Cenococcum geophilum populations show a high degree of genetic diversity in beech forests


  • Jean-Luc Jany,

    Corresponding author
    1. Unité Mixte de Recherche INRA/UHP 1136 ‘Interactions arbres/Micro-organismes’, Centre INRA de Nancy, 54 280 Champenoux, France
      Author for correspondence:Jean-Luc Jany Tel: +33 383 39 40 41Fax: +33 383 39 40 69 Email: jany@nancy.inra.fr
    Search for more papers by this author
  • Jean Garbaye,

    1. Unité Mixte de Recherche INRA/UHP 1136 ‘Interactions arbres/Micro-organismes’, Centre INRA de Nancy, 54 280 Champenoux, France
    Search for more papers by this author
  • Francis Martin

    1. Unité Mixte de Recherche INRA/UHP 1136 ‘Interactions arbres/Micro-organismes’, Centre INRA de Nancy, 54 280 Champenoux, France
    Search for more papers by this author

Author for correspondence:Jean-Luc Jany Tel: +33 383 39 40 41Fax: +33 383 39 40 69 Email: jany@nancy.inra.fr


• The asexual ectomycorrhizal fungus Cenococcum geophilum, known for its wide host and habitat range, has been suggested to provide isolate-dependant drought protection to fine roots. However, little is known about its genetic structure at the fine scale.

• Genetic diversity and population structure of C. geophilum at the regional and stand scales was surveyed in five beech (Fagus silvatica) forests in northeastern France. The stands were selected for their contrasting climatic and edaphic features to assess the effect of environmental factors on population structure.

• The genetic diversity of C. geophilum was estimated using RAPD, PCR/RFLP of the rDNA internal transcribed spacer (ITS) and PCR/RFLP and sequencing of an anonymous sequence characterized amplified region (SCAR) on ectomycorrhizas and sclerotia-derived mycelial cultures.

• A high degree of genetic diversity was observed between and within beech stands in C. geophilum populations. These results suggest the occurrence of a high rate of mitotic or meiotic recombination and an effect of stand features on population structure.


At the soil–root interface, ectomycorrhizal (ECM) fungi are known to play a fundamental role in temperate and boreal forest ecosystems, by enhancing the hydro-mineral nutrition and affecting the host plant growth (Smith & Read, 1997). In these ecosystems, a single host tree may interact with hundreds of different species of ECM fungi (Taylor et al., 2000; Dahlberg, 2001), each species being represented by different genotypes (Bonello et al., 1998; Gherbi et al., 1999; Fiore-Donno & Martin, 2001) differing in their physiological abilities. This is thought to be important to ecosystem functionning (Debaud et al., 1995; Cairney, 1999).

Within the ectomycorrhizal communities, the ubiquitous and cosmopolitan ascomycete Cenococcum geophilum Fr. is one of the dominant and most frequent ectomycorrhizal types. Horton & Bruns (2001) emphasized that there are virtually no studies where this fungus was not detected. C. geophilum was found to be one of the prominent fungi in fennoscandian forests with a relative abundance of 15–18% (Dahlberg et al., 1997; Kåren et al., 1997). Furthermore, Kranabetter & Wylie (1998) have found C. geophilum to be the most frequent ECM fungus on naturally regenerated western hemlock seedlings in northwest British Columbia. Moreover, C. geophilum sclerotia, hypogeous resistant propagules, have been reported to be highly abundant in soils (Dahlberg et al., 1997). In northeastern France, the relative abundance of C. geophilum ectomycorrhizas has been estimated to be 47% in a beech forest (Blaise & Garbaye, 1983).

This abundant fungus could provide a functionnal asset to the colonized fine roots since several studies reported that C. geophilum is more resilient to drought stress than other ECM fungi (Mexal & Reid, 1973; Coleman et al., 1989; Neves Machado, 1995) and may protect roots from drought when it is involved in ectomycorrhizal symbiosis (Pigott, 1982a,b; Neves Machado, 1995). The resistance to drought stress however, varies between different isolates of C. geophilum (Coleman et al., 1989; Neves Machado, 1996) suggesting functionnal heterogeneity among C. geophilum populations. The genetic structure of C. geophilum populations may thus play a fundamental role in the adaptation of the host populations to drought.

Despite this remarkable set of ecological features, little is known about the genetic structure of C. geophilum populations. Surveys of isolates on a worldwide scale revealed a significant and unexpected genetic diversity for an asexual fungus (LoBuglio et al., 1991, 1996; Shinohara et al., 1999). Only a few studies have dealt with genetic diversity of C. geophilum on a detailed scale. Panaccione et al. (2001) showed that C. geophilum populations found in serpentine and nonserpentine soils were polymorphic. These authors distinguished 12 different genotypes among 13 local isolates.

In the present study, we have analysed regional and local genetic variability of C. geophilum and addressed the structuration of this species between different stands of Fagus silvatica L and different soil cores within each stand. Samples were collected in five different beech stands selected for their contrasting climatic and edaphic features in northeastern France. The genetic diversity has been studied by using RAPD, PCR/RFLP of the nuclear rDNA ITS and sequence analysis of a sequence characterized amplified region (SCAR1) on isolates originating from sclerotia. In addition, ectomycorrhizal tips were typed using PCR/RFLP of the ITS and of SCAR1.

Materials and Methods

Beech stands, experimental soil cores and sampling pattern

Investigations were performed on mycorrhizal roots of 80- to 100-yr-old-beech trees (Fagus silvatica L.) and on C. geophilum sclerotia collected from five different mature beech stands managed as regular high forest in northeastern France. These stands (about 1000 m2 each) were located along a 250-km East–West transect of increasing rainfall and contrasting edaphic conditions (Table 1). Sclerotia of C. geophilum and short ectomycorrhizal beech roots were collected in October 1999 and August 2000, respectively, from one dm3 soil cores (10 × 10 × 10 cm) of the A1 horizon (about 10 cm deep) sampled in five randomly localized areas in each stand.

Table 1.  Location and characteristics of the studied beech stands
 La CrêteAmanceCharmoisHennezelTendon
  • a

    Data from the closest meteorology station: Chaumont (05°39

  • 03′′ E. 48°08′47′′ N). bData from the closest meteorology station: Amance (6°20

  • 25′′ E. 48°44′25′′ N). cData from the closest meteorology station: Vittel (5°57

  • 01′′ E. 48°12′09′′ N). dData from the closest meteorology station: La Bresse (6°52

  • 36′′ E. 48°00′24′′ N). eData from the closest meteorology station: Epinal (06°27′09′′ E. 48°10′28′′ N).

Coordinates5°21′25′′ E 48°12′15′′ N6°20′25′′ E 48°44′25′′ N6°16′12′′ E
48°6′40′′ N6°11′45′′ E 48°2′30′′ N6°41′25′′ E 48°07′15′′ N
Elevation370 m280 m400 m420 m570 m
Annual rainfall
1999968 mma859 mmb1207 mmc1207 mmc2539 mmd
2000786 mma855 mmb886 mmc886 mmc2156 mmd
Mean month
Parent materialLimestoneLoessSandstoneSandstoneGranite
TextureClayeyLoamSandy loamSandy loamSandy
Humus typeEutrophic mullMesotrophic mullMesotrophic mullMesotrophic mullMor moder

Preparation of sclerotia and ectomycorrhizas

Soil samples were placed in air-tight plastic bags, transported on ice to the laboratory, and stored at 4°C for less than 7 d. For the preparation of ectomycorrhizas, roots were removed from the soil, gently washed in water, and fine roots were observed under a stereo-microscope. Based on the descriptions of Agerer (1995) and Ingleby et al. (1990), ectomycorrhizas of C. geophilum were identified with thick, straight, unbranched and black hyphae emanating from a black mantle. Ten C. geophilum ectomycorrhizas were picked per soil core (Table 2) and individually stored in a microcentrifuge tube (Eppendorf AG, Hamburg, Germany) at −80°C for 2 months at most. For the preparation of sclerotia, soil samples from the different soil cores were sieved through a 0.4-mm sieve. The fraction remaining on the sieve was observed under a stereo-microscope. A total of 450 sclerotia was collected over the five stands (Table 2). They were kept in distilled water at 4°C for less than 24 h before surface sterilisation.

Table 2.  Samplings from the five beech stands and successful DNA typing analyses of ectomycorrhizas and sclerotia-derived isolates
StandPlot #Ectomycorrhizas sampledEctomycorrhizas typedSclerotia sampledSclerotia-derived isolates typed
La Crête110  4201
210  1201
310  3200
410  0201
510 10201
Amance110 10100
210 10101
310 10101
410 10101
510 10101
Charmois110 10202
210  6201
310 10203
410 10201
510 10202
Hennezel110 10205
210  7203
310 10202
410  5202
510  4207
Tendon110  8202
210  6201
310  5201
410  1202
510  3200
Total 25017345042

Preparation of fungal cultures

Sclerotia were surface-sterilised in a microcentrifuge tube in a 20% solution of a commercial disinfectant containing hypochloride (Domestos, Lever-Fabergé, Zoug, Switzerland) for 30 min, rinsed ten times in sterile water and individually placed on a Pachlewski agar medium (Pachlewski & Pachlewska, 1974). Mycelium arising from sclerotia was subcultured on Pachlewski agar medium every 3 months.

DNA analyses

DNA extraction Total DNA was extracted directly from single ectomycorrhizas or from 10 mg of pure culture mycelium using the DNeasy Plant Mini Kit (Qiagen SA, Courtaboeuf, France) following the manufacturer’s instructions.

RAPD RAPD amplification was carried out in duplicate under conditions specially designed to enhance reproducibility (Selosse et al., 1998). The components of the 25 µl PCR mix were: 10 ng of DNA, 1.84 µM of the 152C primer (5′-CGCACCGCAC-3′) (Invitrogen, Cergy-Pontoise, France), 200 µM dNTP (Qbiogene, Illkirch, France), 1 ×Taq DNA polymerase incubation buffer and three units of 15 U Taq DNA polymerase (Qbiogene). The reactions were done in a Perkin Elmer GeneAmp 9600 thermocycler (Perkin Elmer Instruments, CT, USA) under the following cycling parameters: initial denaturation at 94°C for 1 min, followed by 30 cycles of denaturation at 94°C for 0.5 min, annealing at 55°C for 0.5 min and extension at 72°C for 2 min, with a final extension at 72°C for 10 min. Products were analysed on 2% agarose gels in 1% TBE (Tris buffer-EDTA). Controls with no DNA were included in each series in order to test for the presence of contaminants in reagents and reaction buffers.

RAPD data analysis RAPD intense bands in the 500–1500 bp size range that reproduced consistently for the two replicates were scored for each individual sclerotial isolate. Electrophoregrams of the different samples were compared side by side. RAPD was scored as presence (1) or absence (0) of a band and a matrix of RAPD phenotypes was built. This rectangular matrix of individual RAPD markers was transformed to a resemblance matrix using the Jaccard similarity coefficient and used for a Neighbor-Joining (NJ) analysis to produce a dendrogram for which the robustness of the topology was assessed by bootstrap (500 iterations) (FreeTree program, Pavlicek et al., 1999, available at: http://www.natur.cuni.cz/flegr/programs/freetree).

The effect of spatial separation on genetic structure was tested by a regular Mantel test (Mantel & Valand, 1970) on the resemblance (RAPD pattern similarity) matrix and a Euclidean matrix of metric distance between the sampling plots (MANTEL module in the R Package 4.0d3 Casgrain & Legendre, 2001). The significance of the correlation between the two matrices was tested by 9999 random permutations of the similarity matrix to generate a null distribution of correlation coefficients (z-values). A significant result was inferred if ≥ 95% of the randomly generated statistics were greater than the observed value.

Based on the number of individual RAPD phenotypes, the proportion of distinguishable RAPD phenotypes (PD) in the total number of unique RAPD patterns divided by the isolate number (Ellstrand & Roose, 1987) was calculated as a descriptor of the diversity within the sample. Estimates of phenotypic diversity and partitionning of the RAPD variation were carried out using the Shannon index (Lewontin, 1972). This method does not require the Hardy–Weinberg in order to determine the phenotypic diversity assumptions (Fritsch & Rieseberg, 1996); this is very important considering that there is little information available about reproduction of C. geophilum. The calculation of the Shannon index, originally proposed as a measure of the information content in a code (Shannon & Weaver, 1949), is based on number and frequency of markers for each RAPD phenotype (Lewontin, 1972). Shannon index is defined as: Ho = Σpi log2pi, where pi is the frequency of a given RAPD band. It is used to quantify the degree of within-population diversity. The average diversity over the different populations was calculated from Ho as: Hpop= 1/nΣHo, where n is the number of populations. The diversity in all the populations considered together Hsp, was calculated from the phenotypic frequencies p (–Σp log2p) as: Hsp = –Σps log2ps, where ps is the frequency of presence or absence of the RAPD band in the whole sample (42 individuals in this case). Thus, the component of diversity within populations is Hpop/Hsp and the component between populations is (HspHpop)/Hsp which is also referred to as G′st.

Sequencing of SCAR1 A cloned DNA fragment corresponding to a 620-bp RAPD product occurring on 36 isolates of the 42 sclerotial isolates tested was sequenced using the BigDye Terminator Sequencing Kit (Applied Biosystems, CA, USA) and an ABI genotyper 310 automated sequencer (Applied Biosystems, CA, USA). A couple of primers, OSCAR11 (5′-AGGTATTCAGTGATCCGCGG-3′) and OSCAR12 (5′-GGTGACTTTGGTGTCCTCCC-3′), were designed using the Primer Selection program (http://alces.med.umn.edu/rawprimer.html) for amplifying SCAR1. These primers were synthesised and supplied by Invitrogen (Invitrogen, CA, USA) and used for amplifying and sequencing 11 SCAR1 sequences. These sequences are available from the GenBank database (NCBI), under the following access numbers: AF414847 to AF414857. Sequences were aligned using the MultAlin program (http://www.toulouse.inra.fr/multalin.html) (Corpet, 1988).

Design of C. geophilum ITS specific primers The PCR amplification of the rDNA ITS of mycelium originating from black Cenococcum-like ectomycorrhizas using the universal ITS primers ITS1 and ITS4 (White et al., 1990) indicated that multiple fungal endophytes (e.g. Phialophora sp.) occurred in beech ectomycorrhizal tips (data not shown). As a consequence, we have designed two Cenococcum-specific primers, CGITS1 (5′-TGACGATTGACTCATGTTGC-3′) and CGITS2 (5′-TTCAAAGCGAAAGATTCTGC-3′), based on the multialignment of C. geophilum ITS sequences available in GenBank (NCBI). under the following access numbers: Z11998, Z48521 and Z48523 to Z48537. The two primers CGITS1/CGITS2 were synthesised and supplied by Invitrogen for amplifying a portion of the ITS referred to as CGITS.

PCR/RFLP Amplifications of CGITS, and SCAR1 were performed under the following conditions: unquantified DNA samples were added to a mix containing: 0.6 µM of each primer (Invitrogen), 200 µM dNTP (Qbiogene) and 2.5 units of REDTaq DNA polymerase (SIGMA, MO, USA) and 1 × REDTaq PCR Reaction Buffer solution. PCR was performed in a Perkin Elmer GeneAmp 9600 thermocycler (Perkin Elmer Instruments) under the following cycling parameters: initial denaturation at 94°C for 1 min, followed by 30 cycles of denaturation at 94°C for 0.5 min, annealing at 50°C for 0.5 min and extension at 72°C for 2 min, with final extension at 72°C for 10 min. Twenty microliters of PCR products were digested at 37°C over 5 h using 10 units of HinfI enzyme (Qbiogene), and restriction products were analysed on 8% acrylamide gels.


Genetic variability of C. geophilum assessed by sclerotia typing

RAPD of sclerotia-derived mycelium Cultures from C. geophilum sclerotia, collected from the five beech stands (Table 1), yielded 42 mycelial cultures (average yield = 9.5%). RAPD produced an average of 16 detectable and reliable DNA bands per sample in the 500–1500 bp size range. A total of 18 different multilocus RAPD phenotypes were identified among the 42 isolates analysed; the proportion of distinguishable genets (PD) was 0.42. One RAPD phenotype was prominent and comprised 15 individuals collected in the Charmois and Hennezel beech stands.

Similarity between genets was assessed by NJ analyses of a binary matrix produced from RAPD patterns (Fig. 1). The tree topology indicated that RAPD phenotypes from Hennezel (H) and Charmois (CHA), two stands 12.5 km apart, sharing similar climatic and edaphic features, mostly grouped together, whereas isolates from Amance (AM), Tendon (TD) and La Crête (Cre) were not clearly separated. The overall correlation between physical and genetic distance matrices was r= 0.49 and significant at P < 0.001 (Mantel’s test, 9999 permutations), suggesting that spatial distance between the sampling sites may play a role in structuring the C. geophilum populations. The phenotypic variation between C. geophilum genets can be partitioned into within-stand and between-stand components. Even if the Shannon index has been used for analysis of populations with both small and variable numbers of individuals: seven to 16 in Wolff et al. (1997), or seven to 41 in Cardoso et al. (1998). Bussell (1999) noticed that Shannon index may not be appropriate if the number of individuals per population is small and/or varies. Therefore, we calculated Hpop and H′sp both with the overall data set and with only the three largest populations (excluding La Crête and Amance populations consisting of four individuals only). Hpop and Hsp (Table 3) were found to be similar in both cases and revealed that the total variation was evenly partitionned between interpopulation and intrapopulation variations.

Figure 1.

Dendrogram produced by a Neighbor-Joining analysis of the RAPD patterns of 42 Cenococcum geophilum sclerotia-derived isolates examined among the five studied beech stands: La Crête (Cre), Amance (AM), Charmois (CHA), Hennezel (H) and Tendon (Td). The first number following the stand code indicates the soil core, the second one is the isolate number (e.g. CHA1.2 for the isolate number two, originating from the core #1 in the Charmois stand). The ribotype (Ac) or (Bc) determined by PCR/RFLP of CGITS using the HinfI endonuclase is indicated beside each individuals. The genotype (v, w, x, y or z) determined by DNA sequencing of SCAR1 is indicated beside some of the individuals. Bootstrap values higher than 50 are indicated.

Table 3.  Partitioning of genetic diversity among sclerotia-derived isolates of C. geophilum at different beech stands in north-eastern France revealed by the calculation of the Shannon’s indices
Population numberHpop between standsHsp within stands(Hsp – Hpop)/HspHpop/Hsp

PCR/RFLP of the rDNA ITS The analysis of CGITS variability by PCR/RFLP, yielded two different patterns among the 42 tested sclerotia-derived isolates. The first one (Ac) consisted of two bands of 170 and 150 bp, while the second one (Bc) was a three-band pattern including 150, 90 and 80 bp fragments. Some individuals sharing the same RAPD phenotype can be distinguished considering their ITS-type. For example, isolates CHA1.2 and CHA2.1 were discriminated from the other individuals grouped in a 15 individual RAPD cluster (Fig. 1).

DNA sequencing of SCAR1 Sequencing of SCAR1 was successful on 11 isolates amongst the 42 tested by RAPD. SCAR1 showed no homology to known sequences in the DNA databases. A multialignment of the 11 SCAR1 sequences (available in the PopSet database (NCBI) under the following access number: 15983024) led to the identification of 35 polymorphisms (32 single nucleotide polymorphisms and three insertion/deletion of, respectively, 10, 16 and 17 bp) allowing identification of five different genotypes (v, w, x, y and z). Although the number of SCAR1 polymorphisms was high, SCAR1 sequences of the genotypes v, w and x only differ by one nucleotide. Individuals sharing the same RAPD pattern and the same ITS-type can be further distinguished by their SCAR1 polymorphism. For example, CHA3.5, CHA5.1 and Cre4.1 shared the same RAPD pattern and the same ITS-type (Bc), but exhibit three different SCAR1 sequences (w, x and y, respectively). Amongst the 42 sclerotia-derived isolates, 24 genets were discriminated using RAPD phenotypes, PCR/RFLP of ITS and SCAR1 sequencing; the PD value was 0.57.

Genetic variability of C. geophilum assessed by typing of ectomycorrhizal tips

PCR/RFLP of CGITS was successful for 173 ectomycorrhizal tips among 250 sampled ones. The DNA extracts (77) for which PCR did not yield any product may be of poor quality or may originate from black ectomycorrhizas belonging to species not recognized by CGITS or SCAR1 primers. The analysis of the 173 ectomycorrhizal tips sampled in the five different beech stands, yielded two patterns (Ac and Bc) similar to those observed for sclerotia-derived isolates. Three restriction patterns of SCAR1 were obtained with the HinfI endonuclease. The first one (As) corresponded to the single undigested 450 bp initial fragment. The second pattern (Bs) consisted of two bands of 230 and 180 bp. The third one (Cs) was a three-band pattern that simultaneously displayed the 450 bp uncut fragment and the 230 and 180 bp products. For the isolates exhibiting the three-band pattern, we increased the amount of HinfI enzyme up to 25 units, and/or the duration of the reaction up to 12 h, but the upper 450 bp fragment remained uncut, suggesting the existence of two SCAR1 alleles. Combining the data of PCR/RFLP of CGITS and SCAR1 resulted in the discrimination of five different genotypes (A to E) of C. geophilum forming ectomycorrhizas in the investigated beech stands. Sequencing of CGITS and SCAR1 was not possible due to the low amount of DNA material available.

The five genotypes were unevenly distributed among and within the various beech stands (Fig. 2). A single genotype (A) was detected in the La Crête stand; however, the yield of the PCR analysis was the lowest from our study (36%). The number of detected genotypes was higher for the other stands: two in the Amance (A, 60% and C, 40%) and Hennezel (D, 64% and E, 36%) stands, three in Tendon (A, 13%; D, 61% and E, 26%) and up to four for the Charmois stand (A, 52%; B, 20%; D, 24% and E, 4%). Several genotypes were specific for a stand; C type was only found in the Amance stand and B type was only sampled in the Charmois stand. The D and E types were only sampled in the Tendon, Hennezel and Charmois stands.

Figure 2.

Distribution of the five different genotypes of Cenococcum geophilum ectomycorrhizas within the five studied beech stands: La Crête, Amance, Charmois, Hennezel and Tendon. Genotyping was performed by combining PCR/RFLP of CGITS and SCAR1 using the HinfI endonuclease. A pie-chart is centered on each of the five soil cores (10 × 10 × 10 cm) sampled per beech stand. The different sectors of the pie-chart indicate the number of the five different genotypes (A: red, B: black, C: blue, D: yellow and E: green) among the 10 ectomycorrhizas sampled from the soil core. White sectors are used for samples for which PCR/RFLP of CGITS or SCAR1 was unsuccessful either because they were not Cenococcum geophilum ectomycorrhizas or due to poor quality of the DNA extract.

At the intrastand scale, several ectomycorrhizal genotypes were found to coexist in a single 1 dm3 core (Fig. 2). Two types (E and D) were detected in two different cores in the Hennezel and Tendon stands and in three cores in the Charmois stand. Up to three genotypes were found in one core in the Tendon stand (types A, D and E) and in two cores in the Charmois stand (types A, B and D).


Genetic variability of C. geophilum at the regional scale

Genetic variability has previously been reported for C. geophilum (LoBuglio et al., 1991). These authors discriminated 32 genotypes by analysing RFLP of the whole rDNA repetitive units of 71 C. geophilum isolates from worldwide origin. The same isolates were divided by RFLP into 20 rDNA ITS types (Shinohara, 1994 cited by LoBuglio et al., 1996). Furthermore, a PCR/RFLP study of the ITS of 19 C. geophilum isolates on a worldwide scale led to the discrimination of four genotypes (Farmer & Sylvia, 1998). More recently, Panaccione et al. (2001) reported a high local diversity in serpentine and nonserpentine soils. These authors distinguished 12 different genotypes among 13 local isolates.

In the present study, we also observed a high level of genetic diversity within C. geophilum populations at the regional scale, within five contrasted stands located in mature and homogeneous beech forests in northeastern France. RAPD carried out on 42 sclerotia-derived isolates of C. geophilum identified 18 different RAPD phenotypes. Furthermore, considering the CGITS types and SCAR1 sequence haplotypes, 24 genotypes were finally distinguished. The proportion of distinguishable genets (PD = 0.57) was within the range of values obtained for fungi for which sexual reproduction plays a key role (e.g. PD = 0.61 for Laccaria amethystina;Gherbi et al., 1999). Ectomycorrhizal species relying mostly on vegetative growth, such as Xerocomus chrysenteron (Fiore-Donno & Martin, 2001) show much lower PD values (0.007).

Although the presence of the host-plant DNA in ectomycorrhizal tips precluded the use of RAPD, PCR/RFLP analyses of CGITS and SCAR1 detected a minimum estimate of five C. geophilum genotypes among the 173 ectomycorrhizas collected in the five different beech stands.

The high level of genetic diversity of C. geophilum populations investigated in northeastern France was confirmed by statistical analysis; the Shannon’s diversity index (G′st) was 0.42. As C. geophilum has no known sexual stage, this wide genetic diversity was unexpected. Nevertheless many fungi, thought to spread by asexual means exhibit a high degree of genetic variability, for example plant pathogens such as Magnaporthe grisea (Zeigler et al., 1996) and Botrytis elliptica (Huang et al., 2001) or plant symbionts as Glomus spp. (Lloyd MacGilp et al., 1996). Such a genetic variability could be due whether to cryptic sex, as shown in the human pathogen Coccidioides immitis (Burt et al., 1996), to high-level of somatic mutations (Milgroom, 1996), or to parasexuality events (Milgroom, 1996; Zeigler et al., 1997; Anderson & Kohn, 1998; Taylor et al., 1999). The calculation of linkage disequilibrium between alleles at different codominant loci such as SCAR1, will provide further data to test for the occurrence of recombination events in C. geophilum.

Spatial distribution of genotypes of C. geophilum at the interstand scale

A subdivision in C. geophilum populations was indicated by the interpopulation variation in sclerotia-derived mycelia. This interpopulation variation was as large as the intrapopulation variation, as calculated using Shannon index (Table 3). A significant Mantel’s test (r = 0.49, P= 0.0001) suggested that spatial distance contributes to the structuration of C. geophilum populations. This contention is in accordance with the NJ tree topology, which indicated that genets from the Hennezel and Charmois stands, located 12.5 km apart and sharing similar mid-ranged climatic and edaphic features, were grouped together. However, because the most resembling stands were also the less distant ones (e.g. Hennezel and Charmois), it is not possible to determine whether distance and/or stand ecological features are the major factors affecting genetic variation.

The five genotypes detected by PCR/RFLP of ectomycorrhizal tips were also unevenly distributed among the beech stands. In the most acidic soils, as found in the Tendon and Hennezel stands (42 km apart), the D and E genotypes were prominent, whereas in the less acidic soil of Amance and in the calcareous soil of La Crête (96 km apart) the A genotype was the dominant one. Genet (1999) has already reported morphological and anatomical differences between C. geophilum mycorrhizas colonizing either a calcareous and basic soil or an acidic soil. These data suggest that the soil characteristics might play a role in structuring C. geophilum populations as proposed by Díez et al. (2001) who argued that the soil features are a major ecological factor that may account for the variability of Pisolithus isolates in the native Mediterranean vegetation. Natural selection of genotypes adapted to local conditions (Ennos & McConnell, 1995) can contribute to subdivision in populations as found for C. geophilum by Panaccione et al. (2001). These authors showed that populations of C. geophilum established in serpentine soil and nonserpentine soil were genetically different and proposed that adapted genotypes colonized serpentine soils.

The differences in ecological features of the beech stands, as well as geographic isolation, amplified by the limited capacity to dispersal of C. geophilum soil-borne sclerotia, may contribute to the observed genetic structure. Studying distant beech stands sharing similar edaphic features and estimating gene flow by using a set of codominant markers as SCAR1 would allow to evaluate the relative contribution of these factors.

Spatial distribution of genotypes of C. geophilum at the intrastand scale

The important within-population variance, calculated from RAPD phenotypes supports the interpretation that C. geophilum populations are also structured at the stand level. Among the five PCR/RFLP types of ectomycorrhizas, up to four types were detected within the same stand (e.g. Charmois) and up to three types appeared in the same 1 dm3 core (e.g. soil core 3 in Tendon) (Fig. 2). Therefore, several genotypes may coexist within the same limited soil. Even if the sampling procedure used in the present study precluded the determination of the genet size, the present data suggest that C. geophilum did not fit with the model of low-numbered and exclusive genotypes one should expect for an ectomycorrhizal fungus that lacks meiospore production. However, C. geophilum produces sclerotia which can be transported by water, birds, insects or rodents (Massicotte & Trappe, 1992). We therefore suggest that C. geophilum do not extend exclusively by vegetative mycelial growth, but that sclerotia are responsible to a large extent for its fine-scale dissemination, possibly leading to multiple and mixed genotypes resulting from putative recombination events.

Further analyses are required to explain the origin of the genetic variation observed within C. geophilum. Recombination events are expected, but whether there are due to parasexuality or to the existence of a cryptic sexual state still have to be determined. In addition, population analysis over time and with a spatial scale properly adjusted to define the structuration of genets will shed new light on our understanding of the fine genetic structure of C. geophilum.


This study was part of J. L. Jany’s PhD. project funded by a scholarship from the Office National des Forêts (O.N.F.) and the Région Lorraine. Funding was provided to JG and FM by O.N.F. and INRA (Action Transversale ‘Typage des bio-agresseurs et des symbiotes’) and by the C.N.R.S. program ‘Environnement, vie et sociétés’ managed by the Institut Français de Biodiversité. The research utilized in part the DNA sequencing facilities at INRA-Nancy financed by the INRA and the Région Lorraine. We thank J. L. Churin and P. Vion for their assistance in the field surveys, M. A. Selosse and D. Tagu for critical reading of the manuscript, and J. Díez for stimulating discussions.