Arbutoid mycorrhizas of the genus Cortinarius from Costa Rica

Arbutoid mycorrhizas of Comarostaphylis arbutoides (Arbutoidea, Ericaceae) from neotropical montane forests are rarely described. To date, only mycorrhizal associations with the fungal species Leccinum monticola, Leotia lubrica and Sebacina sp. are known from literature. The genus Cortinarius is one of the most species-rich ectomycorrhizal taxa with over 2000 assumed species. In this study, two sites in the Cordillera de Talamanca of Costa Rica were sampled, where Com. arbutoides is endemic and grows together with Quercus costaricensis. Using a combined method of rDNA sequence analysis and morphotyping, 33 sampled mycorrhizal systems of Cortinarius were assigned to the subgenera Dermocybe, Phlegmacium and Telamonia. Specific plant primers were used to identify the host plant. Here, we present the phylogenetic data of all found Cortinarii and describe four of the arbutoid mycorrhizal systems morphologically and anatomically.


Introduction
Comarostaphylis arbutoides is a tropical woody plant of Central America, occurring in dry oak-pine and cloud forests, as well as in the páramo at an elevation of c. 2500-3430 m a.s.l. Together with Arbutus and Arctostaphylos, it belongs to the ericaceous subfamily Arbutoidea, which are known to form arbutoid mycorrhizas with ectomycorrhizal fungi (Molina and Trappe 1982). Although, Bidartondo and Bruns (2001) infer that Com. arbutoides forms arbutoid mycorrhizas with diverse species of Basidiomycetes and Ascomycetes, only mycorrhizal associations with Leccinum monticola, Sebacina sp. and Leotia cf. lubrica have been described so far (Osmundson et al. 2007;Kühdorf et al. 2014Kühdorf et al. , 2015. However, typical as well as presumable ectomycorrhizal forming species of the genera Cortinarius, Hysterangium, Laccaria, Tricholoma and Phaeocollybia, have also been mentioned from the páramo by Halling and Mueller (1999). Therefore, further mycorrhizal associations with other fungal species for Com. arbutoides can be assumed.
The genus Cortinarius is assumed to be the species-richest genus of Agaricales, containing over 2000 species (Garnica et al. 2005) with a worldwide distribution (Peintner et al. 2004). The taxonomy of Cortinarius is largely based on macromorphological characters, spore morphology as well as on chemical characters (Brandrud 1996). The subdivision of Cortinarius into subgeneric units causes many problems, induced by high morphological variation within species, as well as the different weighting of morphological characters by different taxonomists (Peintner et al. 2004). However, molecular investigation of the genus Cortinarius is just at the beginning (Liimatainen 2013;Zotti et al. 2014). As proposed by Peintner et al. (2004), studies should, first of all, focus on natural units (e.g. sections), bringing DNA sequence data as well as morphological and ecological data in accordance, as already done by several authors (e.g. Garnica et al. 2009Garnica et al. , 2011Suárez-Santiago et al. 2009;Niskanen et al. 2013a, b;Dima et al. 2014;Stensrud et al. 2014;Liimatainen et al. 2015).
Cortinarius is an important ectomycorrhizal fungal genus associated with trees, shrubs and a number of herbaceous plants of many different plant families (Liimatainen 2013), whereby also host specificity occurs (e.g. Brandrud 1996; Garnica et al. 2003;Frøslev et al. 2007;Niskanen et al. 2011;Liimatainen 2013). Based on fruit body collections, Halling and Mueller (1999;, Mueller et al. (2006) and Ammirati et al. (2007) have reported and/or described 18 different Costa Rican Cortinarii. These species were collected in the Talamanca mountain range of Costa Rica, where Comarostaphylis and Quercus trees occur.
In our samples collected in the Cerro de la Muerte (Cordillera de Talamanca, Costa Rica) several different C o r t i n a r i u s s p e c i e s f o r m e d m y c o r r h i z a s w i t h Com. arbutoides and Quercus sp. The genus Cortinarius was identified using molecular methods such as large subunit (LSU) and internal transcribed spacer (ITS) sequencing as well as phylogenetic analysis. Plant primers were used to sequence the ITS region of the host plant from the same mycorrhizal system as used for fungal analysis. According to Agerer (1991), we present a morphological and anatomical description of four cortinarioid mycorrhizal systems associated with Com. arbutoides.
Fine root systems of Com. arbutoides were collected during the rainy seasons in October 2010 and 2011. For this, a soil corer (diameter 3 cm; length 40 cm) was used at distances of 50 and 100 cm from the trunk. At the University of Costa Rica, turgid and apparently healthy morphotypes were sorted out using a stereomicroscope. Systems with the same morphological features (e.g. colour, hydrophobicity presence, emanating elements and rhizomorphs) were assigned to one morphotype. For further analyses, the morphotypes were preserved in 2 % glutaraldehyde with a 0.1 M sodium cacodylate buffer (Münzenberger et al. 2009) for light microscopy or dried on silica gel for DNA extraction, respectively. Identification of each morphotype is based on their respective sequence type. Within these 2 years a total of 60 soil cores were taken and analysed. The genus Cortinarius was proven genetically in 23 soil cores.

Molecular analyses
Genomic DNA was isolated from one unramified root tip per morphotype, using the DNeasy Plant Mini Kit (QIAGEN, Hilden, Germany) following the manufacturer's recommendations. To identify the mycorrhizal fungi at both family and species level, PCR amplification and sequencing of the internal transcribed spacer (ITS) region and the ribosomal nuclear large subunit (LSU) were performed. Here, the primer combinations ITS1F/ITS4 (Gardes and Bruns 1993;White et al. 1990) as well as LR0R/LR5 (Moncalvo et al. 2000) were used. In order to identify the plant from mycorrhizal root tips without coamplifying fungal DNA the angiosperm-specific ITS primer pair ITS-5A/ITS-241r was amplified (Osmundson et al. 2007). Sequencing service was facilitated by GATC Biotech AG (Konstanz, Germany). A total of 399 root tips were analysed genetically, of which 33 were identified as members of the genus Cortinarius. All sequences were deposited in NCBI GenBank under the accession numbers KM456990-KM457022 (ITS), KM457023-KM457055 (LSU), KF419121 (Com. arbutoides) and KM978077 (Quercus sp.).
Sequences were analysed and edited using Chromas Lite v2.01 software (http://technelysium.com.au). Identity of obtained sequence data was confirmed by BLASTn search against the NCBI database (http://www.ncbi.nlm.nih.gov/) and the database UNITE (Kõljalg et al. 2005;http://unite.ut. ee/). For phylogenetic analysis at species level, the datasets of ITS sequences provided by Peintner et al. (2004), Garnica et al. (2005) and Ammirati et al. (2007) were used. The dataset was complemented by best match results obtained by NCBI and UNITE blast search for each sequence. Alignment was performed with MAFFT v7 (Katoh et al. 2002) using the FFT-NS-2 alignment algorithm. To estimate phylogenetic relationships, maximum likelihood and Bayesian approaches were applied. Maximum likelihood analyses were performed using RAxML (v7.7.1;Stamatakis 2006) in a parallelized version supplied by RAxML BlackBox (Stamatakis et al. 2008) with trees inferred from 100 rapid bootstrap analyses as starting trees in a heuristic search for the tree with the highest likelihood. GTRCAT was used in the heuristic search and the final evaluation of the best tree found was based on the GTR + Gamma model. The Bayesian analysis was performed using MrBayes v3.2.1 (Ronquist et al. 2012) on an iMac (2.9 GHz Quad-Core Intel Core i5). The GTR + Gamma model was in effect and four chains in two parallel runs were performed for 2,000,000 generations, sampling every 1000. Analyses were performed until average standard deviation of split frequencies was <0.01 and stationarity was checked using Tracer v1. 6.1 (Rambaut et al. 2014). The first 50,000 trees were discarded before calculating the posterior probabilities. The potential scale reduction factor (PSRF) values for all inferences were ∼1.0, indicating a good posterior probability distribution sample.

Microscopy
The morphological and anatomical description of the mycorrhizas was carried out according to Agerer (1987Agerer ( -2012, and the online key of DEEMY Rambold 2004-2015). Anatomical studies are based on multiple arbutoid mycorrhizal systems. Drawings were performed with an interference contrast microscope (BX50F-3, Olympus Corporation, Tokyo, Japan) connected with a drawing tube. All drawings were done in thousandfold magnification.
For semi-thin sections, the mycorrhizas were fixed with 2 % glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2) at room temperature until further processing. Thereafter, six washes in 0.1 M sodium cacodylate buffer were performed. Samples were postfixed in 1 % osmium tetroxide in the same buffer for 1 h under light exclusion at room temperature. After six washes in double-distilled water, samples were dehydrated by immersion for 15 min in 25, 50, 70 and 95 % acetone and three times for 1 h in 100 % acetone, respectively. The mycorrhizal tips were embedded in Spurr's plastic (Spurr 1969) and sectioned with a diamond knife on an Ultracut Reichert Ultramicrotome (W. Reichert-LABTEC, Wolfratshausen, Germany). The sections (0.5-μm thick) were stained with crystal violet and investigated using a light microscope (Zeiss Axioskop 50, Oberkochen, Germany).

Phylogenetic analysis
A total of 399 root tips were analysed genetically, of which 33 were assigned to the genus Cortinarius after sequence comparison with BLASTn in the NCBI database and UNITE. In NCBI, best matches were mainly received with samples originally from North America, whereas comparison in UNITE almost exclusively resulted in European species (Tab. 1). In NCBI, 11 samples (KKM 109,KKM 117,KKM 149,KKM 167,KKM 198,KKM 204,KKM 298,KKM 335,KKM 407,KKM 429,KKM 437) achieved their highest match with no further identified Cortinarius sp. or Cortinariaceae samples, whereby this was the case only for KKM 432 in UNITE. The highest identity match in NCBI (100 %) is obtained by KKM 144 with Cortinarius comarostaphylii from Costa Rica; in UNITE C. leucophanes from Finland is the closest match (99 %) for this sample. KKM 132 achieved the lowest identity match in both databases (89 as well as 88 %) with Cortinarius exlugubris from New Zealand and Cortinarius terpsichores from Sweden, respectively.
The Bayesian and RAxML phylogenies, generated by ITS sequences are concordant. Both trees show the same grouping structure, supported by mainly higher posterior probabilities (PP) in the Bayesian analysis and lower bootstraps (BS) in the RAxML analysis ( Fig. 1). Within the genus Cortinarius, the 33 samples can be assigned to three different subgenera: Dermocybe, Phlegmacium and Telamonia.
Around half of the analysed cortinarioid sequences belong to the Dermocybe subgenus (PP 1/BS 91), whereas 13 samples are assigned to subgroup BDermocybe I^(PP 1/BS 100), and a further two samples (KKM 149, KKM 429) to subgroup BDermocybe II^(PP 0.99/BS 73) (Fig. 1). The samples KKM 132 and KKM 144 are assigned to subgenus Phlegmacium, but can be found in two different clades. There, KKM 144 is grouped in a very well-supported cluster (PP 1/BS 100), whereas the cluster with KKM 132 received a good PP (0.93), but a weak BS support (53). The remaining 16 cortinarioid samples are assigned to the subgenus Telamonia and are distributed amongst three very well-supported clusters (PP 1/BS 96 to 100) as shown in Fig. 1.
Except for the telamonioid samples KKM 198, KKM 373 and KKM 376, where Quercus sp. was proven as host plant ( Fig. 1), all mycorrhizal systems of the cortinarioid samples are formed with Com. arbutoides.

Morpho-anatomical descriptions of arbutoid mycorrhizas formed by various Cortinarius sp. species with the Ericaceae Comarostaphylis arbutoides
Of the 33 sequence types, four morphotypes were described in detail as, here, sufficient mycorrhizal materials were available. Assignment to the respective subgenus is based on phylogenetic analysis (Fig. 1).
Identification key for the cortinarioid mycorrhizas: 1 Mycorrhizal system usually not silvery, but densely stringy; older parts of mantle not transparent; Morphological characters (Fig. 2a) Mycorrhizal systems irregularly pinnate to dichotomous, with 0-1 orders of ramification, systems abundant and dense, up to 6.8 mm long, strongly hydrophobic, of medium distance fringe exploration type. Main axes 0.3-0.5 mm diameter. Unramified ends sinuous to tortuous, not inflated, cylindric, up to 1.1 (2.1) mm long and 0.2-0.3 mm diameter; mantle ochre to brownish, very tip ochre to yellowish, older parts dark brown and mycorrhizas not carbonising. Surface of unramified ends densely stringy, not smooth, very tip transparent and epidermal cells visible through mantle, older parts of mantle not transparent and occasionally, very tip partly silvery due to enclosed air. Rhizomorphs abundant, up to 0.12 mm diameter, roundish to flat in cross-section, emanating from all parts of the mycorrhiza, connection oblique, distal rhizomorphs connected over a long distance with mantle surface, growing into soil or organic layers, ochre to brownish, repeatedly ramified into smaller filaments, with hairy or even fan-like margins; rhizomorphs appear very wiry to brittle and are frequently found in short broken fragments around the mycorrhizal systems. Cystidia lacking. Sclerotia not observed.
Anatomical characters of the mantle in plan views ( Fig. 2c-e) Mantle lacks cells densely filled with oily droplets or brownish content, blue granules, needle-like contents, matrix, crystals and exudated pigments, as well as cystidia. Outer mantle layers densely plectenchymatous, hyphae irregularly to somewhat star-like arranged, rarely ramified, some hyphae in bundles, without any special pattern (type B, Agerer 1991) Anatomical characters of emanating elements (Fig. 3f-h) Lacking are gelatinized hyphae, matrix, rhizomorphal nodia, simple septa, intrahyphal hyphae, crystals, brownish substances and secreted pigments; elbow-like protrusions not observed . Rhizomorphs undifferentiated, hyphae loosely interwoven and of uniform diameter (type A/B, Rambold 2004-2015); hyphae smooth, cells 60-100 μm long, 2.8-5.2 μm diameter and cell walls 0.3 μm; ramification backwards-oriented or acute, one or two hyphal diameter below the septum or in considerable distance from the septum ramification and ramifications one side branch at septum; septa with clamps, backwards-oriented clamps not observed and septa as thick as hyphal walls; anastomoses are frequent, open with a short bridge, bridge slightly thicker than hyphae and cell walls of anastomoses as thick as remaining walls; surface of peripheral hyphae with few soil particles. Emanating hyphae straight to wavy, smooth, surface occasionally with few soil particles, cells (25)  Morphological characters (Fig. 4a) Mycorrhizal systems irregularly pinnate to dichotomous, with 0-2 orders of ramification, systems solitary or in small numbers to abundant and dense, up to 9.9 mm long, slightly hydrophobic, of medium distance fringe exploration type. Main axes 0.2-0.6 mm diameter. Unramified ends bent or sinuous, not inflated, cylindric, up to 1.5 (2.7) mm long, 0.2-0.3 mm diameter; mantle ochre to yellowish brown, very tip yellowish, older parts dark orange and mycorrhizas not carbonising. Surface of unramified ends loosely stringy to loosely cottony, not smooth, between side branches and main axis sometimes fan-like cottony, slightly silvery by enclosed air, rapidly displaced by water when touched, then mantle generally transparent; epidermal cells visible through mantle. Rhizomorphs frequent, up to 0.11 mm diameter; flat in cross-section, emanating from all parts of the mycorrhiza, connection oblique, distal rhizomorphs connected over a long distance with mantle surface, growing into soil or organic layers, ochre to yellowish brown, repeatedly ramified into smaller filaments, with smooth margins. Cystidia lacking. Sclerotia not observed.
Anatomical characters of the mantle in plan views (Fig. 4c-

KKM 149 (Dermocybe) + Com. arbutoides
Morphological characters (Fig. 5a) Mycorrhizal systems irregularly pinnate to dichotomous, with 0-2 orders of ramification, systems abundant and dense, up to 11.3 mm long, slightly hydrophobic and of medium distance fringe exploration type. Main axes 0.3-0.5 mm diameter. Unramified ends sinuous to tortuous, not inflated, cylindric, up to 1.6 (2.5) mm long, 0.3 mm diameter; mantle brown, very tip brownish to greyish, older parts dark brown and mycorrhizas not carbonising. Surface of unramified ends loosely stringy to loosely cottony, not smooth, between side branches and main axis sometimes fan-like cottony, slightly silvery by enclosed air, rapidly displaced by water when touched, then mantle generally transparent and epidermal cells visible through mantle. Rhizomorphs frequent, up to 0.13 mm diameter, flat in cross-section, emanating from all parts of the mycorrhiza, connection oblique, distal rhizomorphs connected over a long distance with mantle surface, growing into soil or organic layers, ochre to reddish brown, repeatedly ramified into smaller filaments, with smooth margins. Cystidia lacking. Sclerotia not observed.
Anatomical characters of the mantle in plan views (Fig. 5ce) Mantle lacks cells densely filled with oily droplets or brownish content, blue granules, needle-like contents, matrix, crystals and exudated pigments, as well as cystidia. Outer mantle layers loosely plectenchymatous, hyphae irregularly arranged, rarely ramified, without any special pattern (type B, Agerer 1991), colour yellowish to colourless and with few soil particles; hyphae 20-59 (100) μm long, 3.5-4.7 μm in diameter, cells walls 0.3 (0.4) μm thick; hyphae with clamps, constricted at septa and septa as thick as cell walls. Middle mantle layers densely plectenchymatous, some hyphae in bundles or irregularly interwoven, sometimes ampullate at one side of septum, hyphae 20-110 μm long, 2.8-5.8 μm in diameter, cell walls 0.3 μm thick, smooth and colourless; hyphae with simple septa, rarely with clamps, constricted at septa and septa as thick as cell walls.
Inner mantle layers are densely plectenchymatous, hyphae in bundles, forming ring-like structures, sometimes ampullate at one side of septum, hyphae uneven in diameter, hyphae 30-200 μm long, 1.3-5.1 μm in diameter, cell walls 0.3-0.4 μm thick, smooth and colourless; hyphae with simple septa, constricted at septa and septa as thick as cell walls; anastomoses open with short bridge, bridge thinner or as thick as hyphae. Very tip is like other parts of the mantle.

Discussion
Despite the efforts in investigating the biodiversity of Costa Rican Cortinarii, only Ammirati et al. (2007) supplied genetic data for their reported species. This makes it difficult to address mycorrhizal findings to already known species. With one exception, none of the cortinarioid mycorrhizas, found in Costa Rica, can be determined to species level. However, due to sequence identity of 100 %, the mycorrhizal sample KKM 144 can most probably be assigned to C. comarostaphylii. This species occurs in the Costa Rican páramo (Halling and Mueller 1999) and, according to Ammirati et al. (2007), C. comarostaphylii seems to be restricted to the Ericaceae Com. arbutoides. This plant was also confirmed genetically as host plant for the mycorrhiza KKM 144. Due to strict host specificity of many Cortinarius species (e.g. Brandrud 1996;Liimatainen 2013), identification with the European species C. leucophanes (Fig. 1) is excluded, since this species is associated with conifers (Ammirati et al. 2007).
Based on phylogenetic analyses, the sample KKM 132 is placed in the subgenus Phlegmacium, but fitted laboriously into this subgenus during processing. Also, the databases NCBI and UNITE reached comparatively low identity values, which confirms the insufficient sequence data quantity of Costa Rican Cortinarii. However, beside Cortinarius glaucopus, Halling and Mueller (1999) mention another phlegmacioid Cortinarii from the Costa Rican páramo. Since no sequence data is available, it remains unclear, whether or not this unidentified Phlegmacium species correspond to our mycorrhizal sample KKM 132.
Anatomically, KKM 132 resembles "N. reticulosa" in showing similar plectenchymatous mantle structures. However, in contrast to KKM 132, "N. reticulosa" possess anastomoses in the outer mantle layer, a rough hyphal surface in the outer mantle layer as well as in the emanating elements, and simple septa as well as intrahyphal hyphae in the emanating hyphae.
The phlegmacioid mycorrhiza of KKM 132 shows no reaction with KOH, an attribute in common with C. alnobetulae and "N. reticulosa". However, both samples differ from KKM 132 in lacking reaction with cotton blue and toluidine blue, and a positive reaction with sulpho-vanillin, respectively. The likewise absent reaction with NH 4 OH of KKM 132 was not investigated in the other mycorrhizal samples of Phlegmacium.
Anatomically, the KKM 298 differs clearly from the 11 Cortinarius subgenus Telamonia ECMs. C. badiovestitus, C. bibulous, C. hinnuleus and C. obtusus have contact clamps in their emanating elements, whereas KKM 298 possesses no contact clamps at all. The anastomoses of C. cinnabarinus are not only open with a short bridge but are also closed by a clamp with a long bridge. Moreover, together with C. bulliardii and C. malachius and C. cinnabarinus show differentiated rhizomorphs of type D or C, whereas these of KKM 298 are assigned to type A/B. The mycorrhiza of C. saturninus is the only one which offers a matrix in the mantle layers as well as in the rhizomorphs. Together with C. armillatus, both systems feature not only clamps but also secondary septa in the emanating elements which were, in contrast, not observed in KKM 298. None of the Cortinarius species possess open anastomoses in all three mantle layers, whereby, in the case of C. cinnabarinus, C. hinnuleus and C. saturninus no middle mantle layer was described. Secondly, a distinctly ring-like arrangement of the inner mantle layer, as was also shown for KKM 298, was not described in any of the 11 telamonioid Cortinarius ECMs.
For chemical reactions, only KOH was tested in all described Cortinarius subgenus Telamonia ECMs. Here, no reaction was observed for KKM 298, as well as for most of the other telamonioid species. Exceptions are C. bulliardii, C. cinnabarinus and C. malachius which show violet pink, red-violet colour reactions or dissolving incrustations. A likewise negative reaction with NH 4 OH, as shown for KKM 298, cannot be conformed since the other telamoinioid systems were not tested. For C. armillatus, no information on chemical reactions is given at all.
Anatomically, the mycorrhizas of BDermocybe I^and BDermocybe II^do not differ much. Both types show a plectenchymatous mantle throughout undifferentiated rhizomorphs, infrequently with smooth margins (types A or B) and both form short anastomoses with contact clamps. These features are typical for Dermocybe species (Agerer 1995). According to DEEMY, our samples herein resemble D. cinnamomeolutea, D. holoxantha and D. huronensis. Nevertheless, these three species also show clear differences. Hyphae of emanating elements (D. huronensis), as well as of the outer mantle layer (D. cinnamomeolutea) possess a rough surface. D. cinnamomeolutea and D. holoxantha have a matrix in the middle and inner mantle layers and, furthermore, none of these three species show a ring-like inner mantle layer as occurs in both, BDermocybe I^and BDermocybe II^. Anastomoses open or closed by a clamp, are very common in the emanating elements, whereas in BDermocybe II^, open anastomoses are less frequently observed than in BDermocybe I^. Within the described Dermocybe ECM only, D. phoenicea also feature these two types of anastomoses.
The rhizomorphs of BDermocybe I^and BDermocybe IIŝ how a bright pink colour reaction with KOH, as well as with NH 4 OH. A colour reaction with NH 4 OH was only tested in D. cinnamomea, where no reaction was observed. A reaction with KOH was investigated in seven of the nine Dermocybe species, where again, D. cinnamomea, as well as D. palustris showed a reaction, but which only affected the hyphal content or the mantle, respectively.
According to Ammirati et al. (2007), Costa Rica has a limited number of confirmed host plants for ectomycorrhizal fungi, such as Quercus sp., Com. arbutoides and Alnus acuminate. Even though oak species have an extensive distribution and are the most common ectomycorrhizal host plant for Cortinarius, they only occur in small numbers. Additionally, some Cortinarius species have very special habitat preferences (Liimatainen 2013) and, amongst topographical factors, often only exist in small, geographically isolated populations (Frøslev et al. 2005). Beside insufficient recorded biodiversity, this explains why the Cortinarius findings in this study primarily differ from Ammirati et al. (2007), and why species of, e.g. subgenus Dermocybe have not been reported yet for Costa Rica. Moreover, in contrast to Ammirati et al. (2007), who described Costa Rican Cortinarii sporocarps exclusively found in June, we found no fruit bodies at all. This suggests that sampling in October, within the rainy season, is suitable for mycorrhizal investigations but not for Cortinarius fruit bodies in this area.