Mating Type Locus of Chinese Black Truffles Reveals Heterothallism and the Presence of Cryptic Species within the T. indicum Species Complex

Tuber spp. are filamentous ascomycetes which establish symbiosis with the roots of trees and shrub species. By virtue of this symbiosis they produce hypogeous ascocarps, known as truffles. Filamentous ascomycetes can reproduce by homothallism or heterothallism depending on the structure and organization of their mating type locus. The first mating type locus in a truffle species has been recently characterized in Tuber melanosporum and it has been shown that this fungus, endemic in Europe, is heterothallic. The availability of sequence information for T. melanosporum mating type genes is seminal to cloning their orthologs from other Tuber species and assessing their reproductive mode. Here we report on the organization of the mating type region in T. indicum, the black truffle species present in Asia, which is the closest relative to T. melanosporum and is characterized by an high level of morphological and genetic variability. The present study shows that T. indicum is also heterothallic. Examination of Asiatic black truffles belonging to different genetic classes, sorted according to the sequence polymorphism of the internal transcribed spacer rDNA region, has revealed sequence variations and rearrangements in both coding and non-coding regions of the mating type locus, to suggest the existence of cryptic species within the T. indicum complex. The presence of transposable elements within or linked to the mating type region suggests a role of these elements in generating the genotypic diversity present among T. indicum strains. Overall, comparative analyses of the mating type locus have thus allowed us to tackle taxonomical and phylogenetic issues within black truffles and make inferences about the evolution of T. melanosporum-T. indicum lineage. Our results are not only of fundamental but also of applied relevance as T. indicum produces edible fruit bodies that are imported also into Europe and thus may represent a biological threat for T. melanosporum.


Introduction
Sexual reproduction, by virtue of its essential role in determining genetic variability and eliminating deleterious mutations, is at the origin of eukaryotic evolution. In fungi, sexual reproduction is controlled by genomic regions called mating type (MAT) loci [1,2,3]. The filamentous ascomycetes (Pezizomycotina) have a single MAT locus called MAT1 [4,5] with two master regulators of sexual reproduction: the MAT1-1-1 and the MAT1-2-1 genes encoding for an a-box and a high mobility group (HMG) domain protein, respectively [2]. Filamentous ascomycetes are haploid fungi that have two main sexual reproductive modes: homothallism and heterothallism. In homothallic species, both master MAT genes are present in the haploid nucleus of each strain. Thus, these fungi do not have distinctive mating types and are capable of haploid selfing as well as of crossing with any other strain [6]. Conversely, in heterothallic species, the two master MAT genes are located in different strains, thus haploid selfing is prevented and sexual reproduction can occur only between individuals of opposite mating type. The two alternative forms of the MAT locus in these fungi have been termed idiomorphs (MAT1-1 and MAT1-2) rather than alleles because of their dissimilar sequences [7].
Within Pezizomycetes, the fungi belonging to the genus Tuber are organisms that establish symbiosis with the roots of many tree and some shrub species. By virtue of this symbiosis they produce hypogeous fruit bodies, known as truffles. Truffles produced by some species such as Tuber melanosporum Vittad. and Tuber magnatum Pico are edible and highly prized and appreciated by gourmets worldwide for their distinctive aroma. Thus, unearthing the reproductive biology of these species is not only of fundamental but also of applied relevance. The reproduction mode of truffles has been under debate for decades. Gaining direct evidence on their reproductive strategy is hampered by the difficulties of growing and the impossibility of mating these symbiotic fungal species under controlled conditions. Although mating type genes have been isolated from many members of the Pezizomycotina [8], their poor conservation across fungi made the identification of these genes in Tuber spp. by homologous cloning impossible [9]. Despite this, Bertault and colleagues [10] postulated the existence of a selfing reproductive system in Tuber spp. in order to account for the absence of heterozygotes observed when T. melanosporum ascocarps were screened with codominant simple sequence repeat (SSR) markers. However, few years later, an SSR-based screening aimed at investigating the genetic diversity within and among natural populations of T. magnatum showed that outcrossing events in this species were possible. As a matter of fact, two locus and multilocus linkage disequilibrium analyses documented the occurrence of an extensive genetic flow within and between populations [11]. Direct evidence of outcrossing was then gained when protocols were set to separately isolate and analyse the DNA contributed by the spores and the surrounding gleba from the same ascocarps. The evidence that the spores from a subsample of T. magnatum and T. melanosporum ascocarps displayed additional SSR alleles with respect to those showed by the surrounding gleba was interpreted to mean that these fungi can outcross and that gleba of their truffles is of uniparental origin [12,13].
Finally, the sequencing of the T. melanosporum genome, strain MEL28, has revealed the presence of a single mating type (MAT) gene, namely the MAT1-2-1 gene [14]. Because the second MAT gene, MAT1-1-1, has been cloned from a strain that lacked MAT1-2-1 and the use of MAT1-1-1 and MAT1-2-1 specific primer pairs confirmed the presence of either MAT1-1-1 or MAT1-2-1 in the gleba of all T. melanosporum truffle investigated, it has been concluded that this fungus is heterothallic [14,15]. An international patent covering the use for commercial purposes of the truffle MAT genes has been filed [16].
Different Tuber spp. produce edible black truffles. Because T. melanosporum has a distinctive taste, its commercial value is much higher than that of other species. Of these, the Asiatic species T. indicum Cooke and Massee has black ascocarps with morphological characteristics very similar to those of T. melanosporum. However, their flavor is inferior [17,18]. Notably, the ectomycorrhizas formed by these two species are morphologically indistinguishable [19,20]. Because the growing worldwide market for Chinese black truffles and their high level of similarity to T. melanosporum, they have been extensively studied over the last two decades. Indeed, black truffles of Asiatic origin may display many different morphotypes, leading to several authors claiming the presence of several other species such as T. himalayense, T. sinense, T. pseudohimalayense, T. pseudoexcavatum and T. formosanum in addition to T. indicum [21,22,23,24,25]. Extensive morphological and molecular analyses have confirmed the validity of only a few of these taxa: T. pseudohimalayense and T. pseudoexacavatum have been deemed to be conspecific, whereas T. sinense, T. pseudohymalayense and T. formosanum cannot be unambiguously distinguished from T. indicum [26]. With regard to the morphotypes belonging to T. indicum sensu stricto, extensive genetic variability was first revealed by studies based on RFLP analysis of the ITS region with the identification of at least three haplotypes referred to as T. indicum_A, T. indicum_B1 and T. indicum_B2 [27]. Subsequent studies based on ITS region and other phylogenetically informative genes such as 18S and 28S rDNA, b-tubulin and 1aelongation factor have all confirmed the existence of two main clades within T. indicum [28,29,30]. However, due to their high morphological polymorphism, species boundaries within this group of fungi remain elusive [26,31]. Nevertheless, molecular analyses have shown that T. melanosporum is closer to T. indicum than to the sympatric European black species such as T. brumale Vittad. [27,30,31,32,33,34]. Based on phylogenetic analyses and current species distribution it has been also hypothesized that the Asiatic T. indicum and European T. melanosporum likely originated about 25-36 Ma ago by vicariance speciation [34]. The availability of sequence information for the T. melanosporum MAT genes is seminal to cloning their orthologs from Asiatic black truffles and assessing whether or not heterothallism is shared between closely related black truffle species. Even closely related ascomycetes may have different lifestyles and reproductive modes: for example, within the genera Cochliobolus, Neurospora and Aspergillus, both heterothallic and homothallic species have been described [35,36,37].
In addition, being considered rapidly evolving sequences [38], the MAT genes are ideal markers for gaining further insights into the taxonomy of Asiatic black truffles and the phylogeny of the T. indicum -T. melanosporum lineage.
The present study therefore sets out to: (i) isolate and characterize the MAT locus of T. indicum to ascertain the reproductive mode of this species; (ii) evaluate the extent of MAT polymorphism among different strains of T. indicum representative of the three genetic classes identified thus far and (iii) compare the sequences of the MAT genes and the structures of MAT idiomorphs of T. indicum and T. melanosporum. The discovery of the reproductive mode of T. indicum also has repercussions of an applicative nature, due to increasing interest in its cultivation in China [39] and a concern that this species, which is more competitive than T. melanosporum and potentially invasive, may represent a threat for T. melanosporum in Europe [27,40,41].

Fungal Samples and Morphological Analyses
A set of Asian black truffles imported into Italy and in France (Table S1) was analyzed for the morphology of ascocarps and ascospores as previously reported [18,21,26]. Their ascospores and in particular their sporal ornamentations were examined and photographed using a light microscope (Zeiss Axiophot) and a scanning electron microscope (SEM, Philips XL30). For light microscope analysis the spores were mounted in lactic acid; for SEM examination the ascospores, collected from ripened ascocarps, were washed in distilled water and preserved in 95% ethanol. A small drop containing the spores was placed on aluminum stubs covered with double-sided adhesive tape, dried at 25uC and coated with Gold-Palladium using an E1500 SEM Coating Unit (Polaron Equipment Ltd, Watford, UK). Lyophilized specimens of truffles used in this study are conserved at the CNR, Institute of Biosciences and BioResources, Perugia Division.

DNA Isolation, PCR Amplification and Sequencing
Genomic DNA was isolated from ascocarps as described previously [12]. PCR amplification of the ITS region was performed using ITS1 and ITS4 primers [42]. The RFLP analysis of ITS amplicons was performed using the RsaI endonuclease in accordance with the procedure of Paolocci et al. [27].
T. indicum aand HMG-box fragments were amplified by PCR using the T. melanosporum primer pairs p1/p2 and p19/p20 and the cycling conditions reported in Rubini et al. [15] with the following modifications: the annealing temperature was decreased to 50uC and the number of cycles to 25.
The PCR primers used for cloning the T. indicum MAT genes and idiomorphs were designed with the help of PerlPrimer v. 1.1.16 software [43] and are reported in Table S2. PCR amplification of the mating type genes was performed in a 50 ml mixture containing 2.5 mM MgCl 2 , 0.2 mM of each dNTPs, 10 nM of each primer, 1 U of EuroTaq DNA polymerase (Euroclone, Milan, Italy) and 20 ng of DNA. PCR reactions were performed in a 9700 PCR system (Life Technologies, Carlsbad, CA, USA) using the following thermal profile: 2 min and 30 s at 94uC followed by 25-30 cycles of denaturation at 94uC for 30 s, annealing at 60uC for 30 s and extension at 72uC for 30-60 s depending on the length of the amplicons, and a final extension of 7 min at 72uC.
Long distance PCR amplification of MAT1-1 and MAT1-2 fragments was performed with LA-Taq DNA polymerase (Takara Bio Inc., Otsu, Japan) using a 2-step thermal profile: 30 s at 94uC followed by 25-30 cycles of denaturation at 94uC for 30 s and annealing/extension at 68uC for 10-15 min depending on the length of the amplicons, and a final extension of 7 min at 72uC.
To amplify both idiomorphs from a single ascocarp, the longdistance PCR protocol described above was modified by increasing the number of PCR cycles to 50. Sequencing of the PCR products was performed using the Big Dye sequencing kit V. 3.1 (Life Technologies, Carlsbad, CA, USA) in accordance with the manufacturer's protocol either on amplicons purified using a JetQuick PCR purification kit (Genomed, GmbH, Löhne, Germany) or on amplicons previously cloned in E. coli strain JM83 using standard protocols. Sequence visualization, editing and assembly were performed using FinchTV v. 1.3.1 (Geospiza, Inc., Seattle, WA, USA; http://www.geospiza.com) and Geneious v. 4.8.5 (Biomatters Ltd, Auckland, New Zealand; http://www. geneious.com). Gene prediction analyses were performed using FGENESH (http://linux1.softberry.com/berry.phtml).

RNA Isolation and Reverse-transcription Polymerase Chain Reaction Analysis
Total RNA was isolated from MAT1-1 and MAT1-2 ascocarps stored at 280uC following Rubini et al. [15]. To confirm the structure of MAT genes, 5 mg of total RNA was reversetranscribed using the SuperScriptH III Reverse Transcriptase enzyme and 200 ng of random primers (Life Technologies, Carlsbad, CA, USA) according to the supplier's instructions.

Morphological and Molecular Characterization of T. indicum Samples
Given the high genetic variability of Chinese black truffles, a set of 115 ascocarps all displaying the T. indicum morphotypes was analyzed for the polymorphism of the ITS rDNA region. The PCR amplification of the ITS region with universal ITS1/ ITS4 primers produced an amplicon of approximately 600 bp in all samples. The ITS/RFLP analysis with RsaI endonuclease allowed the classification of all the ascocarps into the 3 genetic classes (A, B1, B2) previously identified by Paolocci et al. [27] ( Table S1). In order to evaluate putative morphological differences between truffles belonging to the different RFLP classes, a detailed analysis of ascospore morphology was performed with scanning electron microscope (SEM, Figure 1) and light microscope (Figures S1, S2 and S3). SEM analysis showed a highly variable sporal ornamentation ranging from the aculeate type with free small or large spines (Figure 1a-b) to the aculeate-reticulate type where the spines were more or less enlarged and connected at the base (e.g. Figure 1c  A subsample of eight T. indicum ascocarps for each of the three ITS classes was then used for the analysis of the mating type (MAT) genes and idiomorphs ( Table 1).

Identification of T. indicum MAT Genes
The primer pairs p1/p2 and p19/p20 previously designed on the most conserved regions, namely the HMG-box and a-box domains, of T. melanosporum MAT1-2-1 and MAT1-1-1 genes, respectively [15], were used to amplify, under low-stringent PCR conditions (see M&M), orthologs of the MAT genes from the 24 selected T. indicum samples. As the DNA isolated from each single fruit body was mainly contributed by the gleba, and to a far lesser extent by the spores [15], the number of the PCR cycles for this screening was kept low (less than 30 cycles) to allow the preferential amplification of solely DNA isolated from the gleba. Using this approach, the primer pair p1/p2 produced a single amplicon of approximately 500 bp from 13 out of the 24 T. indicum ascocarps analyzed. The remaining 11 ascocarps yielded an amplicon of about 400 bp with the primer pair p19/p20 only. In every ITS class investigated, the presence of truffles whose gleba was made either by strains harboring the MAT1-1-1 or the MAT1-2-1 gene was balanced ( Table 1).
The PCR fragments from three ascocarps that produced either the MAT1-2-1 specific (Ti_U983, Ti_U986 and Ti_CF10) or the MAT1-1-1 specific (Ti_C61, Ti_LI2, and Ti_C37) amplicon only were cloned and sequenced. Sequence analysis showed that these PCR fragments shared high similarity with HMG-box and a-box sequences of T. melanosporum respectively (data not shown). The finding that only one of the two MAT genes was amplified from all of the samples processed was interpreted to mean that, regardless of the morphotypes, T. indicum was heterothallic.

Amplification of the 59 and 39 Ends of T. indicum MAT Idiomorphs
On the HMG-box and a-box sequences cloned as reported above, two MAT1-2-1 primers, named i5 and i6, and two MAT1-1-1 primers, named i3 and i4, were designed, respectively. For each MAT gene these primers were designed with in opposite orientations and to be combined with the primers i1 and i2 designed on the T. melanosporum 59 and 39 regions flanking the MAT locus and conserved between the strains of different mating types (Table S2, Figure S4). The primer pairs i1/i4 and i2/i3 produced fragments of about 1.4 Kbp and 8 Kbp, respectively, from all of the three T. indicum samples that produced the MAT1-1-1 amplicon when amplified with the primer pair p19/p20 (Ti_LI2, Ti_C61 and Ti_C37). With regard to the samples that gave rise to the MAT1-2-1 amplicon when amplified with the p1/ p2 pair, two (Ti_U986 and Ti_CF10) yielded a fragment of about 8 Kbp and one (Ti_U983) a fragment of 11 Kbp with the primer pair i1/i6. The same samples produced a fragment of 3.5 Kbp (Ti_U983 and Ti_U986) and a fragment of 6.5 Kbp (Ti_CF10) with the primers i2/i5 ( Figure S4).
These PCR-based approaches showed that the MAT1-2-1 and MAT1-1-1 genes had a similar orientation within the MAT loci of both T. indicum and T. melanosporum.

Isolation of T. indicum MAT Idiomorphs
The 59 and 39 ends of the T. indicum MAT amplicons obtained as reported above were sequenced and aligned. The primers i8 and i9 were then designed on the 39 ends of the MAT1-2 and MAT1-1 amplicons, respectively, whereas the primers i7 and i10 were designed on the 59 and 39 flanking regions shared between MAT1-2 and MAT1-1 idiomorphs, respectively (Table S2, Figure S4). Different primer combinations were finally used to amplify the complete T. indicum regions spanning the MAT1-1 and MAT1-2 idiomorphs from the T. indicum samples Ti_U983, Ti_U986 and Ti_CF10, selected on the basis of their different ITS/RFLP patterns (Table 1 and Figure S4). To reach this goal the number of PCR cycles was increased to 50 to allow the amplification from each T. indicum fruit body of the DNA contributed also by the spores.
Using this cycling condition and the primer pair i7/i10, two fragments of approximately 9 and 14 Kbp from ascocarps Ti_U983 and Ti_CF10, and 9 and 11 Kbp from ascocarp Ti_U986 were amplified. The intensity of the two bands produced per sample was different since the DNA contributed by the gleba always yielded the more intense band, as expected ( Figure S5).
Finally, by using the primers i8 or i9, specific to the idiomorphs MAT1-2 and MAT1-1, respectively, in combination with the forward primer i7 common to both, the two MAT regions were separately amplified from each of the three T. indicum selected samples ( Figure S4, Figure 2). The six amplicons obtained were sequenced and deposited in GenBank under accession n. KF318361-KF318366.

Structure of T. indicum MAT Idiomorphs and Comparison with T. melanosporum
Sequence alignment, BLAST and gene prediction analyses using FGENESH were performed to evaluate the length and the structure of MAT1-1 and MAT1-2 idiomorphs and identify the

Structure of MAT1-1 Idiomorph
The MAT1-1 idiomorph was slightly longer in T. indicum (7,583 bp, 7,741 bp and 7,666 bp in Ti_U983, Ti_U986 and Ti_CF10, respectively) than in T. melanosporum (7,430 bp). The structure of MAT1-1 idiomorph was very similar among the three T. indicum samples and T. melanosporum. The major difference resided in a region of about 750 bp, conserved in Ti_U983, Ti_U986 and T. melanosporum, which was replaced in Ti_CF10 by a completely unrelated sequence, ''e'' (Figure 2A, orange hatched box). The ''e'' sequence was also present at the 39 end of the MAT1-2 idiomorph in the same sample ( Figure 2B, orange hatched box). A BlastN search identified several sequences similar to the region ''e'' of Ti_CF10 (about 85% identity) dispersed throughout the T. melanosporum genome and annotated as transposable elements (TEs) in the T. melanosporum database (http://mycor.nancy.inra.fr/IMGC/TuberGenome/blast.php). However, neither BlastN nor BlastX searches against the Genbank database revealed any significant match for this query. Moreover, the sequences of the MAT1-1 of T. indicum differed from each other and from that of T. melanosporum for the presence of several indels ranging from a few bp to approximately 160 bp. All of these insertions are indicated by the grey boxes in Figure 2.
Gene prediction analysis of MAT1-1 revealed the presence of a single gene (Figure 2A, blue arrowed box) exhibiting high similarity (97% sequence identity) with T. melanosporum MAT1-1-1. RT-PCR analysis of RNA isolated from Ti_U983, Ti_U986 and Ti_CF10 fruit bodies confirmed the presence for MAT1-1-1 gene of a single transcript deriving from three exons. On the contrary in T. melanosporum two alternative MAT1-1-1 transcripts were detected [15].
The MAT1-1-1 gene of Ti_U983 differed from Ti_U986 for two nucleotides only, whereas Ti_CF10 differed from Ti_U983 and Ti_U986 for 28 and 30 nucleotides, respectively. The T. indicum MAT1-1-1 sequences differed from that of T. melanosporum for 26, 28 and 27 nucleotides in Ti_U983, Ti_U986 and Ti_CF10, respectively (Table S3, Figure S6). The deduced MAT1-1-1 encoded protein is 319 amino acids long in each of the three T. indicum strains. Ti_U983 and Ti_U986 shared identical protein sequences, whereas Ti_CF10 contains eight amino acid changes compared with the other two strains. The MAT1-1-1 proteins of Ti_U986 and Ti_U983 differed from T. melanosporum MAT1-1-1 for eleven amino acids, whereas for Ti_CF10 it differed for eight amino acids only ( Figure 3A).

Structure of MAT1-2 Idiomorph
The T. indicum MAT1-2 idiomorph was 9,783 bp, 10,096 bp and 12,131 bp long in Ti_U986, Ti_CF10 and Ti_U983, respectively. Overall, this idiomorph was longer in T. indicum than in T. melanosporum (5,550 bp long), due to a large insertion, by about 4,300 bp in Ti_U986 and Ti_CF10 and about 6,650 bp in Ti_U983, between the ''c'' and ''a'' regions ( Figure 2B, green and grey hatched boxes).
In contrast to MAT1-1, in which a single ORF was identified, three ORFs in Ti_U986 and Ti_CF10 and four ORFs in Ti_U983 were predicted within this idiomorph ( Figure 2B). Among these ORFs, RT-PCR analysis confirmed the presence of the MAT1-2-1 gene: this gene exhibited high similarity (94% sequence identity) and the same exon-intron structure as its ortholog from T. melanosporum (Figures 2B and S7). More specifically the MAT1-2-1 gene of Ti_U983 differed from Ti_U986 for one nucleotide only, whereas Ti_CF10 differed from Ti_U983 and Ti_U986 for 38 and 39 nucleotides, respectively. The T. indicum MAT1-2-1 sequences differed from that of T. melanosporum for 46, 48 and 49 nucleotides in Ti_CF10, Ti_U983 and Ti_U986, respectively (Table S3, Figure S7). T. indicum MAT1-2-1 gene encoded a hypothetical protein of 298 amino acids with Ti_U983 differing from Ti_U986 for a single amino acid only. Conversely, Ti_CF10 differed for thirteen and fourteen amino acids from Ti_U983 and Ti_U986, respectively. In all, the T. indicum MAT1-2-1 deduced proteins coded by Ti_U986, Ti_U983 and Ti_CF10 differed from that of T. melanosporum for 19, 18 and 14 aa, respectively and for an insertion of a single serine in the N-terminal region ( Figure 3B).
A second, short ORF was detected within the ''a'' region ( Figure 2B, pink arrowed box) in all T. indicum samples. This ORF was similar to GSTUMT00001089001 previously identified within the same region of the T. melanosporum MAT1-2 idiomorph [15] but was missing within the corresponding region exhibited by the MAT1-1 idiomorphs of both T. indicum and T. melanosporum.
The third putative gene identified in each T. indicum MAT1-2 was named Ti_orf3 and consisted of 4 exons in an inverse orientation with respect to the MAT1-2-1 gene ( Figure 2B). The
Analysis of the sequences flanking the putative Trp1 gene showed the presence of 52 bp long terminal inverted repeats (TIRs) ( Figure S10A). However, the TTAA target site duplications (TSDs) typical of PiggyBac TE [45] were not found in Ti-tr1. Instead, the Ti-tr1 TIRs were integrated at TSDs represented by TA nucleotides ( Figure S10B) which typifies different Class II TE families [46].
A second TE-like element was identified on the 39 flanking region of the T. indicum Ti_CF10 MAT1-2 idiomorph. It was present within a 1.868 bp long insertion specific to the idiomorph of this sample only and was named Ti_tr2 ( Figure 2B, brown box). The Ti_tr2 element contained a 1.462 bp long putative gene (Trp2) coding for an hypothetical protein of 370 aa ( Figure S9). BlastP analysis against the GenBank database showed similarity to transposase genes of ascomycetes such as the Tc1/mariner transposases of Penicillium marneffei (XP_002146479) and Talaromyces stipitatus (XP_002486812.1). Blast analysis against the T. melanosporum database showed low similarity (about 30% aa identity) with gene models (GSTUMT00012613001, GSTUMT00006493001, GSTUMT00007775001, GSTUMT00007054001, GSTUMT00012634001) annotated as transposases. Searches of the Repbase also confirmed the similarity of Ti_tr2 to Tc1/mariner TEs. The Ti_tr2 insertion sites were bordered by two 45 bp long TIRs integrated at TA TSDs ( Figure  S11).
The amino acid alignment of both Trp1 and Trp2 encoded hypothetical proteins with transposases from various organisms showed that they contained the highly conserved DDD/E catalytic triad. In addition, these two transposases showed a set of conserved aa positions ( Figures S8 and S9), which typifies the PiggyBac and Tc1/mariner families, respectively [47].

Discussion
In this study we show that T. indicum, the Asiatic truffle species which is the closest relative to the Périgord black truffle T. melanosporum, is heterothallic since its MAT locus is organized in two idiomorphs harbored by different strains. Because the sequences and the structure of the MAT idiomorphs differ between T. indicum strains belonging to different ITS classes, the presence of a complex of cryptic species is hypothesized. Comparison of MAT sequences between the T. indicum species complex and T. melanosporum also provides us with additional clues for inferring the evolutionary relationship between these black truffle species.

T. indicum is Heterothallic
The scarce conservation of the MAT genes across different classes of fungi has been the main obstacle to identifying sexual genes in truffles by homologous cloning, and thus to deciphering their reproductive modes. Only recently, thanks to the sequencing of the T. melanosporum genome, the first MAT gene, namely TmelMAT1-2-1, has been mined from a truffle species [14]. In turn, the identification of this gene and its flanking genomic regions has allowed the second one, TmelMAT1-1-1, to be cloned in a different strain. This, coupled with the fact that none of the strains analyzed displayed both MAT genes, has provided clear cut evidence in support of the thesis that this black truffle is a heterothallic fungus [15]. Thus, TmelMAT1-1-1 and TmelMAT1-2-1 represent the bedrock for the identification of their orthologs from other fungi within the same genus by homologous cloning. To this end, we employed PCR primers designed on the most conserved domains, namely the a and HMG domains, of the two T. melanosporum MAT genes, to amplify and clone the MAT genes from T. indicum. All T. indicum sequences obtained using MAT domain specific primer pairs showed high similarity to both T. melanosporum MAT, regardless of the morphotypes of the samples analysed. Consistent with a genomic organization that typifies heterothallic fungi, none of the Chinese black truffles screened displayed both MAT genes when PCR was carried out to amplify preferentially the DNA contributed by the gleba, the structure of truffle fruit bodies known to be of uniparental origin [12,13]. Equally consistent with heterothallism is the evidence that amplicons specific to both MAT genes have been conversely obtained when PCR was carried out to allow the amplification of the DNA contributed by the spores (see below). Strikingly, these analyses also prove that the gleba of the Chinese truffles can be formed by strains carrying either the MAT1-2-1 or the MAT1-1-1 gene, as previously observed in T. melanosporum [13].
The MAT Genes and the Structure of MAT Idiomorphs Differed among T. indicum Samples T. indicum is considered a highly polymorphic species, as fruit bodies differing in the morphological traits of their ascospores have been described [21,22,23,24]. Morphological analyses of samples of different geographical origin have, indeed, shown a high level of overlapping characters, making it very difficult to distinguish putative different species within this complex ( [26], this study). A large-scale screening using ITS/RFLP analyses, of T. indicum samples exhibiting different morphotypes has revealed a high level of genetic diversity with the presence of three different genotypes (haplotypes) named T. indicum A, B1 and B2 within these samples [27]. Subsequently, phylogenetic studies based on the sequences of the LSU and ITS of the rDNA region, and of other phylogenetically informative loci such as b-tubulin, 1a-elongation factor and Protein kinase C (PKC), have converged to show the presence within T. indicum of at least two distinctive clusters referred to as T. indicum_A and T. indicum_B [26,28,29,30,31,32,34,48]. However, whether these two T. indicum classes correspond to distinctive cryptic species or merely represent highly-structured populations belonging to the same species has yet to be elucidated [26,31].
The MAT locus is typically a non recombining region and many studies have shown that MAT genes, in particular in heterothallic species, are under relaxed purifying selection or even under positive selection and have high evolutionary rates [49,50,51,52]. Fast-evolving genes in the MAT locus are thus good candidates for tackling taxonomic and phylogenetic questions and for better delineating species boundaries within the Tuber genus. To this end, we have isolated and sequenced the entire T. indicum mating type idiomorphs from ascocarps belonging to the three ITS classes (A, B1 and B2) reported by Paolocci et al. [27]. More specifically, in order to compare their sequences across genetically different samples we have isolated the two idiomorphs from the same ascocarp to ensure the cloning of each pair of idiomorphs from the same biological species. This has been achieved by increasing the number of PCR cycles to allow the DNA isolated from the spores of each ascocarp to be amplified above the detection limit, although to a much lesser extent than the DNA isolated from the surrounding gleba.
The sequences of both MAT genes are very similar in the T. indicum_B samples (Ti_U983 and Ti_U986): only two synonymous substitutions are in fact present in MAT1-1-1, and a single nonsynonymous substitution is present in MAT1-2-1. Conversely, T. indicum_A displays many non-synonymous substitutions in both MAT1-1-1 (8) and MAT1-2-1 (13-14) with respect to their orthologs from the T. indicum_B samples, with only a few (1-2) of them occurring within the a-box and HMG-box regions ( Figure 3). These polymorphisms between the T. indicum_A and T. indicum_B samples are also evident in the genomic regions flanking each of the MAT genes ( Figure 2). Notably, the T. indicum_A MAT1-2 idiomorph differed from those of the T. indicum_B samples for a deletion within the gene named Ti_orf3. This deletion does not disrupt the ORF of the gene but, by changing the number of ANK binding domains, may affect gene function [53]. Length difference of the MAT1-2 idiomorph is also observed in the T. indicum_B samples. However this is due to the insertion of a TE element in Ti_U983 (see below).
With regard to the MAT1-1 idiomorphs, the main polymorphism resides in a region of about 750 bp (Figure 2, region ''e'' of Ti_CF10) with a sequence totally different between T. indicum_A and T. indicum_B samples.
MAT genes are functional markers intimately linked to speciation events. Comparative analyses of the sequence and organization of the MAT genes and idiomorphs suggest significant divergences between T. indicum truffles displaying the ITS class A and those displaying classes B1 and B2. This is therefore interpreted as an additional clue in support of the thesis that there are at least two cryptic species within the T. indicum complex. On these grounds, the present study provides a methodological approach (i.e. the amplification of both MAT genes/idiomorphs from the same ascocarp) and sequence information to carry out a large-scale screening of T. indicum truffles of different ITS classes by MAT-based markers. This screening will be instrumental to further verifying the hypothesis that the T. indicum_A and T. indicum_B samples belong to two different species. Under this assumption, in fact, no T. indicum_A sample is expected to show traits that feature the two MAT idiomorphs of the B classes. In contrast, some T. indicum_B1 samples are expected to display one or both MAT idiomorphs displayed by T. indicum_B2 samples and vice versa.

TE Elements in T. indicum MAT Locus
Extensive length difference has been observed among the MAT1-2 regions of T. indicum samples. More specifically, the MAT1-2 idiomorph of T. indicum_B1 sample (Ti_U983) is longer than that of T. indicum_B2 (Ti_U986) due to the presence of a Class II TE, named Ti_tr1. A class II TE, named Ti_tr2, is also present in the MAT1-2 region of T. indicum_A. However, it is located outside of the idiomorph, in its 39 flanking region.
Class II TEs comprise several superfamilies which are defined on the basis of terminal inverted repeat (TIR) sequences, size and/ or sequence of the TSD and structure and conservation of catalytic domains of the transposase protein [47,54]. Both Ti_tr1 and Ti_tr2 have a ''TA'' TSD which typifies the Tc1/mariner superfamily. Blast similarity searches using the Trp2 encoded putative transposase protein of Ti_tr2 as a query also return TEs of the Tc1/mariner superfamily as best hits. Conversely, the classification of Ti_tr1 remains unclear because its transposase appears to be related to TEs of the PiggyBac superfamily, which is characterized by a different TSD (i.e. TTAA) [55]. The scarce similarity between TIRs sequences of Ti_tr1 and Ti_tr2 ( Figure  S12) and the phylogenetic tree relative to Trp1 and Trp2 encoded putative transposases ( Figure S13) confirm that these two TEs are unrelated to each other.
Tc1/mariner TEs are widespread in fungi [46], conversely PiggyBac TEs are mainly found in animals and insects, more rarely in fungi [45]. However, TEs similar to Ti_tr1 have been identified and annotated as PiggyBac in the T. melanosporum genome [14].
The presence of Class I and II TEs within or linked to the mating type region is a common feature of many filamentous ascomycetes [51,52,56,57,58,59,60]. Because of its low recombination rate, the accumulation of TEs and other mutations in the MAT locus is an expected result [61]. In turn, the accelerated divergence of mating types may promote speciation by determining reproductive isolation between individuals [59]. Thus, TEs within or linked to the MAT locus of T. indicum may have had a crucial role in the diversification of this species complex. In addition, TE insertion, by generating sites for unequal crossing over, may trigger gross rearrangements in fungal genomes, which if targeted on the MAT locus may induce changes in their lifestyle [59]. Indeed, Gioti et al. [62] have attributed the transition from heterothallism to homothallism during the evolutionary history of Neurospora to retrotransposons.
Based on these considerations and the finding that TE elements have been detected within or linked to the MAT locus of black truffle species, the occurrence of events over the evolutionary scale of mating system transition in these fungi cannot be excluded. It will be interesting therefore to assess whether heterothallism and the presence of MAT-linked TEs are common features of all Tuber spp.

Comparison of T. indicum and T. melanosporum MAT Loci Provides Insights into the Evolution of Black Truffles
Comparative analyses reveal a similar structure of both MAT idiomorphs between T. indicum and T. melanosporum. The MAT1-1 idiomorph of both species contains a single gene, MAT1-1-1, in the same position and orientation. The T. indicum and T. melanosporum MAT1-1-1 genes share the same intron/exon structure with the predicted proteins differing for 9-10 amino acids. In a previous study, Rubini et al. [15] showed that the second intron of MAT1-1-1 is differentially spliced in the T. melanosporum fruit body to produce two transcripts coding for proteins of 319 and 399 amino acids. Here we show the presence of a single transcript coding for a putative protein of 319 aa in both the T. indicum_A and T. indicum_B fruit bodies. So, whether alternative splicing of MAT1-1-1 is a stage or species-specific phenomenon has yet to be elucidated.
Compared to MAT1-1, the structure of MAT1-2 idiomorphs of T. indicum and T. melanosporum is more divergent: in addition to MAT1-2-1 gene which shares the same structure and similar sequence, a putative gene (Ti_orf3) coding for an ANK domain containing protein is present in the MAT1-2-1 of T. indicum samples only. T. melanosporum displays a large deletion in this region that encompasses a large portion of the Ti_orf3 gene. However, a highly degenerated relict sequence corresponding to the 39 end of Ti_orf3 can still be identified in the MAT1-2 of T. melanosporum.
It is worth mentioning that an ANK-containing gene has been also identified in the MAT idiomorph of Rhizopus oryzae, a species belonging to a basal group of fungi, the Mucorales [63]. However, this R. oryzae gene contains two additional domains (BTB and RCC1) that are not present in T. indicum Ti_orf3, suggesting that these two ANK-containing genes are not orthologs.
Although additional experiments are required to evaluate whether the Ti_orf3 gene is functional in T. indicum and whether it is involved in any phase of sexual reproduction, it may be helpful for taxonomic and phylogenetic studies. The presence of the Ti_orf3 relict sequence in T. melanosporum suggests, in fact, that it is an ancestral gene that has been likely lost in T. melanosporum after the divergence from the common ancestor of these two black truffle species. This is in agreement with the hypothesis of vicariance for explaining the allopatric isolation of T. indicum from T. melanosporum populations in Asia and Europe [34].
Bonito et al. [64], based on dispersal-vicariance analysis, have recently suggested a North American origin for T. melanosporum, hypothesizing that European and Asiatic black Tuber species originate from long-range dispersal events (i.e. migration via the Beringia and/or Thulean North Atlantic Land Bridges). Sequencing of MAT regions of more Asian, European and North American species (e.g. T. brumale, T. regimontanum, T. pseudoexcavatum, T. formosanum) belonging to the T. indicum -T. melanosporum clade [34,48,64] is required to deepen our knowledge of the evolutionary history of this fungal lineage.
Comparative analyses of the structure and organization of the sex locus in basal lineages of filamentous Ascomycetes (Pezizomycotina), such as members of the Pezizomycetes, may help to investigate the mechanisms underlying fungal mating type evolution. On these grounds, the comparison of the two T. indicum idiomorphs among them and with those of T. melanosporum has revealed the presence of homologous sequences arranged in an opposite orientation. This is especially true in T. indicum where these shared sequences represent approximately 50% of the entire idiomorphic region (Figure 2). The presence in opposite orientation of sequences shared between the idiomorphs fits nicely with the hypothesis that inversions at chromosomal loci containing genes for sex determination may have represented a mechanism that has driven the divergence (and rearrangement) of the sex locus in ascomycetes and other organisms [59]. The presence in the MAT locus of Tuber species of TE elements (see above) reinforces such a thesis.
Are T. indicum and T. melanosporum Potentially Sexually Compatible?
Asiatic black truffles have been imported into Europe since the early 1990s and frequently sold in local markets mixed with the more highly-prized Périgord black truffle [17,40]. Besides the fraudulent commercialization of Asiatic black truffles, the deliberate or accidental use of T. indicum ascocarps to inoculate host plants for establishing truffle orchards in Europe may represent a serious ecological threat for the indigenous populations of T. melanosporum. Incidentally, the introduction of T. indicum in T. melanosporum truffle orchards in Italy as well as in non-European countries (e.g. the USA) has been reported [48,65].
It is worth noting that for both MAT genes the divergence level between the two T. indicum classes (A and B) is similar to that between T. indicum and T. melanosporum (Table S3). Thus, if hybridization between genetically different T. indicum individuals from sympatric populations were discovered, sexual compatibility with T. melanosporum could not be ruled out. It is interesting to note that interspecific mating and hybridization has been reported for fungi such as Ophiostoma spp., Heterobasidion spp., Melampsora spp. and Microbotryum spp. [66,67,68,69,70]. In some of these fungi (e.g. Microbotryum) hybridization has occurred through secondary contact following initial divergence in allopatry [70].
In Ascomycetes, mating compatibility and mutual recognition between strains of opposite mating type are mediated by the pheromone-receptor system [52]. In a wide range of organisms it has been demonstrated that proteins under sex-biased expression such as gamete recognition proteins evolve rapidly [71,72]. A fundamental consequence of the rapid evolution of interacting reproductive proteins is the potential to generate reproductive isolation between populations or species [72,73]. Although rapid evolution appears to be the case for fungal MAT genes, this trend has not been ascertained for ascomycete pheromones [49,74,75]. The distinct pheromones of certain yeast species appear to offer some level of species specificity, it is therefore conceivable that reproductive barriers between some ascomycetes may stem from differences in their pheromone peptide and pheromone receptor sequences.
The a-pheromone and both pheromone receptors have been identified in T. melanosporum genome [14,15]. Work is in progress to isolate their orthologs from T. indicum and perform functional analyses using yeast complementation assays [76]. This will allow us to assess for the level of conservation of these genes across Tuber spp. and test whether the pheromones of one truffle species are recognized by the receptors of the other(s) within T. indicum species complex and between species of this complex and T. melanosporum.

Conclusions
T. indicum is the first truffle species which is non-endemic in Europe for which the MAT locus and thus the reproductive mode have been characterized. This locus has helped us to shed more light on the taxonomy and phylogenetic relatedness within and between black truffles of different geographical provenance. In light of their phylogenetic relatedness and heterothallism, we cannot exclude breeding between T. melanosporum and Chinese truffles. In turn, if geographical barriers have thus far been the only impediment for their crossing, the introduction of one species into the geographical range of the other could have potentially extremely detrimental effects in terms of erosion of their biodiversity and specificity. This is particularly true for the less competitive and highly prized T. melanosporum. In addition, recent studies have documented a skewed representation of both mating types on host plants colonized by T. melanosporum to suggest a possible mating type-dependent distributional pattern of its strains [77,78]. Thus it becomes intriguing and of both fundamental and applied relevance to assess the distributional pattern of strains with opposite mating types also in T. indicum producing sites.