Contrasted Genetic Diversity, Relevance of Climate and Host Plants, and Comments on the Taxonomic Problems of the Genus Picoa (Pyronemataceae, Pezizales)

The species concept within the genus Picoa Vittad. is here revisited in light of new molecular and ecological data obtained from samples collected throughout the Mediterranean basin. Two highly diverse widespread clades and four additional minor lineages were significantly supported by three genes dataset (ITS, 28s LSU and RPB2) inferences for 70 specimens. The two widespread clades occur in very different geographical and ecological areas associated with exclusive host plants in the genus Helianthemum. SEM study of spore surface morphology in these lineages revealed the existence of smooth ascospores in the majority of these clades. However the most frequent lineage in Europe and coastal North Africa displayed either smooth or verrucose spores. Hence this morphological criterion cannot be reliably used to discriminate between the different clades. In addition, SEM observations made on ascospores from several original collections of P. juniperi and P. lefebvrei supported the hypothesis that ornamentation depends on the degree of maturity in some of these lineages. Geographical and ecological, rather than morphological data are here suggested as the most useful characters to separate the different lineages in Picoa. Further studies focusing on these features are needed before the names P. juniperi and P. lefebvrei can be unambiguously linked with the genetic lineages observed.


Introduction
Specimens of Picoa Vittad. produce hypogeous ascomata that can be found in semi-arid and desert ecosystems of most countries around the Mediterranean basin and the Middle East [1,2,3,4,5,6,7,8,9]. They establish mycorrhizal associations with several annual and perennial xerophytic host plants in the Cistaceae, especially those in the genus Helianthemum [10]. These plants and their associated fungi may play a major role in the maintenance of Mediterranean shrublands and grasslands and help prevent erosion and desertification [11].
The genus Picoa was proposed by Vittadini [12] for Picoa juniperi Vittad., a hypogeous ascomycete collected in northern Italy close to Juniperus sp. and characterized by its black flatwarted surface, white gleba filled with pseudoveins, and globose asci with smooth globose spores. Since their first description to the present time, species of the genus Picoa have been subject to important changes in their nomenclatural and taxonomic position, being reassigned into other genera or else being invested as new genera themselves. In 1894, Patouillard [13] described the genus Phaeangium Pat. and its only species Phaeangium lefebvrei Pat. to accommodate a sample collected at Ras-el-Oued (southern Tunisia). He described it as a hypogeous ascomycete characterized by brownish ascoma with hairy unwarted peridium, homogeneously white gleba, and globose to subglobose stipitate asci containing smooth ovoid spores. On the basis of observations made on Hennings specimens, Maire [14] disputed the Patouillard description related to sporocarp surface and gleba structure and recombined the species as Picoa lefebvrei (Pat.) Maire stating the absence of clear distinctive characteristics between Phaeangium and Picoa Alsheikh & Trappe [15] reexamined the original material of Phaeangium lefebvrei from different localities, Tunisia, Algeria, Lybia, Iraq as well as new collections of this species from Kuwait. The spore ornamentation at maturity and the tomentous peridium placed Phaeangium lefebvrei again as the unique species of the Phaeangium genus [15]. These authors considered the smooth spores observed in the type specimen of P. lefebvrei as immature, and synonymised it with the Algerian species Terfezia schweinfurthii Hennings, which was reported to have warty spores. Moreno et al. [1] argued that P. juniperi and P. lefebvrei should be considered congeneric based on their similarities in peridial tomentum, overall colour of the ascoma, and spore ornamentation. Laessøe & Hansen [16] indicated that some earlier unpublished molecular results of O'Donnell et al. [17] already linked P. juniperi with the genus Otidea (Pers.) Bonord. within the Pyronemataceae. The first published sequences by Gutierrez et al. [10] confirmed this taxonomic affiliation. Sbissi et al. [4] agreed with the membership of both species in the genus Picoa close to Geopora cooperi Hark. within Pyronemataceae. In this work, smooth-spored samples from Tunisia were identified as P. juniperi, and those from Europe with minutely warted spores were named P. lefebvrei. The status of a third lineage of African samples with smooth spores could not be resolved. In the works of Ammarellou et al. [5] and Jamali & Banihashemi [7,8], new samples from Iran were sequenced, leading to similar conclusions. Unfortunately, no European samples with smooth spores were included in their analysis. Tedersoo and Smith [18] confirmed one more time the nesting of Picoa lefebvrei within Pyronemataceae next to its neighbor species Geopora cooperi. A third species in Picoa, P. carthusiana Tul & C. Tul, was found to be closely related to the Morchellaceae-Helvellaceae on the basis of morphological [19] and molecular data [17]. This species was combined in the genus Leucangium Quél. by Saccardo [20] as Leucangium carthusianum (Tul. & C. Tul.) Paol. It is currently considered as synonym of the type species of its genus, Leucangium ophtalmosporum Quél [16]. In 1956, Lange described Picoa pachyascus M. Lange, the first American representative species of this genus. However, it has been synonymised with the type species of Imaia Trappe & Kovács, Imaia gigantea (S. Imai) Trappe & Kovács [21]. Finally, Moreno et al. [22] described the last species in the genus, Picoa melospora G. Moreno [23].
In the present work, samples from most Mediterranean countries were analyzed in order to improve our understanding of the biogeographic and phylogenetic relationships between the different lineages of Picoa. SEM was employed to study spore ornamentation in newly collected specimens and original herbarium collections of P. juniperi and P. lefebvrei, and results were compared with phylogenetic data from multilocus sequences to evaluate the taxonomic importance of this morphological feature.

Ethics statement
Truffles of the genus Picoa are not listed in any national or regional law as protected or endangered species. The collection of specimens was not subjected to any restriction or specific permissions. Samples were harvested from open lands that are not privately-owned or protected and no specific permissions were required for these collection sites by the local authorities.

Fungal specimens
Ascomata were collected from a wide variety of habitats and regions across the Mediterranean basin (S1 Text), including semi-arid environments in the European shore (France, Italy, Spain, Greece), semi-arid habitats in the African shore (Algeria) and Middle East (Iran), and more deserted regions near Saharan desert in Africa (Algeria, Tunisia) and Middle East Syrian desert (Israel). Dried samples from Europe, Iran and Israel were preserved at Universidad de Alcalá herbarium (AH). Original collections of Picoa juniperi, P. lefebvrei and Terfezia schweinfurthii were studied and compared with the newly collected samples. Autoptic material from their original authors was kindly loaned by Farlow Herbarium (FH), Botanische Staatssammlung München (M), and Swedish Museum of Natural History (S). Morphological study of asci and ascospores was conducted using an Olympus CX22 microscope. Ascospore ornamentation was examined and photographed with a scanning electron microscopy (SEM) device JEOL JSM-6610LV at University of Science and Technology of Oran, or else with a Zeiss DSM-950 instrument at University of Alcalá.
DNA extraction, PCR amplification and sequencing DNA from Algerian specimens was extracted from approximately 25 mg of dried samples from each sample. Tissues of the gleba were ground in liquid nitrogen and the ascomata Genomic DNAs were isolated using ABIOpure TM Genomic DNA Plant Extraction Kit (Alliance Bio, USA) according to the manufacturer's instructions. Extracts were eluted in 50 μl of Elution Buffer supplied in the kit and stored at -20°C. DNA concentration was estimated using a Nano-Drop spectrophotometer (Thermo Scientific). Four different loci were amplified from DNA template by means of polymerase chain reaction (PCR). ITS rDNA was amplified using the primer pair ITS1 and ITS4 [24], large ribosomal subunit (28S nLSU) was obtained using primers LR0R and LR5 [25], β-tubulin gene was amplified using the couple (Bt2a-Bt2b) [26], and RNA polymerase II second largest subunit using bRPB2-6F and bRPB2-7R [27]. PCR amplifications were performed in a 25 μl final volume. Cycling conditions consisted in an initial denaturation step at 95°C for 2 min, followed by 35 cycles of a 1-min denaturation at 94°C, annealing at 53°C (ITS rDNA) or), or 58°C (β-tubulin), 47°C (LSU rDNA) and 55°C (RPB2) for 1 min, and elongation at 72°C for 1 min, with a final extension step at 72°C for 10 min. Amplification products were analyzed in 1.5% agarose gel in 0.5× TBE buffer (89 mmol l -1 Tris, 89 mmol l -1 borate, 2 mmol l -1 EDTA), stained with ethidium bromide, and visualized under UV light. The PCR products were enzymatically purified by exonuclease and alkaline phosphatase (Exo/SAP) and then bi-directionally sequenced. Sequence reactions were performed using the ABI PRISM TM 3130 Genetic Analyzer with Big Dye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems; HTDS, Tunisia) according to the manufacturer instructions. DNA from European, Iranian and Israelite samples was extracted and amplified following the methods described in previous publications [28]. ITS1 / ITS4; LR0R / LR5; Bt2a / Bt2b and bRPB2-6F / bRPB2-7R were employed for amplification and sequencing purposes. Sequences produced are available in public databases (Table 1).

Phylogenetic analysis
ITS rDNA and LSU nucleotide sequences were first compared with public databases using the BLAST algorithm [29] and then aligned with the sequences retrieved using the ClustalW application [30]. Simultaneously, a manual correction of sequences was conducted. The reference Picoa sequences came from Bidartondo & Doring (unpublished), Gutiérrez et al. [10], Sbissi et al. [4], and Jamali & Banihashemi [7,8]. Aligned loci were independently subjected to MrModeltest 2.3 [31] in PAUP Ã 4.0b10 [32]. The best models (SYM for ITS and RPB2, GTR for LSU) were implemented in MrBayes 3.1 [33], where a Bayesian analysis was performed (two simultaneous runs, six chains, temperature set to 0.2, sampling every 100th generation) until standard deviation of split frequencies was <0.01 after 1 320 000 (ITS) and 310 000 generations (LSU-RPB2-ITS). Finally a full search for the best-scoring maximum likelihood tree was performed in RAxML [34] using rapid bootstrap algorithm and model GTRMIX. Significance thresholds were set above 70% for bootstrap (BP) and 95% for posterior probability (PP).

Morphologic analysis
Macroscopically, samples in lineages I, III and IV were black or blackish, with regularly polygonal or pyramidal warts (Fig 3). European samples in Lineage VI were black, blackish, dark grey or dark brown in color, and were covered with regularly polygonal or pyramidal warts, although some specimens were also minutely warted, giving a papillose appearance. African samples of Lineage VI were dark brown to reddish-brown, nearly smooth or covered with more or less small rounded to angular warts. African and Middle Eastern samples of lineage II were either black with pyramidal warts, brown or dark brown with minutely papillose warts, or sometimes presented widely separated warts, looking yellowish in their interspaces. Spore ornamentation was hardly visible with a light microscope and boundaries among these samples were only perceivable in SEM (Fig 4). Lineage II produces exclusively smooth spores, while Lineage VI produces mostly verrucose, but also smooth spores. All other minor clades, except the Lineage V, have smooth spores.

Discussion
The present work provides a comprehensive view of phylogenetic diversity in Picoa around the Mediterranean basin. A high degree of genetic variability was found in the whole genus and within its major clades with ITS analysis. A combined LSU-RPB2-ITS multigenic analysis largely supported the same clades obtained from ITS inference, while β-tubulin failed to support any distinct lineage at all. Among the four regions analyzed, the internal transcribed spacer (ITS) region has a high amplification and sequencing success rate, and therefore we propose ITS as the suitable barcode for Picoa. β-tubulin was consistently the worst performing marker, with a non-discrimination power. The ITS region has the highest probability of successful identification for the broadest range of fungi and it was formally proposed to the Consortium for the Barcode of Life for adoption as the primary fungal barcode marker [35]. A high cryptic diversity can be observed also in the sister genus Geopora, where a number of distinct genetic lineages with high intraspecific diversity and few characteristic morphological features exist [36], but only one of them, Geopora tolucana, has been proposed as a new independent taxon because of its brownish hymenium, different from the typical whitish one of G. cooperi [37]. Results from the present analyses again provide strong evidence on the close relationships  between Picoa and Geopora species, particularly Geopora cooperi; and clearly supported the nesting of Picoa within the Pyronemataceae. Several studies supported the deep nesting of Picoa within the Geopora lineage [4,5,7,8,18,37,38] as a sister taxon of G. cooperi [4,18,37,38]. Perry et al. [39] pointed out a difficulty in delineating Pyronemataceae family due to the Diversity of the Genus Picoa absence of clear combinations of characters, either macro-or microscopically. These authors suggest that morphological characters traditionally used to segregate this family into subfamilial groups are not phylogenetically informative above the genus level [39]. Of the six genetic lineages identified in Picoa, four of them (I, III, IV and V) are yet too poorly represented to draw reliable hypothesis on their ecology and distribution. The remaining two (II and VI) clearly differ in their bioclimatic origin (arid / Lineage II; semi-arid / Lineage VI) (Fig 6), and putative host plants, but share most morphological features, excepting that Lineage II never produces verrucose spores. Morphological species delimitation within Picoa has always been challenging since diagnostic characters have been considered often ambiguous [13,22]. Spore surface is widely thought to be smooth in P. juniperi and warty in P. lefebvrei [40], despite the fact that P. lefebvrei was originally described with smooth spores [13]. Alsheikh  The distribution areas of linages II and VI seem to be mutually exclusive, actually matching different biogeographical provinces of Paleartic realm as delimited by Urdvardy [45]. This suggests that host plants or edapho-climatic conditions play an important role in the distribution of Picoa [46,47,48,49,50] and could be driving the speciation process in these hypogeous fungi. However, the existence of additional minor lineages suggests that some degree of sympatric evolution can take place, and so putative hosts are here regarded as the most probable factor behind the observed pattern. Both groups of Helianthemum hosts belong in fact to distinct sections and phylogenetic lineages within this genus [51,52]. Unfortunately, the taxonomic treatment of the different ecological lineages in Picoa is not fully clear, due to the lack of information about putative Helianthemum hosts in the original protologues of P. juniperi or Ph. lefebvrei which require more studies. One of the features mentioned by Vittadini in his description of P. juniperi is that this species was collected under Juniperus sp. Only one sample analyzed in the present study matched this condition (AH 39001, Is Arenas, Sardinia, Italy), and its molecular profile linked it to Lineage VI. Interestingly, another feature mentioned by Vittadini was that P. juniperi specimens fruit in autumn and winter along with other edible truffles, and they are even sold mixed with them in the markets. In the present work, only two samples AH 38893 and AH 39139 (both in Lineage VI) were collected in autumn or early winter (November-December), all others being found in late winter or spring (February-May), excepting samples in Lineage IV, which were all collected even later, in summer (June-July). It is hence tempting to associate the name P. juniperi with lineage VI (which is also the most widespread in Europe). On the other hand, P. lefebvrei type was collected at Ras-el-Oued, south of Gabes [13], where the dominant Helianthemum host is H. lippii, a host with uncertain presence in Europe [53]. Because of this host distinction, this name can only be applied to Lineages II or V, and it is tempting to use it for Lineage II, the most widespread in Africa and Middle East. Unfortunately, we lack conclusive ecological information about the putative hosts and fruiting season of the minor lineages (I, III, IV and V), and so we cannot reject the possibility that either P. juniperi or P. lefebvrei were proposed for specimens of these clades. It is even possible that some or all subclades within lineage VI (VI-1 to VI-9) should be considered independent taxa if no intermediate specimens can be found and some apomorphic features can be identified (e.g. affinity for a specific Helianthemum host). Further studies including richer sampling and providing more accurate data on these key ecological parameters are needed to understand the processes of speciation in these conspicuous species.
Supporting Information S1 Text. Fungal samples used in this study. (PDF)