Introduction

Species of the genus Inocybe (Inocybaceae, Basidiomycota) are widely distributed ectomycorrhizal fungi occurring in all climatic zones and continents (Matheny et al. 2009). Approximately 500 species are known (Kirk et al. 2008), but this number might increase in the near future when previously unexplored areas, especially in the Southern Hemisphere, are studied more intensively (Larsson et al. 2009). Inocybe species traditionally are characterised and identified by macroscopical and microscopical features of basidiocarps. However, these characters strongly overlap between species, which renders morphological identification of species difficult. Currently, ectomycorrhizas (ECM) of only ten Inocybe species have been described in detail: I. appendiculata (Beenken 1996a; Beenken et al. 1996a), I. avellana (Ingleby 1999), I. fuscomarginata (Beenken 1996b; Beenken et al. 1996b), I. heimii (Magyar et al. 1999), I. lacera (Cripps and Miller 1995; Cripps 1997), I. lanuginella (Agerer 1995), I. nitidiuscula (Ilyas et al. 2013), I. obscurobadia (Beenken 1996c; Beenken et al. 1996c), I. petiginosa (Agerer 1995; Ingleby et al. 1990) and I. terrigena (Beenken 1996d; Beenken et al. 1996d). The morpho-anatomical characteristics of ECM of Inocybe species are relatively similar, and all have plectenchymatic mantles (types A, B and E, and some with a gelatinous matrix, type C sensu Agerer 1991). The similarity of morpho-anatomical characters of Inocybe ECM may be one reason why the ECM of so few species have been described in detail. Inocybe species form ECM with diverse plant hosts belonging to the families Betulaceae, Casuarinaceae, Cistaceae, Dipterocarpaceae, Fabaceae, Fagaceae, Myrtaceae, Nothofagaceae, Pinaceae and Salicaceae (Agerer 1987–2008; Agerer et al. 1996–2012; Glen et al. 2001; Horak 1977; Matheny and Watling 2004; Matheny and Ammirati 2003; Ryberg et al. 2008; Matheny et al. 2009). Inocybe species also form orchid and arbutoid mycorrhizas (Ryberg et al. 2008; Wang and Qiu 2006).

Similar to other fungal genera, the majority of Inocybe sequences in public databases (e.g., GenBank) are unidentified or have been assigned incorrectly to species (Ryberg et al. 2008), which makes ectomycorrhiza identification even more difficult.

The semiarid sandy habitats of the Danube–Tiscia interfluves represent the westernmost limit of the Eurasian steppe belt. Fruitbodies of a number of Inocybe species have been collected from this region (Babos 1999; Nagy 2004; Nagy and Gorliczai 2007). Several of these species are reported from semiarid sandy grassland with woody steppe patches near Fülöpháza, from which the ectomycorrhiza of Inocybe heimii, described in detail previously (Magyar et al. 1999) was collected.

As part of an investigation of the community of ectomycorrhizal fungi in semiarid woody steppe, we aimed to study the anatomical characteristics of Inocybe ECM collected from different plant hosts and compare them with the Inocybe ECM described previously. In addition, we used nucleotide sequences for the internal transcribed spacer (ITS) region of nuclear DNA (nrDNA) to aid with identification of the fungi and attempted to locate corresponding species hypotheses (SHs) sensu the UNITE Database (Kõljalg et al. 2013).

Materials and methods

Sampling

Samples were collected in a strictly protected semiarid sandy grassland with woody steppe patches near Fülöpháza, Hungary (46° 52′ N, 19° 24′ E). The area is a part of the Kiskunság National Park. The sampling site was described in detail in previous studies (Knapp et al. 2012; Kovács and Szigetvári 2002). Five ectomycorrhizal plant species inhabit this woody steppe/grassland: Fumana procumbens, Helianthemum ovatum, Populus alba, Salix rosmarinifolia and the adventive Pinus nigra.

Two morphotypes (Mt 4 and 6) were collected from samples collected as follows. An approximately 15 × 15 × 15-cm cube of soil under basidiocarps of Inocybe species was cut with a sharp knife and transferred to the laboratory in plastic bags. From these samples, the morphotypes resembling ECM of Inocybe species were collected. Five additional morphotypes (Mt 1, 2, 3, 5 and 7) were collected from soil samples as part of a larger-scale study of the ectomycorrhizal fungal community in the area (Seress, Geml, Németh, Lukács, Dima, Nagy and Kovács, unpublished results). During this investigation, soil samples (altogether 159 contained ECM) were taken from patches of a specific host plant using a 1 1/8″ × 12″ Plated Soil Recovery Probe soil sampler (AMS, American Falls, ID, USA). When morphotypes resembling ECM of Inocybe species were found in sufficient numbers for subsequent anatomical study, they were documented and fixed for such examination. Each morphotype was represented by 20–50 root tips per sample. Ectomycorrhizal root tips were washed in a sieve with water, and a soft paintbrush was used to remove soil particles under a stereomicroscope. The frequency of each morphotype in a soil sample was roughly estimated as rare, moderately frequent or frequent. Ectomycorrhizal root tips were fixed in formalin–ethanol–acetate (FEA) for anatomical examination or 70 % ethanol for both anatomical and molecular analyses.

Morphological characterisation

The morphology and anatomy of the seven Inocybe morphotypes were characterised following Agerer’s method (Agerer 1991). The ramification systems were examined under a stereomicroscope (Nikon SMZ 1000), and mantle structure and hyphal characteristics were studied using differential interference contrast (Nomarski) optics (Olympus BX51 and Zeiss Axioskop 2 Plus microscopes) and photodocumented (using Olympus C-4040Zoom and Zeiss AxioCam ICc5 cameras).

Unramified ends of the ECM were embedded using the Leica Historesin Embedding Kit following the manufacturer’s instructions. Longitudinal sections (10 μm thickness) were observed microscopically (Nikon Eclipse 80i) and photodocumented (POT RT-KE 1055862–0 camera). Photos of the same section were merged using the ‘Photomerge’ function of Adobe PhotoShop CS5.

Molecular identification

Total genomic DNA was extracted from root tips using the E.Z.N.A. Fungal DNA Extraction Kit (Omega Bio-Tek, Norcross, GA, USA) following the manufacturer’s instructions. The ITS region of the fungal nrDNA was amplified and sequenced using the ITS1F/ITS4 (Gardes and Bruns 1993; White et al. 1990) primer pair as described previously (Kovács et al. 2008).

Sequencing was carried out by LGC GmbH (Berlin, Germany). The electropherograms were checked using the Pregap4 and Gap4 programs of the Staden package (Staden et al. 2000). The ITS sequences obtained from each morphotype and sporocarp were deposited in GenBank (KT630088–KT630097, Fig. 1).

Fig. 1
figure 1

Fifty percent majority rule consensus phylogram derived from Bayesian inference analysis of nrDNA ITS sequences of Inocybe species. Conocybe species served as the outgroup. The seven new morphotypes collected from the Hungarian semiarid sandy steppe and described in the present study are highlighted in bold. Species names are given only if a sequence originates from a sporocarp. Sequence origin, host and provenance are based on GenBank entries and the original publications. Thick branches indicate that Bayesian posterior probabilities and ML bootstrap values are 0.90 and 70 % or higher, respectively. Scale bar indicates 0.1 expected changes per site per branch

Full size image

The sequences were compared with those deposited in the GenBank (http://www.ncbi.nlm.nih.gov/genbank/) and UNITE (https://unite.ut.ee/) public databases using the BLAST search tool (Altschul et al. 1990), and with our database of ectomycorrhizal fungal ITS sequences from the area (930 obtained from ectomycorrizal root tips and 60 from sporocarps) using the BLAST tool in BioEdit (Hall 1999).

We also checked how frequently the operational taxonomic units (OTUs) containing the ITS obtained in the present study were found in the soil samples collected during the ectomycorrhizal fungal diversity study of the area (Seress, Geml, Németh, Lukács, Dima, Nagy and Kovács, unpublished results). During that study, altogether, 19 OTUs were identified as representing Inocybe ECM; they were detected in 50 soil samples. Five OTUs were found only in one soil sample, 4 OTUs were in 2, 1 OTU was in 3 samples, while 2–2 OTUs were in 4 and 5 samples. Only 5 Inocybe OTUs were present in more than 5 soil samples (Seress, Geml, Németh, Lukács, Dima, Nagy and Kovács, unpublished results).

Short, thin roots of Salix and Populus species cannot be identified unambiguously from their morphology, so based on ITS sequence data in GenBank, we designed host-specific PCR primers targeting the nrDNA ITS region: Salix (Salix_ITSrevB: 5′-GCAACGAGAGCATCCTTGAA-3′) and Populus (Populus_ITSrevB: 5′-CAACGAGAGCATCCTAGAATTT-3′). We used both primers in combination with the eukaryotic primer ITS1 (White et al. 1990) with the following PCR program: denaturation at 94 °C for 5 min, then 35 cycles of denaturation for 30 s at 94 °C, annealing for 30 s at 52 °C and extension for 90 s at 72 °C, and a final extension for 10 min at 72 °C. Molecular identification of the host was not needed when ECM of Pinus nigra were collected.

For the phylogenetic analysis, one ITS sequence for each of the seven morphotypes studied herein and the ITS sequences obtained from the sporocarps collected were analysed together with sequences obtained from public databases (GenBank and UNITE). Species of genera Conocybe were chosen as the outgroup. The 87 ITS sequences were aligned using MAFFT (online version 7) with the E-INS-i alignment strategy under the default settings (Katoh and Standley 2013) and adjusted manually in SeaView (Gouy et al. 2010). The alignment used in the final analyses was 834 characters long. The indel positions were coded following the simple indel coding algorithm (Simmons et al. 2001) with the program GapCoder (Young and Healy 2003). Adding indel characters to the nucleotide alignment of ITS sequences increases the robustness of phylogenetic analyses (Nagy et al. 2012). Bayesian inference (BI) analyses were performed with MrBayes 3.1.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). For BI, the ITS sequences and indel characters were split into two partitions to which the GTR+G and two-parameter Markov (Mk2 Lewis) models, respectively, were applied. Four Markov chains were run for 10,000,000 generations, sampling every 1000 steps, and with a burn in at 5000 sampled trees. The Markov Chain Monte Carlo (MCMC) convergence was checked with AWTY online (Nylander et al. 2007).

Maximum likelihood (ML) phylogenetic analyses were carried out with RAxML (Stamatakis 2014) as implemented in the raxmlGUI (Silvestro and Michalak 2012) with GTRGAMMA for DNA and the default set for binary (indel) characters. Rapid bootstrap analysis with 1000 replicates was used to test the support of the branches. Phylogenetic trees were visualised and edited in MEGA6 (Tamura et al. 2011).

Results

Phylogenetic analyses and identification of the morphotypes

Five species (I. appendiculata, I. curvipes [=I. lanuginella], I. lacera, I. obscurobadia and I. petiginosa) with ECM described previously are phylogenetically distant from Inocybe taxa studied herein, so their ITS sequences were omitted after preliminary analyses and were not included in the final data set. Although ITS sequence-based identification was carried out during anatomical characterisation of the ectomycorrhiza of I. nitidiuscula (Ilyas et al. 2013), we did not include sequences of this species in the analyses owing to taxonomic ambiguity (see ‘Discussion’). No ITS sequences of I. heimii and I. avellana were located in public databases.

The ITS region was successfully amplified and sequenced from one root tip in the case of Mt 4 and 6. In the case of Mt 1 and 7, the ITS region from two root tips, and for Mt 2, 3 and 5, the ITS region from five root tips per morphotype were successfully amplified and sequenced; the sequences obtained from different root tips of one morphotype were identical.

In total, 83 Inocybe ITS sequences and four outgroup sequences were included in the final data set. The Inocybe sequences formed eight large clades, most of which contained well-supported smaller subclades (Fig. 1). The taxonomic names used in public databases and the possible assignations to SHs (Kõljalg et al. 2013) illustrate well the taxonomic ambiguities and problems in the genus. Several clades contained intermixed names, whereas some clades showed high species diversity with very short branch lengths.

No species name could be assigned to the fungal partner of two morphotypes (Mt 6 and 7). We could assign Inocybe species names to five ectomycorrhizal morphotypes, of which four corresponded with one–one SH: I. phaeoleuca (Mt 1), I. subporospora (Mt 3), I. semifulva (Mt 4) and I. psammophila (Mt 5). The clade containing Mt 4 included sequences identified with the name I. flocculosa; however, we considered the clade to represent I. semifulva because the ITS sequence of the type specimen of the latter was among the sequences (Fig. 1). More than one SH could be associated with the clade containing Mt 2 and sequences for I. splendens. The clades in which the seven Inocybe morphotypes were nested contained sequences of ectomycorrhizal samples from different regions and hosts from around the world (Fig. 1).

Characterisation of seven Inocybe ectomycorrhizal morphotypes

Morphotype 1. Inocybe phaeoleuca (host: Populus alba)

Morphology

(Supplementary Material Fig. S1e) Ectomycorrhizal system 2–2.5 mm long, irregularly pinnate, two or three side branches per 10 mm. Axes 0.2–0.3 mm diameter, the unramified ends sinuous and tapering, 0.8–1.5 mm long and 0.1–0.3 mm diameter. Mycorrhizal surface yellow-ochre, slightly darker ochre in old ECM, cottony. Distinct mantle surface visible, mantle semi-transparent.

Anatomy

Outer layer of mantle plectenchymatous, with hyphae irregularly arranged, no specific pattern discernible, gelatinous matrix present (type C; Agerer 1991). Hyphae colourless, 2–3 μm diameter, clamps not observed. Inner mantle layer (Fig. 2e) densely plectenchymatous, transient to pseudoparenchymatous mantle type. Hyphae 3–4 μm diameter, colourless, clamps and matrix lacking. Emanating hyphae straight or bent, colourless, cylindrical, not constricted at septa, 2–3 μm diameter, clamps present. Hyphae sometimes with frizzled or inflated ends and Y-shaped ramifications. Longitudinal section (Supplementary Material Fig. S2g) Fungal mantle 13–50 μm thick, plectenchymatous, different layers not discernible. Hyphal cells 1–3 μm radially. Mantle of ectomycorrhizal tip plectenchymatous, 25–30 μm thick. Epidermal cells of host plant rectangular or radially oval or elliptic, oriented parallel or obliquely to root axis, 15–30 μm tangentially, 23–75 μm radially. Hartig net paraepidermal. Hyphal cells around epidermal cells cylindrical, 1–1.5 μm thick, in one row. Hartig net in plan view of palmetti type, with 1–3 μm broad lobes.

Fig. 2
figure 2

Mantle structure of the seven Inocybe ectomycorrhizal morphotypes (Mt). Mantle structure of a Inocybe sp. (Mt 6), b I. semifulva (Mt 4), c I. psammophila (Mt 5) and d I. splendens (Mt 2). e Inner mantle layer of I. phaeoleuca (Mt 1). Mantle structure of f I. subporospora (Mt 3) and g Inocybe sp. (Mt 7). Scale bars = 20 μm

Full size image

Occurrence

Rare in the original sample. Based on ITS sequence data, I. phaeoleuca Mt 1 was also present in six Populus alba, one Salix rosmarinifolia and two Pinus nigra soil sample (frequent in samples).

Morphotype 2. Inocybe splendens (host: Populus alba)

Morphology

(Supplementary Material Fig. S1a) Ectomycorrhizal system 8 mm long, monopodial-pyramidal, two or three side-branches per 10 mm, axes 0.3–0.4 mm diameter. The unramified ends straight, bent or sinuous with tapering tips, 1–2.5 mm long, 0.2–0.3 mm diameter. Mycorrhizal surface very loosely cottony, slightly silvery, sometimes slightly pinkish brown with characteristic black spots which also occur in young ECM. Distinct mantle surface visible, mantle sometimes semi-transparent.

Anatomy

Outer layer of mantle plectenchymatous, hyphae in net-like arrangement, repeatedly and squarrosely branched (type E; Agerer 1991), mantle cells colourless. Inner mantle layer (Fig. 2d) transitional between plectenchymatous and pseudoparenchymatous mantle types, hyphae irregularly shaped. Hyphae 3–4 μm diameter in outer mantle layer, 4–5 μm diameter in inner mantle layer, clamps and matrix lacking in both layers. Hyphae sometimes inflated, ampullate on both sides of septum. Emanating hyphae straight, colourless, cylindrical, constricted at septa, 2–4 μm diameter, clamps present. Longitudinal section (Supplementary Material Fig. S2f) Mantle 20–40 μm thick, plectenchymatous, different layers not discernible, hyphae 1–5 μm radially. Mantle of ectomycorrhizal tip plectenchymatous, 80–100 μm thick. Epidermal cells of host plant tangentially oval to elliptic or cylindrical, oriented obliquely to root axis, 25–50 μm tangentially, 8–20 μm radially. Hartig net epidermal, hyphal cells around epidermal cells roundish, 2–3 μm thick, in one row. Hartig net in plan view of palmetti type, with 2–4 μm wide lobes (Supplementary Material Fig. S3d).

Occurrence

Frequent in original sample. Based on ITS sequence data, I. splendens Mt 2 was present in three additional Populus alba samples (also frequent).

Morphotype 3. Inocybe subporospora (host: Pinus nigra)

Morphology

(Supplementary Material Fig. S1f) Ramification of ectomycorrhizal system irregularly pinnate, dichotomous-like. Axes 0.2–0.3 mm diameter, the unramified ends 2–3 mm long, 0.2–0.3 mm diameter. Distinct mantle surface sometimes visible, mantle slightly transparent. Mycorrhizal surface loosely stringy, brown and lighter, yellowish-, orange-brown coloured at the tapering ends. Old ECM dark brown, almost black colour. ECM sinuous or tortuous shaped, constricted between older and younger parts.

Anatomy

Mantle plectenchymatous, gelatinous matrix between the hyphae (type C; Agerer 1991) (Fig. 2f). Hyphae 5–6 μm diameter, colourless, cylindrical, not constricted at septa, without clamps. Emanating hyphae straight, colourless, 2–3 μm diameter, not constricted at septa, clamps present. Anastomoses not observed. Longitudinal section Supplementary Material Fig. S2c Fungal mantle thin, 8–35 μm thick, shows plectenchymatous–pseudoparenchymatous transition form but different layers not discernible. Hyphae 2–4 μm radially. Mantle of ectomycorrhizal tip plectenchymatous, 30–40 μm thick. Cortical cells of host plant irregularly angular, oriented obliquely or rarely parallel to root axis, 8–20 μm tangentially, 25–33 μm radially. Hartig net protrudes towards endodermis, hyphal cells around cortical cells cylindrical, 4–5 μm thick, in one row (Supplementary Material Fig. S3f). Hartig net in plan view of palmetti type, with 3–4 μm broad lobes.

Occurrence

Frequent in original sample. Based on ITS sequence data, I. subsporospora Mt 2 was present in one additional Pinus nigra sample (also frequent).

Morphotype 4. Inocybe semifulva (host: Salix rosmarinifolia)

Morphology

(Supplementary Material Fig. S1c) Ramification of ectomycorrhizal system monopodial-pinnate, two or three side-branches per 10 mm. Axes 0.2–0.3 mm diameter, the unramified ends 0.2–1.2 mm long, 0.1–0.2 mm diameter. Mycorrhizal surface loosely stringy, yellow, distinct mantle surface visible, mantle not transparent. ECM straight or sinuous, lacking inflated, cylindrical tips.

Anatomy

Fungal mantle plectenchymatous, hyphae rather irregularly arranged, no specific pattern discernible (type B; Agerer 1991) (Fig. 2b). Hyphae 2–3 μm diameter, cylindrical, not constricted at septa, colourless, without matrix. Emanating hyphae straight or wavy, colourless, 2–3 μm diameter, clamps present, Y-shaped ramifications observed. Anastomoses not observed. Longitudinal section (Supplementary Material Fig. S2e) Fungal mantle 10–50 μm thick, plectenchymatous, different layers not discernible, hyphae 1–5 μm radially. Mantle of ectomycorrhizal tip plectenchymatous, thinner than other parts, 5–8 μm thickness. Epidermal cells of host plant tangentially oval to elliptic or cylindrical, oriented obliquely to root axis, 29–40 μm tangentially, 41–60 μm radially. Hartig net (Supplementary Material Fig. S3b) periepidermal. Hyphal cells around epidermal cells cylindrical, 2–3 μm thick, in one row. Hartig net in plan view of palmetti type, with 2–5 μm broad lobes.

Occurrence

Very frequent in the original sample. It was not observed in any other soil sample.

Morphotype 5. Inocybe psammophila (host: Populus alba)

Morphology

(Supplementary Material Fig. S1d) Ectomycorrhizal system irregularly pinnate, two or three side-branches per 10 mm. Axes 3–4 mm long, 0.2–0.3 mm in diameter. The unramified ends tortuous and cylindrical, 1.3–2.5 mm long, 0.1–0.3 mm diameter. Mycorrhizal surface brown, stringy, covered with soil particles, distinct mantle surface visible, mantle semi-transparent.

Anatomy

Fungal mantle plectenchymatous, gelatinous matrix between hyphae (type C; Agerer 1991) (Fig. 2c). Hyphae 3–5 μm diameter, colourless, cylindrical, constricted at septa, without clamps. Emanating hyphae straight, colourless, 2–4 μm diameter, constricted at septa, clamps present. Anastomoses not observed. Longitudinal section (Supplementary Material Fig. S2b) Fungal mantle very thin, 5–12 μm thick, plectenchymatous, and different layers not discernible, hyphae 1–2 μm radially. Mantle of ectomycorrhizal tip plectenchymatous, 20–25 μm thick. Epidermal cells of host plant radially oval to elliptic, oriented obliquely to root axis, 8–23 μm tangentially, 25–50 μm radially. Hartig net (Supplementary Material Fig. S3e) paraepidermal, hyphal cells around epidermal cells roundish, 2–7 μm thick, in one row. Hartig net in plan view with infrequent, 4–6 μm broad lobes.

Occurrence

Very frequent in the original sample. Based on ITS sequence data, I. psammophila Mt 5 was present in four additional Populus alba and two Salix rosmarinifolia samples (also frequent).

Morphotype 6. Unidentified Inocybe sp. (host: Salix rosmarinifolia)

Morphology

(Supplementary Material Fig. S1b) Ectomycorrhizal tips solitary or arranged in small numbers, without ramification. Mycorrhizal ends sinuous with inflated, club-shaped tips, 1–3 mm long, 0.1–0.2 mm diameter. Mycorrhizas brown, older mycorrhizas usually darker brown. Infrequent emanating hyphae render mycorrhizal surface very loosely cottony, distinct mantle surface visible, mantle not transparent.

Anatomy

Fungal mantle plectenchymatous, arrangement of hyphae rather net-like, repeatedly and squarrosely branched (type E; Agerer 1991) (Fig. 2a). Hyphae 3–5 μm diameter, colourless, cylindrical, constricted at septa, clamps and matrix lacking. Emanating elements have open anastomoses with short bridge and cell walls as thick as remaining walls. Emanating hyphae straight or wavy, colourless, 2–3 μm diameter, clamps present. Longitudinal section (Supplementary Material Fig. S2a) Fungal mantle 8–15 μm thick, plectenchymatous, different layers not discernible, hyphae 1–1.5 μm radially. Mantle of the ectomycorrhizal tip plectenchymatous, 15–20 μm thick. Epidermal cells of host plant tangentially oval to elliptic or cylindrical, oriented obliquely to root axis, 15–33 μm tangentially, 30–67 μm radially. Hartig net paraepidermal. Hyphal cells around epidermal cells roundish or cylindrical, 2–3 μm thick, in one or rarely two rows. Hartig net in plan view of palmetti type, with 2–4 μm broad lobes (Supplementary Material Fig. S3a).

Occurrence

Very frequent in the original sample. Based on ITS sequence data, Inocybe sp. Mt 6 was also present in three Salix rosmarinifolia and two Populus alba samples (also frequent).

Morphotype 7. Unidentified Inocybe sp. (host: Salix rosmarinifolia)

Morphology

(Supplementary Material Fig. S1g) Unramified ends solitary or in small numbers, sinuous-shaped, 2–4 mm long, 0.1–0.2 mm diameter, with tapering tips. Mycorrhizas brownish-ochre with yellowish-brown tips, old mycorrhizas brown. Mycorrhizal surface cottony, distinct mantle surface visible, mantle not transparent.

Anatomy

Outer layer of mantle plectenchymatous (Fig. 2g), hyphae are irregularly arranged, no specific pattern discernible (type B; Agerer 1991), mantle cells colourless. Inner mantle layer plectenchymatous, arrangement of hyphae rather net-like, repeatedly and squarrosely branched (type E; Agerer 1991). Hyphae 3–4 μm diameter, clamps and matrix lacking in both layers. Emanating elements have open anastomoses with short bridge, bridge as thick as hyphae. Emanating hyphae straight or wavy, colourless, cylindrical, constricted at septa, 2–3 μm diameter, Y-ramifications and clamps present. Longitudinal section (Supplementary Material Fig. S2d) Fungal mantle 8–20 μm thick, plectenchymatous, two different layers discernible. Outer mantle layer hyphae 1–3 μm radially. Inner mantle layer hyphae 2–4-μm diameter, arranged more tightly. Mantle of ectomycorrhizal tip plectenchymatous, more coarse (40–50 μm) than other parts. Epidermal cells of host plant tangentially oval to elliptic or cylindrical, oriented obliquely to root axis, 5–13 μm tangentially, 18–43 μm radially. Hartig net (Supplementary Material Fig. S3c) paraepidermal. Hyphal cells around epidermal cells roundish or cylindrical, 2–3 μm thick, in one row. Hyphae of Hartig net in plan view create infrequent bigger lobes, which are 4–5-μm broad.

Occurrence

Relatively rare in the original sample. Based on ITS sequence data, Inocybe sp. Mt 7 was also present in three additional Salix rosmarinifolia samples.

Discussion

Here, we described seven morphotypes of Inocybe ECM collected from two native tree species (Salix rosmarinifolia and Populus alba) and one exotic host (Pinus nigra) growing in sandy semiarid woody steppe in Hungary. Two of the morphotypes were formed by unidentified Inocybe species, whereas the other five morphotypes, on the basis of the results of molecular phylogenetic analyses, were formed by I. phaeoleuca, I. psammophila, I. semifulva, I. splendens and I. subporospora. However, one should bear in mind the taxonomic and nomenclatural problems that plague the genus and ambiguous database records when considering the results together with previous morpho-anatomical descriptions of Inocybe species.

The seven Inocybe ECM described here and the ones described previously, albeit there are differences, share common anatomical characteristics. The clamped emanating hyphae render the surface of the ECM loosely or densely cottony. On this basis, all Inocybe ECM described belong to the short exploration type (Agerer 2001). Inocybe ECM have a plectenchymatous mantle formed by hyphae with no clamp connections, and the mantle structure shows no discernible pattern or slightly organised plectenchymatous types A, B or E. In some cases, a gelatinous matrix is present around the hyphae (type C). Neither cystidia nor rhizomorphs have been reported from any Inocybe ectomycorrhiza.

Although the two species are phylogenetically distant, the ECM of I. phaeoleuca and I. psammophila are very similar. Both ECM are pale brown, ochre, slightly cottony, with tortuous mycorrhizal tips, and both have gelatinous mantle (type C). Mantle with gelatinous material is reported for three other Inocybe ECM: I. fuscomarginata (Beenken 1996b; Beenken et al. 1996b), I. terrigena (Beenken 1996d; Beenken et al. 1996d) and I. heimii (Magyar et al. 1999). However, I. fuscomarginata and I. terrigena ECM are yellow-white and yellow-orange, respectively, densely cottony and have a ring-like organised inner mantle structure; no such characters were detected in the ECM described here. As these three species belong to Inocybe subgenus Mallocybe, the two ECM described here are the first Inocybe ECM with gelatinous mantle known from the subgenus Inocybe.

The ectomycorrhiza of I. heimii formed with Fumana procumbens described previously was collected from the sampling area of the present study (Magyar et al. 1999). The ECM of both I. phaeoleuca and I. psammophila are very similar to the ECM described as I. heimii (Magyar et al. 1999). They are brownish, loosely cottony and have the C-type gelatinous plectenchymatous mantle. Both I. phaeoleuca and I. heimii have a dense plectenchymatous mantle layer transient to a pseudoparenchymatic structure, but the former has Y-shaped and the latter 90° angle hyphal ramifications of emanating elements. Inocybe heimii was identified by comparison of the ITS RFLP profile obtained with one restriction enzyme (EcoRI) from a sporocarp and the ECM isolated from the soil under the sporocarp. We checked the EcoRI restriction map of the ITS region of I. phaeoleuca and I. psammophila (data not shown). Both species have a single restriction site and might show the same profile as detected in I. heimii, especially considering that different primers were used for amplification of I. heimii (ITS1/ITS4, Magyar et al. 1999) from those used in the present study. Unfortunately, following the description of I. heimii (Magyar et al. 1999), neither a voucher of the sporocarp nor an ITS sequence are available, so checking the identification is impossible.

The Inocybe ectomycorrhizal morphotypes described here with the similar type-B or -E plectenchymatic mantle are not easily distinguishable especially when they are derived from the same host. Inocybe semifulva and the two unidentified Inocybe species form very similar ECM with Salix rosmarinifolia; only slight differences, such as the yellow colour and more robust ends of the ECM of I. semifulva, are discernible. Although the ECM might represent different mantle types, their plectenchymatic mantles are very similar, and there might be transient structures or different representative parts of the same ectomycorrhiza, such as in the case of Mt 7 described here or in the case of I. obscurobadia (types A, B and E) (Beenken et al. 1996c) and I. appendiculata (types B and E) (Beenken et al. 1996a). The former is white or whitish and differs in colour from the ECM described here, whereas the latter is brown and loosely cottony and is very similar to other Inocybe ECM described here with non-gelatinous mantle.

As mentioned in the ‘Introduction’, the morphology-based species identification of Inocybe is problematic because of overlapping and barely distinguishable characters. Based on the morpho-anatomical characterisation of the seven Inocybe ECM described here and comparison with the ten descriptions published previously, we could discern slightly differing or no distinguishing features among the ECM of different species. Even characteristic features, such as the gelatinous C-type pseudoparenchymatic mantle, cannot be associated with monophyletic groups or specific phylogenetic lineages. Although there is high species diversity and considerable phylogenetic distances within the genus, the known ECM of Inocybe species are similar from a morpho-anatomical perspective. All ITS sequences obtained from root-tips considered to be from one morphotype were identical, which demonstrates that the morphotypes, if carefully investigated, can be differentiated. Based on our experiences, it seems that the ectomycorrhizal colour of Inocybe species is useful to distinguish morphotypes and anatomical features can refine this sorting. Nevertheless, when ECM are studied, molecular phylogenetic analysis should be used for precise identification of Inocybe species.

The I. nitidiuscula ectomycorrhiza was described in association with Alnus nitida from Pakistan (Ilyas et al. 2013). The ectomycorrhiza is creamy white, with frequent emanating hyphae and a plectenchymatic A-type mantle (Ilyas et al. 2013). Although the mantle documented also shows similarities to B- and E-type mantles formed by Inocybe species, the whitish colour of the ectomycorrhiza is different from the ECM described here. However, the identity of the species (I. nitidiuscula) forming the ectomycorrhiza described (Ilyas et al. 2013) is questionable. Inocybe nitidiuscula was originally described from Germany (Britzelmayr 1891) and is known to be associated mainly with coniferous trees (Picea) in Europe (Jacobsson and Larsson 2012). Browsing public sequence databases and using UNITE SHs in the PlutoF cloud database (Abarenkov et al. 2010), we established that ITS sequences assigned to I. nitidiuscula belong to at least five different, phylogenetically separated lineages with more than 3 % sequence differences (data not shown). As an example, the Central/North European sequences of I. nitidiuscula (SH092815.07FU) show only 89 % similarity with sequences assigned to the same species from Pakistan (SH003766.07FU). Therefore, we included only a European sequence of I. nitidiuscula in our analysis and omitted the ectomycorrhizal sequence from Pakistan.

Ryberg and Matheny (2012) observed in their analyses that plant association was a conserved feature of Inocybaceae. It is also hypothesised that the Inocybaceae, as a young group, diversified in the Paleotropics and was originally associated with angiosperms (Matheny et al. 2009; Ryberg and Matheny 2012). Only one of the major lineages of Inocybe (the subgenus Mallocybe) is hypothesised to be associated with conifers (Matheny et al. 2009). Our phylogenetic analyses show that Inocybe species have a wide host and geographic range. The mycorrhizal formation of I. subporospora and I. phaeoleuca with Pinus nigra might be a secondary trait, not only at an evolutionary scale because of the gymnosperm host but also at a local scale. Pinus nigra was introduced into the area in the 1950s when, with the aim of stabilising the sand, intensive planting of the species started (Tamás 2003). Given the gymnosperm/angiosperm association discussed above, we can exclude the possibility that the fungi are co-invaders (sensu Nuñez and Dickie 2014). Thus, both I. phaeoleuca and I. subporospora are probably native to the area but are capable of forming novel associations (sensu Nuñez and Dickie 2014) with the invasive Pinus species.