The black Perigord truffle (Tuber melanosporum) is recognized worldwide as an icon of European gastronomy. Its edible fruitbody is a hypogeous fleshy structure producing meiospores (Fig. 1). This ascomycete fungus is ectomycorrhizal, that is, symbiotically associates with tree roots. The demand for this highly appreciated delicacy and the decrease of its production over the twentieth century (Savignac et al., 2012) have fuelled intense efforts at its cultivation, and at sequencing its genome (Martin et al., 2010). In this issue of New Phytologist, Murat et al. (pp. 176–187) analyze T. melanosporum population genetics in two truffle plantations. They reveal genotypes extending over a few meters, displaying a strong genetic structure at fine scale, with a pattern of isolation-by-distance within the plantation, and a striking spatial segregation of genotypes according to their mating type.
‘“Dwarf males” exist in animal and plant species, such as dioecious mosses … in truffles, their existence remains an appealing speculation.’
Tuber melanosporum is heterothallic: mating can only occur between haploid cells of different mating types, that is, carrying different alleles at the MAT locus (Billiard et al., 2012). In ascomycetes such as T. melanosporum, vegetative hyphae are haploid, and fruiting first requires mating (Fig. 1). In fact, marker segregation has been observed when extracting DNA from meiospores in T. melanosporum, confirming that same-clone mating is prevented (Paolocci et al., 2006; Riccioni et al., 2008). Truffle flesh, from which most if not all DNA is extracted by standard protocols, arises only from the female parent that builds and feeds the fruitbody. The male genotype can be deduced, by difference, from the sporal genotype (Rubini et al., 2011).
Spatial segregation of mating types
Murat et al. observed that spatially close individuals differing in their genotypes, according to microsatellite markers, carry the same mating type. Rubini et al. (2011) observed such spatial segregation, and showed that it emerged secondarily: on each nursery-grown seedling, ectomycorrhizas from multiple individuals with different mating types initially co-occurred, and the dominance of mycelia carrying the same mating type emerged after several months. Such spatial segregation may be due to competitive exclusion between different genotypes, with use of the MAT locus as a marker for self-recognition, probably in addition to other polymorphic loci.
Although the MAT locus is rarely used as a marker for self-recognition in fungi, in Neurospora crassa, Sordaria brevicollis, Ascobolus stercorarius and A. heterothallicus the MAT locus is one of the loci controlling vegetative incompatibility (Glass et al., 2000). Vegetative incompatibility is a common phenomenon in filamentous fungi that often results in death of the hyphal cells that have fused between individuals carrying different alleles at the loci involved in self-recognition (Glass et al., 2000). Vegetative incompatibility is thought to protect resources within hyphae from exploitation by non-kin (Debets & Griffiths, 1998), whereas kin cooperate by sharing space and resources. This is selected for because the allele present in an individual controlling for altruism towards related individuals is often present in these related individuals: the allele controlling altruism therefore benefits from the altruism for its transmission to the next generation (Hamilton, 1964). The loci controlling vegetative incompatibility often display high degrees of polymorphism and even trans-specific polymorphism (Debets & Griffiths, 1998; Wu et al., 1998), both of which are footprints of balancing selection expected at markers used for kin recognition. Competitive exclusion of non-kin is expected in fungi, and particularly in symbiotic fungi that would benefit from monopolizing the host root system, and in fact, vegetative incompatibility has been reported in plant parasites (López-Villavicencio et al., 2011) and for ectomycorrhizal basidiomycetes (Worrall, 1997). Competitive exclusion of non-kin would have further consequences: fewer genetic conflicts are expected in this case on a root system, due to kin selection among genets (Buckling & Brockhurst, 2008), which should lead to more cooperative behaviors, and thereby to more benefits for the symbiotic association, including the tree – an intriguing prediction that deserves testing.
Why would selection recruit the MAT locus for vegetative incompatibility? Vegetative incompatibility typically involves multiple, non-homologous loci, which increases the precision of kin recognition (Glass et al., 2000), and recruiting MAT would increase the number of discriminating loci. However, the MAT locus is biallelic in ascomycetes (Billiard et al., 2012), and thus poorly efficient for kin recognition. The coupling of sexual and vegetative incompatibility furthermore complicates mating, where cell fusion occurs between two different mating types, while vegetative fusion requires identical alleles at the vegetative incompatibility loci. Developmental switch must then evolve to differentiate the compatibility at the sexual vs vegetative stages, with changes in gene expression, as described in Neurospora crassa (Shiu & Glass, 1999). One possible evolutionary explanation for the link between mating type and vegetative incompatibility may be that the resulting spatial segregation decreases the probability that compatible cells meet and therefore lowers the odds of mating; this is beneficial under the assumption that sex is costly (Otto & Lenormand, 2002). The best strategy for reproduction is indeed the clonal propagation of the fittest allelic combinations, with only some rare recombination events for purging deleterious mutations and producing new, possibly beneficial, allelic combinations. If a link between vegetative incompatibility and mating type evolved to avoid ‘sexual harassment’, this would be bad news for truffle production!
Looking for a father
However, avoiding sexual harassment may not be simple: the steps toward truffle fruitbodies remain unknown (Fig. 1), and as recognized by Murat et al., we ignore what structure plays the role of the male partner. We therefore ignore whether gamete limitation actually occurs in truffle grounds. In fact, Linde & Selmes (2012) found no difference in fruitbody production between root systems displaying a single vs both mating types, and Murat et al. did find fruitbodies within large areas dominated by a single mating type.
In related ascomycetes, the male partner can produce a differentiated hyphal structure, an undifferentiated hypha, or a passively mobile cell, called a ‘microconidium’, or more appropriately a spermatium (Fig. 1). In this case, dispersal of spermatia may alleviate gamete limitation. Urban et al. (2004) and Healy et al. (2012) reported occurrence of candidate spermatia in ectomycorrhizal Pezizales and Tuber spp., whose failures to germinate on sterile media or form ectomycorrhizas support a role of spermatia. Their small size (5 μm in diameter), thin cell walls and lack of reserves are unusual features for asexual multiplication. If spermatia were described in T. melanosporum, their dispersal range would be of interest: indeed, Murat et al. and Rubini et al. (2011) did not identify the males in the plantations, suggesting a migration from quite far.
Alternatively, mating may require close contact, either because spermatia do not disperse, or because direct hyphal contact is required (Fig. 1). But even this does not necessarily imply gamete limitation. Indeed, PCR amplifications from soil sometimes detect both mating types, as reported by Murat et al. and Rubini et al. (2011). Such DNA may issue from the spore bank expected for hypogeous fruitbodies, due to specimens not removed by dispersers (Grubisha et al., 2007; Fig. 1). However, spore DNA is notoriously difficult to extract: for example, DNA extracted from fruitbodies reveals only maternal alleles and a single mating type. Instead, very small mycelia, attached to a few roots, or even nonmycorrhizal germinations, may exist and act as males (Douhan et al., 2011). They cannot act as female, since they would not have enough resources to sustain fruitbody growth. ‘Dwarf males’ (Fig. 1) exist in animal and plant species, such as dioecious mosses (Hedenäs & Bisang, 2011); in truffles, their existence remains an appealing speculation.
A modern protodomestication
We thus do not control truffle reproduction, and ignore whether some gamete limitation occurs. This fits the definition of protodomestication, where the harvest is enhanced by empirical treatments favoring establishment and persistence (without control of reproduction, which is the hallmark of a true domestication). Protodomestication is a difficult step as long as the biology of the target species remains unknown, during which some practices can slow down the emergence of interesting traits. In wheat and barley, for example, archaeological data show that the emergence of grain indehiscence, a major trait that allows the harvesting of all mature seeds, took more than one millennium (Tanno & Willcox, 2006). This is far more than requested using modern selection. Facing wild populations where dehiscence is the rule to allow seed dispersal, early farmers ignored this possibility and harvested cereals before ripening to avoid seed loss (Tanno & Willcox, 2006), so that indehiscent mutants were only weakly selected for. Furthermore, selection for smaller grains for sowing vs larger ones for human feeding likely delayed the augmentation of grain size.
It is thus still preliminary to recommend practices for truffle producers and to comment about existing empirical methods. As highlighted by Murat et al., European truffle producers often disseminate pieces of mature truffles to re-inoculate the soil in already established plantations. While this may increase population density and counter-act gamete limitation if any, the use of lower-quality fruitbodies for this purpose, to sell the best ones, may induce a huge genetic load for future generations. It may favor genetic traits determining low flavor and/or size, or even high sensitivity to pathogens. Similarly, the use of frozen truffles (M-A. Selosse & E. Taschen, pers. obs.) found after the harvesting season, may select for slow maturating or late initiated truffles, maladapted to local conditions. The same applies to fruitbodies used for nursery inoculation.
Since inoculated plantations now produce > 80% of the fruitbodies (Savignac et al., 2012), their genetic quality becomes determinant for future production and inoculants. Truffle producers should avoid inadequate practices reminiscent of those of early cereal domestication; the antagonism between selling the best truffles immediately, and using some of them as inoculants to improve the truffle grounds in the long term should be explained.
Uncertainties about the truffle life cycle lead us to ancestral problems in the domestication process. This also calls for more studies of populations in natural vs planted stands, to check the impact of the ongoing protodomestication. Tuber melanosporum populations are mostly studied in plantations, where trees had been inoculated in nurseries, and where the observed patterns are partly of anthropogenic origin. ‘Wild’ populations should be compared for genetic diversity, strength of the spatial segregation of mating types and other loci, at various scales – although anthropic genetic disturbance has likely already occurred. The large diffusion of nursery-inoculated trees may have impacted T. melanosporum genetic structure at regional scales. Many empirical traditions typical of European truffle production (Savignac et al., 2012) also await studies to validate them: for example no genetic study has so far rigorously validated the inoculation method by monitoring the persistence of the inoculants in the long term.
Population genetics of ectomycorrhizal ascomycetes remains too rarely investigated (Douhan et al., 2011): Murat et al. and Rubini et al. (2011) have contributed to change this. The possibility, already described, of accessing the male genotype raises exciting possibilities to unravel the range of male gene dispersal, the mate choice, and the nature of the male contribution (Fig. 1). Identifying the location of fathers, by increasing sample sizes and spatial scale of sampling, would also be enlightening. Last, many other ectomycorrhizal ascomycetes (Healy et al., 2012) also await similar investigations.