core Team, 2008). A community matrix was built from the number of trees colonized per MOTU and per site (Table S3). Sites were described by both qualitative factors (host identity, region, soil and site categories) and quantitative parameters (altitude, GPS coordinates, climatic parameters; Table S1). Additional soil parameters were measured for sites located in continental France (Table S2). Detection of spatial autocorrelation and geographical structure To reject the hypothesis of spatial autocorrelation of sites, both Geary-C randomization tests for each MOTU (Thioulouse et al., 1995) and a multivariate spatial autocorrelation test for communities (spdep package; Smouse & Peakall, 1999) were processed. Isolation by distance was tested through correlation between geographical distances (calculated from GPS coordinates; see Fig. 1 and Table S1) and Bray–Curtis distances through a Mantel test (Mantel, 1967). To ensure that residuals of all generalized linear models (GLMs) were not spatially autocorrelated, a vector of spatial weight for neighboring sites was computed, and then used to measure the maximum likelihood of a spatial simultaneous autoregressive lag model (LeSage & Pace, 2010). This vector was integrated in the variance partition tests (used to test host and habitat effect on beta diversity). Additionally, a vector of neighboring sites according to their altitude was used to perform autocorrelation tests for species and communities on an elevation gradient (Fig. 2). Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
New Phytologist
Research 1231
Fig. 2 Ectomycorrhizal (ECM) fungal species distribution range along an altitudinal gradient over the whole sampling. Only species occurring on more than one tree are represented. Stars indicate species whose distribution was correlated with altitude. Red squares, species associated with the Alnobetula subgenus; green circles, species associated with the Alnus subgenus; black circles, no or unknown specificity. Solid lines, elevation range reported from ectomycorrhizas; dotted lines, elevation range reported from sporocarp observations (P-A. Moreau, pers. obs., herbarium Lille Institute of Pharmacy (LIP), France).
Host and habitat effects on alpha diversity To estimate alpha diversity, the least sampled sites (with less than three trees producing sequences) were removed from the dataset. This threshold was determined by testing the sample size effect on species richness in a preliminary analysis. Species richness (number of MOTUs per site) and Simpson diversity index (Simpson, 1949; described as the community evenness) were computed (diversity function of the Vegan package in R 2.3.4.4; Dixon, 2009). The species richness was square-root-transformed to meet the assumption of homoscedasticity. A GLM was built to test host, climate, geographic or habitat effect on alpha diversity indices (Table 1). Diversity indices were assumed to follow a Gaussian distribution, and normality of residuals was tested through a Shapiro–Wilk test (Royston, 1982). A Fisher F-test tested the significance of each factor included in the GLM. A principal component analysis (PCA) on soil, geographic and climatic parameters, and diversity indices was used to test if any of these parameters were significant drivers of alpha diversity. Soil, geographic and climatic parameters, and host/habitat factors were included in a GLM, and their significance was tested through an F-test. Host and habitat effects on beta diversity To estimate the beta diversity, which describes the dissimilarity between communities, singletons (MOTUs found only on one Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
Table 1 Effect of host or site characteristics on alpha diversity (molecular operational taxonomic units (MOTUs), richness) or beta diversity (Bray– Curtis distance and nonmetric multidimensional scaling (NMDS) structure) for all MOTUs Diversity index
df
Deviance
F-statistics
P-value
Species richness (square root), all MOTUs, 23 sites Type of soil 1 1.688 9.20 0.023 Residuals 22 8.705 Community evenness (square root), all MOTUs, 23 sites Type of soil 1 0.011 19.25 0.004 Region 2 0.007 6.94 0.027 Altitude 1 0.005 9.81 0.020 Longitude 1 0.011 19.53 0.004 Residuals 22 0.058
Significance * ** * * **
Significance: *, P < 0.05; **, P > 0.01; ***, P > 0.001.
tree in the whole dataset) were removed from the analyses, and Bray–Curtis distances were computed between communities per site (Bray & Curtis, 1957). Variance analysis of these distances was performed through a multivariate permutation test (Adonis test, Vegan package in R; Table 2), including host, habitat, geographic and climatic parameters, and sampling period as factors (Table 2). The effect of host species on site ordination was illustrated by a nonmetric multidimensional scaling (NMDS, metaMDS function; Fig. 3) on Bray–Curtis distances. Correlation of New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
New Phytologist
1232 Research Table 2 Effect of host and site characteristics on beta diversity (Bray–Curtis distance) for MOTUs recorded from at least two trees
Table 3 Effect of host and site characteristics on community structure (NMDS structure), including soil characteristics measured on 19 sites
Adonis on Bray–Curtis distances, all MOTUs > 1, 39 sites
Correlation with NMDS structure (19 sites)
R2
P-value
Significance
df
Deviance
F-statistics
P-value
Significance
Subgenus Host species Type of soil Type of site Longitude Residuals
1 3 1 5 1 38
1.011 1.695 0.409 1.580 0.474 11.553
4.368 2.441 1.769 1.365 2.050
0.001 0.001 0.028 0.029 0.010
*** *** * * **
Subgenus Host species Type of soil Type of site pH N C OM P K
0.485 0.316 0.189 0.522 0.255 0.333 0.408 0.449 0.558 0.404
0.001 0.001 0.021 0.001 0.085 0.032 0.010 0.006 0.003 0.018
*** *** * ***
Significance: *, P < 0.05; **, P > 0.01; ***, P > 0.001.
* ** ** ** *
Significance: *, P < 0.05; **, P > 0.01; ***, P > 0.001. C, carbon; K, potassium; N, nitrogen; P, phosphate; OM, organic matter soil content.
Fig. 3 Multivariate analysis of community structure based on nonmetric multidimensional scaling (NMDS) on molecular operational taxonomic units (MOTUs) occurring on more than one tree. Ellipse delimits the 95% confidence interval around centroids for each class of host species. The grid size is 0.5 for both axes.
NMDS structure with soil and climate parameters was tested by permutation tests (environmental fitting test, envfit function, 999 permutations; Table 3, Fig. 4). The variance of the beta diversity was partitioned with respect to significant factors and parameters identified by Adonis and environmental fitting tests, and distance between sites (spatial weights for sites, computed for spatial autocorrelation test). The variance partition test (Varpart function in Vegan; Dixon, 2009) relies on a redundancy analysis ordination of Hellinger distances between sites (Legendre & Gallagher, 2001; Fig. 5). Specificity and distribution pattern of ‘specific’ Alnus-associated ECM fungi The MOTU specificity pattern was quantified by the number of trees colonized per MOTU and per host species (Fig. 6). For Alnicola, Lactarius and Alpova genera, preference of their species for a given host subgenus was tested by a v2 test (tables not shown, computed from Table S3). To test the respective influence of host and environmental parameters on ECM fungi distribution, the same community analysis was run on reduced New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
Fig. 4 Nonmetrical multidimensional scaling (NMDS) on molecular operational taxonomic units (MOTUs) occurring on more than one site, for 19 sites, and correlation with host, habitat factor and soil parameters. Only significant correlations between NMDS structure and soil parameters are shown (P < 0.05). OM, organic matter content; C, carbon content; P, phosphate; K, total potassium content. The significance of host and habitat factor is reported.
datasets (NMDS and environmental fitting test, analysis of Bray–Curtis distances by Adonis): a highly host-specific genus (Alnicola, Fig. S1a), and a generalist genus (Tomentella, Fig. S1b). The results of spatial and altitudinal autocorrelation tests (performed on each MOTU) were used to describe MOTU distribution on a regional scale and along an elevation gradient (Fig. 2). Estimation of unseen diversity The Chao diversity index (Chao et al., 2004) and the number of unseen MOTUs per species were estimated (specpool and Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
New Phytologist
Fig. 5 Variance partition of Hellinger distances between 19 sites, comparing host factors (host species and subgenus), soil parameters (pH, nitrogen (N), phosphorus (P), sulfur (S), potassium (K) and organic matter (OM) soil content) and geographical position effects (altitude and longitude). The nonsignificant interactions are not represented, such as the interaction with distance between sites (measured by the spatial weights vector). The results of the same analysis performed on Alnicola spp. communities are reported into brackets.
(a)
(b) Fig. 6 Molecular operational taxonomic unit (MOTU) distribution and degree of specificity. (a) Number of trees colonized per MOTU, and list of MOTUs colonizing together 50% of root tips (black) or 75% of root tips (black and grey); (b) relative proportion of trees colonized per MOTU and per host species (red, Alnus alnobetula ssp. alnobetula; dark blue, A. alnobetula ssp. suaveolens; cyan: Alnus cordata; red, Alnus glutinosa; black, Alnus incana).
estimateR function, Vegan package in R 2.3.4.4; Dixon, 2009) to determine if more ECM fungi could be expected than observed for each Alnus sp. (Table 4).
Results
Research 1233
contaminants (nonECM) and 175 sequences belonged to nonAlnus roots. A total of 1178 sequences were kept for further analysis. Overall, a mean of 4.23 trees per site produced sequences, and on average 7.1 sequences were recorded per tree. Eighty-six MOTUs were identified in the whole dataset. Forty-five were named at the species level by similarity with sporocarp sequences from the LIP herbarium (including three unpublished taxa), 41 at the genus level only (as ‘sp.’) by similarity with environmental Genbank sequences. Names, accession numbers and maximum identity are reported in Table S3. The unidentified taxa are represented in only one or very few mycorrhizas, and mostly belong to the genera Tomentella, Inocybe, Clavulina, Russula, and to Pezizales (Geopora, Humaria, Pachyphloeus, Tarzetta). Among the 86 MOTUs, 27 occurred only on one tree (singletons). Autocorrelation test between sites and geographical structure The whole dataset was not spatially autocorrelated (based on co-occurrence) according to a multivariate randomization test (P = 0.067). In addition to this, the test was not significant when rare species were removed (P = 0.062), showing that the sites were distant enough from each other to be considered as independent. Bray–Curtis distances were not correlated with geographic distance, either for the whole community (Mantel test on geographic distances, P = 0.663; on log-transformed distances, P = 0.772) or for A. glutinosa communities (Mantel test on logtransformed geographic distances, P = 0.074). At the MOTU level, the spatial autocorrelation hypothesis was rejected (GearyC permutation test), except for three rare species (Inocybe muricellata, P = 0.007; Inocybe sp.2, P = 0.047; Melanogaster rivularis, P = 0.027) that were sampled on two or three sites < 35 km apart. Considering that Inocybe muricellata basidiocarps were also collected elsewhere in France (P-A. Moreau, unpublished data), and that Melanogaster rivularis is considered as a Corsican endemic species only known from the two sampled localities, no significance is given to these positive correlations. Along the altitudinal gradient, communities were not autocorrelated (multivariate randomization test, P = 0.928). However, the distribution of 13 MOTUs was autocorrelated with altitude (P < 0.05 for individual autocorrelation tests, Fig. 2). Nine species, such as Alpova alpestris and Alnicola pallidifolia, associated with the subgenus Alnobetula, were mainly located at high altitude (over 1400 m, see Fig. 2). Four species, Alnicola alnetorum, A. badiofusca/umbrina and Alnicola citrinella, associated with the subgenus Alnus, were restricted to lower sites (< 1400 m; Fig. 2). The same figure is completed by unpublished sporocarp observations (from herbarium LIP), which enlarge the distribution range of most sporocarp-forming species, especially in the genera Cortinarius and Paxillus (Fig. 2, dotted lines).
Gamma diversity and sample overview A total of 4419 ectomycorrhizas were collected, of which 3717 were extracted and used for PCR (84% of the collected ectomycorrhizas). Out of the total pool of amplified DNAs, 1438 good-quality ITS sequences were obtained. Eighty-five were Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
Comparison of host species and habitat effects on community alpha and beta diversities Variance analyses were used to test possible influences of host identity and habitat parameters on the two selected descriptors of New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
New Phytologist
1234 Research Table 4 Comparison of community alpha diversity among host species Host species
No. of sites
All MOTUs
Chao
Species/Chao
Species richness per site
Alnus alnobetula ssp. alnobetula A. alnobetula ssp. suaveolens A. cordata A. glutinosa A. incana
5 2 5 23 4
24 13 18 59 30
49.6 9.0 19.4 3.0 30.1 5.9 96.6 8.5 58.6 8.8
0.48 0.67 0.60 0.62 0.51
7.4 4.2 9.0 0.0 5.6 3.8 7.8 4.6 11.0 4.6
Host species
No. of sites
MOTUs on more than one tree
Chao
Species/Chao
Species richness per site
A. alnobetula ssp. alnobetula A. alnobetula ssp. suaveolens A. cordata A. glutinosa A. incana
5 2 5 23 4
21 13 14 44 24
37.9 6.4 19.4 3.0 22.0 3.7 52.6 2.9 38.0 4.9
0.55 0.67 0.64 0.84 0.63
11.6 2.5 14.5 0.7 5.6 1.8 12.4 2.5 14.2 2.5
Diversity indices per site (species richness), extrapolated richness per host (Chao) and proportion of observed taxa (species richness/Chao), for all molecular operational taxonomic units (MOTUs) or only MOTUs colonizing more than one tree.
community alpha diversity (species richness and Simpson index). The species richness (number of MOTUs per site) was significantly different between organic and mineral microenvironments (19% of the deviance was explained by soil categories), but not among host species or subgenera (Table 1). The Simpson diversity index – more sensitive to the abundance of dominant species – was significantly different between soil categories (17% of the deviance explained, see Table 1), but was also influenced by geographical parameters such as the longitude (17% of the deviance explained), the region (8.2%) and the altitude (5.8%). For the two alpha diversity indices, GLM residuals were normal (P > 0.05 for both Shapiro–Wilk tests), and not spatially autocorrelated (all P-value > 0.05). PCA on diversity indices, climate and soil parameters pointed out a similar projection of species richness with precipitation (coldest and warmest quarter, data not shown) and temperature of the wettest quarter. When included in the GLMs, climate, soil parameters, and sampling period did not significantly explain alpha diversity variations. An Adonis multivariate permutation test was used to measure the effect of host identity and habitat on Bray–Curtis dissimilarities (a measure of beta diversity). Host, habitat and longitude shaped beta diversity – host species accounting for 14.6% of Bray–Curtis distance variance between sites, host subgenus for 8.7%, site categories for 16.3%, and longitude for 6.9% (Table 1). The host effect detected by the Adonis test was illustrated by the NMDS (Fig. 3). Communities of A. alnobetula (subgenus Alnobetula), A. incana and A. glutinosa (subgenus Alnus) were significantly separated. By contrast, A. cordata communities partly overlapped those of A. glutinosa and were highly variable between sites (Fig. 3). Among climatic parameters, only the minimum temperature of the coldest month and the annual mean temperature had a significant effect on dissimilarities when tested without host and soil parameters (P = 0.005 and P = 0.012, respectively, Adonis test). However, none of climatic parameters correlated with NMDS structure according to the environmental fitting test. New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
Abiotic soil parameters such as organic matter (OM) and carbon content, phosphorus, potassium and nitrogen concentrations were all significantly correlated with NMDS structure (Fig. 4, Table 3). The same analysis pointed out strong differences between host subgenera and host species (Fig. 4, Table 3), and also confirmed the differences between the site and soil microenvironment categories. An NMDS restricted to A. glutinosa (not shown) revealed a significant correlation of NMDS structure with pH (r2 = 0.47, P = 0.04). To compare the respective roles of host identity, habitat parameters and distance on community structure, a partition of Hellinger distances was computed on 19 sites. In all, only 26% of the variance was explained (Fig. 5), mainly by host and soil parameters (13% of variance explained for each), and, to a lesser extent, by longitude and altitude (0.08% all together), and distance between sites (0.03%). The interaction between host, soil and altitude accounted for 8% of the variance, confirming the overlap between these factors (Fig. 5). Specificity and distribution pattern of ‘specific’ Alnus-associated ECM fungi The most abundant MOTUs were shared among host species such as Tomentella aff. sublilacina, A. citrinella and the six other most frequent MOTUs (Fig. 6). Nineteen out of the 20 most frequent MOTUs were shared among host species (Fig. 6). Fortyone species were recorded from one single host, among which 30 were recorded from a single tree. Chi-squared tests confirmed the host subgenus-specificity (P < 0.001 for Alnicola matrix and for Lactarius matrix). No test was performed on Alpova, as mycorrhizas were rarely found in our study. To test the respective influence of host and environmental parameters on ECM fungi distribution, a partial community analysis was attempted on the two dominant fungal genera: Alnicola, mainly host-specific at the subgenus level, and Tomentella, in which the dominant species (T. cf. ellisii 1, Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
New Phytologist T. cf. ellisii 2, T. cf. sublilacina, T. cf. stuposa) did not show preferences between alder subgenera or species. For the Alnicola dataset, communities associated with different host subgenera (16.0% of the Bray–Curtis differences explained, Adonis test, Table 2), host species (11.7%) and category of site (16.3%). NMDS structure was also correlated with an altitudinal gradient (r2 = 0.2480, P = 0.006; Table 1) and with several soil parameters (Table 2, Fig. S1b), such as soil pH, organic matter content, carbon, sulfur and nitrogen concentrations. The partition of variance of Hellinger distances between 19 sites revealed the significant effect of host identity (23% of the variance explained; Fig. 5), followed by soil parameters (12% all together), geographical parameters (8%) and their interaction (13%). For the Tomentella dataset (Fig. S1b), nitrogen, carbon, sulfur and organic matter content were significantly correlated with the NMDS structure (P < 0.05 for all, environmental fitting test on NMDS). Tomentella communities were significantly different between organic and more mineral sites (Adonis test on Bray– Curtis distances between sites, F = 3.1910, P = 0.014 and 8.6% of the variance explained) but not among sites, regions, host species and subgenera, confirming the absence of host specificity below genus level for this genus. Estimation of unseen diversity When taking all MOTUs into account, sampling was not saturated, as the number of observed species was excluded from Chao confidence interval, and 57.7 7.7% of MOTUs were detected for most host species (Table 4). When removing singletons, the number of observed species per host was inside the Chao confidence interval, showing that the dominant species were sampled. Singletons were mainly recorded from A. glutinosa (Table 2).
Discussion Gamma diversity and outlines of Alnus-associated ECM communities A large diversity of ecological conditions and five different host species and subspecies were investigated. A relatively high diversity of fungal MOTUs was expected; however, only 86 MOTUs were detected. This is twice to four times higher than former studies of Alnus-associated ECM communities from Estonia (Tedersoo et al., 2009 (40 taxa)) and North America (Kennedy & Hill, 2010 (14 taxa); Kennedy et al., 2011 (23 taxa)). Chao estimates suggested that nearly half of the potential MOTUs were detected (Table 2), and singletons were relatively few (27 MOTUs), supposing that the sampling effort was sufficient to detect most Alnus-associated ECM fungi present in France. Only five Alnus-associated ECM species reported from the myco-ecological literature were not found: Amanita friabilis, Alnicola badia, Alnicola suavis, Alpova corsicus, and Cortinarius kuehneri (Brunner & Horak, 1990). Moreover, some species thriving on several sampling sites, such as Gyrodon lividus, Paxillus spp., Alpova alpestris, and Alnicola salabertii, were only detected on a few Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
Research 1235
mycorrhizas. It is likely that additional sampling would better detect some of these ‘specific’ species, possibly overlooked as a result of irregular spatial or temporal distribution. Among the singletons, wide-ranged species such as Clavulina spp. were also recorded (e.g. from the Sainte Croix Volvestre site, where A. glutinosa is growing along with conifers and other broad-leaved trees). Most singletons were not documented by sporocarps (Table S3) and do not belong to usual alder-associated genera. Moerover, a few clues from experimental in vitro studies also show that Alnus spp. can form ectomycorrhizas with generalist fungi (e.g. A. glutinosa with Pisolithus tinctorius; Molina, 1981). Reciprocally, Alnus ECM fungi can colonize Betula occidentalis in experimental plots (Bogar & Kennedy, 2013). However, such associations rarely occur in nature, supporting the hypothesis of a high specificity of the whole alder–ECM community. All sites were dominated by a few species of Tomentella and Alnicola, as already revealed in previous studies of alder-associated ECM communities in both Europe and America (Becerra et al., 2005; Tedersoo et al., 2009; Kennedy & Hill, 2010; Kennedy et al., 2011). Phylogenetic studies by Rochet et al. (2011) of Alnicola, as well as by Kennedy & Hill (2010) of Tomentella, have already demonstrated that ECM species found on alders do not occur on other host trees. This observed dominance of a few specific ECM fungi, belonging to few lineages, is also a unique feature of the alder lineage worldwide. Hence, the dominance of such fungi shows that Alnus–ECM fungi mutualism is an isolated system, highly conserved at a large geographical scale and probably not connected to common mycorrhizal networks involving other local tree species. The drivers of Alnus-associated ECM community diversity and specificity In Europe, Alnus populations often grow in patches. Thus, Alnus associated ECM fungi might be highly dependent on the Alnus host, and isolated between sites. If dispersal is not limited for these ECM fungi, then community alpha diversity may be shaped either by neutral processes or by a combination of host, habitat and competition. Our sample, covering a broad range of habitats and hosts at a regional scale, was adequate for testing these alternative hypotheses, although competition and other interspecific interactions would require experimental settings. At the regional scale (southern France), spatial autocorrelation tests and correlation between Bray–Curtis distances and geographic distance were nonsignificant. At the MOTU level, dominant Tomentella spp. and most Alnicola spp. occurred nearly everywhere, such as A. citrinella, present in the Alps, southwest France and Corsica. Interestingly, sequences from A. citrinella were 100% identical to Genbank sequences recorded for fungi found in other Alnus populations, in Europe and the Middle East (e.g. JN198094). Even though gene flow should be measured with markers other than ITS, these results suggest that distance itself might not limit the dispersal of alder–ECM fungi. Furthermore, the community evenness was lower for high longitude and in Corsica, and changes of beta diversity correlated New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
New Phytologist
1236 Research
with longitude (Table 1). Corsica as an island was shown to host some endemic hypogeous species such as A. corsicus (Moreau et al., 2013), but no (or a few) ectomycorrhizas were found for these species. Moreover, communities were slightly less diverse in Corsica, and characterized by a lower evenness. Therefore, the geographic position of Corsica may explain the observed effect of longitude and shows that geography, more than distance itself, shapes these communities. The topography also played an important role, as altitude influenced both evenness and beta diversity variation. Altitude appeared to be a key determinant of the distribution of several ECM fungi, such as Lactarius and Alnicola spp. (Fig. 2). While Bahram et al. (2012) found a lower species richness at high altitude in the Hyrcanian forest, our study rather suggests a shift of ECM communities over 1400 m, with the occurrence of the subgenus Alnobetula, whatever the region. Therefore, the gap between habitats naturally colonized by the two Alnus subgenera in western Europe is also illustrated by fungal species distribution (Fig. 2). Among sites, the ECM species richness was variable, but not determined by the tree species. The alpha diversity was low (maximum of 16 MOTUs per site, 8.4 on average), as previously reported (Tedersoo et al., 2009). This common trait of alder species suggests a similar constraint on ECM communities, which should be investigated by comparing alders with other host trees. The differences between host subgenera and species were revealed by community composition. However, most MOTUs were shared among hosts, at least within a subgenus, such as Alnicola spp. and Lactarius spp. (Fig. 6). Then Alnus ‘specific’ ECM fungi showed preferences towards the host rather than strict host-species specificity. This pattern is relatively common for ECM fungi (Dickie, 2007) and is also suggested by meta-analyses on several host trees (Ishida et al., 2007; Tedersoo et al., 2012). In this study, focused on an exceptional model of specific fungi community interaction, by comparing several congeneric host species at a regional scale, both the ‘host species’ effect on communities and the ‘preferences’ of ECM fungi were confirmed, as for other host trees. These results show that the host-species effect cannot result only from host-species specificity, possibly because, at a regional scale, long-distance dispersal may limit the selection for a higher specificity. Habitat was described both qualitatively and quantitatively, but none of the tested climate and soil parameters correlated with the alpha diversity variations. A larger number of species were generally found on organic soils (e.g. in peat bogs) than in mineral sites (e.g. stream sites), as already reported in the published literature (Tedersoo et al., 2009). Evenness was also higher on organic soils in forest and peat bogs than in mineral substrates (effect of site, Table 1). Thus, other relevant parameters should be measured to identify those that may explain alpha diversity variations between organic and mineral soils, such as soil water content or anoxia. Focusing on factors favoring a high alpha diversity, interpretable as a local patrimonial richness, is of interest in nature conservancy programs, such as riparian forest conservancy in the European Habitat Directive (e.g. Council Directive 92/43/EEC Art. 17 (910E)). New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
Beta diversity was partly explained by several soil parameters such as organic matter content (Fig. 3, Table 1). However, for each host species, driving soil factors were different. The A. glutinosa communities were shaped by pH, carbon, nitrogen, phosphorus and organic matter content, as shown by Tedersoo et al. (2009), whereas A. alnobetula communities were partly explained by high potassium concentration on most sampled sites (Fig. 4). The distribution of Tomentella species, shown to be indifferent to host identity (Fig. S1b), appeared to be shaped by nitrogen, sulfur, carbon and soil organic matter content. The distribution of Alnicola spp. was mainly influenced by host identity (23% of variance, Fig. 5), but also by soil factors such as pH, organic matter, nitrogen, carbon and sulfur concentration (Figs 5, S1b). Curiously, climatic parameters, although variable in our sample, failed to explain alpha and beta diversity variations. The minor impact of climate in our study stresses the major role of host identity and soil parameters as drivers of these communities at a regional scale. All in all, host identity, soil parameters, and geographical position accounted for 13, 13 and 9% of the beta diversity variance, respectively (Fig. 5). Interestingly, these factors all explained beta diversity variation in similar proportions, and do not exclude each other. In the future, competition or neutral processes should also be considered, as a large part of beta diversity variation remained unexplained. Moreover, by comparing Alnicola and Tomentella, two dominant genera, our study shows that different processes may explain ECM fungi distribution and that specificity of Alnicola spp. may rely not only on host but also on habitat preferences. Disentangling host and habitat effects on alders To confirm host effect independently of altitude and soil in situ, different alder species should be sampled in naturally mixed populations. In the Alps, A. alnobetula and A. incana grow together in some sites, but these situations are often located at the limits of their respective distributional range. A better situation is found in Corsica where the ecology of A. cordata matches that of A. glutinosa along rivers or roadsides. Eight MOTUs were found on both A. cordata and A. glutinosa in Corsica – four were shared locally and four were shared among sites (differences between hosts were still significant; Adonis test on Bray–Curtis distance, F = 2.52, P = 0.004). The shared fungi were dominant MOTUs without host preference (Tomentella spp.) or associated with the Alnus subgenus (Lactarius cyathuliformis, A. citrinella). To extrapolate these few cases, when co-occurring, each Alnus sp. seems to share dominant MOTUs such as Tomentella spp., but keeps a few specific MOTUs. Interestingly, except for Alnicola salabertii and Clavulina spp., the ‘nonshared’ MOTUs of A. cordata are Pezizales, that is, fungi not considered as host-specific to date (Tedersoo et al., 2006, 2009). The occurrence of these Pezizales suggests that A. cordata, which exhibited a low diversity of MOTUs, hosts less specific associates than other alders. To understand this particularity, further investigations on A. cordata in its native areas (Corsica and Italy) are required, as well as in its increasingly numerous plantations in continental Europe. Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
New Phytologist Conclusion By investigating one of the most specific plant–ECM fungal interactions, at a regional scale, this study demonstrates that: host-species specificity is rare and ECM fungi exhibit preferences for a host subgenus; host identity, altitude and soil drive ECM community diversity and structure; specific ECM fungi exhibit a habitat preference within host subgenus distribution range; climate has a minor influence at a regional scale; and still a large part of the variability remains unexplained. These findings can potentially be applied to most ECM fungi, but a few aspects are particular to Alnus, such as a low alpha diversity and high similarity between distant communities. The conserved dominance of a few specific ECM fungi across continents may suggest high gene flow between populations, a constant and uniform selective pressure, or a drastic bottleneck at the beginning of the ECM– Alnus association. These hypotheses have to be tested to refine the comprehension of processes favoring or limiting the selection for more specificity in ECM mutualism at a regional scale.
Acknowledgements The authors thank Dr Teodoro Mara~ non and Luis V. Garcıa at (IRNAS) – CSIC (Spain) for technical help with the soil analyses, Gilles Corriol (Conservatoire Botanique National des Pyrenees et de Midi-Pyrenees) and Nicolas Cayssiols (ADASEA Aveyron) for helping us locate sites in the Pyrenees and Aveyron, respectively, and La€etitia Hugot (Conservatoire Botanique National de Corse) and Franck Richard (Centre d’Ecologie Fonctionnelle et Evolutive, Montpellier) for their help during sampling in Corsica. We thank Anders Dahlberg (Uppsala University), Christophe Thebaud and Jer^ome Chave (Laboratoire Evolution et Diversite Biologique, Toulouse), Marc-Andre Selosse (Centre d’Ecologie Fonctionnelle et Evolutive, Montpellier), Henry Beker, Peter Kennedy, Leho Tedersoo and two anonymous referees for discussions around this study. Financial support was provided by the Universite Paul Sabatier Toulouse 3, the Region Midi-Pyrenees (grant APP RRTT 2008-2011 to M.G.), the CNRS (grant EC2CO 2012 to M.R.) and the Labex TULIP.
References Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic alignment search tool. Journal of Molecular Biology 215: 403–410. Baar J, Groenendael JM, Roelofs JGM. 2000. Are ectomycorrhizal fungi associated with Alnus of importance for forest development in wet environments? Plant Biology 2: 505–511. Bahram M, P€ olme S, K€oljalg U, Zarre S, Tedersoo L. 2012. Regional and local patterns of ectomycorrhizal fungal diversity and community structure along an altitudinal gradient in the Hyrcanian forests of northern Iran. New Phytologist 193: 465–473. Becerra A, Pritsch K, Arrigo N, Palma M, Bartoloni N. 2005. Ectomycorrhizal colonization of Alnus acuminata Kunth in northwestern Argentina in relation to season and soil parameters. Annals of Forest Science 62: 325–333. Bogar LM, Kennedy PG. 2013. New wrinkles in an old paradigm: neighborhood effects can modify the structure and specificity of Alnusassociated ectomycor-rhizal fungal communities. FEMS Microbiology Ecology. 83: 767–777. Ó 2013 The Authors New Phytologist Ó 2013 New Phytologist Trust
Research 1237 Bray JR, Curtis JT. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecological Monographs 27: 325. Brunner I, Horak E. 1990. Mycoecological analysis of Alnus-associated macrofungi in the region of the Swiss National Park as recorded by J. Favre (1960). Mycologia Helvetica 4: 111–139. Bruns TD, Bidartondo MI, Taylor DL. 2002. Host specificity in ectomycorrhizal communities: what do the exceptions tell us? Integrative and Comparative Biology 42: 352–359. Chao A, Chazdon RL, Colwell RK, Shen TJ. 2004. A new statistical approach for assessing similarity of species composition with incidence and abundance data. Ecology Letters 8: 148–159. Dickie IA. 2007. Host preference, niches and fungal diversity. New Phytologist 174: 230–233. Dilly O, Bach H-J, Buscot F, Eschenbach C, Kutsch W, Middelhoff U, Pritsch K, Munch J. 2000. Characteristics and energetic strategies of the rhizosphere in ecosystems of the Bornh€oved Lake district. Applied Soil Ecology 15: 201–210. Dixon P. 2009. VEGAN, a package of R functions for community ecology. Journal of Vegetation Science 14: 927–930. Gardes M, Bruns TD. 1993. ITS primers with enhanced specificity for basidiomycetes – application to the identification of mycorrhizae and rusts. Molecular Ecology 2: 113–118. Helm DJ, Allen EB, Trappe JM. 1996. Mycorrhizal chronosequence near Exit Glacier, Alaska. Canadian Journal of Botany 74: 1496–1506. Hijmans RJ, van Etten J. 2012. raster: Geographic data analysis and modeling. R package version 2.0-31. [WWW document] URL http://CRAN.R-project.org/ package=raster [accessed on 24 December 2012]. Hoeksema JD. 1999. Investigating the disparity in host specificity between AM and EM fungi: lessons from theory and better-studied systems. Oikos 84: 327–332. Hoeksema JD. 2010. Ongoing coevolution in mycorrhizal interactions. New Phytologist 187: 286–300. Horton T, Bruns T. 1998. Multiple host fungi are the most frequent and abundant ectomycorrhizal types in a mixed stand of Douglas fir (Pseudotsuga menziesii D. Don) and bishop pine (Pinus muricata D. Don). New Phytologist 139: 331–339. Ishida TA, Nara K, Hogetsu T. 2007. Host effects on ectomycorrhizal fungal communities: insight from eight host species in mixed conifer broadleaf forests. New Phytologist 174: 430–440. Kennedy PG, Garibay-Orijel R, Higgins Logan M, Angeles-Arguiz R. 2011. Ectomycorrhizal fungi in Mexican Alnus forests support the host co-migration hypothesis and continental-scale patterns in phylogeography. Mycorrhiza 21: 559–568. Kennedy PG, Hill LT. 2010. A molecular and phylogenetic analysis of the structure and specificity of Alnus rubra ectomycorrhizal assemblages. Fungal Ecology 3: 195–204. Kretzer A, Li Y, Szaro TM, Bruns TD. 1996. Internal transcribed spacer sequences from 38 recognized species of Suillus senu lato: phylogenetic and taxonomic implications. Mycologia 88: 776–785. Legendre P, Gallagher ED. 2001. Ecologically meaningful transformations for ordination of species data. Oecologia 129: 271–280. LeSage JP, Pace RK. 2010. Spatial conometric models. In: Fischer MM, Getis A, eds. Handbook of applied spatial analysis. Berlin, Heidelberg: Springer Berlin Heidelberg, 355–376. Mantel N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209–220. Miller MA, Pfeiffer W, Schwartz T. 2010. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop (GCE). New Orleans, 1–8. Molina R. 1981. Ectomycorrhizal specificity in the genus Alnus. Canadian Journal of Botany 59: 325–334. Moreau P-A, Peintner U, Gardes M. 2006. Phylogeny of the ectomycorrhizal mushroom genus Alnicola (Basidiomycota, Cortinariaceae) based on rDNA sequences with special emphasis on host specificity and morphological characters. Molecular Phylogenetics and Evolution 38: 794–807. Moreau P-A, Welti S, Peric B, Jargeat P, Manzi S, Vizzini A. 2013. Alpova komoviana (Boletales, Paxillaceae), a new sequestrate fungus from Montenegro, with a revised phylogeny of the genus in Europe. Mycological Progress 12: 109–119. New Phytologist (2013) 198: 1228–1238 www.newphytologist.com
New Phytologist
1238 Research Nilsson RH, Kristiansson E, Ryberg M, Hallenberg N, Larsson K-H. 2008. Intraspecific ITS variability in the kingdom fungi as expressed in the international sequence databases and its implications for molecular species identification. Evolutionary Bioinformatics 4: 193–201. Pritsch K, Boyle H, Munch JC, Buscot F. 1997. Characterization and identification of black alder ectomycorrhizas by PCR/RFLP analyses of the rDNA internal transcribed spacer (ITS). New Phytologist 137: 357–369. R Development Core Team. 2008. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. URL: http://www.R-project.org. Rameau J-C. 1994. Typologie phytosociologique des habitats forestiers et associe s – types simplement repre sentatifs ou remarquables sur le plan patrimonial, Vol. 5. Paris, France: Ministere de l’Agriculture et de la P^eche; Nancy, France: ENGREF. Rochet J, Moreau P-A, Manzi S, Gardes M. 2011. Comparative phylogenies and host specialization in the alder ectomycorrhizal fungi Alnicola, Alpova and Lactarius (Basidiomycota) in Europe. BMC Evolutionary Biology 11: 40. Royston JP. 1982. An extension of Shapiro and Wilk’s W test for normality to large samples. Journal of the Royal Statistical Society. Series C (Applied Statistics) 31: 115–124. Simpson EH. 1949. Measurement of diversity. Nature 163: 688. Smith SE, Read DJ. 2008. Mycorrhizal symbiosis, 3rd edn. London, UK: Academic Press, 800. Smouse PE, Peakall R. 1999. Spatial autocorrelation analysis of individual multiallele and multilocus genetic structure. Heredity 82: 561–573. Stamatakis A, Hoover P, Rougemont J. 2008. A rapid bootstrap algorithm for the RAxML Web servers. Systematic Biology 57: 758–771. Tedersoo L, Bahram M, Toots M, DieDhiou AG, Henkel TW, KjøLler R, Morris MH, Nara K, Nouhra E, Peay KG et al. 2012. Towards global patterns in the diversity and community structure of ectomycorrhizal fungi. Molecular Ecology 21: 4160–4170. Tedersoo L, Hansen K, Perry B, Kjøller R. 2006. Molecular and morphological diversity of pezizalean ectomycorrhiza. New Phytologist 170: 581–596. Tedersoo L, Suvi T, Jairus T, Ostonen I, P~olme S. 2009. Revisiting ectomycorrhizal fungi of the genus Alnus: differential host specificity, diversity and determinants of the fungal community. New Phytologist 182: 727–735. Thioulouse J, Chessel D, Champely S. 1995. Multivariate analysis of spatial patterns: a unified approach to local and global structures. Environmental and Ecological Statistics 2: 1–14.
Thompson JN. 2001. Coevolution. In: Encyclopedia of life sciences. London, UK: Nature Publishing Group, 1–5. Timms R, Read AF. 1999. What makes a specialist special? Trends in Ecology and Evolution 14: 333–334. White TJ, Bruns TD, Lee S, Taylor JW. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols a guide to methods and applications. San Diego, CA, USA: Academic Press, 315–322.
Supporting Information Additional supporting information may be found in the online version of this article. Fig. S1 Nonmetrical multidimensional scaling (NMDS) of communities from 19 sites, showing the correlation with soil parameters, host and habitat factors, for Alnicola (a) and Tomentella (b) communities, according to environmental fitting tests. Table S1 Sampling site location and description, including climatic parameters Table S2 Soil chemical composition for a subset of sites Table S3 MOTU identification and occurrence among sites, including BLAST results Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.
New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-profit organization dedicated to the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews. Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged. We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average time to decision is
