Mycoscience 59 (2018) 38e48

Contents lists available at ScienceDirect

Mycoscience

journal homepage: www.elsevier.com/locate/myc

Full paper Generalism in the interaction of mycobionts with orchids characterizes a biodiversity hotspot in the tropical Andes of Southern Ecuador

* Paulo Herrera a, c, , Ingrid Kottke b, M. Carmen Molina c, Marcos Mendez c, Juan Pablo Suarez a a Departamento de Ciencias Biologicas, Universidad Tecnica Particular de Loja, San Cayetano Alto S/n C.P, 11 01 608, Loja, Ecuador b Evolutionary Ecology, Institute of Evolution and Ecology, Eberhard-Karls-University Tübingen, Auf der Morgenstelle 1, 72076 Tübingen, Germany c Area de Biodiversidad y Conservacion, Departamento de Biología, Geología, Física y Química Inorganica, Universidad Rey Juan Carlos, C/ Tulipan S/n., E-28933 Mostoles, Madrid, Spain article info abstract

Article history: Biotic interactions play an important role in the assembly and stability of communities. All orchids Received 28 September 2016 depend on mycobionts for early establishment, but whether individual orchid species depend on a Received in revised form specific or broad spectrum of mycobionts is still a matter of debate. Tulasnellaceae () is the 1 August 2017 richest and most widespread mycobiont worldwide. We assessed Tulasnellaceae richness in epiphytic Accepted 2 August 2017 and terrestrial orchids in different habitats, and evaluated the degree of generalism in orchid- Available online 1 December 2017 Tulasnellaceae interactions and the robustness of this mutualistic system to the extinction of myco- biont partners. We sampled 114 orchid individuals including all common and rare species in 56 plots of Keywords: 2 Bipartite network 1m in 3 habitats: pristine forest, regenerating forest and a landslide site in a tropical montane rainforest in Southern Ecuador. We found 52 orchid and 29 Tulasnellaceae species. The composition of Tulasnel- Orchid laceae OTUs was moderately to highly similar across habitats and between orchid growth forms. A Tropical montane rain forest significantly nested network architecture indicated the existence of a core of generalist Tulasnellaceae OTUs interacting with both rare and common orchids. Terrestrial and epiphytic orchids showed signif- icant differences in robustness to the extinction of their Tulasnellaceae mycobionts. Thus, generalist mycobionts may be relevant for the preservation of hyperdiverse orchid communities in the tropics. © 2017 The Mycological Society of Japan. Published by Elsevier B.V. All rights reserved.

1. Introduction supplement of carbon and minerals supplied by mycobionts to establish protocorms in the wild (Rasmussen, 1995). Biotic interactions such as mutualism play an important role in Recently, much attention has been given to documenting the the assembly, persistence, and stability of communities (Fortuna diversity of the mycobionts associated with orchids and their de- et al., 2010). Fungi interact with over 80% of land , forming gree of specificity (Jacquemyn, Honnay, Cammue, Brys, & Lievens, a mutualistic interaction known as mycorrhizae (Smith & Read, 2010; Kartzinel, Trapnell, & Shefferson, 2013; McCormick, 2008). Mycorrhizal symbionts (mycobionts) are a vital compo- Whigham, & O’Neill, 2004; Pellegrino, Luca, & Bellusci, 2014). nent of plant ecosystems (Bonfante & Anca, 2009), and they Among these fungi, members of Tulasnellaceae (Basidiomycota) are improve plant growth by promoting nutrient uptake especially one of the richest and most widespread mycobionts forming when the availability of soil nutrients is low (Yang et al., 2016; van mycorrhizae with a broad spectrum of photosynthetic Orchidaceae der Heijden, Bardgett, & van Straalen, 2008). As orchids produce in temperate and tropical ecosystems (Dearnaley, 2007; Dearnaley, tiny seeds lacking stored nutrients, they require an obligate Martos, & Selosse, 2012; Jacquemyn, Waud, Merckx, Lievens, & Brys, 2015a; Martos et al., 2012; Suarez et al., 2006). As Tulasnel- laceae mycobionts are widespread on rotten wood and humus-rich substrates (Roche et al., 2010; Cruz, Suarez, Kottke, Piepenbring, & * Corresponding author. Departamento de Ciencias Biologicas, Universidad Oberwinkler, 2011, 2014), they are available to associate with or- Tecnica Particular de Loja, San Cayetano Alto s/n C.P, 11 01 608, Loja, Ecuador. E-mail address: [email protected] (P. Herrera). chids in terrestrial and epiphytic habitats of the tropics. https://doi.org/10.1016/j.myc.2017.08.003 1340-3540/© 2017 The Mycological Society of Japan. Published by Elsevier B.V. All rights reserved. P. Herrera et al. / Mycoscience 59 (2018) 38e48 39

For many years, there has been debate on whether individual Hülber, & Hietz, 2009). This kind of asymmetric abundances is orchid species depend on a specific or broad spectrum of myco- characteristic of extremely species-rich, mutualistically dependent bionts (Cozzolino & Widmer, 2006; Dearnaley, 2007; Shefferson communities (Medan et al., 2007). Orchid species richness in the et al., 2007; Warcup, 1981; Waud, Busschaert, Lievens, & tropical area may be supported by quite distinct habitats in the Jacquemyn, 2016). McCormick and Jacquemyn (2014) found spe- neighborhood such as pristine forests, young regenerating forests cific orchid-mycobiont relationships with distinct mycorrhizal and frequent disturbance by landslides combined with a strong communities of coexisting orchids in temperate grasslands. Simi- altitudinal gradient (Beck, Bendix, Kottke, Makeschin, & Mosandl, larly, Jacquemyn, Brys, Waud, Busschaert, and Liebens (2015b) 2008). However, we are lacking the corresponding information on found that groups of phylogenetically related orchid species were orchid mycobionts. associated with specific groups of Tulasnellaceae in species-rich The main objective of this study was to assess the richness of Mediterranean grasslands. When single orchid genera were stud- Tulasnellaceae orchid mycobionts and the degree of specialization ied in megadiverse tropical systems, a high degree of specialism of or generalism in orchid-Tulasnellaceae interactions in the orchid-Tulasnellaceae was also found (Otero, Flanagan, Herre, montane rainforest area of Southern Andes, Ecuador. We specif- Ackerman, & Bayman, 2004, 2007; Riofrío, Naranjo, Iriondo, & ically aimed to answer the following questions: (1) What is the Torres, 2007). In contrast, pioneer studies at the community level richness of Tulasnellaceae mycobionts associated to epiphytic and in biodiversity hotspots of tropical systems provide a different terrestrial orchids in the tropical montane rainforest? (2) Does picture. For instance, Martos et al. (2012) found a broad sharing of Tulasnellaceae richness and abundance differ among orchid life- mycobionts among highly diverse orchid species in rainforests and forms (terrestrial versus epiphytic) or among sites (pristine for- sub-alpine heathlands from La Reunion Island. Similarly, multiple est, regenerating forest and regenerating landslide site)? (3) Are epiphytic orchid species in Neotropical cloud forests were associ- there any indications of generalism or specificpartnershipsbe- ated with a high number of Tulasnellaceae and Serendipitaceae tween the species included in the mutualistic network? (4) Is the (previously Sebacinales subgroup B) (Suarez, Weiss, Abele, Tulasnellaceae-orchids network robust against the loss of Tulas- Oberwinkler, & Kottke, 2006, 2008, 2016; Kottke et al., 2013; nellaceae mycobiont species? If so, does it differ between Weiss, Waller, Zuccaro, & Selosse, 2016). epiphytic and terrestrial orchids? The study of mycobiont-orchid interactions based on the network theory (Bascompte & Jordano, 2007) is gaining mo- 2. Materials and methods mentum in ecological research on complex and megadiverse sys- tems (Jacquemyn et al., 2015b; Kottke et al., 2013; Martos et al., 2.1. Study sites 2012). Understanding the processes that structure these networks will help explain if and how hyperdiverse sets of interacting species This study was carried out in the Reserva Biologica San Francisco coevolve and are maintained in their habitats (Jordano, Vazquez, & (RBSF) located half -way between Loja and Zamora, Zamora- Bacompte, 2009). This network approach provides a basic Chinchipe Province, Ecuador (3º5801700S, 79º0404300W), adjacent to description of the general pattern of assembly and interactions Podocarpus National Park on the steep eastern slope of the (Jordano, Bascompte, & Olesen, 2003), such as nestedness Cordillera El Consuelo, Southern Ecuadorian Andes. RBSF spans (Bascompte, Jordano, Melian, & Olesen, 2003) or modularity about 1000 ha over an altitudinal gradient of 1500 me3400 m (Olesen, Bascompte, Dupont, & Jordano, 2007). Significant nested- above sea level (a.s.l.). Permanent plots were established in 1997, ness and modularity may indicate the importance of niche-based and since then the montane rainforest has been comprehensively processes on community assembly, affecting the diversity, stabil- studied (Beck et al., 2008; Bendix et al., 2013), including the in- ity and coevolutionary dynamics between symbiotic species ventorying of mycorrhizal fungi (Haug, Setaro, & Suarez, 2013; (Chagnon, Bradley, & Klironomos, 2012). For example, the network Kottke et al., 2013, 2008). The evergreen forest is mostly approach revealed contrasting nestedness and modularity of the composed of undisturbed, partly regenerating natural sites (Brehm orchid-symbiont relationship between a temperate system et al., 2008; Weber, Günter, Aguirre, Stimm, & Mosandl, 2008). (Jacquemyn et al., 2011) versus a tropical (Martos et al., 2012) and a Average annual precipitation is about 2200 mm and mean tem- Mediterranean (Jacquemyn et al., 2015b) system. While the perature is 15.5 C at 1900e2170 m a.s.l. (Bendix, Rollenbeck, temperate European Orchis spp. and their mycobionts formed a Richter, Fabian, & Emck, 2008). Approximately 280 tree species, significantly nested network, the Mediterranean and La Reunion including 52 angiosperms and 1 gymnosperm, have been identified Island networks were highly modular, epiphytic and terrestrial so far, and most of them are rare in the area (Homeier, Breckle, associations forming distinct modules in La Reunion. In a small area Günter, Rollenbeck, & Leuschner, 2010). Forest cover is close to of a tropical mountain rainforest in Southern Ecuador, Kottke et al. 100% in the pristine forest and 73% in the 40-y-old regenerating (2013) found a significantly nested network architecture among the forest included in this study. The relatively thin tree stems highly diverse orchids and 3 groups of diverse mycorrhizal fungi. (5e25 cm diam) act as phorophytes for many orchids and other However, they did not consider the effect of habitats, life forms or epiphytes, which is characteristic of the cloud forest (Homeier et al., sites on the architecture of the fungi-orchids network based on 2013; Riofrío et al., 2007). The altitudinal gradient and the steep- Tulasnellaceae as mycobionts. ness of the forest-covered slopes further promote the frequent Orchidaceae are extraordinarily speciose and abundant, and landslides caused by human activities like road construction, form an essential part of the tropical mountain rainforest in the leading to high species turnover rates (Homeier, Werner, Gradstein, Southern Andes in Ecuador (Homeier & Werner, 2007). This area is Breckle, & Richter, 2008; Oesker, Daliz, Günter, Homeier, & Matezki, considered a hotspot of plant biodiversity (Barthlott, Mutke, 2008). Rafiqpoor, Kier, & Kreft, 2005; Brehm et al., 2008), but the habitat of Orchidaceae were sampled at 4 sites which had is constantly threatened by anthropogenic activities, making its different elevations and disturbance, from pristine forests to a preservation a priority in global conservation. Most of its orchid landslide area. Sites 1 and 4 were located in pristine forests at 2000 species are endemic and rare (Homeier & Werner, 2007). Only 2 to m and 2170 m a.s.l., respectively. Site 3 was located in a 40-y-old 5 species are abundant (>2000 individuals in 0.1 ha), while most regenerating forest at 2170 m a.s.l. close to Site 4, and Site 2 was others are doubletons or singletons (Kottke et al., 2013). However, located in an area at 1900 m a.s.l. where a human-caused landslide many species may still remain undescribed (Endara, 2000; Winkler, occurred 40 years ago (Supplementary Figs. S1 and S2). 40 P. Herrera et al. / Mycoscience 59 (2018) 38e48

The 40-y-old regenerating forest is a product of anthropogenic database (https://www.ncbi.nlm.nih.gov) using the option mega- deforestation (Weber et al., 2008) and is structurally characterized blast (Altschul et al., 1997). Only sequences of our strains similar to by a lower proportion of trees (>10 cm diam at breast height (dbh)) Tulasnellaceae species were aligned using the G-INS-I option in and a greater proportion of bushes and tree ferns, thus appearing MAFFT Ver. 6.602b (Katoh, Kuma, Toh, & Miyata, 2005). Alignments less homogeneous and more open than the primary forest sur- were checked to eliminate chimeric sequences using the software rounding it (Martinez, Mahecha, Lischeid, & Beck, 2008). The Bellerophon (Huber, Faulkner, & Hugenholtz, 2004) and also human-caused landslide produced a large gap which has vegeta- manually copying and comparing suspicious parts of sequences on tion in a young succession stage and is characterized by the growth BLAST (<200 bp by part). of pioneer species that are not found in the primary forest A phylogenetic analysis was performed using the entire nrDNA (Bussmann, Wilcke, & Richter, 2008). Soil properties and fertility in ITS-5.8 S region. Due to the known difficulties in aligning Tulas- landslide areas can vary considerably, providing less favorable nellaceae sequences (Cruz, Suarez, Kottke, & Piepenbring, 2014; conditions for plant growth than undisturbed soils (Bussmann Suarez et al., 2006), we carried out a two-step procedure. Firstly, et al., 2008). Ecological factors such as the amount of solar radia- a neighbour-joining analysis was performed in PAUP* 4.0b10 tion, minimum and maximum temperatures at the forest ground (Swofford, 2002) using the BioNJ modification with Kimura 2-dis- and rainfall are higher in the regenerating forest and the landslide tances of the sole 5.8 S region (160 bp). The resulting tree displayed site than in the primary forest. In contrast, amount of litter fall, seed 4 prominent groups (Groups IeIV), and new alignments of the quantity and seed diversity are lower in these two habitats. whole nrDNA ITS-5.8 S region were performed for each group Permanent plots of 1 m2 were randomly established at least (Group I ¼ 604 bp, group II ¼ 611 bp, group III ¼ 576 bp and group 10 m apart at each site. Eight plots at each site were terrestrial and 8 IV ¼ 927 bp). Phylogenetic reconstructions were carried out using a plots were on tree stems (1 m in length, between 1 and 2 m above maximum likelihood (ML) analysis for each group in MEGA5 ground), except at Site 2 where only terrestrial plots could be (Tamura et al., 2011) under the general time reversible DNA sub- established as there were no trees. Each plot included at least 5 stitution model with 1000 bootstrap replicates, eliminating gaps orchid individuals of presumably different species. In 2008 the and missing data. Additionally, a maximum parsimony (MP) anal- plots were visited several times throughout the flowering period to ysis was performed for each group in MEGA5, using the close- collect orchid samples to identify them at the species level. neighbour-interchange (CNI) on random trees method with 1000 bootstrap replicates. All trees were midpoint-rooted. 2.2. Mycorrhizae sampling, fungal DNA isolation, PCR, cloning and Operational taxonomic units (OTUs) were defined based on sequencing sequence similarity of the nrDNA ITS-5.8 S region. A table of p- distance was calculated in PAUP 4.0b10 and sequences were Roots were sampled from 2 individuals of randomly selected grouped with OPTSIL (Goker,€ García-Blazquez, & Voglmayr, 2009) orchids at all plots except for two, where 3 individuals were at a 95% and 97% threshold of sequence similarity and 0.5 of linkage sampled. Three root samples were collected per individual from a fraction. For further analysis, we chose the 95% threshold, which total of 114 orchid individuals. Roots considered to be mycorrhizal was calibrated using molecular and morphological data of Tulas- based on their appearance (color and health status) were collected nellaceae by Cruz et al. (2011, 2014). These OTUs were mapped into from the pure humus layer in terrestrial orchids or in direct contact the ML trees. with the bark of tree stems for epiphytic orchids. At the landslide site, mycorrhizal roots were collected from the mineral soil. Sam- 2.4. Morphological criteria and DNA barcoding for orchid ples were wrapped in aluminum foil, brought to the laboratory, and identification further handled within 24 h. Fungal colonization in root cortical tissue was verified by microscopic observation of fungal pelotons in The orchids in bloom at the time of sampling were subjected to thin sections stained by 0.05% methyl-blue (C.I. 42780) (Merck, morphological identification referring to Dodson et al. (unpub- Darmstadt, Germany) in lactic acid using an Axiostar plus micro- lished work), Dodson and Escobar (1993), Dodson (2001, 2002, scope (Carl Zeiss, Gottingen,€ Germany). Genomic DNA was extrac- 2003, 2004) and available web resources (e.g., http://www. ted from 2 cm segments of well-colonized roots using a DNeasy pleurothallids.com). Plant Mini Kit (Qiagen, Hilden, Germany) following the manufac- Young leaves were collected from each individual sampled for turers’ instructions. The internal transcribed spacers 1 and 2, mycorrhizas. DNA was extracted from fresh or dry (60 C overnight) including the 5.8 S region (nrDNA ITS-5.8 S) were targeted for PCR leaf segments using the DNeasy Plant Mini Kit following the amplification using the universal eukaryotic primers ITS1 (50- manufacturer's instructions. To identify orchids, we amplified 2 TCCGTAGGTGAACCTGCGG-30)(White, Bruns, Lee, & Taylor, 1990) genes, i.e., the commonly used matK (Pridgeon, Solano, & Chase, and TW14 (50-GCTATCCTGAGGGAAACTTC-30)(Cullings, 1994) with 2001; Pridgeon & Chase, 2003; van den Berg et al., 2005) and the Phusion High-Fidelity PCR Mastermix (Finnzymes, Espoo, ycf1, which according to Neubig et al. (2009) is phylogenetically Finland). PCR products were cloned with the Zero Blunt TOPO PCR informative at the species level in orchids. We also tried to amplify Cloning Kit (Invitrogen, Carlsbad, USA). Eight positive clones the nrDNA ITS-5.8 S gene but were unable due to the poor quality inserted with the targeted sequences were grown in liquid LB of the sequences. Firstly, the plastid gene matK was amplified us- Broth, Miller (Difco, Detroit, USA), purified with S.N.A.P. miniprep ing 2 primer combinations, i.e., 19F (50-CGTTCTCATATTGCACTATG- kit (Invitrogen, Carlsbad, USA) and sequenced bi-directionally with 30)(Whitten, Williams, & Chase, 2000) and 881R (50-TMTTCATCA- ABI 3730xl in Macrogen (Seoul, Korea) using universal primers GAATAAGAGT-30)(Pridgeon et al., 2001), and 731F (50- M13F and M13R (For further details see Kottke et al., 2010). TCTGGAGTCTTTCTTGAGCGA-30)(Pridgeon et al., 2001) and trnK-2R (50-AACTAGTCGGATGGAGTAG-30)(Johnson & Soltis, 1995). PCR 2.3. Fungal phylogenetic analysis and OTU delimitation conditions were as follows: initial denaturation at 98 C for 30 s followed by 30 cycles of 98 C for 10 s, 52 C for 20 s, and 72 C for Sequences were edited and assembled into a consensus 30 s and then a final elongation step at 72 C for 7 min. An sequence using Sequencher 4.6 (Gene Codes, Ann Arbor, USA). annealing temperature of 58 C was applied when the 731F and Homology searches for each nucleotide sequence were carried out trnK-2R primer combination was used. The PCR reaction volume with BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) in GenBank was 25 mL with concentrations of 0.2 pmol of each primer, 10% P. Herrera et al. / Mycoscience 59 (2018) 38e48 41 bovine serum albumin (BSA) (Sigma, St. Louis, USA) with a final partition with the largest modularity (Haug et al., 2013). The concentration of 0.8 mg/mL(Iotti & Zambonelli, 2006) and 1U Phu- analysis was based on the interaction of each orchid individual sion High-Fidelity PCR Mastermix. A negative control included PCR with individual Tulasnellaceae OTUs. We used an iteration factor mix without DNA template. of 1.0, a cooling factor of 0.995 and 1000 randomizations. The The 30 end of ycf1 plastid gene was amplified using the primers significance of modularity was evaluated by calculating a Z score 3720F (50-TACGTATGTAATGAACGAATGG-30) and 5500R (50- obtained from the modularity of the orchid-fungal network minus GCTGTTATTGGCATCAAACCAATAGCG-30)(Neubig et al., 2009). PCR the mean modularity of the randomizations (M rand) divided by conditions were as follows: initial denaturation at 98 C for 3 min the standard deviation of the randomizations (sigma M rand) followed by 35 cycles of 98 C for 30 s, 58 C for 1 min, and 72 C for (Haug et al., 2013). The Z score was converted to a p-value using 3 min, ending with a final elongation step at 72 C for 3 min. PCR the free online calculator QuickCalcs (GraphPad Software; http:// products were sequenced bi-directionally by ABI 3730xl in Mac- www.graphpad.com/quickcalcs/). rogen using the same primers as for PCR amplification. Sequence chromatograms were edited using the software Sequencher 4.6. 2.7. Evaluating the robustness of the mutualistic network to species BLAST search (Altschul et al., 1997) was used to find published loss sequences with a high similarity to the named orchid species. As most of our orchid species have not been sequenced before, naming To evaluate the robustness of the bipartite orchid-Tulasnellaceae was generally only successful at the level (Supplementary network to species loss, we applied the index R50, which indicates Table S1). the proportion of symbiotic fungi that should become extinct to cause a secondary extinction of 50% of plants (Dunne, Williams, & 2.5. Estimating richness and evaluating similarity of Tulasnellaceae Martinez, 2002). The relative robustness of orchids to - OTUs across orchid life forms and sites ceae OTUs extinction was evaluated for the complete network using the function second. extinct (modified by Santamaría, Galeano, Potential richness and inventory completeness of Tulasnellaceae Pastor, & Mendez, 2014) in the software R. Three scenarios of OTUs were evaluated using the program EstimateS Ver. 8.2.0 species extinction were simulated: (1) random extinction, (2) or- (Colwell et al., 2012). Individual-based species accumulation curves dered extinction from the most-to the least-linked species, and (3) were calculated for the epiphytic and terrestrial life-form orchids, ordered extinction from the least-to the most-linked species. We as well as for the sites. The obtained curves and the curves also partitioned the network into groups to assess the relative describing their 95% confidence intervals were fitted to a Clench robustness of epiphytic or terrestrial orchids to Tulasnellaceae curve (Soberon & Llorente, 1993) using Statistica Ver. 7.0.61.0 OTUs extinction, following Santamaría et al. (2014). The possibility (StatSoft, Tulsa, Oklahoma, USA) following Jimenez and Hortal of secondary extinction of symbiotic fungi in response to the pri- (2003). The asymptote of each curve was calculated as a/b, where mary extinction of orchids was not simulated due to the asymmetry a and b are the two parameters of the Clench curve. The sampling in dependence in this mutualistic system. Based on the bipartite effort needed to record 95% of the estimated proportion of OTUs pollination networks suggested by Santamaría et al. (2014),R50 was calculated according to Jimenez and Hortal (2003). ranged from 1 (maximum robustness) to 1 divided by the number Similarity in the composition of Tulasnellaceae OTUs between of pollinators or, in our case, Tulasnellaceae fungi (minimum epiphytic and terrestrial orchids and across sites was evaluated by robustness to partner extinction). Chao-Jaccard and Chao-Sørensen similarity indices based on inci- Differences in the robustness of the network to the simulated dence and abundance (Chao, Chazdon, Colwell, & Shen, 2005) using Tulasnellaceae extinction scenarios were evaluated in R with a one- the program EstimateS Ver. 8.2.0. These indices account for shared, way general linear model (GLM) using the 300 iterations as repli- but unobserved species among samples. cates. The orchid life forms (epiphytic and terrestrial orchids) were considered factors in the GLM analyses. Significance of the differ- 2.6. Description of the mutualistic network ences in robustness was calculated by means of a posteriori Tukey's honesty significant difference test according to Santamaría et al. Qualitative interactions based on presence/absence data be- (2014). tween orchid species and Tulasnellaceae OTUs were assembled in an ordered binary matrix (Fig. 1). Degree distributions, i.e., the 3. Results number of links per Tulasnellaceae OTU and orchid species, were displayed and network connectance was calculated to further 3.1. Fungi sampling characterize the structure of the mutualistic network using R software Ver. 3.1.1 (R Core Team, Vienna, Austria) with the 'bipar- All collected roots were colonized by mycobionts. Staining tite' package (function plotweb for visualization and function net- showed coils of young hyphae and structures that denoted worklevel for connectance values) (Dormann, Gruber, & Fründ, collapsed or degenerating pelotons in the cortical tissue. DNA was 2008). To determine if the structure of the orchid-Tulasnellaceae isolated and sequenced from the colonized roots of all 114 sampled network was significantly nested, a formal nestedness index was orchid individuals. A total of 418 mycorrhizal fungal DNA sequences calculated as NODF (nested overlap and decreasing fill), and its were obtained from the complete nrDNA ITS-5.8 S and partial significance was tested against 2 null models with 1000 randomi- nrDNA 28 S region (approximately 1500 bp length). The most zations (ER model assigning interactions completely at random and frequently amplified mycorrhizal fungi belonged to members of CE model randomizing each species similar to the observed data Tulasnellaceae (310 sequences) and were obtained from 89 orchid matrix), implemented in ANINHADO Ver. 3.0 Bangu (Almeida-Neto, individuals (4.13 sequences on average per orchid species). Mem- Guimaraes,~ Guimaraes,~ Loyola, & Ulrich, 2008; Guimaraes~ & bers of Atractiellales (3 OTUs) and Serendipitaceae (12 OTUs) were Guimaraes,~ 2006). Significant nestedness would indicate a non- also obtained from 32 to 19 orchid individuals, respectively. Ser- random assemblage of the network. endipitaceae and Atractiellales sequences were removed from Modularity as a descriptor of network structure was tested further analyses in this study. A total of 100 of the original 310 using the program NETCARTO (Guimera & Amaral, 2005), which Tulasnellaceae sequences were discarded due to poor sequence assesses compartments by simulated annealing to find the quality, inappropriate length or chimera formation (14 chimeras 42 P. Herrera et al. / Mycoscience 59 (2018) 38e48

Fig. 1. Binary matrix presenting the relationship between the studied Tulasnellaceae and orchids. The 29 Tulasnellaceae OTUs (rows) and 52 orchid species (columns) are ordered by decreasing number of interacting partners (bars and numbers to the left and below the matrix). The grayscale check marks the number of interactions: one (light grey); two (dark grey); three or more (black); no interaction (white). detected). Thus, we used 210 high-quality Tulasnellaceae se- morphological identification of the orchids at the generic level. A quences for phylogenetic analysis and OTU definition. All sequences total of 89 orchid individuals containing Tulasnellaceae as myco- were deposited in GenBank (accession numbers are shown in bionts were delimitated to 52 orchid species by matK and ycf1 Supplementary Table S1). barcoding and morphotyping. A few species could be named when they were in flower or when sequences were available, but most of 3.2. Phylogenetic analysis and OTU delimitation of Tulasnellaceae these species were not successfully sequenced. Orchids belonged to 9 genera (Supplementary Table S1) in subfamily Epiden- The neighbour-joining analysis of the sole 5.8 S region obtained droideae. Stelis (18 spp.), Pleurothallis (13 spp.), Maxillaria (9 spp.), 4 well-supported bootstrap groups (Groups IeIV). Group I had the Epidendrum (5 spp.) and Elleanthus (3 spp.) were represented by highest number of sequences with a total of 78 sequences, while multiple species, while Artorima, Prosthechea, Sobralia and Groups II, III and IV had 56, 32 and 44 sequences, respectively. Oncidium were only represented by a single species each. Indi- Alignments of the complete nrDNA ITS-5.8 S region within these 4 vidual species were mostly present as singletons (25 spp.), a few groups resulted in well-resolved clades with high support values as doubletons (12 spp.) and a few were encountered 3 or more for ML and MP analyses (Supplementary Figs. S3eS6). When a times (15 spp.) (Fig. 1). fraction linkage of 0.5 was used, we obtained 29 OTUs based on a Seven orchid species were found at the landslide plots, while threshold of 95% sequence similarity and 33 OTUs were 45 species were found on the 48 epiphytic and terrestrial forest obtained at a threshold of 97%. The OTUs obtained at the 95% plots. Species composition was quite different between the threshold are mapped in the 4 phylogenetic trees of Tulasnellaceae landslide and the forest sites as well as between terrestrials and (Supplementary Figs. S3eS6). As the results of ML and MP analyses epiphytes. Artorima, Prosthechea, Sobralia, Epidendrum sp. 2, were consistent with the monophyly of most of the 29 delimited Maxillaria sp. 4 and Maxillaria sp. 7 only occurred at the landslide OTUs (Supplementary Figs. S3eS6), OTU delimitation was phylo- site, while the Pleurothallidinae genera Pleurothallis and Stelis genetically supported. were restricted to the forest and were the most commonly sampled individuals in the pristine (Sites 1 and 4) and in the 3.3. Orchid identity and occurrence regenerating forest (Site 3). Maxillaria sp. 5 was present both in the forest and landslide plots. Only Maxillaria sp. 22, Pleurothallis High-quality sequences were obtained for 87 of 114 sampled coriacardia and 4 species of Stelis (S. superbiens and Stelis sp. 3, 11 orchid individuals which were BLAST-searched to confirm the and 30) were found in both epiphytic and terrestrial forest P. Herrera et al. / Mycoscience 59 (2018) 38e48 43 habitats (Supplementary Table S1). sites (Fig. 2; Supplementary Table S2). Richness of Tulasnellaceae OTUs was considered significantly different across life forms or 3.4. Richness and similarity of Tulasnellaceae OTUs across life forms sites when the 95% confidence intervals of the corresponding and sites accumulation curves did not overlap. Broad overlap between the 95% confidence intervals of the accumulation curves indicated no The higher number of Tulasnellaceae OTUs occurred in terres- differences in OTU richness across habitats or sites (Fig. 2; trial orchids (Fig. 2; Supplementary Tables S2) and 10 OTUs were Supplementary Table S2)(Ellison & Gotelli, 2004). Chao-Jaccard common to both life forms. The accumulation curves of Tulas- and Chao-Sørensen indices showed high similarity in the nellaceae OTUs in epiphytic and terrestrial orchids indicated an composition of Tulasnellaceae OTUs between epiphytic and inventory completeness of 66% and from 38 to 71% at the different terrestrial habitats (Table 1). However, similarity was lower

Fig. 2. Accumulation curves of Tulasnellaceae OTUs obtained by accumulating orchid samples across (A) orchid growth forms (epiphytic, dashed line; terrestrial, continuous line) or (B) sites (Site 1, continuous black line; Site 2, dashed black; Site 3, continuous grey; Site 4, dashed grey lines), of a tropical mountain rain forest located in the Reserva Biologica San Francisco, Zamora-Chinchipe Province, Southern Ecuador. Dots show the asymptotic values for each accumulation curve and vertical bars indicate their confidence intervals at 95% (with the same color and type of line). 44 P. Herrera et al. / Mycoscience 59 (2018) 38e48

Table 1 Similarity indices of Tulasnellaceae OTUs at 95% similarity of the epiphytic and terrestrial orchid communities and at the different sites located in the Reserva Biologica San Francisco, Southern Ecuador.

Comparison Incidence-based Abundance-based

Chao-Jaccard Chao-Sørensen Chao-Jaccard Chao-Sørensen

Epiphytic versus terrestrial 0.74 0.85 0.74 0.85 Pristine versus landslide 0.5 0.66 0.5 0.67 Pristine versus regenerating 1 1 1 1 Landslide versus regenerating 0.33 0.49 0.33 0.49

between the landslide (Site 2) and the regenerating forest (Site 3) robustness of epiphytic and terrestrial orchids to Tulasnellaceae (Table 1). extinction showed R50 values between 0.32 and 1.00 (Fig. 3). One- way GLM showed a significant effect of orchid life form on 3.5. Structure of the orchid-Tulasnellaceae network and stability robustness to Tulasnellaceae OTUs extinction (Fig. 3). In the random against species loss extinction sequence, terrestrial orchids were more robust than epiphytic orchids, while the opposite was found in the most-to The bipartite network formed by the 29 Tulasnellaceae OTUs least-connected sequence (Fig. 3). and the 52 orchid species (Supplementary Figs. S7 and S8) had a connectance of 7.2% and was significantly nested (NODF ¼ 16.2; 4. Discussion p < 0.05 for both ER and CE null-models). NODF values were 18.2 for orchids and 9.9 for Tulasnellaceae. The number of singletons and We found that the interactions between orchids and Tulasnel- doubletons was high for Tulasnellaceae and orchids (55% and 71%, laceae fungi in the tropical forest of Southern Ecuador were char- respectively) (Fig. 1). The network had a few generalist fungi (OTUs acterized by a high richness of Tulasnellaceae, shared between T1, T9, T13, T16) frequently linked with a few common orchid epiphytic and terrestrial orchids across sites (landslide, pristine and species (S. superbiens, Stelis sp. 11, Pleurothallis sp. 25). OTU T1 was regenerating forest). The structure of this mutualistic network was the most common phylogenetic species associated with 26 in- nested, and its robustness differed only slightly between the orchid dividuals and 19 orchid species, followed by 3 others frequently life-forms (epiphytic or terrestrial) of this network. Tulasnella linked with OTUs T16, T13 and T9 (Fig. 1). Fungi and orchids with richness in European temperate habitats has reached asymptotic few links were mainly linked to partners with many links, with one values near 10e12 OTUs (Jacquemyn & Deja, 2012; Jacquemyn et al., exception (T21 linked to Maxillaria sp. 7). No significant modularity 2011). Not surprisingly, this diversity is higher in tropical forests, as was found for the bipartite network (M ¼ 0.6, p ¼ 0.9). indicated by previous studies in La Reunion (58 OTUs) (Martos Index R50, indicating the robustness of the network to Tulas- et al., 2012) and the preliminary results for our study area (33 nellaceae OTUs extinction, ranged from 0.45 when the loss of OTUs at 3% sequence divergence in Kottke et al., 2013). Here, we Tulasnellaceae proceeded from the most-to the least-connected, to obtained 29 Tulasnella OTUs, and the non-saturated accumulation 0.75 when extinction was random and 0.93 when extinction pro- of fungal OTUs suggests that more locally-distributed rare species ceeded from the least-to most-linked Tulasnellaceae. The may be present at our study sites.

Fig. 3. Mean robustness (R50) to Tulasnellaceae extinction under scenarios of extinction at random, most to least connected and least to most connected. 300 interactions were used in all cases. Robustness is indicated by circles for epiphytic orchids and squares for terrestrial orchids. Error bars indicate standard deviation. Chi-squared (c2) and significance (P) of a one-way general linear model (GLM) comparing differences in robustness between the groups of orchids to the extinction of Tulasnellaceae are shown at the top of each simulation strategy. P. Herrera et al. / Mycoscience 59 (2018) 38e48 45

If differences in sampling effort among studies are not consid- random and the most-to least-linked extinction scenarios. This ered, Tulasnella richness may depend on the threshold (95% in this pattern has also been described in pollination networks (Burgos study, 97% in Martos et al., 2012) of sequence similarity for OTU et al., 2007; Memmott, Waser, & Price, 2004; Santamaría et al., delimitation. Thresholds of 95e97% sequence similarity (of ITS re- 2014). In our network, this reveals that generalist fungi and or- gion) have been widely used to delimitate Tulasnellaceae species chids play an important role in the stability of the mutualistic (Cruz et al., 2014, 2011; Girlanda et al., 2011; Jacquemyn & Deja, system. Terrestrial and epiphytic orchids showed similar robust- 2012, 2014; Jacquemyn et al., 2011; Linde et al., 2017). Further- ness to the extinction of their mycobionts, and although differences more, Linde, Phillips, Crisp, and Peakall (2014) supported the use of were statistically significant, the effect was minor. This means that our chosen threshold as well as the use of the ITS region to mycobionts are equally important for the persistence of both sub- congruently delimitate species within Tulasnellaceae based on an sets of orchids. A broad generalization of interactions is important extensive multi-locus analysis. However, even when a more con- for network robustness (Pocock, Evans, & Memmott, 2012). Loss of servative threshold (99%) was used to delimitate OTUs, results generalist keystone species (most-connected species) may drive remained consistent with our main conclusions (Supplementary cascades of biodiversity loss. Furthermore, high specificity is un- Table S3). On the other hand, Tulasnella richness may also depend likely in mutualistic interactions, as species that can associate with on the primers selected to amplify fungal DNA. To increase the a greater variety of species have a wider range of ecological options possibility of identifying all fungi diversity in the root, we used a for survival (Timms & Read, 1999). pair of universal eukaryotic primers (ITS1 and TW14) which target Our results indicate a high diversity of Tulasnellaceae myco- highly-conserved sites in the SSU and LSU, respectively (Lindahl bionts and at least a core of generalist Tulasnellaceae OTUs. Irreg- et al., 2013). Although we detected a broad diversity of Tulasnel- ular distribution of mycorhizal fungi is considered a key factor for laceae, other primer combination methods could possibly have the persistence of small orchid population sizes as found in tropical detected additional species (Jacquemyn & Deja, 2012, 2015b; forests (Otero & Flanagan, 2006). Our findings suggest that the Jacquemyn et al., 2011; Waud et al., 2016). We suggest using mul- availability of common mycorhizal fungi supports the persistence tiple primer alternatives in future studies to detect a more complete of a wide range of orchids in the area, by at least two mechanisms: diversity of Tulasnellaceae. (1) simultaneous association with several mycobionts may maxi- Tulasnellaceae communities in the study area were similar mize nutrient absorption, especially for epiphytic habitats where across habitats despite a high turnover of orchid species (Werner, nutrients are limited (Jacquemyn et al., 2010), and (2) obligate 2011; Werner, Homeier, & Gradstein, 2005). One may argue that mutualisms are stabilized by the simultaneous association with pooling terrestrial orchids (from the humus layer and in mineral facultative mutualists (Takimoto & Suzuki, 2016). These conditions soil) in one group could potentially hide differences, but these two of Tulasnellaceae may support the development and coexistence of kinds of substrates commonly have very similar components in the high orchid richness in the tropical montane rainforest of tropical forests. In fact, orchids at the landslide site and the Southern Ecuador. terrestrial orchids in the pristine forest shared 7 of the 9 Tulas- nellaceae OTUs, including the most frequent ones. These results Acknowledgments were unexpected and contrast with the findings for La Reunion (Martos et al., 2012), where significant differences were found in This research was funded by the Deutsche For- mycobiont composition between epiphytic and terrestrial orchids. schungsgemeinschaft (DFG RU 816). We thank Nature and Culture These contrasting findings may be due to several factors, such as International (NCI) for providing research facilities at the Reserva climatic and edaphic niche differences (Beck & Richter, 2008), the Biologica San Francisco in Ecuador. We would also like to thank phylogenetic and biogeographic history of these regions, or simply Alberto Mendoza for his help in orchid identification, Silvia San- to different sampling strategies (plots versus -based tamaría for sharing the R code for the robustness analysis and Lori sampling, and small areas versus large landscapes). De Hond for correcting the English text. The use of complex network tools in our analysis allowed us to disentangle the structure of this orchid-mycobiont system domi- Appendix A. Supplementary data nated by rare species. The network showed significant nestedness but no modularity, contrary to findings in another tropical system Supplementary data related to this article can be found at (Martos et al., 2012). A nested structure, which is widespread in https://doi.org/10.1016/j.myc.2017.08.003. pollinator and seed-disperser webs (Bascompte & Jordano, 2007), is of fundamental importance for community formation and biodi- References versity maintenance (Bastolla et al., 2009; Gomez, Perfectti, & ~ ~ Jordano, 2011). A significantly nested mutualistic network in- Almeida-Neto, M., Guimaraes, P., Guimaraes, P. R., Loyola, R. D., & Ulrich, W. (2008). A consistent metric for nestedness analysis in ecological systems: Reconciling dicates that rare orchids associate with common mycobionts and concept and measurement. Oikos, 117, 1227e1239. https://doi.org/10.1111/ rare mycobionts associate with generalist orchids (Bascompte, j.0030-1299.2008.16644.x. Jordano, & Olesen, 2003, 2006; Okuyama & Holland, 2008), Altschul, S. F., Madden, T. L., Schaffer,€ A. A., Zhang, J., Zhang, Z., Miller, W., et al. (1997). Gapped BLAST and PSI-BLAST: A new generation of protein database resulting in marked asymmetry of interactions (Bascompte et al., search programs. Nucleic Acids Research, 25,3389e3402. https://doi.org/ 2003). A nested structure also means that there is a core of 10.1093/nar/25.17.3389. generalist orchids and mycobionts, as in other non-tropical Barthlott, W., Mutke, J., Rafiqpoor, D., Kier, G., & Kreft, H. (2005). Global centers of e mycorrhizal interactions (Valyi, Rillig, & Hempel, 2015). Speciali- diversity. Nova Acta Leopoldina, 92,61 83. Bascompte, J., & Jordano, P. (2007). Plant-animal mutualistic networks: The archi- zation can sometimes be overestimated by low sampling effort, tecture of biodiversity. Annual Review of Ecology, Evolution, and Systematics, 38, particularly in small ecological networks (Fründ, Mccann, & 567e593. https://doi.org/10.1146/annurev.ecolsys.38.091206.095818. Williams, 2016). Although our network was undersampled and Bascompte, J., Jordano, P., Melian, C. J., & Olesen, J. M. (2003). The nested assembly of plant animal mutualistic networks. Proceedings of the National Academy of Sci- caution is needed in interpreting these results, we consider that the ences of the United States of America, 100, 9383e9387. https://doi.org/10.1073/ nested structure is reliable because this network feature has been pnas.1633576100. found to be robust to undersampling (Nielsen & Bascompte, 2007). Bascompte, J., Jordano, P., & Olesen, J. M. (2006). Asymmetric coevolutionary net- works facilitate biodiversity maintenance. Science, 312,431e433. https:// In the Tulasnellaceae-orchids network, robustness progressively doi.org/10.1126/science.1123412. decreased from the least-to most-linked extinction scenario to the Bastolla, U., Fortuna, M. A., Pascual-García, A., Ferrera, A., Luque, B., & Bascompte, J. 46 P. Herrera et al. / Mycoscience 59 (2018) 38e48

(2009). The architecture of mutualistic networks minimizes competition and Endara, L. (2000). Orchidaceae. In R. Valencia, N. Pitman, S. Leon, & P. M. Jorgensen increases biodiversity. Nature, 458,1018e1020. https://doi.org/10.1038/ (Eds.), Libro rojo de las plantas endemicas del Ecuador. Publicaciones del Herbario nature07950. QCA (pp. 257e372). Quito: Pontificia Universidad Catolica del Ecuador. Beck, E., Bendix, J., Kottke, I., Makeschin, F., & Mosandl, R. (2008). Gradients in a Fortuna, M. A., Stouffer, D. B., Olesen, J. M., Jordano, P., Mouillot, D., Krasnov, B. R., tropical mountain ecosystem of Ecuador. Ecological Studies 198. Berlin: Springer et al. (2010). Nestedness versus modularity in ecological networks: Two sides of Verlag. the same coin? Journal of Animal Ecology, 79,811e817. https://doi.org/10.1111/ Beck, E., & Richter, M. (2008). Ecological aspects of a biodiversity hotspot in the j.1365-2656.2010.01688.x. Andes of southern Ecuador. In S. Gradstein, J. Homeier, & D. Gansert (Eds.), The Fründ, J., Mccann, K. S., & Williams, N. M. (2016). Sampling bias is a challenge for tropical mountain forest e patterns and processes in a biodiversity hotspot. quantifying specialization and network structure: Lessons from a quantitative Biodiversity and Ecology Series 2 (pp. 197e219). Gottingen:€ Universitatsverlag€ niche model. Oikos, 125, 502e513. https://doi.org/10.1111/oik.02256. Gottingen€ . Girlanda, M., Segreto, R., Cafasso, D., Liebel, H. T., Rodda, M., Ercole, E., et al. (2011). Bendix, J., Beck, E., Brauning,€ A., Makeschin, F., Mosandl, R., Scheu, S., et al. (2013). Photosynthetic Mediterranean meadow orchids feature partial mycohetero- Ecosystem services, biodiversity and environmental change in a tropical mountain trophy and specific mycorrhizal associations. American Journal of Botany, 98, ecosystem of South Ecuador. Ecological Studies 221. Berlin: Springer-Verlag. 1148e1163. https://doi.org/10.3732/ajb.1000486. Bendix, J., Rollenbeck, R., Richter, M., Fabian, P., & Emck, P. (2008). Climate. In Goker,€ M., García-Blazquez, G., & Voglmayr, H. (2009). Molecular taxonomy of E. Beck, J. Bendix, I. Kottke, F. Makeschin, & R. Mosandl (Eds.), Gradients in a phytopathogenic fungi: A case study in Peronospora. PLoS One, 4, e6319. https:// tropical mountain ecosystem of Ecuador. Ecological studies 198 (pp. 63e73). doi.org/10.1371/journal.pone.0006319. Berlin: Springer Verlag. Gomez, J. M., Perfectti, F., & Jordano, P. (2011). The functional consequences of van den Berg, C., Goldman, D., Freudenstein, J., Pridgeon, A., Cameron, K., & mutualistic network architecture. PLoS One, 6, e16143. https://doi.org/10.1371/ Chase, M. (2005). An overview of the phylogenetic relationships within Epi- journal.pone.0016143. dendroideae inferred from multiple DNA regions and recircumscription of Guimaraes,~ P. R., & Guimaraes,~ P. (2006). Improving the analyses of nestedness for and Arthuseae (Orchidaceae). American Journal of Botany, 92, large sets of matrices. Environmental Modelling and Software, 21,1512e1513. 613e624. https://doi.org/10.3732/ajb.92.4.613. https://doi.org/10.1016/j.envsoft.2006.04.002. Bonfante, P., & Anca, I. A. (2009). Plants, mycorrhizal fungi, and bacteria: A network Guimera, R., & Amaral, L. (2005). Functional cartography of complex metabolic of interactions. Annual Review of Microbiology, 63, 363e383. https://doi.org/ networks. Nature, 433, 895e900. https://doi.org/10.1038/nature03288. 10.1146/annurev.micro.091208.073504. Haug, I., Setaro, S., & Suarez, J. P. (2013). Reforestation sites show similar and nested Brehm, G., Homeier, J., Fiedler, K., Kottke, I., Illig, J., Noske,€ N. M., et al. (2008). AMF communities to an adjacent pristine forest in a tropical mountain area Mountain rain forests in southern Ecuador as a hotspot of biodiversitydlimited of south Ecuador. PLoS One, 8, e63524. https://doi.org/10.1371/ knowledge and diverging patterns. In E. Beck, J. Bendix, I. Kottke, F. Makeschin, journal.pone.0063524. & R. Mosandl (Eds.), Gradients in a tropical mountain ecosystem of Ecuador. van der Heijden, M. G. A., Bardgett, R. D., & van Straalen, N. M. (2008). The unseen Ecological studies 198 (pp. 15e23). Berlin: Springer Verlag. majority: Soil microbes as drivers of plant diversity and productivity in Burgos, E., Ceva, H., Perazzo, R., Devoto, M., Medan, D., Zimmermann, M., et al. terrestrial ecosystems. Ecology Letters, 11, 296e310. https://doi.org/10.1111/ (2007). Why nestedness in mutualistic networks? Journal of Theoretical Biology, j.1461-0248.2007.01139.x. 249,307e313. https://doi.org/10.1016/j.jtbi.2007.07.030. Homeier, J., Breckle, S., Günter, S., Rollenbeck, R., & Leuschner, C. (2010). Tree di- Bussmann, R. W., Wilcke, W., & Richter, M. (2008). Landslides as important versity, forest structure and productivity along altitudinal and topographical disturbance regimesdcauses and regeneration. In E. Beck, J. Bendix, I. Kottke, gradients in a species-rich Ecuadorian montane rain forest. Biotropica, 42, F. Makeschin, & R. Mosandl (Eds.), Gradients in a tropical mountain ecosystem of 140e148. https://doi.org/10.1111/j.1744-7429.2009.00547.x. Ecuador. Ecological studies 198 (pp. 319e330). Berlin: Springer Verlag. Homeier, J., & Werner, F. A. (2007). Spermatophyta checklistdreserva Biologica san Chagnon, P. L., Bradley, R. L., & Klironomos, J. N. (2012). Using ecological network Francisco (prov. Zamora-chinchipe, S. Ecuador). In S. Liede-Schumann, & theory to evaluate the causes and consequences of arbuscular mycorrhizal S. W. Breckle (Eds.), Vol. 4. Provisional check-list of flora and fauna of San Fran- community structure. New Phytologist, 194,307e312. https://doi.org/10.1111/ cisco valley and its surroundings (estacion científica san Francisco), southern j.1469-8137.2011.04044.x. Ecuador. Ecotropical monographs (pp. 15e58). Ulm: Society for Tropical Ecology. Chao, A., Chazdon, R. L., Colwell, R. K., & Shen, T. J. (2005). A new statistical approach Homeier, J., Werner, F. A., Gawlik, J., Peters, T., Diertl, K. H., & Richter, M. (2013). Plant for assessing similarity of species composition with incidence and abundance diversity and its relevance for the provision of ecosystem services. In J. Bendix, data. Ecology Letters, 8,148e159. https://doi.org/10.1111/j.1461- E. Beck, A. Brauning,€ F. Makeschin, R. Mosandl, S. Scheu, et al. (Eds.), Ecosystem 0248.2004.00707.x. services, biodiversity and environmental change in a tropical mountain ecosystem Colwell, R., Chao, A., Gotelli, N., Lin, S. Y., Mao, C., Chazdon, R., et al. (2012). Models of South Ecuador. Ecological studies 221 (pp. 93e106). Berlin: Springer-Verlag. and estimators linking individual-based and sample-based rarefaction, Homeier, J., Werner, F. A., Gradstein, S. R., Breckle, S. W., & Richter, M. (2008). Po- extrapolation, and comparison of assemblages. Journal of Plant Ecology, 5,3e21. tential vegetation and floristic composition of Andean forests in South Ecuador, https://doi.org/10.1093/jpe/rtr044. with a focus on the RBSF. In E. Beck, J. Bendix, I. Kottke, F. Makeschin, & Cozzolino, S., & Widmer, A. (2006). Response to Otero and Flanagan: Orchid R. Mosandl (Eds.), Gradients in a tropical mountain ecosystems of Ecuador. diversityebeyond deception. Trends in Ecology and Evolution, 21,65e66. https:// Ecological studies 198 (pp. 87e100). Berlin: Springer-Verlag. doi.org/10.1016/j.tree.2005.11.017. Huber, T., Faulkner, G., & Hugenholtz, P. (2004). Bellerophon: A program to detect Cruz, D., Suarez, J. P., Kottke, I., & Piepenbring, M. (2014). Cryptic species revealed by chimeric sequences in multiple sequence alignments. Bioinformatics, 20, molecular phylogenetic analysis of sequences obtained from basidiomata of 2317e2319. https://doi.org/10.1093/bioinformatics/bth226. Tulasnella. Mycologia, 106, 708e722. https://doi.org/10.3852/12-386. Iotti, M., & Zambonelli, A. (2006). A quick and precise technique for identifying Cruz, D., Suarez, J., Kottke, I., Piepenbring, M., & Oberwinkler, F. (2011). Defining by PCR. Mycological Research, 110,60e65. https://doi.org/ species in Tulasnella by correlating morphology and nrDNA ITS-5.8S sequence 10.1016/j.mycres.2005.09.010. data of basidiomata from a tropical Andean forest. Mycological Progress, 10, Jacquemyn, H., Brys, R., Merckx, V., Waud, M., Lievens, B., & Wiegand, T. (2014). 229e238. https://doi.org/10.1007/s11557-010-0692-3. Coexisting orchid species have distinct mycorrhizal communities and display Cullings, K. (1994). Molecular phylogeny of the monotropoideae (Ericaceae) with a strong spatial segregation. New Phytologist, 202,616e627. https://doi.org/ note on the placement of the pyroloideae. Journal of Evolutionary Biology, 7, 10.1111/nph.12640. 501e516. https://doi.org/10.1046/j.1420-9101.1994.7040501.x. Jacquemyn, H., Brys, R., Waud, M., Busschaert, P., & Liebens, B. (2015b). Mycorrhizal Dearnaley, J. D. W. (2007). Further advances in orchid mycorrhizal research. My- networks and coexistence in species-rich orchid communities. New Phytologist, corrhiza, 17,475e486. https://doi.org/10.1007/s00572-007-0138-1. 206,1127e1134. https://doi.org/10.1111/nph.13281. Dearnaley, J. D. W., Martos, F., & Selosse, M. A. (2012). Orchid Mycorrhizas: Mo- Jacquemyn, H., Deja, A., De hert, K., Cachapa, B., & Lievens, B. (2012). Variation in lecular ecology, physiology, evolution and conservation aspects. In B. Hock (Ed.), mycorrhizal associations with tulasnelloid fungi among populations of five The mycota IX: Fungal associations (pp. 207e230). Berlin: Springer. Dactylorhiza species. PLoS One, 7, e42212. https://doi.org/10.1371/ Dodson, C. (2001). Native ecuadorian orchids (Vol. 2). Sarasota: Dodson Trust. journal.pone.0042212. DresslerellaeLepenthes. Jacquemyn, H., Honnay, O., Cammue, B., Brys, R., & Lievens, B. (2010). Low specificity Dodson, C. (2002). Native ecuadorian orchids (Vol. 3). Sarasota: Dodson Trust. and nested subset structure characterize mycorrhizal associations in five LepantheseOdontoglossum. closely-related species of the genus Orchis. Molecular Ecology, 19, 4086e4095. Dodson, C. (2003). Native ecuadorian orchids (Vol. IV). Sarasota: Dodson Trust. https://doi.org/10.1111/j.1365-294X.2010.04785.x. OncidiumeRestrepiopsis. Jacquemyn, H., Merckx, V., Brys, R., Tyteca, D., Cammue, B., Honnay, O., et al. (2011). Dodson, C. (2004). Native ecuadorian orchids (Vol. V). Sarasota: Dodson Trust. Analysis of network architecture reveals phylogenetic constraints on mycor- RodrigueziaeZygosepalum. rhizal specificity in the genus Orchis (Orchidaceae). New Phytologist, 192, Dodson, C., & Escobar, R. (1993). Native ecuadorian orchids (Vol. 1). Medellin: 518e528. https://doi.org/10.1111/j.1469-8137.2011.03796.x. Editorial Colina. AaeDracula. Jacquemyn, H., Waud, M., Merckx, V., Lievens, B., & Brys, R. (2015a). Mycorrhizal Dormann, C. F., Gruber, B., & Fründ, J. (2008). Introducing the bipartite package: diversity, seed germination and long-term changes in population size across Analysing ecological networks. R News, 8,8e11. nine populations of the terrestrial orchid Neottia ovata. Molecular Ecology, 24, Dunne, J. A., Williams, R. J., & Martinez, N. D. (2002). Network structure and 3269e3280. https://doi.org/10.1111/mec.13236. biodiversity loss in food webs: Robustness increases with connectance. Ecology Jimenez, V., & Hortal, J. (2003). Las curvas de acumulacion de especies y la neces- Letters, 5, 558e567. https://doi.org/10.1046/j.1461-0248.2002.00354.x. idad de evaluar la calidad de los inventarios biologicos. Revista Iberica de Ara- Ellison, G. N., & Gotelli, N. J. (2004). A primer of ecological statistics. Sunderland, cnología, 8,151e161. Massachusetts: Sinauer. Johnson, L., & Soltis, D. (1995). Phylogenetic inference in Saxifragaceae sensu stricto P. Herrera et al. / Mycoscience 59 (2018) 38e48 47

and Gilia (Polemoniaceae) using matK sequences. Annals of the Missouri Otero, J. T., Flanagan, N. S., Herre, E. A., Ackerman, J. D., & Bayman, P. (2007). Botanical Garden, 82,149e175. https://doi.org/10.2307/2399875. Widespread mycorrhizal specificity correlates to mycorrhizal function in the Jordano, P., Bascompte, J., & Olesen, J. M. (2003). Invariant properties in coevolu- neotropical, epiphytic orchid Ionopsis utricularioides (Orchidaceae). American tionary networks of planteanimal interactions. Ecology Letters, 6,69e81. Journal Botany, 94, 1944e1950. https://doi.org/10.3732/ajb.94.12.1944. https://doi.org/10.1046/j.1461-0248.2003.00403.x. Pellegrino, G., Luca, A., & Bellusci, F. (2014). Relationships between orchid and Jordano, P., Vazquez, D., & Bacompte, J. (2009). Redes complejas de interacciones fungal biodiversity: Mycorrhizal preferences in mediterranean orchids. Plant mutualistas planta-animal. In R. Medel, M. A. Aizen, & R. Zamora (Eds.), Ecología Biosystems, 150,180e189. https://doi.org/10.1080/11263504.2014.940071. y evolucion de interacciones planta-animal (pp. 17e41). Santiago de Chile: Pocock, M. J., Evans, D. M., & Memmott, J. (2012). The robustness and restoration of Editorial Universitaria. a network of ecological networks. Science, 335,973e977. https://doi.org/ Kartzinel, T. R., Trapnell, D. W., & Shefferson, R. P. (2013). Critical importance of large 10.1126/science.1214915. native trees for conservation of a rare Neotropical epiphyte. Journal of Ecology, Pridgeon, A., & Chase, M. (2003). Phylogenetics of the subtribe Pleurothallidinae 101, 1429e1438. https://doi.org/10.1111/1365-2745.12145. (Epidendreae: Orchidaceae) based on combined evidence from DNA sequences. Katoh, K., Kuma, K., Toh, H., & Miyata, T. (2005). MAFFT version 5: Improvement in Lankesteriana, 7, 49. https://doi.org/10.15517/lank.v3i2.23013. accuracy of multiple sequence alignment. Nucleic Acids Research, 33,511e518. Pridgeon, A., Solano, R., & Chase, M. (2001). Phylogenetic relationships in Pleuro- https://doi.org/10.1093/nar/gki198. thallidinae (Orchidaceae): Combined evidence from nuclear and plastid DNA Kottke, I., Haug, I., Setaro, S., Suarez, J. P., Weiß, M., Preußing, M., et al. (2008). Guilds sequences. American Journal of Botany, 88, 2286e2308. https://doi.org/10.2307/ of mycorrhizal fungi and their relation to trees, ericads, orchids and liverworts 3558390. in a neotropical mountain rain forest. Basic and Applied Ecology, 9,13e23. Rasmussen, H. N. (1995). Terrestrial orchids from seed to mycotrophic plant. Cam- https://doi.org/10.1016/j.baae.2007.03.007. bridge: Cambridge University Press. Kottke, I., Setaro, S., Haug, I., Herrera, P., Cruz, D., Fries, A., et al. (2013). Mycorrhiza Riofrío, L., Naranjo, C., Iriondo, J., & Torres, E. (2007). Spatial structure of Pleuro- networks promote biodiversity and stabilize the tropical mountain rain forest thallis, Masdevallia, Lepanthes and Epidendrum epiphytic orchids in a fragment ecosystem: Perspectives for understanding complex communities. In J. Bendix, of montane cloud forest in South Ecuador. Lankesteriana, 7,102e106. https:// E. Beck, A. Brauning,€ F. Makeschin, R. Mosandl, S. Scheu, et al. (Eds.), Ecosystem doi.org/10.15517/lank.v7i1-2.18447. services, biodiversity and environmental change in a tropical mountain ecosystem Roche, S. A., Carter, R. J., Peakall, R., Smith, L. M., Whitehead, M. R., & Linde, C. C. of South Ecuador. Ecological studies 221 (pp. 187e203). Berlin: Springer-Verlag. (2010). A narrow group of monophyletic Tulasnella (Tulasnellaceae) symbiont Kottke, I., Suarez, J., Herrera, P., Cruz, D., Bauer, R., Haug, I., et al. (2010). Atractiel- lineages are associated with multiple species of Chiloglottis (Orchidaceae): lomycetes belonging to the ‘rust’ lineage (Pucciniomycotina) form mycorrhizae Implications for orchid diversity. American Journal of Botany, 97,1313e1327. with terrestrial and epiphytic neotropical orchids. Proceedings of The Royal So- https://doi.org/10.3732/ajb.1000049. ciety B, 277, 1289e1298. https://doi.org/10.1098/rspb.2009.1884. Santamaría, S., Galeano, J., Pastor, J. M., & Mendez, M. (2014). Robustness of alpine Lindahl, B. D., Nilsson, R. H., Tedersoo, L., Abarenkov, K., Carlsen, T., Kjøller, R., et al. pollination networks: Effects of network structure and consequences for (2013). Fungal community analysis by high-throughput sequencing of amplified endemic plants. Arctic, Antarctic, and Alpine Research, 46, 568e580. https:// markersda user's guide. New Phytologist, 199, 288e299. https://doi.org/10.1111/ doi.org/10.1657/1938-4246-46.3.568. nph.12243. Shefferson, R. P., Taylor, D. L., Weiß, M., Garnica, S., McCormick, M., Adams, S., et al. Linde, C. C., May, T. W., Phillips, R. D., Ruibal, M., Smith, L. M., & Peakall, R. (2017). (2007). The evolutionary history of mycorrhizal specificity among lady's New species of Tulasnella associated with terrestrial orchids in Australia. IMA slipper orchids. Evolution, 61, 1380e1390. https://doi.org/10.1111/j.1558- , 8,27e47. https://doi.org/10.5598/imafungus.2017.08.01.03. 5646.2007.00112.x. Linde, C. C., Phillips, R. D., Crisp, M. D., & Peakall, R. (2014). Congruent species Smith, S., & Read, D. (2008). Mycorrhizal symbiosis (3rd ed.). San Diego: Academic delineation of Tulasnella using multiple loci and methods. New Phytologist, 201, Press. 6e12. https://doi.org/10.1111/nph.12492. Soberon, J., & Llorente, J. (1993). The use of species accumulation functions for the Martinez, A., Mahecha, M. D., Lischeid, G., & Beck, E. (2008). Succession stages of prediction of species richness. Conservation Biology, 7, 480e488. https://doi.org/ vegetation regeneration: Secondary tropical mountain forests. In E. Beck, 10.1046/j.1523-1739.1993.07030480.x. J. Bendix, I. Kottke, F. Makeschin, & R. Mosandl (Eds.), Gradients in a tropical Suarez, J. P., Eguiguren, J. S., Herrera, P., & Jost, L. (2016). Do mycorrhizal fungi drive mountain ecosystem of Ecuador. Ecological studies 198 (pp. 409e415). Berlin: speciation in Teagueia (Orchidaceae) in the upper Pastaza watershed of Springer Verlag. Ecuador? Symbiosis, 69,161e168. https://doi.org/10.1007/s13199-016-0399-6. Martos, F., Munoz, F., Pailler, T., Kottke, I., Gonneau, C., & Selosse, M. A. (2012). The Suarez, J. P., Weiss, M., Abele, A., Garnica, S., Oberwinkler, F., & Kottke, I. (2006). role of epiphytism in architecture and evolutionary constraint within mycor- Diverse tulasnelloid fungi form mycorrhizas with epiphytic orchids in an An- rhizal networks of tropical orchids. Molecular Ecology, 21, 5098e5109. https:// dean cloud forest. Mycological Research, 110, 1257e1270. https://doi.org/ doi.org/10.1111/j.1365-294X.2012.05692.x. 10.1016/j.mycres.2006.08.004. McCormick, M. K., & Jacquemyn, H. (2014). What constrains the distribution of Suarez, J. P., Weiss, M., Abele, A., Oberwinkler, F., & Kottke, I. (2008). Members of orchid populations? New Phytologyst, 202, 392e400. https://doi.org/10.1111/ Sebacinales subgroup B form mycorrhizae with epiphytic orchids in a nph.12639. neotropical mountain rain forest. Mycological Progress, 7,75e85. https://doi.org/ McCormick, M. K., Whigham, D. F., & O'Neill, J. (2004). Mycorrhizal diversity in 10.1016/j.mycres.2006.08.004. photosynthetic terrestrial orchids. New Phytologyst, 163, 425e438. https:// Swofford, D. (2002). PAUP*4.0 phylogenetic analysis using parsimony (*and other doi.org/10.1111/j.1469-8137.2004.01114.x. methods). Sunderland: Sinauer. Medan, D., Perazzo, R. P., Devoto, M., Burgos, E., Zimmermann, M. G., Ceva, H., et al. Takimoto, G., & Suzuki, K. (2016). Global stability of obligate mutualism in com- (2007). Analysis and assembling of network structure in mutualistic systems. munity modules with facultative mutualists. Oikos, 125, 535e540. https:// Journal of Theoretical Biology, 246,510e521. https://doi.org/10.1016/ doi.org/10.1111/oik.02741. j.jtbi.2006.12.033. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., & Kumar, S. (2011). Memmott, J., Waser, N. M., & Price, M. V. (2004). Tolerance of pollination networks MEGA5: Molecular Evolutionary Genetics Analysis using maximum likeli- to species extinctions. Proceedings of The Royal Society B, 271, 2605e2611. hood, evolutionary distance, and maximum parsimony methods. Molecular https://doi.org/10.1098/rspb.2004.2909. Biology and Evolution, 28, 2731e2739. https://doi.org/10.1093/molbev/ Neubig, K., Whitten, W., Carlsward, B., Blanco, M., Endara, L., Williams, N., et al. msr121. (2009). Phylogenetic utility of ycf1 in orchids: A plastid gene more variable than Timms, R., & Read, A. (1999). What makes a specialist special? Trends in Ecology and matK. Plant Systematics and Evolution, 277,75e84. https://doi.org/10.1007/ Evolution, 14,333e334. https://doi.org/10.1016/S0169-5347(99)01697-3. s00606-008-0105-0. Valyi, K., Rillig, M. C., & Hempel, S. (2015). Land-use intensity and host plant identity Nielsen, A., & Bascompte, J. (2007). Ecological networks, nestedness and sampling interactively shape communities or arbuscular mycorrhizal fungi in roots of effort. Journal of Ecology, 95,1134e1141. https://doi.org/10.1111/j.1365- grassland plants. New Phytologist, 205,1577e1586. https://doi.org/10.1111/ 2745.2007.01271.x. nph.13236. Oesker, M., Daliz, H., Günter, S., Homeier, J., & Matezki, S. (2008). Spatial hetero- Warcup, J. H. (1981). The mycorrhizal relationships of Australian orchids. New geneity patternsda comparison between Gorges and Ridges in the upper part Phytologist, 87, 371e381. https://doi.org/10.1111/j.1469-8137.1981.tb03208.x. of an evergreen lower montane forest. In E. Beck, J. Bendix, I. Kottke, Waud, M., Busschaert, P., Lievens, B., & Jacquemyn, H. (2016). Specificity and F. Makeschin, & R. Mosandl (Eds.), Gradients in a tropical mountain ecosystem of localised distribution of mycorrhizal fungi in the soil may contribute to co- Ecuador. Ecological studies 198 (pp. 267e274). Berlin: Springer. existence of orchid species. Fungal Ecology, 20,155e165. https://doi.org/ Okuyama, T., & Holland, J. N. (2008). Network structural properties mediate the 10.1016/j.funeco.2015.12.008. stability of mutualistic communities. Ecology Letters, 11, 208e216. https:// Weber, M., Günter, S., Aguirre, N., Stimm, B., & Mosandl, R. (2008). Reforestation of doi.org/10.1111/j.1461-0248.2007.01137.x. abondoned pastures: Silvicultural means to accelerate forest recovery and Olesen, J. M., Bascompte, J., Dupont, Y. L., & Jordano, P. (2007). The modularity of biodiversity. In E. Beck, J. Bendix, I. Kottke, F. Makeschin, & R. Mosandl (Eds.), pollination networks. Proceedings of the National Academy of Sciences of the Gradients in a tropical mountain ecosystem of Ecuador. Ecological studies 198 (pp. United States of America, 104, 19891e19896. https://doi.org/10.1073/ 431e441). Berlin: Springer Verlag. pnas.0706375104. Weiss, M., Waller, F., Zuccaro, A., & Selosse, M. A. (2016). Sebacinalesdone thousand Otero, J. T., Ackermann, J. D., & Bayman, P. (2004). Differences in mycorrhizal and one interactions with land plants. New Phytologist, 211,20e40. https:// preferences between two tropical orchids. Molecular Ecology, 13, 2393e2404. doi.org/10.1111/nph.13977. https://doi.org/10.1111/j.1365-294X.2004.02223.x. Werner, F. A. (2011). Reduced growth and survival of vascular epiphytes on isolated Otero, J. T., & Flanagan, N. S. (2006). Orchid diversitydbeyond deception. Trends in remnant trees in a recent tropical montane forest clear-cut. Basic and Applied Ecology and Evolution, 21,64e65. https://doi.org/10.1016/j.tree.2005.11.016. Ecology, 12,172e181. https://doi.org/10.1016/j.baae.2010.11.002. 48 P. Herrera et al. / Mycoscience 59 (2018) 38e48

Werner, F. A., Homeier, J., & Gradstein, S. R. (2005). Diversity of vascular epiphytes molecular evidence. American Journal of Botany, 87, 1842e1856. https://doi.org/ on isolated remnant trees in the montane forest belt of Southern Ecuador. 10.2307/2656837. Ecotropica, 11,21e40. Winkler, M., Hülber, K., & Hietz, P. (2009). Population dynamics of epiphytic orchids White, T., Bruns, T., Lee, S., & Taylor, J. (1990). Amplification and direct sequencing of in a metapopulation context. Annals of Botany, 104, 995e1004. https://doi.org/ fungal ribosomal RNA genes for phylogenetics. In M. A. Innis, H. Gelfand, 10.1093/aob/mcp188. J. S. Sninsky, & T. J. White (Eds.), PCR protocols and applications: A laboratory Yang, G., Yang, X., Zhang, W., Wei, Y., Ge, G., Lu, W., et al. (2016). Arbuscular manual (pp. 315e322). San Diego: Academic Press. mycorrhizal fungi affect plant community structure under various nutrient Whitten, W., Williams, N., & Chase, M. (2000). Subtribal and generic relationships of conditions and stabilize the community productivity. Oikos, 125,576e585. Maxillarieae (Orchidaceae) with emphasis on Stanhopeinae: Combined https://doi.org/10.1111/oik.02351.