SEARCH

SEARCH BY CITATION

Keywords:

  • climate change;
  • ectomycorrhizal symbiosis;
  • elevation;
  • precipitation;
  • species richness;
  • temperature

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Altitudinal gradients strongly affect the diversity of plants and animals, yet little is known about the altitudinal effects on the distribution of microorganisms, including ectomycorrhizal fungi.
  • By combining morphological and molecular identification methods, we addressed the relative effects of altitude, temperature, precipitation, host community and soil nutrient concentrations on species richness and community composition of ectomycorrhizal fungi in one of the last remaining temperate old-growth forests in Eurasia.
  • Molecular analyses revealed 367 species of ectomycorrhizal fungi along three altitudinal transects. Species richness declined monotonically with increasing altitude. Host species and altitude were the main drivers of the ectomycorrhizal fungal community composition at both the local and regional scales. The mean annual temperature and precipitation were strongly correlated with altitude and accounted for the observed patterns of richness and community.
  • The decline of ectomycorrhizal fungal richness with increasing altitude is consistent with the general altitudinal richness patterns of macroorganisms. Low environmental energy reduces the competitive ability of rare species and thus has a negative effect on the richness of ectomycorrhizal fungi. Because of multicollinearity with altitude, the direct effects of climatic variables and their seasonality warrant further investigation at the regional and continental scales.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Climate is the main factor determining biodiversity patterns at the global scale (Hawkins et al., 2003; Currie et al., 2004). Changes in the global climate have shifted the geographical distribution of plant and animal species both latitudinally and altitudinally (Parmesan & Yohe, 2003; Parmesan, 2006), causing extinctions in taxa that fail to migrate (Thomas et al., 2004). Information on diversity patterns along climatic gradients may enable the prediction of the response of species and entire communities to climate change (Grytnes & McCain, 2007).

Because the same climatic factors largely explain both the latitudinal and altitudinal gradients of diversity, altitudinal patterns are commonly used to predict large-scale latitudinal patterns (Lomolino et al., 2006) to avoid the confounding historical and geographical factors. Compared with latitudinal gradients, seasonality and temperature fluctuate less strongly along altitudinal gradients (Rahbek, 1995). Changes in ecosystems are more abrupt along altitudinal gradients which permit the migration of species within a few generations (Hewitt, 1996). However, low oxygen concentration, low air pressure and high UV radiation constrain climatic effects at high altitude. In addition, the land area at higher altitudes is severely reduced and may promote the extinction of critically small populations. Although the species diversity of various organisms usually declines with altitude, contrasting patterns occur in certain vascular plants, bryophytes and lichens (e.g., Bhattarai & Vetaas, 2003; Bruun et al., 2006; Desalegn & Beierkuhnlein, 2010). Therefore, the extrapolation of altitudinal diversity patterns across taxa seems premature.

Most of the current knowledge on altitudinal and latitudinal diversity patterns relies on macroorganisms, whereas microorganisms have received little attention. The few studies that have been performed have revealed contrasting altitudinal diversity patterns in soil and aquatic microorganisms: a decline in Acidobacteria (Bryant et al., 2008) but an increase in aquatic bacteria and diatoms (Wang et al., 2011) at higher altitudes. These studies suggest that different processes may drive the altitudinal diversity of microorganisms: for example, soil temperature has the strongest effect on Acidobacteria, whereas nutrients influence diatoms. Changes in climate may affect microorganisms directly or indirectly through shifts in vegetation. There is increasing evidence that biological interactions, including associations between micro- and macroorganisms, can be influenced by climate change (reviewed in Tylianakis et al., 2008).

Ectomycorrhizal (EcM) fungi play an essential role in nutrient cycling and ecosystem functioning in temperate forest ecosystems. In contrast with the general latitudinal pattern, EcM fungal diversity seems to decline from temperate to tropical ecosystems (Tedersoo & Nara, 2010). However, researchers have used different sampling strategies and molecular techniques, which elevate the error term in comparisons among studies. Thus, information on altitudinal patterns may shed additional light on the global climatic effects on the diversity and community composition of EcM fungi. It has been hypothesized that topography at local scales and climate at large scales may drive EcM fungal diversity and community composition (Lilleskov & Parrent, 2007). At the local scale, EcM fungi respond to various biotic and abiotic factors, such as host (Ishida et al., 2007; Tedersoo et al., 2008a), competition (Kennedy, 2010), dispersal limitation (Peay et al., 2007), soil nutrients (Toljander et al., 2006) and pollution (Parrent et al., 2006; Andrew & Lilleskov, 2009).

In this study, we aimed to characterize EcM fungal communities along the altitudinal gradient in the old-growth Hyrcanian forest of northern Iran, which is listed as a United Nations Educational, Scientific and Cultural Organization (UNESCO) World Natural Heritage site (http://whc.unesco.org/en/tentativelists/5214/). By sampling in three altitudinal transects, we tested the hypothesis that climatic factors underlying the altitudinal gradient drive species richness and community composition of EcM fungi at both the local (i.e. transect) and regional (northern Iran) scales, when differences in host communities, soil variables and spatial autocorrelation are taken into account. We further predicted that EcM fungal richness declines with increasing altitude, being particularly sensitive to temperature.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Sampling area

The Hyrcanian forests cover c. 18 500 km2 of mountainous slopes along the southern coasts of the Caspian Sea, extending from southeast Azerbaijan to the Golestan province of northeast Iran. The area contains refugia of the broad leaf forests that once covered most of the North Temperate Zone 25–50 million yr ago, in the early Cenozoic era (Leestmans, 2005; Ramezani et al., 2008). Sampling was performed during the autumn of 2008 and 2009 along three roughly straight transects in the Hyrcanian forests spanning the submontane and montane deciduous forests, and subalpine shrubs. The three transects were separated up to 410 km from east to west (Fig. 1; Table 1). Several non-EcM trees, such as Parrotia persica C.A. Mey., Acer velutinum Boiss., Buxus hyrcana Pojark., Taxus baccata L. and Ulmus glabra Huds., are integral components of the Hyrcanian forests. Dominant EcM host trees in different altitudes include Fagus orientalis Lipsky (600–2200 m above sea level), Carpinus betulus L. (200–2200 m), Quercus castaneifolia C.A.Mey. (50–1000 m), Q. macranthera Fisch. & C.A. Mey. (2000–2400 m), Tilia caucasicaRupr. (0–800 m) and Betula pendula Roth (2500–2700 m). Fagus orientalis and C. betulus are the most dominant trees in the area. Betula pendula is only present in Savadkuh at > 2500 m above sea level. Quercus macranthera dominates the transition zone between forest and grassland in Savadkuh and Nowshahr. All transects are located in protected natural sites.

image

Figure 1. Locations of the transects in the Hyrcanian forests of Iran.

Download figure to PowerPoint

Table 1.   Geography, altitudinal range, climate, number of ectomycorrhizal (EcM) host trees and the range of soil variables in the three transects
 AsalemNowshahrSavadkuh
Latitude37°36′, 37°41′36°27′, 36°36′36°22′, 36°59′
Longitude48°44′, 48°51′51°06′, 51°42′52°52′, 53°20′
Altitudinal range sampled (m a.s.l.)400–2000100–2400400–2700
Mean annual temperature (°C)11.21 ± 2.5614.2 ± 2.7313.58 ± 2.15
Mean annual precipitation (mm)552 ± 221542 ± 246237 ± 116
Number of EcM host trees245
pHKCl3.98–6.614.01–6.394.86–6.08
N (%)0.13–0.540.19–0.750.27–0.64
P (lactate dissolved; mg kg−1)0.12–66.730.12–101.875.05–49.30
K (mg kg−1)135.39–576.34139.71–564.31220.15–431.44
Ca (mg kg−1)1339.50–4576.30710.50–5154.201379.00–4351.38
Mg (mg kg−1)185.16–507.81152.89–1171.24105.59–398.09
Organic matter (%)4.20–13.665.30–20.646.87–15.79

Sampling design

Plots were established along transects with 200 m vertical intervals. To account for the co-occurrence of different vegetation types, replicate plots were established at certain altitudes. From each plot, seven soil samples (15 cm × 15 cm to 10 cm depth) were taken 0.5–2 m distant from the tree bases along a horizontal line at c. 20 m distant from each other to minimize the effect of spatial autocorrelation (Lilleskov et al., 2004; Bahram et al., 2011). Geographical coordinates and altitude were recorded using a GPS Garmin 60CSx (Garmin International Inc., Olathe, KS, USA). All samples were processed within < 48 h: roots were washed in tap water and separated from soil and debris. EcM root tips were separated into morphotypes based on color, shape, transparency, abundance of hyphae, rhizomorphs and cystidia under a stereomicroscope. Subsequently, two to five root tips from each morphotype per sample were collected and transferred to CTAB buffer (1% cetyltrimethylammonium bromide, 100 mM Tris-HCl (pH 8.0), 1.4 M NaCl, 20 mM ethylenediaminetetraacetic acid) for further analysis. Rhizosphere soil from all samples per plot was pooled and subjected to analysis of total nitrogen (N), Ca2+, Mg2+, K+, extractable phosphorus (P), pH and organic matter, as described in detail in Tedersoo et al. (2009).

Molecular analysis

One healthy root tip from each morphotype per sample was selected for DNA extraction with a Qiagen DNeasy 96 Plant Kit (Qiagen, Crawley, West Sussex, UK) according to the manufacturer’s instructions. Primer ITSOF-T (5′-cttggtcatttagaggaagtaa-3′), in combination with reverse primers LB-W (5′-cttttcatctttccctcacgg-3′), ITS4 (5′-tcctccgcttattgatatgc-3′) or ITS2 (5′-gctgcgttcttcatcgatgc-3′), was used to amplify the internal transcribed spacer (ITS) region. For low-quality sequences, PCR was re-performed using taxon-specific primers ITS4-Tom (5′-aactcggacgaccagaggca-3′), ITS4-Seb (5′-tcagcgggtartcctactc-3′), ITS4-Russ (5′-agcgggtagtctcaccc-3′), ITS4-Cg (5′-cacatggcaar- ggcaaccg-3′), ITS4-Clavu (5′-ggtagtcccacctgattc-3′) or LR3-Pez (5′-cmtcrggatcggtcgatgg-3′). If no reliable sequence was obtained, the nuclear 28S (nuLSU) rRNA gene was amplified using primers LR0R (5′-acccgctgaacttaagc-3′) and TW13 (5′-ggtccgtgtttcaagacg-3′). Morphotypes yielding no sequence were re-extracted and re-amplified up to twice more to identify as many species as possible. To confirm morphological identification of the host trees, the plant plastid trnL region was amplified and sequenced from one to two root tips per sample using primers TrnC (5′-cgaaatcggtagacgctacg-3′) and TrnD (5′-ggggatagagggacttgaac-3′). PCRs were performed based on Tedersoo et al. (2006). Primers ITS5 (5′-ggaagtaaaagtcgtaacaagg-3′), ITS4, ctb6 (5′-gcatatcaataagcggagg-3′) and TrnD were used for sequencing of the fungal ITS, nuLSU and plant trnL regions, respectively. PCR products were visualized on 1% agarose gels under UV light and purified using Exo-Sap enzymes (Sigma, St. Louis, MO, USA). Sequencher 4.9 software (GeneCodes Corp., Ann Arbor, MI, USA) was used to edit, trim and assemble raw sequences. Sequences were assigned to species based on a 97% similarity threshold and to EcM fungal lineages according to Tedersoo et al. (2010). All unique sequences are publicly available in EMBL under sequence accessions FR851992FR852374. The identification of fungi was perceived by running MegaBLAST and BLASTn searches of sequences against the International Sequence Databases (INSD) and UNITE (Abarenkov et al., 2010a), as implemented in the massblaster function in the PlutoF workbench (Abarenkov et al., 2010b).

Data analysis

Climatic data (i.e. mean annual temperature and precipitation) for each plot were extracted from a high-resolution interpolated database (Hijmans et al., 2005) using the software ArcGIS (ESRI, Redlands, CA, USA). Likewise, from bioclimatic variables, precipitation seasonality (coefficient of variation of monthly average rainfall) and temperature seasonality (coefficient of variation of monthly average temperature) were included in the analyses. It should be mentioned that there are limitations with the use of this database because of the small number of stations in mountainous areas, which may lead to inaccurate approximations of local climatic differences. Unfortunately, data with better resolution are unavailable for this and most other mountainous areas. Soil variables were log transformed, whereas slope was arcsine transformed. For aspect, sine (aspect) and cosine (aspect) were used following Legendre et al. (2009). All statistical analyses were performed within transects (local scale) and over all transects (regional scale).

Bray–Curtis and Euclidean dissimilarity indices were used to generate community distance matrices (i.e. EcM fungal and host communities) and environmental distance matrices (i.e. geographical, altitudinal distance and climatic variables), respectively. The effect of rare species was down-weighted by applying Hellinger transformation for community data (Legendre & Gallagher, 2001). To identify the main factors influencing EcM fungal community structure, community dissimilarity of EcM fungi among the plots was visualized in nonmetric multidimensional scaling (NMDS) graphs, and the environmental factors were fitted to the ordination plots using the ‘envfit’ function of the Vegan package in R (R Core Development Team, 2008). To analyze the community variation among plots and transects, Mantel tests were performed as implemented in the ecodist package of R as based on the same dissimilarity matrices. Partial Mantel tests were further applied to partial out the effect of geographical or altitudinal distance and host community when testing for the effects of the remaining significant variables. The effect of host was addressed using the dominant host species of the plot as a single categorical predictor and a distance matrix calculated on the basis of the relative abundance of host species.

EstimateS (Colwell, 2006) was applied to calculate rarefied species accumulation curves among transects and to estimate the minimum total species richness of the region based on Chao2, ICE and Jacknife2 minimum species richness estimators. To account for differences in sampling effort resulting from the absence of EcM roots in some samples, species richness was rarefied to four samples (i.e. the smallest sample size among plots). These rarefied values were used in the analyses of species richness. These values were strongly related to the average species richness per sample (R2 = 0.795, < 0.001); therefore, only the results of rarefied richness are presented.

The relationship between species richness and experimental factors was tested by partial general least-squares (GLS) analysis using the nlme package of R. Multicollinearity between environmental variables was evaluated on the basis of the variance inflation factor (VIF). As a result of a strong correlation between altitude and climatic variables (Supporting Information Table S1), the model selection procedure was performed by excluding climatic variables and by substituting altitude with climatic variables. Polynomial regression analysis was used to examine species richness patterns along altitudinal gradients by including altitude and its quadratic term in the models. The best models were selected on the basis of corrected Akaike information criterion (AICc) values. Using the model averaging technique, averaged coefficient and confidence intervals were calculated for each variable according to the weighted AICc across all models. Variables were considered significant when confidence intervals excluded zero-values. To account for the effect of spatial autocorrelation at the regional scale, spatial eigenvectors were extracted based on the principal coordinates of neighbor matrices (PCNM) (Borcard & Legendre, 2002) using spatial eigenvector mapping (SEVM), and were used as a response variable for model selection. In addition, residuals of the spatial model were extracted and further applied as a response variable in the regression analysis using SAM (Rangel et al., 2010). We also tested the effects of altitude, climatic variables, host and soil variables on the proportion of singletons to estimate whether any changes in richness are related to the loss of rare species. Three plots comprising the relict B. pendula occurring in Savadkuh were excluded from the statistical analyses of species richness because of the low replication of this host.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Species richness

In total, 261 (82.8%) soil samples contained living EcM roots. ITS sequences of 1090 (76.8%) unique morphotypes per sample and 1755 (60.9%) root tips subjected to molecular analysis were successfully identified, revealing 367 EcM fungal species from 9, 21 and 15 plots in Asalem, Nowshahr and Savadkuh, respectively (Tables S2; S3; Fig. S1). On average, 4.0 ± 1.03 (mean ± SD) and 21.7 ± 5.44 species were recovered per sample and per plot, respectively. On a plot basis, rare species that only occurred once or twice (singletons and doubletons, respectively) accounted for 87.5% of all species at the local scale and 75.7% at the regional scale. The minimum richness estimators suggested the presence of at least 602–686 species in the region.

At the local scale, EcM fungal species richness declined significantly with altitude in Asalem (R2 = 0.619, P = 0.012) and Nowshahr (R2 = 0.384, P = 0.003), but not in Savadkuh (R2 = 0.008, P = 0.78; Fig. 2). The best GLS models revealed that altitude was the main determinant of the species richness in Nowshahr and Asalem (Table 2). In Savadkuh, none of the variables was significantly related to the species richness. Species richness was not spatially autocorrelated in any of the transects.

image

Figure 2. Variation in the total species richness of plots along the altitudinal gradient in different transects and the whole region. Solid lines represent linear regression lines. (a) Asalem (R2 = 0.619, P = 0.012); (b) Nowshahr (R2 = 0.384, P = 0.003); (c) Savadkuh (R2 = 0.008, P = 0.780); (d) whole region (R2 = 0.283, < 0.001). 1Residuals of the spatial model have been used instead of species richness.

Download figure to PowerPoint

Table 2.   Best general least-squares (GLS) analysis models explaining the variation in species richness at local (transect) and regional scales
 Excluding climate variablesExcluding altitude
Variablest valueP valueVariablest valueP value
  1. 1Significant variables determining the species richness based on averaged β coefficient.

AsalemAltitude1− 3.3750.012Mean annual temperature12.8480.025
NowshahrAltitude1− 3.444< 0.001Mean annual temperature13.2880.004
SavadkuhAspect− 1.0160.333Aspect− 1.0160.333
Whole regionAltitude1− 3.919< 0.001Mean annual precipitation12.7240.010
Spatial component1− 3.1610.003Spatial component1− 3.2480.002
Temperature seasonality− 2.3060.027

At the regional scale, altitude had a significant negative effect on species richness (R2 = 0.243, < 0.001). The linear model explained the relationship between altitude and species richness better than any of the polynomial functions. The best GLS model revealed that altitude (R2 = 0.243, < 0.001) and the spatial component (R2 = 0.160, = 0.009) explained c. 40% of the variation in species richness (P = 0.001; Table 2). Partialing out the effect of the spatial component increased the explanatory power of altitude (R2 = 0.283, < 0.001; Fig. 2d). Substituting the climatic variables with altitude in the model selection procedure revealed that the mean annual temperature was the main driver of species richness in two of the transects, whereas the mean annual precipitation played the strongest role at the regional scale (Table 2). Together with species richness, the proportion of singletons and doubletons within each plot decreased significantly with increasing altitude (R2 = 0.095, = − 2.059, P = 0.046).

Community structure

Across three transects, the /tomentella–thelephora (85 spp.), /inocybe (50 spp.), /russula–lactarius (47 spp.), /sebacina (35 spp.) and /cortinarius (32 spp.) lineages dominated the EcM fungal community in terms of species richness. Cenococcum geophilum (45 plots), Cortinarius Ir1 (15 plots), Clavulina Ir1 (14 plots), Cortinarius Ir2 (13 plots) and Tomentella Ir2 (12 plots) were the most frequent EcM fungal species across all transects.

Shifts in EcM fungal community were related to altitude in all transects (Table 3; Fig. 3). In Nowshahr, the effect of altitude on EcM fungal community remained significant after partialing out the effects of host and geographical distance. In Savadkuh and Asalem, the effect of altitude became nonsignificant after accounting for these factors (Table 3).

Table 3.   Results of different analyses of putative factors affecting ectomycorrhizal (EcM) fungal community composition in the three transects and whole region
 TestAltitudeMean annual temperatureMean annual precipitationHost1Geographical distance
  1. nd, not determined.

  2. Envfit fits the environmental factors onto Non-metric Multidimensional Scaling (NMDS) ordination plots; Mantel test finds the correlation between two distance matrices; and Partial Mantel test finds the Mantel statistics while controlling for a third distance matrix.

  3. Values in bold are statistically significant (P < 0.05).

  4. 1Dominant host was used as a single categorical factor in Envfit. Distance matrices generated based on host community were used in Mantel and Partial Mantel tests.

  5. 2For altitude and climate variables, geography and host community; for host community, geography and altitude; and for host community, altitude and geography matrices were partialed out.

  6. 3The site effect was tested based on a categorical factor with three levels.

AsalemEnvfitR2 = 0.899R2 = 0.861R2 = 0.756R2 = 0.069nd
P = 0.001P = 0.001P = 0.008P = 0.589
Mantel= 0.335= 0.294= 0.358= 0.230= 0.225
P = 0.023P = 0.047P = 0.038P = 0.124P = 0.134
Partial Mantel2= 0.189= 0.101= 0.209= 0.101r = − 0.045
P = 0.154P = 0.306P = 0.137P = 0.315P = 0.604
NowshahrEnvfitR2 = 0.375R2 = 0.338R2 = 0.337R2 = 0.231nd
P = 0.012P = 0.029P = 0.027P = 0.032
Mantel= 0.264= 0.187= 0.313= 0.188= 0.139
P = 0.001P = 0.01P = 0.001P = 0.008P = 0.098
Partial Mantel= 0.211= 0.27= 0.271= 0.048= 0.135
P = 0.008P = 0.001P = 0.002P = 0.264P = 0.103
SavadkuhEnvfitR2 = 0.248R2 = 0.460R2 = 0.046R2 = 0.333nd
P = 0.198P = 0.025P = 0.755P = 0.124
Mantel= 0.186= 0.206= 0.045= 0.168= 0.041
P = 0.037= 0.026P = 0.335P = 0.051P = 0.352
Partial Mantel= 0.121= 0.144= − 0.006= 0.099 = − 0.009
P = 0.121P = 0.083P = 0.509P = 0.149P = 0.530 
Whole regionEnvfitR2 = 0.188R2 = 0.003R2 = 0.232R2 = 0.214R2 = 0.2553
P = 0.011P = 0.722P = 0.006P = 0.010P = 0.001
Mantel= 0.177= 0.091= 0.166= 0.167= 0.154
P = 0.001P = 0.026P = 0.001P = 0.001P = 0.002
Partial Mantel= 0.138= 0.011= 0.132= 0.113= 0.161
P = 0.001P = 0.376P = 0.004P = 0.007P = 0.003
image

Figure 3. Community dissimilarity among the plots as revealed by nonmetric multidimensional scaling (NMDS) ordination, reflecting the effects of dominant host, altitude, mean annual precipitation (MAP) and mean annual temperature (MAT) in different transects: (a) Asalem, (b) Nowshahr and (c) Savadkuh. Host taxa are represented by symbols: Fagus orientalis (triangles), Carpinus betulus (squares), Quercus spp. (circles) and Betula pendula (reverse triangles). The altitudinal gradient is illustrated by the gray scale from white (minimum) to black (maximum).

Download figure to PowerPoint

At the regional scale, transect, host species and altitude were the main determinants of the EcM fungal community (Fig. 4; Table 3). Among the experimental factors, altitude (rMantel = 0.177, P = 0.001) and precipitation seasonality (rMantel = 0.223, P = 0.001) showed the strongest correlation with the community composition. After partialing out the effects of altitude/host community and geographical distance, the EcM fungal community variation was correlated with precipitation seasonality (rMantel = 0.161, P = 0.001), altitude (rMantel = 0.138, P = 0.001) and host community (rMantel = 0.113, P = 0.007).

image

Figure 4. Community dissimilarity among the plots at the regional scale as revealed by nonmetric multidimensional scaling (NMDS) ordination, illustrating site effect (a), host effect (b) and altitude effect (c) (a ≤ 500; 500 <b ≤ 1000; 1000 < c ≤ 1500; 1500 <d ≤ 2000; e > 2000 m). Ellipses indicate 95% confidence intervals around centroids of each category. MAP, mean annual precipitation; MAT, mean annual temperature.

Download figure to PowerPoint

The abundance of a few lineages was correlated significantly with altitude: /paxillus–gyrodon (= − 0.418, = 0.004) and /pachyphloeus–amylascus (= − 0.352, = 0.017) decreased, whereas /hebeloma–alnicola (= 0.325, = 0.029) increased, along the elevational gradient (Fig. S2).

Climatic variables were strongly correlated with altitude, whereas soil variables were strongly multicollinear, and only the P concentration increased significantly along the altitudinal gradient (Table S1).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Species richness along the altitudinal gradient

Our major finding is that EcM fungal species richness declines linearly with increasing altitude, which is in agreement with the well-known pattern of species richness–altitude relationships (e.g. Stevens, 1992; Allen et al., 2002; Lomolino et al., 2006). In macroorganisms, a hump-shaped (unimodal) relationship with altitude is, however, the most common pattern (Rahbek, 2005). Recent studies have reported contrasting altitudinal diversity patterns in macro- and microorganisms. For example, Bryant et al. (2008) determined unimodal and linear declines for the diversity patterns of angiosperms and Acidobacteria, respectively, along an altitudinal gradient.

Altitude was the main factor determining EcM fungal richness at both the regional and local scales. Among the environmental variables at the local scale, the mean annual temperature had the strongest effect on richness and was most strongly correlated with altitude, which is a common phenomenon (Körner, 2007). Based on the metabolic theory of ecology, temperature is the main driver of biodiversity (Allen et al., 2002) and may affect soil productivity, microbial heterotrophic activities and ecological interactions (Brown et al., 2004; Tylianakis et al., 2008). The lower proportion of singletons and doubletons at higher elevation may account for the declining species richness pattern in this study. Taken together, these results suggest that rare species are relatively less competitive in harsh climatic conditions, and increasing dominance reduces overall diversity. Other studies have also shown that the lower fungal species richness resulting from elevated N deposition is mostly ascribed to the greater sensitivity and disappearance of rare species (Avis et al., 2008; Cox et al., 2010). Alternatively, higher energy results in greater production and hence carbon availability which may support more EcM fungal richness at low altitudes (Druebert et al., 2009). These results are consistent with the studies of Bacteria which demonstrated that climate is the most plausible factor explaining bacterial diversity patterns along altitudinal gradients (Bryant et al., 2008; Wang et al., 2011).

Given that altitudinal gradients largely reflect latitudinal trends, the declining pattern of EcM fungal richness along the altitudinal gradient concurs with the general increase in richness from cold to warm temperate regions and, further, to tropical ecosystems (Lomolino et al., 2006). However, our altitudinal gradient lack tropical sites, for which relatively lower phylogenetic and species-level richness are reported in EcM fungi (Tedersoo & Nara, 2010). Here, we avoid historical and biogeographical constraints that are characteristic of global studies, and methodological biases that are inherent to comparisons among studies. However, local and regional studies suffer from relatively limited climatic and geographical scales that usually exclude truly arctic/alpine and tropical ecosystems.

Changes in community structure

The regional community of EcM fungi comprises a large proportion of species from the /tomentella–thelephora, /inocybe and /russula–lactarius lineages, which is consistent with other studies in northern temperate ecosystems (e.g. Parrent et al., 2006; Ishida et al., 2007; reviewed in Tedersoo & Nara, 2010). Host species and altitude were the main drivers of the EcM fungal community at both the regional and local scales. The mean annual temperature and mean annual precipitation were significantly correlated with the fungal community composition at the local and regional scales, respectively. The presence of some common species in both the high- and low-altitude plots suggests that several EcM fungi have a wide temperature optimum. However, other studies suggest that many species of EcM fungi display differential tolerance to temperature (Rygiewicz et al., 2000; Robinson, 2001; López-Gutiérrez et al., 2008). For example, arctic strains of Hebeloma spp. have several physiological adaptation strategies to tolerate cold (Tibbett et al., 1998). Temperature has also been determined as the principal driver of the community composition of soil microfungi (Widden, 1987) and wood-decomposing fungi (Meier et al., 2010) along altitudinal gradients. Temperature can affect the fungal community directly via physical tolerance and enzymatic processes, or indirectly through its effect on the host plant community or soil nutrients (Heinemeyer et al., 2004; Clemmensen et al., 2006; Craine et al., 2009).

Despite strong gradients in altitude and climatic variables, host tree species had the strongest effect on EcM fungal community composition. Because the dominant trees, Fagus orientalis and Carpinus betulus, differentially modify soil properties, much of the host species effect could be ascribed to soil factors, although soil pH and nutrient concentrations had no significant effect on the EcM fungal community composition. At the local scale, both the host species and soil nutrients are among the key factors driving the community composition of EcM fungi (Toljander et al., 2006; Ishida et al., 2007; Tedersoo et al., 2008a,b).

Our results indicate that biotic and abiotic variables may differentially affect the species richness at local and regional scales, indicating that their effects on biodiversity are scale dependent (Willis & Whittaker, 2002). Within a site, host species and different nutrient concentrations related to topography and microsites play a substantial role in structuring the fungal communities (see previous paragraph). In this study, the roles of host and altitude in structuring the EcM fungal community were more pronounced at the local scale, whereas the effect of geographical distance increased at the regional scale. Together with geographical distance and dispersal barriers, dispersal limitation plays a more important role in the differentiation of EcM fungal communities (Peay et al., 2007), potentially ameliorating the overall effects of preference for host and soil environments at the regional and global scales. At the continental to global scales, historical factors (i.e. isolated distribution of host plant families) and long-term co-evolution between hosts and EcM fungi become increasingly important in determining the community composition at both the species and higher phylogenetic levels (Pritsch et al., 2010; Tedersoo et al., 2010).

Conclusions

Along the altitudinal gradients, temperature and precipitation affect the species richness of EcM fungi, suggesting that water–energy dynamics is the primary driver in the diversity patterns of EcM fungi, which is consistent with most other organisms (Hawkins et al., 2003). These climatic factors, together with host plant community, also affect EcM fungal community composition. The strong effect of climate on the EcM fungal community and richness suggests that EcM fungi are vulnerable to climate change, especially rare members of the community, because of strong dispersal limitation in fragmented landscapes (Peay et al., 2007) and sensitivity to other stressors, such as air pollution (Parrent et al., 2006; Andrew & Lilleskov, 2009).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Amin Fatahi and Hamed Shoubi for assistance in the field; Margit Nõukas and Katren Mikkel for assistance in the laboratory; Eveli Sisas and Sander Tint for help with ArcGIS software; Kessy Abarenkov, Vilmar Veldre, Ivika Ostonen and Märt Toots for useful discussions; and three anonymous reviewers for their valuable comments. We are also grateful to Jafar Fathi and the Kheirudkenar Educational Forest of the University of Tehran, as well as the Office of Forest and Rangelands Organization in Sari, Savadkuh and Sangdeh, for cooperation during field work. Financial support was provided by The European Social Fund (ESF) grants 7434, 8235, Doctoral Studies and Internationalisation Programme DoRa and Frontiers In Biodiversity Research (FIBIR).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1 Rarefied species accumulation curves at local and regional scales.

Fig. S2 Relative abundance of ectomycorrhizal fungal lineages along the altitudinal gradient in the Hyrcanian forests.

Table S1 Correlation matrix between variables used in this study

Table S2 List of ectomycorrhizal fungal species in the Hyrcanian forests

Table S3 Species richness, dominant ectomycorrhizal fungal species and lineages, and the diversity of host trees in the three transects

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.

FilenameFormatSizeDescription
NPH_3927_sm_FigS1-S2.doc180KSupporting info item
NPH_3927_sm_TableS1-S3.doc645KSupporting info item