SEARCH

SEARCH BY CITATION

Keywords:

  • ectomycorrhizal fungi;
  • genetic differentiation;
  • nested clade analysis;
  • phylogeography;
  • ribosomal DNA;
  • truffle

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information
  • • 
    Glaciations and postglacial migrations are major factors responsible for the present patterns of genetic variation we see in natural populations in Europe. For ectomycorrhizal fungi, escape from refugia can only follow range expansion by their specific hosts.
  • • 
    To infer phylogeographic relationships within Tuber melanosporum, sequences of internal transcribed spacers (ITS) and the 5.8S coding region of the ribosomal DNA repeat were obtained for 188 individuals sampled over the entire distribution of this species in France, and in north-western Italy and north-eastern Spain.
  • • 
    Ten distinct ITS haplotypes were distinguished, mapped and treated using F- and NST-statistics and nested clade (NCA) analyses. They showed a significant genetic differentiation between regional populations. NCA revealed a geographical association of ITS haplotypes, an old fragmentation into two major groups of populations, which likely colonized regions on different sides of the French Central Massif.
  • • 
    This re-colonization pattern is reminiscent of the one observed for host trees of the Perigord truffle, such as oaks and hazelnut trees. This suggests that host postglacial expansion was one of the major factors that shaped the mycobiont population structure.

Introduction

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

The Quaternary climatic fluctuations dramatically influenced the distribution of the flora and fauna in Europe (Hewitt, 1999). Repeated climate changes and the advance of glaciers forced animal and plants to retreat to southern refugia. The biota expanded after the last glaciations when the climate became warmer (Hewitt, 1996). Extensive studies on the postglacial expansion of animals and plants (Taberlet et al., 2001), including forest trees (Petit et al., 2003), contrast with the lack of knowledge on the populations of rhizospheric microorganisms. Ectomycorrhizal fungi are root-associated mutualistic symbionts of trees in boreal and temperate forests (Smith & Read, 1997) and their populations have probably been shaped by the reduction and expansion of the forests.

The Perigord truffle (Tuber melanosporum Vittad.) is an ascomycete endemic to calcareous soils in southern Europe and is found in symbiotic association with roots of deciduous trees, mostly oaks (Quercus spp.) and hazelnut trees (Corylus avellana) (Delmas, 1978). The fruiting body of T. melanosporum is an edible truffle (= hypogeous ascocarp), which is highly appreciated for its delicate organoleptic properties (i.e. taste and perfumes) (Hall et al., 2003). The high prize of the Perigord truffle has prompted the development of its culture through man-made inoculation of seedlings (Chevalier & Grente, 1978; Chevalier & Dupré, 1990). It can be assumed that the natural distribution and genetic structure of populations of this black truffle species have been structured by at least five major factors: first the distribution of its host plant species (i.e. ectomycorrhizal deciduous trees); second the spore dispersal by mycophagous animals; third limiting ecological factors (calcareous soils and a temperate climate); fourth geographical barriers (i.e. Mediterranean Sea, which limits its expansion towards North Africa), and fifth historical events (i.e. northward re-colonization routes from glacial refugia in southern Europe). The genetic diversity of T. melanosporum is strikingly low (Gandeboeuf et al., 1997) and a population bottleneck probably occurred during the last, and coldest, glaciation (c. 10 000 yr ago), when the broadleaved forest of Europe was considerably reduced and restricted mainly to the Mediterranean coastal zone (Bertault et al., 1998, 2001). To our knowledge, the analysis of genetic differences among southern populations of T. melanosporum and those occurring at the northern limits of this truffle species expansion (i.e. north-eastern France) has not yet been analyzed. Information on the genetic differentiation among these populations may shed light on postglacial phylogeography of this edible fungus.

We assessed the genetic variability of T. melanosporum isolates by sequencing the rDNA ITS and five sequenced characterized amplified regions (SCAR). Methods based on F- and NST-statistics showed a genetic differentiation between populations. As these methods did not detect historical events (e.g. fragmentation or range expansion) that affected the populations, we implemented our genetic study with nested clade analysis (NCA) (Templeton et al., 1995; Templeton, 1998, 2004) to shed light on the possible migration processes after the last glaciation. These analyses strengthened the ‘glaciation hypothesis’ (Bertault et al., 1998) with a rapid colonization of Western Europe from relict Italian populations of T. melanosporum. The identified migration routes in France are similar to those observed for oaks species (e.g. Quercus pubescens) and other hardwood tree species (Petit et al., 2002a, 2003) which can host Tuber species.

Materials and Methods

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

Sampling sites

The study organism, Tuber melanosporum Vittad. (Ascomycota, Pezizomycotina, Pezizomycetes, Pezizales, Tuberaceae), is endemic to France and certain areas of Southern Europe in Italy and Spain, and occurs only in calcareous soils in association with deciduous trees. This black truffle does not tolerate the cold temperatures of high mountains or northern Europe. These ecological requirements cause a fragmented distribution of T. melanosporum in France (Callot et al., 1999). For instance, this fungus is absent from the acid and coldest regions of the Central Massif and the Alps, and the acidic sandy soils of Landes (south-western France). Our sampling strategy aimed to cover the natural geographical range of T. melanosporum in France to investigate whether genetic variation among regional populations occurs. Fruiting bodies of T. melanosporum were collected in natural truffle grounds in 17 geographical regions of France, northern Italy (Piedmont) and north-eastern Spain (Iberian Mountains) (Supplementary material Table 1). For the analysis of molecular variance (amova) and nested clade analysis (NCA), the 17 sampled regional populations were pooled to have sufficient sample sizes for meaningful comparisons in seven larger geographical areas corresponding to: first Lorraine, the northernmost known limits of expansion; second Burgundy; third the calcareous Pre-Alps (Isere, Drome) and southern Jura; fourth the coastal southern France (Rhone Provence, Inner Provence, and Languedoc) and the Pyrenees (Roussillon and Ariege); and fifth the western France regions including the foothills of the Central Massif range (Tarn, Perigord-Quercy, Lower Quercy) and Charentes and Touraine (Fig. 1). Ascocarps from north-western Italy (Piedmont, region 6) and north-eastern Spain (region 7) were pooled and included in our study. Most Spanish samples were from a series of valleys along the Iberian Mountains, a calcareous mountain system that extends c. 400 km along the north-eastern edge of the Meseta (Central Plateau) in Spain. The Italian truffles were sampled in the Val Curone valley (Alessandria area), a sedimentary-calcareous region, which extends at the southern edge of the Po Plain. Owing to the low polymorphism of T. melanosporum and its selfing reproduction, a few individuals can reflect the regional populations better than extensive sampling per locality (Bertault et al., 2001). We thus sampled one or two fruit bodies per truffle ground in 120 localities over a wide geographical range (Supplementary material Table 1) for ITS sequencing. A single ITS sequence was analyzed per truffle ground (except when several ITS haplotypes were found) to avoid bias linked to the repeated use of ascocarps originating from the same clonal mycelium. Fruit bodies were collected with the help of local pickers and trained dogs in natural truffle grounds during winter from December 1998 to February 2003. Precise geographical information is not provided to protect the truffle grounds from furtive harvesting. Each ascocarp was washed, its peridium peeled and its inner part (= gleba) was conserved at −80°C pending DNA extraction.

Table 1.  Single nucleotide polymorphism (SNP) sites in the 10 rDNA ITS haplotypes of Tuber melanosporum found among 188 specimens collected throughout its natural range in 17 regions in France, north-western Italy (Piedmont) and north-eastern Spain (Iberian Mountains)
HaplotypesRegional populationsNucleotide position0 1 1 2 3 4 4 5 5 0 4 7 1 6 2 9 0 37 3 2 5 9 9 7 4 5
  1. The most frequent haplotype (haplotype I) was used as the reference sequence. Identical nucleotides are indicated by dots. Regions where haplotypes were sampled are indicated (see Supplementary Material Table 1 for details).

IAll, except LanguedocT C C C C C C T G
IIAll, except Jura, Roussillon, Ariege & TouraineG . . . T . . . C
IIIBurgundy, Piedmont, Inner Provence & Languedoc. . . . . . . . C
IVTarnG . . . . . . . .
VAriege. . T . . . . . .
VIAriege. . . . . T . . .
VIITarnG . . . T . T . C
VIIIJuraG . . . T . G . C
IXDromeG . . A T . . . C
XLanguedoc. G . . . . . . C
image

Figure 1. Map of France showing the regions were accessions of Tuber melanosporum were sampled. Borders of the French administrative departments are outlined. The number of samples for each population is given in brackets (n). See Supplementary material Table 1 Data for details on sampling sites. The groupings of regional populations used in the analysis of molecular variance (amova) are indicated.

Download figure to PowerPoint

DNA manipulations

Total DNA was extracted from 20 mg of gleba with the DNeasy Plant Mini Kit (Qiagen SA, Courtaboeuf, France) following the manufacturer's instructions. Single nucleotide polymorphisms (SNP) were sought in the rDNA internal transcribed sequences (ITS) and in randomly amplified genomic sequences (SCAR). Amplification and sequencing of the nuclear rDNA ITS from 188 specimens were carried out as previously described (Henrion et al., 1994; Martin et al., 2002). Sequencing of rare ITS haplotypes was carried out at least twice on different DNA extracts to rule out any amplification and sequencing artefacts. RAPD amplifications were carried out with primers E4, E20, and G14 (Operon Technology, Alameda, CA, USA), using the conditions described by Gandeboeuf et al. (1997). Five nonpolymorphic RAPD bands were cloned and sequenced from the specimen MEst collected in Bauduen (Var, France). RAPD products were ligated into a pCR4-TOPO plasmid vector of a TOPA-TA Cloning Kit (Invitrogen, Groningen, The Netherlands). Both strands of inserts were sequenced with a Taq Big DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA), and analyzed with an ABI 310 DNA sequencer (Applied Biosystems). Sequences were assembled using Sequencher 3.1.1 (GeneCodes, Ann Arbor, MI, USA) and SCAR primers were designed (available upon request). For a set of specimens representative of the different regional populations, we amplified the SCAR sequences using primers herein designed and PCR conditions used for ITS amplifications. The sequences of ITS and SCARs were deposited in GenBank (National Center for Biotechnology Information, NCBI). Accession numbers are provided in Supplementary material Table 1. SNPs were detected by sequence multialignments using the ‘assembly contig’ routine of Sequencher 3.1.1.

Data analyses

We used three approaches to investigate genetic differences among regional populations: first the index of fixation (Wright, 1978) implemented with amova (Excoffier et al., 1992); second NST-statistics (Pons & Petit, 1996); and third NCA (Templeton, 1998, 2004). The amova analyses was carried out with arlequin 2.001 (Excoffier et al., 1992), which calculated ØST analogous to Wright's FST. We partitioned the genetic variation between three hierarchical components: among seven groups of regional populations (i.e. geographical areas), among 17 regional populations and within regional populations (Fig. 1). Fixation index was calculated for the whole geographical range, between the seven groups of regional populations, and between all region pairs (Schneider et al., 2000). A Neighbor-Joining (NJ) dendrogram was constructed using Slatkin's genetic distance calculated in arlequin, to analyze the relationship between genetic distance among regional populations with their relative geographical positions. The Neighbor-Joining analysis was carried out using paup4.08b (Swofford, 2002). Genetic differentiation among the seven groups of regional populations and among the 17 regional populations was also tested applying NST-statistics (Pons & Petit, 1996) using the programme PERMUT (available from http://www.pierroton.inra.fr/genetics/labo/Software/). This test compares NST with values of GST. NST estimates consider not only differences in the frequencies of haplotypes between populations, as with GST, but also genetic distances between haplotypes. In cases of correspondence between haplotype phylogenies and their geographical distribution, estimates for NST will be greater than GST values (Pons & Petit, 1996).

We used a Mantel test to analyze a potential correlation between geographical (Km) and genetic distance and to estimate the effect of isolation-by-distance (Rousset, 1997; Hutchison & Templeton, 1999). A matrix of distances among the seven groups of regional populations or the 17 regional populations was calculated with the module geodistances of the Package R. 4.0d3 (Casgrain & Legendre, 2001), which used latitudes and longitudes. Mantel statistics were calculated in arlequin on the matrices of geographical distances [Ln (distance in Km)] and genetic distance as Slatkin's linearized pairwise ØSST/1 − ØST; Slatkin, 1995). The signification of the correlations was tested by 9999 random permutations.

The null hypothesis that there is no geographical association between sequence types and geographical localities (i.e. populations) was tested by permutation (10 000 replicates) using the program CHIPERM version 1.2 (Posada, 2000; available at: http://bioag.byu.edu/zoology/crandall_lab/programs.htm). We then estimated a sequence-type network using the programme TCS (Clement et al., 2000), which uses statistical parsimony (Crandall, 1996) for network estimation. The maximum number of mutational steps that constitute a parsimonious connection between two sequence types was calculated with 95% confidence. The resulting network was used to construct the nested clade design with TCS (Clement et al., 2000). NCA has become a common tool in intraspecific phylogeography (Templeton, 2004) and has been used to investigate phylogeographic relationships of rDNA ITS haplotypes (James et al., 2001; Rodríguez-Lanetty & Hoegh-Guldberg, 2002; Wörheide et al., 2002). After the nesting procedure, we employed the NCA of geographical distances (Templeton et al., 1995; Templeton, 1998, 2004) using the program geodis version 2.0 (Posada et al., 2000), with 10 000 permutations. geodis addresses the sampling strategy adequacy and quantifies the degree of confidence in the quantitative distance measure by testing the null hypothesis that the haplotypes (or clades) nested within a high nested clade show no geographic associations given their overall sample numbers.

The NCA of geographical distances, as implemented in the program GEODIS, has two parts. Part one looks for statistically significant geographical associations of sequence types by permutational Χ2 statistics for each clade. Then part two, using different geographical information, estimates two different geographical parameters for each clade in the nested design. Dc(x), the clade distance, measures the average distance of each member of a clade from its geographical centre and measures how geographically widespread are the individuals that bear haplotypes from this clade, that is Dc(x) is a measure for the geographical range of clade x. Dn(x), the nested clade distance, measures the average geographic distance of all members of a clade from the geographical centre of its higher-level nesting clade, which is also estimated by averaging the coordinates of all members of the higher-level nesting clade. (I–T)Dc and (I–Tn)Dn give the average Dc and Dn values for all the interior clades within a nesting clade minus the average Dc and Dn values for all the tip clades in the same nesting clade. All calculated distances can either be significantly large, small or nonsignificant. Significant values from the second part of the analysis are then used to work with the inference key (Templeton, 2004) (available at http://bioag.byu.edu/zoology/crandall_lab/geodis.htm) to infer patterns of population structure, population history or a combination of both. The last version of the inference key reduces the incidence of false positive to a minimum at a cost of a slight reduction in power (Templeton et al., 1995; Templeton, 1998; Cruzan & Templeton, 2000; Templeton, 2004). However, one must be cautious in inferring cause and effect in NCA analyses (Knowles & Maddison, 2002). Criticisms of Knowles & Maddison (2002) and Hey & Machado (2003) regarding NCA do not, however, apply to our study as NCA performed well in the case of range expansion (Templeton, 2004). The criticism of Petit & Grivet (2002), that genetic fixation within populations can bias the outcome of NCA, should not apply to our data, as we analyzed single individuals from isolated truffle grounds.

Results

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

SNP analyses and haplotype frequencies in regional populations

Sequencing 50 kbp of the PCR-amplified SCAR fragments from a representative set of T. melanosporum fruit bodies from the different geographical areas revealed no SNP. By contrast, there were nine variable sites corresponding to 10 different haplotypes (Table 1) across all 188 ITS sequences. For the statistical analyses, only one fruiting body was considered from each truffle ground to avoid the possibility of sampling ascocarps originating from the same fungal individual mycelium. A total of 148 ITS sequences were finally analyzed. ITS haplotypes (variants) I and II were the most common haplotypes (60% and 28%, respectively), whereas haplotype III was less frequent (7%) (Supplementary material Table 1). The remaining haplotypes presented a very low frequency (< 2%). These 10 haplotypes showed a differential geographical distribution (Supplementary material Table 1 ;Fig. 2). Haplotypes I and II were distributed widely in most sampling sites in France, Italy and Spain, whereas the other haplotypes presented a geographically restricted range. Haplotype I was more frequent in the western French Atlantic regions (e.g. 90% of ascocarps in Charentes) than in populations in eastern and north-eastern France (e.g. 20% in Burgundy). By contrast, haplotype II was more frequent in eastern France and north-western Italy (e.g. 71% in Isere and 54% in Italian Piedmont) (Fig. 2). Haplotype III presented a disjunct distribution and was found only in Burgundy (32%), Inner Provence (6%), Languedoc (25%), and Piedmont (18%). Haplotypes IV to X were distributed locally and were restricted to single truffle grounds (Supplementary material Table 1; Fig. 2). Additional surveys are, however, needed to confirm the regional ITS haplotype frequencies in the smaller populations.

image

Figure 2. Distribution of ITS haplotypes of Tuber melanosporum in France. The pie chart diameters are proportional to the number of ascocarps analyzed per region. Black lines delimit areas of distribution of the chlorosplastic DNA (cpDNA) haplotypes of oaks in France (Petit et al., 2002a): haplotype 1 was found in the southern corner of France; haplotype-7 in the Rhone valley, and haplotypes 10–12 in western France. Arrowed lines show potential postglacial re-colonization routes for the Perigord truffle: the Atlantic route and the Rhone valley route.

Download figure to PowerPoint

Genetic differences among regional populations

A FST value of 0.20 (P < 0.001) indicates a significant genetic differentiation among T. melanosporum populations (Wright, 1978). A test of correspondence, carried out using permut (Pons & Petit, 1996), confirmed this genetic differentiation (NST = GST = 0.22). Since the large geographical areas (eastern and western France) were further subdivided into regions occupied predominantly by a few haplotypes (Supplementary material Table 1; Fig. 2), hierarchical partitioning of variation (amova) was conducted within the 17 regional populations, among regional populations, and between geographical areas (i.e. groups of regional populations). The bulk of the total haplotype diversity was found within populations (c. 80%). Additional variation was attributable to differences among populations within a geographical area (c. 3%) (Table 2). A significant variation (c. 17%) was attributable to the difference between geographical areas (groups of regions). The higher pairwise genetic distance value was found for populations found on the eastern and western sides of the Central Massif (i.e. Atlantic regions vs Rhone valley). A NJ dendrogram of FST genetic distances reflected the geographical distribution of the haplotypes described above (Fig. 3). Two main clusters were characterised: populations of the ‘western’ cluster have a frequency of haplotype I higher than haplotype II and presented rare haplotypes (e.g. IV & VII in Tarn). The populations within the ‘western’ cluster were characterized by a low frequency of haplotype II, and included populations found in the Pyrenees, western Atlantic regions and southern France (e.g. haplotype V and VI in Ariege). Populations of the ‘eastern’ cluster have a higher frequency of haplotype II and were found in north-eastern France (Lorraine), and in the Alps (Drome, Isere, Piedmont). Burgundy presented similar percentages of haplotypes I and II and nested outside the main clusters. Thus, the F-statistics showed that the position of the regional populations related to the Central Massif range impacted the haplotype frequencies in the populations and the genetic distances among them.

Table 2.  Analysis of molecular variance (amova) of ITS haplotypic diversity for 17 regional populations within France, north-western Italy and north-eastern Spain
Source of variationd.f.SDDVariance components% Total
  1. The analyses were carried out using 148 collections from different truffle grounds (see sampling localities in Supplementary Material Table 1). Variance was partitioned among groups of regional populations (i.e. geographical areas), among regional populations within these areas, and within regional populations (see Fig. 1). Degrees of freedom (d.f), sums of square deviations (SSD), variance component estimates, and the percentages of the total variance (% Total) contributed by each component. FST was = 0.20, P-value < 0.001.

Among groups of regional populations  620.4900.1213 16.52
Among populations within groups 1016.9820.02480  3.38
Within populations13111.3440.58816 80.10
Total14777.0480.73427100.0
image

Figure 3. Neighbour-joining tree of the 17 regional populations sampled in France, north-western Italy (Piedmont) and north-eastern Spain (Iberian Mountains) based on Slatkin's genetic distances. Frequent and rare haplotypes are indicated.

Download figure to PowerPoint

According to NST-statistics (NST = GST), there is no or only weak overall correspondence between the haplotype phylogenies and geographical distribution (Pons & Petit, 1996). This result is presumably the result of the wide geographical codistribution of the most numerous haplotypes (I & II). However, the NJ tree based on the FST genetic distances between ITS haplotypes (Fig. 3) joined adjacent subregions. The increase in genetic differentiation with geographical distance was marginally significant (Mantel test, P = 0.036). However, the genetic difference drastically increased for population pairs located on different sides of the Central Massif suggesting a potential geographical isolation by this mountainous range. The amova (Table 2), the NST test (data not shown), the NJ tree (Fig. 3), and the Mantel test (data not shown) were not significantly affected by the small size of several of the analyzed populations and the low frequencies of the rare alleles as shown by analyses performed using only subsamples of the entire populations (data not shown). For example, FST, NST and GST values of 0.18, 0.16 and 0.12 were found when the analyses were carried out on the seven larger geographical areas (Fig. 1) confirming the significant genetic differentiation among groups. Most of the results were also obtained even when the analyses were limited to the three common alleles.

Haplotype network and nested design

In this study, we applied a nested clade analysis approach (NCA; Templeton et al., 1995, Templeton, 1998, 2004), to investigate the roles of contemporary ecological processes and population histories in shaping the natural populations of T. melanosporum in France. The utility and applicability of ITS rDNA sequences for phylogeography and NCA has been demonstrated in a study of the basidiomycetous species Schizophyllum commune (James et al., 2001). The initial test for geographical association of haplotypes revealed a significant association between sequence types and geographical locations (Monte Carlo significance < 0.001); therefore we proceeded with the NCA. The nested cladogram comprises 11 0-step clades (haplotypes), four one-step clades (1–1, 1–2, 1–3 and 1–4) and one 2-step clade (2–1) (Fig. 4). The 95% connection limit was established at nine steps, far from the largest number of connections detected between two sequences. Ambiguities were evident in only one area of the sequence-type network, where a closed loop between four haplotypes (I, III, IV and one hypothetical one which was not observed) was found. This ambiguous loop did not affect subsequent nesting procedures. The nesting routine of the programme TCS generated three 1-step nested clades. The 1-step nested clades 1–1, 1–2 and 1–3 are tip clades, and clade 1–4 is an interior clade. Haplotypes located at the tips of the cladogram tended to have restricted geographical distributions, whereas ubiquitous, and presumably ancestral haplotypes, were on interior nodes (i.e. haplotypes I to IV) (Templeton et al., 1995). Clade 1–1 comprised the haplotype III found only in northern Italy, and eastern and southern France (Burgundy, Languedoc, and Inner Provence) and the rare haplotype X collected in Languedoc. In clade 1–2 nested the haplotype II (highly frequent in northern Italy and eastern France) and the tip haplotypes VII (Tarn), VIII (Jura) and IX (Drome). The ancestral haplotype I, predominant in western France, nested in clade 1–3 with the rare haplotypes V and VI (Ariege). Finally, a nonsampled haplotype and haplotype IV (Tarn) nested in clade 1–4. The null hypothesis was rejected (P < 0.0001) for the overall tree (and for all the clades) showing that our sampling strategy was adequate for the phylogeographic analysis.

image

Figure 4. Nested clade design used for nested contingency analysis of geographical associations of 10 ITS haplotypes of Tuber melanosporum. The black dot denotes a hypothetical (not observed) intermediate haplotype. Each line in the network represents one mutational step. Nested design was based on the sequence type network generated by the TCS programme and inferred using the rules defined by Templeton (2004). The Templeton's inference series are given at the bottom of the figure.

Download figure to PowerPoint

The NCA inference analysis (Templeton, 2004) for the two-step clade 2–1 (Fig. 4) indicated a past fragmentation. Such a past fragmentation could account for the different frequency of haplotypes I and II in western and eastern France (see Discussion). The inference series for the more recent outer part of the haplotype tree indicated an isolation of the populations (Fig. 4). In particular, clade 1–2 indicated a restricted gene flow with genetic isolation of populations having the rare haplotypes VII to IX. The analyses of clade 1–1 indicated as well an isolation of the populations, which restricted the dispersion of haplotype X (Languedoc) after arising from haplotype III. The series of inference for clades 1–3 and 1–4 were not conclusive (Fig. 4).

Discussion

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

The moderate variation of the nuclear rDNA ITS from Tuber melanosporum revealed a strong pattern of geographic differentiation. Within 188 isolates collected over the whole geographical distribution of the Perigord truffle, including the northern limits of expansion of this species (Burgundy and Lorraine), 10 distinct haplotypes were distinguished (Table 2; Supplementary material Table 1). One of these, haplotype I, was particularly common, predominant in every population except in eastern France where haplotype II was the most frequent. The presence of one very common haplotype in most populations confirms that this species went through a population bottleneck, after which new allelic haplotypes have originated in low frequencies (Bertault et al., 1998). It appears that some of the southern populations (e.g. Languedoc) had the largest number of haplotypes, while Touraine and Roussillon had only one haplotype each (Fig. 2; Table 1; Supplementary material Table 1), although further surveys (i.e. larger sample size) may modify this distribution. Ascocarps were harvested in different truffle grounds ruling out the sampling of fruiting bodies belonging to the same mycelium.

Index of fixation approach (FST = 0.20), NST (0.22), and amova showed a significant geographic differentiation between populations. We also found a significant association between genetic (FST) and geographical distances (Mantel test) when we clustered the distance values between populations in two groups (one with populations from western France and the other formed by the populations of the Rhone Valley). This suggests that geographical factors (e.g. the Central Massif range) may have had an important impact on the gene flow among populations. These analyses were based on a single locus that can represents, at the best, only a snapshot of the evolutionary history of T. melanosporum (Beerli & Felsenstein, 1999). The occurrence of significant allelic variation in ITS contrasts with the lack of sequence polymorphisms within the SCAR sequences. It is thought that concerted evolution limits variation in the ITS and other regions of ribosomal DNA repeat (Ganley & Scott, 2002) and the current data was therefore unexpected. Additional analyses using the vegetative incompatibility gene sequence and microsatellite loci are currently underway to confirm the geographic differentiation deduced from the ITS polymorphism.

Truffle postglacial history suggested by nested clade analysis

Using NCA, it may be possible to infer the factors that generated genetic differentiation between populations and to predict migration routes (Templeton et al., 1995; Templeton, 2004), although its validity and limitations are debated (Knowles & Maddison, 2002; Petit & Grivet, 2002; Templeton, 2004). NCA incorporates the geographical distribution of the haplotypes to the nested phylogenetic relationships among haplotypes to infer population events in historical order (Templeton, 1998). The geographical contingency analysis supported the differences among populations, as Monte Carlo's test for the association of haplotypes with regions was highly significant (P < 0.001).

If we accept the ‘interior equals oldest’ predictions from coalescent theory (Templeton, 1998), then we can infer that the interior haplotypes (I, II and III) of the network are the oldest ones (Fig. 4). All these haplotypes are present in northern Italy, and they likely represented the main haplotypes present in the glacial refugia. This is in agreement with the hypothesis by Bertault et al. (1998) of a drastic glacial bottleneck as origin of the low polymorphism of the Perigord truffle, and the reduced number of ITS haplotypes of T. melanosporum is consistent with the survival of small and isolated populations in southern Europe refugia. The bottleneck of T. melanosporum populations likely occurred during the last and the coldest glacial period (10 000–16 000 yr ago) (Bertault et al., 1998). Such a population bottleneck probably fixed most of the loci, accounting for the lack of SNPs within the SCAR markers (present work). Subsequent expansion from these smaller populations would carry different combinations of ITS haplotypes to colonized areas. The presence of the three ancestral haplotypes of T. melanosporum in Piedmont supports the ‘Italian peninsula origin’ hypothesis (Bertault et al., 1998) with a migration pathway through the Po plain to the western Alps and a subsequent northward expansion along the western and eastern France migration routes. Following Templeton's inference key, we found a statistically significant past fragmentation in the two step clade 2.1. This past fragmentation probably occurred during the colonisation of calcareous foothills along the Central Massif (Fig. 2). Regions having a higher level of diversity (e.g. Languedoc) may correspond to the separation point of the two colonisation routes, whereas regions at the edge of migration routes, such as Touraine and Lorraine, contain only a subset of the lineages described and therefore exhibit lower levels of allelic richness. After the initial rapid colonization, the Central Massif mountainous range, the mosaic pattern of limestone soils in France and the strict soil requirement of the Perigord truffle may have represented effective barriers limiting the gene flow between regions and maintaining the genetic differences until today. Due to the unfavorable acidic sandy soils of Landes (south-western France), a plausible route to colonize Spanish calcareous regions may have been the eastern coastline of the Pyrenean region, from which the black truffle could reach the calcareous Iberian Mountains in north-eastern Spain (Fig. 2). On the other hand, T. melanosporum probably survived in refugia in Spain with haplotype I and II moving out of Spain to colonize western France and types I, II and III coming out of Italy to colonize Eastern France. A more extensive data set from the Iberian Mountains and other Spanish regions will surely provide further clues to support these contrasted scenarii.

Black truffle and its hosts shared the routes of postglacial re-colonization in France

During the maximum expansion of the glaciers, the deciduous forest of Europe was restricted to the Mediterranean costal zone (Bennet et al., 1991). Climatic and fossil data support the hypothesis that three regions hold the main glacial refugia for the host trees of T. melanosporum: the Iberian and Italian Peninsulas, and the Balkans (Comes & Kadereit, 1998; Taberlet et al., 2001). Since T. melanosporum is an ectomycorrhizal symbiont of oaks and other temperate deciduous trees, such as Tilia and Corylus species, this symbiotic fungus was likely restricted to some of these regions. Molecular and fossil (i.e. pollen) data have revealed that the two main northwards re-colonization routes for oaks and many hardwood species (including Tilia and Corylus species) were the Rhone valley and the western plain of France (Brewer et al., 2002; Petit et al., 2002a; 2002b; 2003). They determined the current distribution of the haplotypes of oak species (e.g. the thermophilous Quercus pubescens) in France (Fig. 3 in Petit et al., 2002a). In Provence, the oak haplotype 1 is the most prominent haplotype (Fig. 2). In the Rhone valley, haplotype 7 dominates, whereas the western France populations are dominated by the closely related oak haplotypes 10, 11 and 12. Although edaphic (e.g. calcareous soils) and paleoclimatic factors likely influenced T. melanosporum postglacial recolonization patterns, the NCA-suggested routes of expansion for the Perigord truffle in France closely resemble that of Q. pubescens (Petit et al., 2002a). It is therefore tempting to speculate that the oak re-colonization routes have impacted on the distribution of T. melanosporum populations during their northward migration (Fig. 2). It has already been suggested that the northwards expansion of plant and animal species from southern refugia likely followed the establishment of deciduous woodland. For example, oak and hedgehog species followed similar expansion patterns (Seddon et al., 2001). Mycophagous animals (e.g. squirrels, wild boar) known to dispersed truffles spores probably favored this expansion. Oak species can withstand the cold conditions of northern Europe, while T. melanosporum is restricted to mid-latitudes. It is therefore possible that this truffle species migrated northwards later than oaks and thus oaks could have been widespread by the time T. melanosporum reached France. These various hypotheses will be investigated using additional methods to infer population histories, and larger fruit body surveys.

Acknowledgements

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

This project has benefited from a Ph.D. fellowship of the Italian ISASUT (University of Torino) to Claude Murat. J. Díez is a receipt of a ‘Ramon y Cajal’ Researcher Contract from the Spanish ‘MCyT’. We thank Drs J-L. Jany (Laval University, Quebec) and A. Mello (University of Torino, Italy) for stimulating discussions, and Dr Rémy Petit (INRA-Pierroton) for discussion and running PERMUT on the current data. The provision of samples by M. L. Ballureau, C. Bazin, F. Beaucamp, P. Berthet, C. Beneteau, C. Bezombes, O. Bois, M. Bombled, I. Calvo, R. Chinouilh, A. Clare, A. Daniel, L. Danglade, J. P. Ducret, Mr Fayolle, B. Finot, H. Frochot, C. Galoger, A. Goyon, L. Genola, Mr Gillot, F. Gobeau, J. Goubier, A. Guillén, M. Honrubia, H. Huguet, G. Jallifier, Mr Junique, J. L. Lacam, A. Lauriac, Mr Lespinasse, J. Manreza, P. Materne, A. Meunier, M. Mondion, P. Moulin, S. Pallares, J. L. Prune, C. Raquillet, P. Rejou, J. F. Rey, R. Ribes, L. & G. Riousset, J. M. Rocchia, J. J. Roux, T. Sanchez, P. Sourzat, Y. Soulas, J. P. de Santis, P. Tabouret, A. Terrade, O. Viguier, J. Verbinski and L. Vergé is greatly appreciated. We also would like to thank all authors of freely available genetics software. We are grateful for the useful comments of the referees on an earlier version of the manuscript. This project was funded by grants from the Bureau des Ressources Génétique (research project BRG 69–68), the Institut Français de la Biodiversité (CNRS programme ‘Environnement, vie et sociétés’), INRA and the Italian Project ‘Tuber: Biotecnologia della micorrizazione’ funded by CNR.

Supplementary material

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

The following material is available as Supplementary material at http://www.blackwellpublishing.com/products/journals/suppmat/NPH/NPH1189/NPH1189sm.htm

Table S1 List of Tuber melanosporum specimens and sampling localities used in the present study.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information
  • Beerli P, Felsenstein J. 1999. Maximum-likelihood estimation of migration rates and effective population numbers in two populations using a coalescent approach. Genetics 152: 763773.
  • Bennet KD, Tzedakis PC, Willis KJ. 1991. Quaternary refugia of North European trees. Journal of Biogeography 18: 103115.
  • Bertault G, Raymond M, Berthomieu A, Callot G, Fernandez D. 1998. Trifling variation in truffles. Nature 394: 734.
  • Bertault G, Rousset F, Fernandez D, Berthomieu A, Hochberg ME, Callot G, Raymond M. 2001. Population genetics and dynamics of the black truffle in a man-made truffle field. Heredity 88: 451458.
  • Brewer S, Cheddadi R, De Beaulieu JL, Reille M. 2002. The spread of deciduous Quercus throughout Europe since the last glacial period. Forest Ecology and Managment 156: 2748.
  • Callot G, Byé P, Raymond M, Pargney JC, Parguey-Leduc A, Janex-Favre MC, Moussa R, Pagès L. 1999. La Truffe, la Terre, la Vie. Paris, France: INRA Press.
  • Casgrain P, Legendre P. 2001. The R package for multivariate and spatial analysis, Version 4.0d3 – User's Manual. URL.http://www.fasumontreal.ca/BIOL/legendre/.
  • Chevalier G, Dupré C. 1990. Recherche et expérimentation sur la Truffe et la trufficulture en France. In: BencivengaM, GranettiB, eds. 2nd Congresso Internazionale Sul Tartufo. Spoleto, Italy: Comunità Montana dei Monti Martani e del Serano,157166.
  • Chevalier G, Grente J. 1978. Applications pratiques de la symbiose ectomycorhizienne: production à grande échelle de plants mycorhizés par la truffe (Tuber melanosporum). Mushroom Science X: 483505.
  • Clement M, Posada D, Crandall KA. 2000. TCS: a computer program to estimate gene genealogies. Molecular Ecology 9: 16571659.
  • Comes H, Kadereit JW. 1998. The effect of Quaternary climatic changes on plant distribution and evolution. Trends in Plant Science 3: 432438.
  • Crandall KA. 1996. Multiple interspecies transmissions of human and simian T-cell leukemia/lymphoma virus type I sequences. Molecular Biology and Evolution 13: 115131.
  • Cruzan MB, Templeton AR. 2000. Paeloecology and coalescence phylographic analysis of hypothesis from the fossil record. Trends in Ecology and Evolution 15: 491496.
  • Delmas J. 1978. Tuber spp. In: ChangST, HayesWA, eds. The biology and cultivation of edible mushrooms, London, UK: Academic Press, 645681.
  • Excoffier L, Smouse PE, Quattro JM. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131: 479491.
  • Gandeboeuf D, Dupré C, Roecket-Drevel P, Nicolas P, Chevalier G. 1997. Grouping and identification of Tuber species using RAPD markers. Canadian Journal of Botany 75: 3645.
  • Ganley ARD, Scott B. 2002. Concerted evolution in the ribosomal RNA genes of an Epichloe endophyte hybrid: comparison between tandemly arranged rDNA and dispersed 5S rRNA Genes. Fungal Genetics and Biology 35: 3951.
  • Hall IR, Yun W, Amicucci A. 2003. Cultivation of edible ectomycorrhizal mushrooms. Trends in Biotechnology 21: 433438.
  • Henrion B, Chevalier G, Martin F. 1994. Typing truffle species by PCR amplification of the ribosomal DNA spacers. Mycological Research 98: 3743.
  • Hewitt GM. 1996. Some genetic consequences of ice ages and their role in divergence and speciation. Biology Journal of the Linnean Society 58: 247276.
  • Hewitt GM. 1999. Post-glacial re-colonization of European biota. Biology Journal of the Linnean Society 68: 87112.
  • Hey J, Machado CA. 2003. The study of structured populations – new hope for a difficult and divided science. Nature Reviews Genetics 4: 535543.
  • Hutchison DW, Templeton AR. 1999. Correlation of pairwise genetic and geographic distance measures: inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution 53: 1891914.
  • James TY, Moncalvo JM, Li S, Vilgalys R. 2001. Polymorphism at the DNA spacers and its relation to breeding structure of the widespread mushroom Schizophyllum commune. Genetics 157: 149161.
  • Knowles LL, Maddison WP. 2002. Statistical phylogeography. Molecular Ecology 11: 26232635.
  • Martin F, Díez J, Dell B, Delaruelle C. 2002. Phylogeography of the ectomycorrhizal Pisolithus species as inferred from the nuclear ribosomal DNA ITS sequences. New Phytologist 153: 345358.
  • Petit RJ, Aguinagalde I, De Beaulieu JL, Bittkau C, Brewer S, Cheddadi R, Ennos R, Fineschi S, Grivet D, Lascoux M, Mohanty A, Muller-Starck G, Demesure-Musch B, Palme A, Martin JP, Rendell S, Vendramin GG. 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300: 15631565.
  • Petit RJ, Brewer S, Bordacs S, Burg K, Cheddadhi R, Coart E, Cottrell J, Csaikl UM, Van Dam B, Deans JD, Espinel S, Finschi S, Finkeldey R, Glaz I, Giocoechea PG, Jensen JS, Könìg AO, Lowe AJ, Madsen SF, Matyas G, Munro MC, Popescu F, Slade D, Tabbener H, De Vries SGM, Ziegenhagen B, De Beaulieu JL, Kremer A. 2002b. Identification of refugia and post-glacial colonisation routes of European white oaks based on chloroplast DNA and fossil pollen evidence. Forest Ecology and Management 156: 4974.
  • Petit RJ, Grivet D. 2002c. Optimal randomization strategies when testing the existence of a phylogeographic structure. Genetics 161: 469471.
  • Petit RJ, Latouche-Hallé C, Pemonge MH, Kremer A. 2002a. Chloroplast DNA variation of oaks in France and the influence of forest fragmentation on genetic diversity. Forest Ecology and Management 156: 115129.
  • Pons O, Petit RJ. 1996. Measuring and testing genetic differentiation with ordered versus unordered alleles. Genetics 144: 12371245.
  • Posada D. 2000. Testing geographic association – Chiperm, version 2.1. Provo, UT, USA: Department of Biology, Brigham Young University.
  • Posada D, Crandall KA, Templeton AR. 2000. GeoDis: a program for the cladistic nested analysis of the geographical distribution of genetic haplotypes. Molecular Ecology 9: 487488.
  • Rodríguez-Lanetty M, Hoegh-Guldberg O. 2002. The phylogeography and connectivity of latitudinally widespread sceractinian coral Plesiastrea versipora in the western Pacific. Molecular Ecology 11: 11771189.
  • Rousset F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145: 12191228.
  • Schneider S, Roessli D, Excoffier L. 2000. ARLEQUIN, Version 2.001: a software for population genetics data analysis. Geneva, Switzerland: Genetics and Biometry Laboratory, University of Geneva.
  • Seddon JM, Santucci F, Reeve NJ, Hewitt GM. 2001. DNA footprints of European hedgehogs, Ericacevs europaeus and E. Concolor: Pleistocene refugia, postglacial expansion and colonization routes. Molecular Ecology 10: 21872198.
  • Slatkin M. 1995. A measure of population subdivision based on microsatellite allele frequences. Genetics 139: 457462.
  • Smith SE, Read DJ. 1997. Mycorrhizal symbiosis, 2nd edn. New York, NY, USA: Academic Press.
  • Swofford DL. 2002. paup: Phylogenetic analysis using parsimony (and other methods), Version 4.08b10 for Macintosh. Sunderland, MA, USA: Sinauer Associates Inc.
  • Taberlet P, Fumagalli L, Wust-Saucy AG, Cosson JF. 2001. Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology 7: 475464.
  • Templeton AR. 1998. Nested clade analyses of phylogenetic data: testing hypothesis about gene flow and population history. Molecular Ecology 7: 381397.
  • Templeton AR. 2004. Statistical phylogeography: methods of evaluating and minimizing inference errors. Molecular Ecology 13: 789809.
  • Templeton AR, Routman E, Phillips CA. 1995. Separating population structure from population history: a cladistic analysis of the geographical distribution of mitochondrial DNA haplotypes in the tiger salamander, Ambystoma tigrinum. Genetics 140: 767782.
  • Wörheide G, Hooper JN, Degnan MB. 2002. Phylogeography of western pacific Leucetta ‘chagosensis’ (Porifera: Calcarea) from ribosomal DNA sequences: implications for populations history and conservation of the Great Barrier Reef World Heritage Area (Australia). Molecular Ecology 11: 17531768.
  • Wright S. 1978. Evolution and the genetics of populations, Vol. 4. Variability within and among natural populations. Chicago, IL, USA: University of Chicago Press.

Supporting Information

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

Table S1. List of Tuber melanosporum specimens and sampling localities used in the present study. The code number of the sampling locality is indicated in brackets. Regions are mapped in Fig. 1. The ITS haplotypes and the GenBank accession numbers are indicated for each accession.

FilenameFormatSizeDescription
NPH_1189_sm_TableS1.doc105KSupporting info item