Ribosomal ITS diversity among the European species of the genus Hydnum (Hydnaceae)

Grebenc, T., Martín, M.P. & Kraigher, H. 2009. Ribosomal ITS diversity in the European species of the genus Hydnum (Hydnaceae). Anales Jard. Bot. Madrid 66S1: 121-132. Several morphological species of the genus Hydnum L. are known to occur in Europe, but little molecular evidence exists to confirm the exact number and delimitation of the species. The present study seeks to investigate the genus Hydnum through sequence analysis of the nuclear ribosomal ITS regions and through morphological studies. The DNA sequences phylogenetic analysis revealed high diversity among the ITS region sequences in H. repandum (two clades) and H. rufescens (six clades) while the specimens of H. albidum, H. umbilicatum and H. ellipsosporum formed one and clearly separated clade per morphological species. Phylogenetic distances among the recognised species and the obtained morphologically unsupported clades are comparable and support the idea of several new, yet undescribed species. The intraspecific variability in the sequence data among phylogenetic species is generally low. Detailed morphological analysis of putative informative morphological characteristics could not support any of the observed non-monophyletic DNA-sequences clades within H. repandum or H. rufescens, and the proper use of names is not yet clear. Similar intraspecific variation has also been observed in many other ectomycorrhizal genera and could be explained by intensive speciation within variable groups under the influence of various factors (niche effect, ectomycorrhizal partner selection).


Introduction
Members of the family Hydnaceae Chevallier (1826) are primarily identified by the presence of posi -tive geotropic spines, ranging from small granular warts to clear individual spines (Ainsworth & al., 1973).Recent literature cites six valid genera in the family.To our knowledge, five of them were never included in any molecular analyses: Corallofungus Kobayasi, Dentinum Gray, Gloeomucro R.H. Petersen, Nigrohydnum Ryvarden, and Phaeoradulum Pat.(Kirk & al. 2001); while Hydnum as the type genus, was the only genus represented in phylogenetic studies.At higher taxonomic rank Hydnum was placed in Cantharellales first by Kreisel (1969) and later confirmed with molecular data by Pine & al. (1999) and subsequent papers.
Hydnum rufescens and H. repandum are distributed over an exceptionally wide area and are even recognised in the Far East (Asia) although several synonyms from different areas and for local populations were published and many local names were conspecific with European species (Maas Geesteranus, 1971).Hydnum elatum Massee and two more unnamed Hydnum species were recognised for Asia and Australia in addition to H. umbilicatum (Maas Geesteranus, 1971).The latter was described in North America by Peck (1902).Despite it was commonly found on several continents (Hall & Stuntz, 1971) its presence in Europe was only confirmed for Finland (Huhtinen & Ruotsalainen, 2006).
Clear delimitation of species cited in Europe is not always easy.Morphological characters can vary with the developmental stage of pileus and environmental conditions during the growth period (Hall & Stuntz, 1971;Maas Geesteranus, 1975).Spore size and shape can well separate H. albidum and H. ellipsosporum from the others (Ostrow & Beenken, 2004) while taxonomical position of H. rufescens within the genus is confusing, not only after classical identification but also after molecular data have become available.Molecular identification of H. repandum and H. rufescens ectomycorrhizae on Norway spruce showed distinct restriction patterns of amplified ITS region in genomic rDNA.Additionaly the variability of the restriction pattern within H. rufescens was observed after digestion of the amplified PCR product with HinfI endonuclease.The observed additional differences indicate possible variability of collections from different sites (Agerer & al., 1996).Ostrow & Beenken (2004) found a good correlation for selected morphological and molecular characters for four European species with only few samples sequenced for each species.They reported no sequence diversity within H. rufescens, although only for H. repandum and H. ellipsosporum was the absence of any such intraspecific variability clearly stated.
Comparison of rDNA ITS sequences is a valuable tool in phylogenetic studies, and to provide more accurate species delimitation (Taylor & al., 2000).Currently there is a poor overlap between morphological and molecular species concept based on the variability of the rDNA ITS sequences in studied Hydnum collections.To support the results obtained at the molecular level, selected morphological characters indicative of taxonomic affiliation in the genus Hydnum were measured and correlated to the clades retrieved in the DNA-sequences phylogenetic analyses.Multivariate statistics were employed for these analyses.

Materials and methods
DNA analyses were undertaken in the laboratories in Slovenia (SFI) and in Spain (RJB).The different protocols were standardized at both sites, such that the final results obtained from the same sample were equal.Thin layer cromatography (TLC) analysis was carried out in RJB in Spain.

Fungal material
Specimens from the genus Hydnum included in the study (Table 1) were either collected from various localities in the years 1999-2002 and stored in the herbarium at Slovenian Forestry Institute (LJU) or obtained from herbarium MA-Fungi (Madrid, Spain).We have tried to locate the type or representative material for European species in different institutional herbaria (UPS, MSB, and PC; Holmgren & al., 1998).However, according to the curators, the material either does not exist or was not possible to locate.Even though H. umbilicatum has been cited in Slovenia, no reference material was available from the area; the two collections included in the study were kindly sent by Lorelei L. Norvell from the Pacific Northwest Mycology Service.
Specimens with fully developed basidiomata and spores were used for examination of macro-and mi-croscopic morphological characters.Fifty spores per basidiomata were measures to calculate an average and extreme values for length, width, and spore volume (Ostrow & Beenken, 2004) and to assess several other potential informative characters (cap colour, stem diameter and position, cap diameter, spine position, pigmented content, and potential ectomycorrhizal partners on the collection site).Nomenclature followed the morphological species concepts of Maas Geesteranus (1975) (H.repandum and H. rufescens), Focht (1996) and Ostrow & Beenken (2004) for H. albidum, and Harrison & Grund (1987) for H. umbilicatum, the only non-European species included in the study.

Molecular methods
DNA extraction: Twenty milligrams of the hymenium from fresh or dried material were used for the DNA extraction following standard protocols after Whiting & al. (1997) or using 2% CTAB (Rogers & Bendlich, 1985;Doyle & Doyle, 1990).From older herbarium material DNA was extracted by E.Z.N.A. Fungi DNA Miniprep Kit (Omega Biotek) as described in Martín & García-Figueres (1999).For both methods DNA was re-suspended in pre-warmed, sterile milli-Q water to the approximate final concentration 100 ng/ml.
PCR amplification: Primers ITS1F (Gardes & Bruns, 1993) and ITS4 (White & al., 1990) or ITS4b (Gardes & Bruns, 1993) were used for PCR amplification of the ITS region, including 5.8 S rDNA.Amplification reactions were obtained using two methods: a) standard procedure described in White & al. (1990) in a total reaction volume of 40 µl with AmplyTaq polymerase (Perkin Elmer) and/or b) individual reactions in a final volume of 25 µl with Ready-To-Go PCR Beads (GE Healthcare Life Sciences) as mentioned in Winka & al. (1998).The PCR reactions were performed after Kraigher & al. (1995) in a PE 9700 DNA thermocycler with an annealing temperature 55°C.Negative controls, lacking fungal DNA, were run for each experiment to check for any contamination of the reagents.Amplified DNA was separated and analysed as described in Grebenc & al., 2000.Sequencing and cloning: Prior to sequencing, the amplification products were cleaned using the E.Z.N.A. Clean kit.When only weak PCR products were obtained the products were cleaned from the gel using QIAquick Gel (QIAGEN Inc.), cloned with pGEM®-T Easy Vector Systems (Promega), and purified with QIAPrep Spin Mini prep.Three clones were selected for sequencing with vector specific primers T7 and SP6 (QIAGEN Inc.).Sequence Navi -Ribosomal ITS diversity in Hydnum gator Software (Applied Biosystems) was used to identify the consensus sequence from the two strands of each isolate.When the sequences obtained from the cloned products were identical, only one sequence was included in the alignment.The sequences were submitted to EMBL database with the accession numbers indicated in Table 1.
A first maximum parsimony analysis (MP) was inferred using the heuristic search option of the 100 most parsimonious trees in PAUP*4.0b10.Gaps were treated as missing data.Branch lengths equal to zero were collapsed to polytomies.Nonparametric bootstrap support (Felsenstein, 1985) for each clade was tested based on 10 000 replicates, using the fast-step option.The consistency index, CI (Kluge & Farris, 1969), retention index, RI (Farris, 1989), and rescaled consistency index, RC (Farris, 1989) were obtained.
A second analysis was carried out using a Bayesian approach (Huelsenbeck & al., 2000;Larget & Simon, 1999).Posterior probabilities were approximated by sampling trees using a Markov Chain Monte Carlo (MCMC) method.The posterior probabilities of each branch were calculated by frequency of trees that were visited during the course of the MCMC analysis.The analysis was performed assuming the general time reverse model (Rodriguez & al., 1990) including estimation of invariant sites and assuming a discrete gamma distribution with six categories (GTG+I+G).
No molecular clock was assumed.A run with 10 000 000 generations starting with a random tree and employing 12 simultaneous chains was executed.Every 100 th tree was saved into a file of total of 100000 trees.We plotted the log-likelihood scores of sample points against generation time using TRACER 1.0 (http://evolve.zoo.ac.uk/software.html?i=tracer) and determined that stationarity was achieved when the log-likelihood values of the sample points reached a stable equilibrium value (Huelsenbeck & Ronquist, 2001).The initial 1000 trees were discarded as burning before stationarity was reached.Using the "sumt" command of MrBAYES, majority-rule consensus trees were calculated from 19 000 trees sampled after reaching likelihood convergence to calculate the posterior probabilities of the tree nodes.DNA-sequences phylogenetic trees were drawn in TREEVIEW (Page, 1996).

Results
A total of 32 new ITS nrDNA sequences were generated.The sequences were aligned with 17 ITS nrDNA sequences available from Genbank and UNITE to produce a matrix of 675 unambiguously aligned nucleotide position characters of which 464 were constant, 71 variables are parsimony uninformative, Ribosomal ITS diversity in Hydnum and 140 parsimony informative.The alignment with 52 sequences is available at TreeBASE (http://www.treebase.org/,as SN2182).
In order to cast additional light on the molecular differences among observed H. rufescens and H. repandum clades, we applied a TLC chromatography of the pigments in the basidiomata of one collection from each clade (see Table 1).Under visual light no spots of pigment were recognised after the chromatography.Under the illumination with UV (wave length 302 nm) spots appeared at Rf: 87.5 (close to ergosterol or L-DOPA), Rf: 73.5, and a longer spot with Rf: 65-59.All samples gave the same pigment composition, though only the intensity of coloration of sporocarps was represented.
Several other potential taxonomic informative characteristics were assessed for samples within each of the clades for possible separation of observed terminal clades in DNA-sequences phylogenetic trees at the morphological level (Table 2).Multiple range tests for  1 and 2, followed by the Accession Number from the GenBank or UNITE databases.spore size and volume clearly separated H. albidum, H. umbilicatum, and H. ellipsosporum from repandum and H. rufescens while clades within the latter two morphological species could not be clearly separated, except based on cap colour.We have observed no significant statistical difference between characters within either H. repandum or H. rufescens clades.

Discussion
The last comprehensive revision of the family Hyd naceae was published by Maas Geesteranus (1975) based on morphological characteristics of basidiomata and spores.Unequivocally defined taxa are a prerequisite for a comprehensive ecological, physiological, or molecular analysis of a taxon.The aim of this study was: to support the established taxa as recognised after the morphological concept of the species within the European Hydnum species, employing molecular tools, and to clarify the observed mole cular differences in the previous study (Agerer & al., 1996).The value of the presented results would be greater if type material existed and was available at least for H. repandum and H. rufescens.
The basidiomata included in the study were primarily identified based on morphological characteristics.Spore size and shape can be a good criterion to separate Hydnum albidum, H. ellipsosporum (Ostrow & Beenken, 2004), and H. umbilicatum (Hall & Stuntz, 1971) from the rest of Hydnum species from Europe.Huhtinen & Ruotsalainen (2006) examined the material from Finland and were able to separate all together three taxa within "H.rufescens" specimen: H. ellipsosporum, H. umbilicatum and H. ru fescens s. str.with at least two separate populations in the latter species based on spore shape.However, the data from several identification books/keys (Maas Geesteranus, 1975;Jülich, 1984;Courtecuisse & Duham, 1995;Ostrow & Beenken, 2004), as well as the results of the present study indicate that spores of H. rufescens and H. repandum do not differ significantly in size or shape.The basidiome size, position of the stipe, distribution of the spines, and shape and colour of the basidiome were relatively reliable morphological criteria to distinguish these two morphological species without the need to employ molecular tools, yet our results indicate much higher diversity within these two species.An umbilicate pileus, otherwise typical for H. umbilicatum (Hall & Stuntz, 1971), was observed for some collections of H. rufescens from Europe but observed morphological characteristics do not indicate the presence of H. umbilicatum in Europe since the spore size is within the range of H. rufescens and the umbilicate T. Grebenc & al. pilei were distributed among different DNA-sequences phylogenetic clades obtained (data not shown).The exception is a collection from Finland which was microscopically identified as H. umbilicatum but not anaysed at DNA level (Huhtinen & Ruotsalainen, 2006).In specimens of H. umbilicatum collected in North America (Harrison & Grund, 1987) or Asia (Maas Geesteranus, 1971), the spores are larger than for other species and form a unique clade within the DNA-sequences phylogenetic tree, similary distant from several H. rufescens clades and H. ellipsosporum.Solely using morphological criteria a misidentification of H. rufescens with H. umbilicatum would only be possible for samples with extremely large spores which might explain the use of the name "umbilicatum" in some European literature.
Hydnum ellipsosporum was first described only recently, based on many collections from Germany (Ostrow & Beenken, 2004).The distribution of this species seems to be broader since two collections analysed in our study were from Spain (RUFHYD8) and Slovenia (RUFHYD1), at both locations collected from sites dominated by broadleaved trees, which is not common for other Hydnum collections.The species was also confirmed from various locations in Finland.After the DNA-sequences phylogenetic analyses, we confirmed Ostrow & Beenken's well supported molecular and morphological separation of the species from other European species of Hydnum.
Based on DNA-sequences, the phylogenetic separation within H. repandum and H. rufescens indicates higher DNA-sequences variability than that observed solely at the morphological level.Within DNAsequences phylogenetic trees based on ITS1, 5.8S and ITS2 sequences nrDNA, H. repandum specimens formed two non-monophyletic but well supported clades (RE1 and RE2) after both maximum parsimony and Bayesian analysis.The only morphological criterion to separate these two clades is the size of pileus, which appeared smaller for collections distributed in clade RE2.Additionally, clade RU1 (samples identified as H. rufescens, but with mixed morphological characteristics of H. rufescens and H. repandum; see Table 2) appeared as neighbour clade.In this case, despite being uncommon in fungi (Spiers and Hopcroft, 1994), hybridisation cannot be excluded.The resolution obtained from the ITS region was not enough to distinguish H. repandum and H. repandum f. amarum (HYDREP1), a locally recognised form reported to be non-edible due to its bitter taste at all developmental stages of the fruitbody (Stropnik & al., 1988;Petkovšek & Vrščaj, 1998).A somewhat bitter taste of some basidiomata of H. repandum has been reported by other authors for collections from Asia.
A similar taste has also occasionally been noticed in Europe by Jaccottet in 1948(cit. in Maas Geesteranus, 1971) but it never proposed to form a rate taxonomical unit.No data is available for constant and uniform occurrence of basidiomata with a bitter taste.This character may be due to the influence of ecological conditions of the site.Hydnum rufescens, the only species within the genus with previously confirmed intraspecific variability (Agerer & al., 1996), appeared to be the most variable.The specimens were found to fall into six well supported but non-monophyletic clades after DNAsequences phylogenetic analysis.The DNA-sequences phylogenetic tree distances between each of the neighbouring H. rufescens clades and other sister clades (e.g.well established species H. umbilicatum and H. ellipsosporum) indicate that each of the six H. rufescens clades can be recognised and treated as separate species.The morphological information assessed did not correlate well with the molecular results.Rough statistical analysis can only separate group RU3, with generally larger sporocarps but no significant difference in spore size, and no other assessed characteristics could be found for any of the other clades.
The main evolution force for ribosomal region is a concerted evolution which should lead to homogenisation of individual repeats and produce a uniform sequence in all repeats of a given phylogenetic species (Vogler & DeSalle, 1994).Differences in sequences within one morphological species, as observed in H. repandum and H. rufescens, may indicate the presence of more than one phylogenetic species (cryptic species) or variation within the species on a molecular level which cannot be explained by the concerted evolution theory.Similar variation and presence of more phylogenetic species was earlier observed and proposed within Rhizopogon roseolus (Martín & al., 2000), Tricholoma flavovirens (Pers.)S. Lundell.(Horton, 2002), Tuber rufum (Iotti & al., 2007;Grebenc & al., in prep), and in other ectomycorrhizal genera (Leccinum, Lactarius, Inocybe, Tricholoma, and Russula) (Kåren & al., 1997;Horton, 2002).This could be the case in Hydnum rufescens and H. repandum as well.Relatively high abundance of such variability seriously challenges the morphological species concept for these taxa.Several studies demonstrated that DNA-sequences based phylogenetic species recognition and concept were advantageous in mycology and seem likely to become popular among mycologists (Taylor & al., 2000).It could well be applied in the case of H. rufescens, H. repandum, yet raising the question about the use of current names for several phylogenetic species.
Despite relatively high genetic distance in the DNA-sequences phylogenetic tree, the geneflow between clades cannot be excluded, but the presence of putative heteroduplex in rDNA was rejected using DGGE analysis of ITS region (Grebenc & al., 2006).
Based on the high molecular diversity of presumably homogeneous rDNA region in H. rufescens and H. repandum these two species may be in a process of intensive speciation, not correlated to the geographical distances between the different clades obtained.There are several possible triggers which could lead to a possible diversification at the molecular level in H. rufescens and H. repandum.Harrington & Rizzo (1999) reported a high importance of niche in determining the development and maintenance of fungal species which is not necessarily correlated to geographical distances.H. rufescens and H. repandum are common species in Europe growing next to one or more different ectomycorrhizal partners (Table 1) which could lead to a possible diversification at the molecular level, as observed in the H. ellipsosporum samples analysed.Other ecological variables not quantified in the present study, for example soil and other environmental parameters, possible niche specialist character of the species (Giraud & al., 2008), or ectomycorrhiza characteristics and partners, mating types, etc., should be evaluated to explain the variability at the molecular level.

Fig. 1a .
Fig. 1a.DNA-sequences phylogenetic tree for the Hydnum specimens under study.Strict consensus tree with bootstrap values for heuristic search of the 100 most parsimonious trees.DNA-sequences phylogeny clades/morphological species: RU, H. rufescens; RE, H. repandum; UM, H. umbilicatum; EL, H. ellipsosporum; AL, H. albidum and ALM, H. albomagnum.OTUs names using codes in Tables1 and 2, followed by the Accession Number from the GenBank or UNITE databases.
Summary data (values are summarized for all measurements and collections available, averaged and rounded) and statistical analysis (Multiple Range Test, p<0.05;A-D -similarity groups) for potential informative morphological characters assessed on all available collections distributed among each DNA-sequences phylogenetic clade obtained.

Table 1 .
Hydnum.Collections included in the sequence analyses.Species were determined after morphological characteristics.Taxon names, location, potential host (s), herbarium voucher, DNA isolation code and GenBank accession numbers (Acc.Num.) are given.
1 Collections included in the TLC analyses are marked with * next to the code. 2 Sequences obtained after cloning are marked with * next to the Accession Number.