part of a highlight on orchid biology...in temperate forests, mh orchids usually associate with...

PART OF A HIGHLIGHT ON ORCHID BIOLOGY

The importance of associations with saprotrophic non-Rhizoctonia fungi among

fully mycoheterotrophic orchids is currently under-estimated: novel

evidence from sub-tropical Asia

Yung-I Lee1,2, Chih-Kai Yang3,4 and Gerhard Gebauer5,*1Biology Department, National Museum of Natural Science, No 1, Kuan-Chien Rd, Taichung, Taiwan, 2Department of LifeSciences, National Chung Hsing University, Taichung 40227, Taiwan, 3The Experimental Forest, College of Bio-Resources

and Agriculture, National Taiwan University, 12 Chienshan Rd., Sec. 1, Chushan Township, Nantou 55750, Taiwan,4Department of Life Science, National Taiwan Normal University, 88 Tingchow Rd., Sec. 4, Taipei 11677, Taiwan

and 5Laboratory of Isotope Biogeochemistry, Bayreuth Center of Ecology and EnvironmentalResearch (BayCEER), University of Bayreuth, D-95440 Bayreuth, Germany

* For correspondence. E-mail [email protected]

Received: 18 December 2014 Returned for revision: 4 February 2015 Accepted: 27 April 2015 Published electronically: 25 June 2015

� Background and Aims Most fully mycoheterotrophic (MH) orchids investigated to date are mycorrhizal withfungi that simultaneously form ectomycorrhizas with forest trees. Only a few MH orchids are currently known to bemycorrhizal with saprotrophic, mostly wood-decomposing, fungi instead of ectomycorrhizal fungi. This study pro-vides evidence that the importance of associations between MH orchids and saprotrophic non-Rhizoctonia fungi iscurrently under-estimated.� Methods Using microscopic techniques and molecular approaches, mycorrhizal fungi were localized and identi-fied for seven MH orchid species from four genera and two subfamilies, Vanilloideae and Epidendroideae, growingin four humid and warm sub-tropical forests in Taiwan. Carbon and nitrogen stable isotope natural abundances ofMH orchids and autotrophic reference plants were used in order to elucidate the nutritional resources utilized by theorchids.� Key Results Six out of the seven MH orchid species were mycorrhizal with either wood- or litter-decaying sapro-trophic fungi. Only one orchid species was associated with ectomycorrhizal fungi. Stable isotope abundancepatterns showed significant distinctions between orchids mycorrhizal with the three groups of fungal hosts.� Conclusions Mycoheterotrophic orchids utilizing saprotrophic non-Rhizoctonia fungi as a carbon and nutrientsource are clearly more frequent than hitherto assumed. On the basis of this kind of nutrition, orchids can thrive indeeply shaded, light-limiting forest understoreys even without support from ectomycorrhizal fungi. Sub-tropicalEast Asia appears to be a hotspot for orchids mycorrhizal with saprotrophic non-Rhizoctonia fungi.

Key words: Orchids, Orchidaceae, mycoheterotrophy, mycorrhiza, ectomycorrhiza, Vanilloideae, Epidendroideae,Gastrodia, stable isotopes, carbon, nitrogen, saprotrophic fungi.

INTRODUCTION

In nature, orchids are known to begin their life cycle as myco-heterotrophs (Rasmussen, 1995; Leake, 2004). Because of therudimentary embryo and the lack of endosperm in seeds, thegermination of orchid seeds is dependent on the formation of amycorrhizal association, which supplies young seedlings, i.e.protocorms, with all carbon (C) and mineral nutrients until theseedlings develop green leaves and become putatively autotro-phic (Leake, 1994; Merckx, 2013). Mycorrhizal partners of themajority of these adult green orchids are widely distributedfungi of the polyphyletic Rhizoctonia group, includingTulasnella, Ceratobasidium, Thanatephorus and Sebacina cladeB (Dearnaley et al., 2012). In contrast to the putatively autotro-phic nutritional mode of chlorophyllous orchids, a few orchidsremain achlorophyllous and depend on their mycorrhizal part-ners for C and mineral nutrient supplies throughout their entirelife cycle. These achlorophyllous orchids are known as myco-heterotrophic (MH) plants (Leake, 1994; Merckx, 2013).

In temperate forests, MH orchids usually associate with nar-row clades of ectomycorrhizal (ECM) fungi and obtain photo-synthates from neighbouring trees through undergroundmycorrhizal networks (Taylor and Bruns, 1997; Hynson et al.,2013). In tropical and sub-tropical forests, a few MH orchidshave been reported to associate with saprotrophic (SAP) non-Rhizoctonia fungi and obtain nutrients through the ability of thefungi to cause wood or litter decay. For example, Gastrodiaspp. have been shown to associate with the litter- and/or wood-decomposing fungi Armillaria, Mycena, Resinicium andCampanella or Marasmius (Kusano, 1911; Kikuchi et al.,2008; Martos et al., 2009; Ogura-Tsujita et al., 2009;Dearnaley and Bougoure 2010), Epipogium roseum associateswith a litter-decomposing species of Coprinaceae in cultureconditions (Yamato et al., 2005), Wullschlaegelia aphylla asso-ciates with litter-decaying species of Gymnopus and Mycena(Martos et al., 2009), Eulophia zollingeri associates with an-other litter decomposer, Psathyrella cf. candolleana (Ogura-Tsujita and Yukawa 2008), and Erythrorchis spp. associate

VC The Author 2015. Published by Oxford University Press on behalf of the Annals of Botany Company.All rights reserved. For Permissions, please email: [email protected]

Annals of Botany 116: 423–435, 2015

doi:10.1093/aob/mcv085, available online at www.aob.oxfordjournals.org

with a wide range of wood-rotting fungi of Hymenochaetaceaeand Polyporaceae (Umata, 1995, 1997a; Dearnaley, 2007). Inaddition, some MH orchids in tropical and sub-tropical forestsalso associate with ECM fungi, but lack mycorrhizal specificityas in temperate forests (Roy et al., 2009). A recent molecularapproach indicates that Lecanorchis associates with diverseECM fungi, e.g. Lactarius, Russula, Atheliaceae and Sebacinaclade A (Okayama et al., 2012).

The analysis of 13C and 15N (nitrogen) isotope natural abun-dances has been extensively used to elucidate the nutritional re-sources utilized by organisms in ecosystems. Along food chainsmost organisms have isotope values similar to their food re-sources (Fry, 2006). ECM fungi are significantly enriched in13C and 15N as compared with autotrophic plants (Gebauer andDietrich, 1993; Gleixner et al., 1993), and consequently, MHplants associated with ECM fungi have isotope signatures closeto those of ECM fungi (Gebauer and Meyer, 2003; Trudellet al., 2003). As compared with ECM fungi, wood-decayingSAP fungi are even more enriched in 13C, but less enriched in15N, and, therefore, MH plants associated with wood-decayingSAP fungi should have isotope signatures similar to this groupof SAP fungi. Unfortunately, little information about the isoto-pic composition of MH plants associated with non-RhizoctoniaSAP fungi is available to date (Ogura-Tsujita et al., 2009;Martos et al., 2009; Dearnaley and Bougoure, 2010; Sommeret al., 2012; Hynson et al., 2013). Furthermore, the currentlyavailable knowledge provides no clue about whether MH plantsassociated with wood-decaying or litter-decaying fungi are dif-ferent in terms of their isotopic composition. Such a distinctionshould be expected from the different isotopic composition ofthe substrates on which these fungi live (Gebauer and Schulze,1991; Cernusak et al., 2009).

A great diversity of MH orchids (>120 species) occurs intropical and sub-tropical Asia. In Taiwan, among the approx.400 native orchids, >50 fully MH orchids in 15 genera havebeen recorded (Su, 2000). The Xitou Experimental Forest lo-cated in central Taiwan has long humid seasons with warmtemperatures that obviously favour the growth of MH orchids.According to the report by Yang et al. (2010), nine fully MHorchids, including representatives of the genera Cyrtosia,Galeola and Lecanorchis (subfamily Vanilloideae) andGastrodia and Epipogium (subfamily Epidendroideae) occur inthis misty forest with abundant litter and dead wood. Amongthese MH orchids in the Xitou Experimental Forest, threeGastrodia spp., G. appendiculata, G. fontinalis and G. nantoen-sis, occur sympatrically in a bamboo forest, whereas in anotherbamboo forest, two vanilloid orchids, C. javanica and L. thalas-sica, grow sympatrically.

The richness of MH orchids in the sub-tropical forests in cen-tral Taiwan allows us to test the following questions. (1) Whatare their mycorrhizal partners? Current knowledge about theidentity of mycorrhizal fungi in the vanilloid genera Cyrtosiaand Erythrorchis, close relatives of Galeola, are primarilybased on only in vitro isolations (Hamada, 1939; Umata, 1995,1997a). Here we identify the fungal associates of seven MH or-chids using molecular methods. (2) Gastrodia spp. occurmainly in Asia, Africa and Australia. How much diversity isthere in mycorrhizal partners over the range of Gastrodia spp.?We compare the fungal composition in mycorrhizas of sympat-ric and allopatric species. (3) Although the mycorrhizal partners

of Cyrtosia, Galeola and Lecanorchis of subfamilyVanilloideae have already been investigated, their nutritionalresources are still not clear. Cyrtosia and Galeola appear to as-sociate with SAP fungi, whereas Lecanorchis presumably asso-ciates with ECM fungi. However, Lyophyllum shimeji, an ECMfungus, could stimulate seed germination in vitro ofErythrorchis, a close relative genus of Galeola (Umata, 1997b),suggesting the possible recruitment of an ECM mycorrhizalpartner in the natural environment. In this study we analyse forthe first time the C and N stable isotope abundances of threevanilloid orchids and four Gastrodia spp. to reveal their nutri-tional resources, either ECM fungi or wood-decaying or litter-decaying non-Rhizoctonia SAP fungi.

MATERIALS AND METHODS

Sample collection and locations

Specimens of seven fully MH orchids (Figs 1 and 2) were sam-pled from four sites in Central Taiwan from 2011 to 2012(Supplementary Data Table S1). The four sites are locatedapprox. 500–3000 m from each other in the Xitou ExperimentalForest (College of Bio-resources and Agriculture, NationalTaiwan University), Nantou County, Taiwan at 1000 m abovesea level. The climate is sub-tropically moist, with a mean an-nual temperature of 16�6 �C and a mean annual precipitation of2600 mm. Site A (23�6902900N, 120�7901200E) consists of abroadleaf forest on organic soil (pH 3�7) dominated by Phoebeformosana and Machilus japonica (Lauraceae) with someunderstorey plants (see Table S2). The target plant at this sitewas the MH orchid Galeola falconeri. Site B (23�4004400N,120�4505300E) consists of a dense bamboo (Phyllostachysedulis) forest mixed with Cryptomeria japonica trees on or-ganic soil (pH 4�4) with only few understorey plants (see TableS2). The MH orchids Cyrtosia javanica and Lecanorchis tha-lassica grow sympatrically at this site. Site C (23�4002600N,120�4704500E) consists of a coniferous forest on organic soil(pH 4�0) dominated by Cryptomeria japonica with some under-storey plants (see Table S2). The target orchid at this site wasthe MH Gastrodia flabilabella. Site D (23�4004100N,120�4704300E) consists of a dense bamboo (Phyllostachysedulis) forest on organic soil (pH 4�4) with only few under-storey plants (see Table S2). The MH orchids Gastrodia appen-diculata, G. fontinalis and G. nantoensis co-occur at this site.Voucher specimens of G. falconeri (Yung-I Lee 201225),C. javanica (Yung-I Lee 201223), L. thalassica (Yung-I Lee201224), G. appendiculata (Yung-I Lee 201122), G. fontinalis.(Yung-I Lee 201217), G. nantoensis (Yung-I Lee 201121) andG. flabilabella (Yung-I Lee 201117) have been deposited in theherbarium of the National Museum of Nature and Science,Taichung, Taiwan. Light climate data of the four sites weremeasured with a LM-8000 Lux meter (Lutron ElectronicEnterprise Co., Ltd, Taipei, Taiwan) at 20 cm from ground levelat three different points in each site. Site A received a mean of3670 lux (¼ 6 %), site B a mean of 980 lux (¼ 2 %), site C amean of 3050 lux (¼ 5 %) and site D a mean of 950 lux (¼2 %), whereas outside of the forests at the same time a mean of57 800 lux (¼ 100 %) was measured.

424 Lee et al. — Associations with saprotrophic fungi among fully mycoheterotrophic orchids

http://aob.oxfordjournals.org/lookup/suppl/doi:10.1093/aob/mcv085/-/DC1






Microscopy

Mycorrhizal roots were collected and fixed in 2�5 % glutar-aldehyde and 1�6 % paraformaldehyde buffered with 0�05 M

phosphate buffer, overnight at 4 �C. After fixation, the sampleswere dehydrated using an ethanol series, and embedded inTechnovit 7100 (Kulzer & Co., Wehrheim, Germany). Sectionsof 3mm thickness were obtained using Ralph knives on aReichert-Jung 2040 Autocut rotary microtome. Sections werestained with 0�05 % (w/v) toluidine blue O (TBO) in benzoatebuffer for general histology (Yeung, 1984). The sections wereexamined and the images were captured using a digital cameraattached to a microscope (Axioskop 2, Carl Zeiss AG, Jena,Germany).

Molecular identification of mycorrhizal fungi

After checking for fungal colonization by free-hand sectionsunder the microscope, mycorrhizal roots were washed in waterand kept at –80 �C until use. DNA was extracted from eachsample by using a DNeasy Plant Mini Kit (Qiagen, Hilden,Germany). The internal transcribed spacer (ITS) region of thefungal nuclear rRNA gene was amplified with the primer com-binations ITS1F/ITS4 or ITS1F/ITS4B (White et al., 1990;Gardes and Bruns, 1993). The large subunit (LSU) nuclear ribo-somal DNA (nrDNA) sequences were amplified using primercombinations LR0R/LR5 (Moncalvo et al., 2000) or LR0R/LR3 (Vilgalys and Hester, 1990). PCR amplification and se-quencing were carried out as described by Ogura-Tsujita et al.(2009). PCR products that were difficult to sequence directlywere cloned using the pGEM-T Vector System II (Promega,Madison, WI, USA). Sequences were identified(Supplementary Data Table S3) using a BLAST search against

the NCBI sequence database (National Center forBiotechnology Information, GenBank) to find the closest se-quence matches in the database. For phylogenetic analysis,LSU marasmioid sequences from GenBank were added to theanalysis by referring to Moncalvo et al. (2000, 2002), Wilsonand Desjardin (2005), Matheny et al. (2006), Martos et al.(2009) and Ogura-Tsujita et al. (2009), and sequences ofCyphella digitalis, Nia vibrissa and Henningsomyces candiduswere used as outgroup taxa. LSU sequences of Polyporalesfrom GenBank were added to the analysis by referring to Justoand Hibbett (2011) and Binder et al. (2013), and sequences ofColtriciella oblectabilis was used as outgroup taxa. ITS se-quences of Russula from GenBank were added to the analysisby referring to Okayama et al. (2012), and sequences ofArcangeliella camphorata and Lactarius quietus were used asoutgroup taxa. DNA sequences were aligned usingCLUSTALX (Thompson et al., 1997), followed by manual ad-justment. Phylogenetic relationships were analysed by a model-based Bayesian approach using MrBayes 3�2.1 (Ronquist andHuelsenbeck, 2003). The ‘best-fit’ model of evolution was se-lected under the Akaike information criterion test (Akaike,1974) as implemented in MrModeltest 2�2 (Nylander, 2004).The general time reversal plus invariant rates and a gammadistribution (GTRþ IþC) was selected for the analyses. Twoseparate runs of four Monte Carlo Markov chains (MCMCs;Yang and Rannala, 1997) were performed for 10 000 000 gener-ations until the mean deviation of split frequency dropped be-low 0�01, and a tree was sampled every 1000th generation.Trees from the first 25 % of generations were discarded usingthe ‘burn-in’ command, and the remaining trees were used tocalculate a 50 % majority-rule consensus topology and to de-termine the posterior probability (PP) for individual branches.The alignment data sets were further analysed by maximum

A C

B

FIG. 1. Flower morphology of vanilloid species. (A) Galeola falconeri, scale bar¼ 5 cm; (B) Cyrtosia javanica, scale bar¼ 1 cm; (C) Lecanorchis thalassica, scalebar¼ 1 cm.

Lee et al. — Associations with saprotrophic fungi among fully mycoheterotrophic orchids 425


parsimony (MP) using PAUP* version 4�0b10 (Swofford,2002). Support for groups was evaluated using the bootstrapmethod (Felsenstein, 1985) with 1000 replicates. The treesobtained in these analyses were drawn with the TreeGraph 2software (Stover and Muller, 2010).

Stable isotope abundance analysis

Five 1 m2 plots were selected at each site; each plot includedfully MH orchids and four to five autotrophic reference plantspecies. Flower stalks of seven fully MH orchids, leaves of fourto five autotrophic reference plants and soil samples from theorganic layer were taken from each of the five plots at eachsite. The reference plants collected at the respective sites arelisted in Supplementary Data Table S2. In addition, on site B,fruit bodies of a litter-decaying SAP fungus (Marasmius sp.)were found and collected in five replicates.

Samples were dried at 105 �C, ground to a fine powder andstored in a desiccator with silica gel until analysed. Relative Nand C isotope abundances of the samples were measured using

a dual-element analysis mode with an elemental analyser cou-pled to a continuous flow isotope ratio mass spectrometer as de-scribed in Bidartondo et al. (2004). Measured abundances aredenoted as d values that were calculated according to the givenequation d15N or d13C¼ (Rsample/Rstandard� 1)� 1000 [%],where Rsample and Rstandard are the ratios of heavy isotope tolight isotope of the samples and the respective standard.Standard gases (N2 and CO2) were calibrated with respect to in-ternational standards by using the reference substances N1 andN2 for N isotopes and ANU sucrose, NBS 18 and NBS 19 forC isotopes, provided by the International Atomic EnergyAgency (Vienna, Austria). Reproducibility and accuracy of theisotope abundance measurements were routinely controlled bymeasuring the test substance acetanilide (Gebauer and Schulze,1991). At least six test substances with varying sample weightwere routinely analysed in each batch of 50 samples. The maxi-mum variation of d13C and d15N within and between batcheswas always <0�2 %. To compare isotope abundances of or-chids and reference plants from different sites, the data werenormalized. Enrichment factors (e) were calculated per plot:e¼ dS� dREF, with S as a single d13C or d15N value of an adult

A C

B D

FIG. 2. Flower morphology of Gastrodia species. (A) Gastrodia appendiculata, scale bar¼ 1 cm; (B) Gastrodia fontinalis, scale bar¼ 1 cm; (C) Gastrodia flabila-bella, scale bar¼ 1 cm; (D) Gastrodia nantoensis, scale bar¼ 1 cm.



orchid and REF as the mean value of non-orchid referenceplants from the respective plot (Preiss and Gebauer, 2008). Theoriginal d13C and d15N values of orchids, the respective refer-ence plants, fungi and soil samples are available inSupplementary Data Table S2. Total N concentrations in leaf,stem, fungus and soil samples were calculated from sampleweights and peak areas using a six-point calibration curve persample run based on acetanilide measurements (Gebauer andSchulze, 1991). Acetanilide has a constant N concentration of10�36 %. Enrichment factors and total N concentrations of or-chids and reference plants were tested for normal distribution.Enrichment factors e13C and e15N were not normally distrib-uted and therefore these data were tested for statistical differ-ences using the Kruskal–Wallis non-parametric test followedby a post-hoc Mann–Whitney U-test with an adjusted signifi-cance level according to Holm (1979). The autotrophic refer-ence plants were treated as one group after confirminginsignificant differences among the data of each species. TotalN concentrations were normally distributed and thus theStudent’s t-test was used to test total N concentrations in or-chids and reference plants for statistical differences.

RESULTS

Histological studies

In the three vanilloid study species, the below-ground structureof Galeola falconeri had a long rhizome with a few thick roots.Cyrtosia javanica had a short rhizome with a few thick roots.Lecanorchis thalassica had a slender rhizome with a number ofthick roots (Supplementary Data Fig. S1). The exodermal cellsof roots in the three vanilloid orchids were characterized by thethickened outer and lateral walls. Colonization by fungal hy-phae in the middle cortex cells could be observed (Fig. 3), andtheir exodermal and outer cortex layers were occasionally colo-nized (Fig. 3B).

The below-ground structures of Gastrodia appendiculata,G. fontinalis and G. nantoensis were similar, having thick rhi-zomes with a few slim roots (Supplementary Data Fig. S1E).The epidermal cells remained intact or became collapsed with-out fungal colonization. The outer and inner cortex cells wereusually uncolonized, whereas the middle cortex cells were filledwith fungal hyphae (Fig. 4A, B, D). Gastrodia flabilabella hada tuberous rhizome with several coralloid roots (SupplementaryData Fig. S1D). As observed in the other three Gastrodia spp.,the epidermal, outer and inner cortical layers were rarely colo-nized. The middle cortex cells were heavily colonized by fungalhyphae. It is worth noting that several papillae-like cell wallthickenings could be observed at the adjoining walls betweenthe outer and middle cortex cells (Fig. 4A, C, D).

Molecular identification of mycorrhizal fungi

Two of the three MH vanilloid study species were associatedwith SAP non-Rhizoctonia fungi known to be wood-decaying.The ITS sequences obtained from seven G. falconeri individ-uals (22 samples) and those obtained from five C. javanica in-dividuals (20 samples) had a high DNA sequence homologywith species of Meripilaceae (order Polyporales) by BLAST

analysis. Thus, wood-decaying Meripilaceae have to be consid-ered as the exclusive fungal associates of G. falconeri andC. javanica (Fig. 5). In contrast, for the third MH vanilloid or-chid L. thalassica, the ITS sequences obtained from seven indi-viduals (24 samples) demonstrate a high DNA sequencehomology with species of the ECM fungus Russula (Fig. 6).

For all Gastrodia spp. studied, associations with SAP non-Rhizoctonia fungi were found. For the three Gastrodia spp.growing sympatrically in a bamboo forest (G. appendiculata,15 samples of five individuals; G. fontinalis, 22 samples of sixindividuals; G. nantoensis, 20 samples of six individuals), theITS sequences demonstrate a high DNA sequence homologywith Mycena spp. The ITS sequences of Mycena obtained fromthe roots of G. appendiculata, G. fontinalis and G. nantoensis(57 root samples) are grouped into three types (Supplementary

EX

OC

MC

OC

OC

MC

MC

IC

H

H

EX

EX

H

A

B

C

FIG. 3. Histology of mycorrhizas in three vanilloid species. In their epidermalcells, the outer and lateral walls become thickened. (A) Light micrograph show-ing a transverse section of a root of G. falconeri. Fungal colonization in the mid-dle cortex cells could be observed. (B) Light micrograph showing a transversesection of a root of C. javanica. A few cells in the exodermal, outer and innercortex cells are filled with fungal hyphae. (C) Light micrograph showing a trans-verse section of a root of L. thalassica. Fungal colonization is mainly found inthe middle cortex cells. EX, exodermal cells; H, fungal hyphae; IC, inner cortex;

MC, middle cortex; OC, outer cortex. Scale bar¼ 100mm.








Data Table S4). Type I was only detected in G. fontinalis, andtype II was only detected in G. appendiculata. Type III was de-tected in both G. appendiculata and G. nantoensis. In addition,the mycorrhizal roots of G. fontinalis (eight samples) were alsocolonized by Gymnopus spp. For G. flabilabella (21 samples ofseven individuals), collected in a coniferous forest, the gener-ated ITS sequences indicate homology with species of the fun-gal genus Hydropus (Fig. 7).

Stable isotope natural abundance and total N concentrations

Comparisons of enrichment factors e13C (H¼ 77; d.f.¼ 7;P< 0�001) and e15N (H¼ 77; d.f.¼ 7; P< 0�001) among theseven orchid species and the set of reference plants revealedhighly significant differences in our data set. All study orchidswere enriched by 7�9 6 0�2 % (G. appendiculata) to 12�1 6 0�5% (G. flabilabella) in 13C and by 4�7 6 0�6 % (G. flabilabella)to 8�8 6 2�7 % (L. thalassica) in 15N in comparison with auto-trophic reference plants growing at the same sites (Fig. 8).Based on post-hoc tests, this enrichment in 13C and 15N washighly significant for all orchid species (in all cases U¼ 0;P< 0�001).

The orchids themselves fall into three distinct groups. TheECM-associated L. thalassica was significantly more enrichedin 15N than all other orchids associated with SAP fungi (U¼ 0to U¼ 3; P< 0�001 to P¼ 0�003; mean difference 3�7 %) andsignificantly less enriched in 13C than the group composed of

G. falconeri, C. javanica and G. flabilabella (U¼ 0; P< 0�001;mean difference 3�0 %). However, L. thalassica was not signif-icantly distinguishable in its 13C enrichment from the orchidgroup composed of G. nantoensis, G. appendiculata and G. fon-tinalis (U¼ 25; P¼ 0�295). The group composed of G. falco-neri, C. javanica and G. flabilabella was more enriched in 13Cthan the group composed of G. nantoensis, G. appendiculataand G. fontinalis (U¼ 0; P< 0�001; mean difference 3�5 %).However, both of these groups were not significantly differentin their 15N enrichments (U¼ 107; P¼ 0�836).

The investigated litter-decaying SAP fungus Marasmius sp.was enriched by 9�4 6 0�5 % in 13C and by 4�3 6 0�2 % in 15Ncompared with autotrophic reference plants (Fig. 8) and thushad similar enrichments in 13C and 15N to the orchid groupcomposed of G. nantoensis, G. appendiculata and G. fontinalis.

Orchid flower stems had a slightly higher mean total N con-centration (2�88 6 0�50 mmol g d. wt–1; n¼ 35) than referenceplant leaves (2�67 6 0�67 mmol g d. wt–1; n¼ 90;Supplementary Data Table S2). However, this slight differencewas not significant (t¼ 1�671; P¼ 0�097; d.f.¼ 123).

DISCUSSION

Fungal colonization in the roots

In three MH vanilloid orchids, mycorrhizal colonization ofwood-decaying and ECM fungi was mainly observed in thecortical layers of the old roots, a finding similar to reports for

OC

OC OC

OCE

E

E

E

MC

MC

MC

MC

H

H

H

HIC

ICIC

A C

B D

FIG. 4. Histology of mycorrhizas in four Gastrodia species. Their epidermal, outer and inner cortical layers are rarely colonized. (A) Light micrograph showing atransverse section of a root of G. appendiculata. The middle cortex cells are heavily colonized by fungal hyphae. The thickened papillae-like cell walls could be ob-served at the adjoining walls between the outer and middle cortex cells (arrows). (B) Light micrograph showing a transverse section of a root of G. fontinalis. (C)Light micrograph showing a transverse section of a root of G. flabilabella. (D) Light micrograph showing a transverse section of a root of G. nantoensis. A few papil-lae-like cell walls occur at the adjoining walls between the outer and middle cortex cells (arrows). E, epidermal cells; H, fungal hyphae; IC, inner cortex; MC, middle

cortex; OC, outer cortex. Scale bar¼ 50mm.




mycorrhizal colonizations in chlorophyllous Vanilla spp. byPorras-Alfaro and Bayman (2007). The outermost layers of oldroots in MH vanilloid orchids are characterized by thickenedexodermal layers, whereas the epidermal cells are no longeralive. Colonization of fungal hyphae was also found in thethickened exodermal layers (Fig. 3B), suggesting their roles formaintenance of nutrient uptake by older roots (Esnault et al.,1994).

In the Gastrodia spp. investigated, fungal colonization wasrestricted to a few cortical layers of root systems, but was notcommonly observed in rhizomes. In the colonized corticallayers, the papillae-like cell wall thickenings were abundant atthe adjoining walls between the outer and middle cortex cells,corresponding to pathways for fungal hyphae (Fig. 4A, C, D).The presence of papillae-like cell wall thickenings could be po-tentially underdeveloped structures of wall ingrowths in spe-cialized transfer cells as described in symbiotic associations byPate and Gunning (1972), suggesting a specific nutrient trans-port network in the mycorrhiza (Martos et al., 2009).

Mycorrhizal partners

Subfamily Vanilloideae contain a number of non-photosyn-thetic genera (40 % of the 15 genera of vanilloid orchids), e.g.Cyrtosia, Erythrorchis, Galeola, Pseudovanilla andLecanorchis (Cameron, 2009). Chlorophyllous Vanilla spp.have been shown to associate with a wide range of Rhizoctoniafungi, including Ceratobasidium, Thanatephorus and

Tulasnella (Porras-Alfaro and Bayman, 2007), whereas the MHvanilloid taxa Cyrtosia and Erythrorchis mainly associate withwood-decaying fungi, such as Armillaria and species ofHymenochaetaceae and Polyporaceae (Umata, 1995, 1997a;Cha and Igarashi, 1996; Dearnaley, 2007). In this study, G. fal-coneri and C. javanica were identified to associate exclusivelywith wood-decaying fungi of Meripilaceae (order Polyporales).According to the findings of Okayama et al. (2012),Lecanorchis spp. in Japan associate with a broad range of ECMfungi, including Lactarius, Russula, Atheliaceae and Sebacina,with Lactarius and Russula dominating. In this study, Russulawas identified as the preferred mycorrhizal partner of L. thalas-sica. In phylogenetic analyses, MH vanilloid orchids, i.e.Cyrtosia, Erythrorchis, Galeola, Pseudovanilla andLecanorchis, are closely related to chlorophyllous Vanilla spp.in tribe Vanilleae (Cameron, 2009). All these results togethersuggest that the nutritional shift from autotrophy to mycohe-terotrophy in vanilloid orchids correlates with shifts in fungalpartners from Rhizoctonia fungi to wood-decaying non-Rhizoctonia fungi or to ECM fungi. A shift towards either awood-decaying fungus or an ECM fungus can even happen forsympatrically growing closely related species, as in our case forC. javanica and L. thalassica. Liebel et al. (2015) recently ar-gued that this shift in fungal partners is essential for the MHmode of nutrition. The annual C and N flux from Rhizoctoniafungi to their orchid partners is rather low (Stockel et al.,2014). This low matter flux is obviously sufficient to supportgrowth of the tiny initially MH orchid protocorms, but appearsto be insufficient to support growth of adult MH orchids.

87/1·0

Polyporales sp. (AB470242)

Meripilus giganteus (AF287874)

Physisporinus vitreus (JQ031129)

Rigidoporus vinctus (AY333794)

Rigidoporus microporus (AY333795)

Hypochnicium michelii (JN939579)

Hypochnicium erikssonii (DQ677508)Hypochnicium lyndoniqe (JX124704)

Mycoacia cf. columellifera (JN710572)

Nigroporus vinosus (JN710576)

Steccherinum cf. murashkinskyi (JN710586)

Antrodiella foliaceodentata (JN710515)

Climacocystis borealis (JN710527)

Coltriciella oblectabilis (KC155387)

Hypochnicium eichleri (AJ406508)

Podoscypha elegans (JN649356)

Podoscypha sp. (JN649365)

Polyporus brumalis (AF347108)

Trametes aff. maxima (JN164802)

–/0·63

57/0·99

84/1·0

74/1·0

100/1·0

0·1

60/0·54

71/0·99

71/0·78

–/0·53

90/1·0

90/1·0

100/1·0

58/0·99

Mycobiont of Cyrtosia javanica (KP238184)

Mycobiont of Galeola falconeri (KP238181)

Mycobiont of Cyrtosia javanica (KP238182)

FIG. 5. Phylogenetic relationships of the mycorrhizal fungi of G. falconeri and C. javanica based on the Bayesian analysis of LSU ribosomal DNA sequences ofPolyporales available in GenBank (Justo and Hibbett, 2011; Binder et al., 2013). GenBank accession numbers are shown in parentheses. The values above branches

are bootstrap percentages and Bayesian posterior probabilities (>50 %), respectively. ‘–’ indicates that the node was not supported in MP analysis.


Lactarius quietus (AF096982)Arcangeliella camphorata (EU644702)Russula postiana (AF230898)

Russula turci (EF530935)Russula solaris (AF418627)Russula firmula (AF418631)

Russula claroflava (AY061665)Russula occidentalis (AY534206)Russula integra (AY061683)

Russula nitida (AY061696)

Russula azurea (AY061660)Russula lilacea (AY061731)

Russula fellea (AF418616)Russula ochroleuca (EU350580)

Russula raoultii (AF418621)

Russula mairei (AM113959)

Russula gracillima (AY061678)

Russula emetica (AY061673)Russula betularum (EU598183)

Russula raoultii (AY061712)

Russula vesca (AM113965)Russula vesca (AY606965)

Russula cremoricolor (DQ974755)

Russula heterophylla (DQ422006)Russula parazurea (DQ422007)

Russula pallescens (DQ421987)Russula pallidoapora (AY061701)Russula littoralis (AY061702)

Russula ilicis (AY061682)

Russula laurocerasi (AY061735)Russula illota (DQ422024)

Russula nigricans (EU597075)

Russula compacta (EU598172)

Russula nigricans (EF534352)

Russula romellii (AY061714)

Russula peckii (EU598174)

Russula rosea (AY061715)

Russula sphagnaphila (AY061719)Russula paludosa (AJ971402)

Russula brunneoviolacea (AM113956)Russula cuprea (AY061667)

Russula xerampelina var. xerampelina (AJ971402)

Russula decolorans (DQ367913)Russula xerampelina (AF418632)Russula favrei (EF530944)

Russula pascua (AF061705)Russula aeruginea (EU819421)

Russula rubra (AY061717)

Russula roseipes (AY061716)Mycobiont of L. virella VI-01M Kumanoe (AB597714)

Mycobiont of L. nigricans NI-11M Kiyosumi (AB597704)

Mycobiont of L. japonica var. japonica JJ-10M Hachioji (AB597663)

Mycobiont of L. japonica var. japonica JJ-10M Hachioji (AB597664)Mycobiont of L. japonica var. japonica JJ-10M Nichinan (AB597665)Mycobiont of L. japonica var. japonica JJ-13M Nichinan (AB597666)

Mycobiont of L. japonica var. japonica JJ-14M Kiyosumi (AB597667)Mycobiont of L. kiusiana var. kiusiana KK-08M Mugi (AB597682)

Mycobiont of L. trachycaula TR-02M Shishikui (AB597706)Mycobiont of L. trachycaula TR-05M Shishikui (AB597709)

Mycobiont of L. kiusiana var. kiusiana KK-06M Mugi (AB597680)

Mycobiont of L. japonica var. hokuriluensis JH-13M Jindai (AB597651)

Mycobiont of L. japonica var. hokuriluensis JH-04M Joetsu (AB597642)


Mycobiont of L. japonica var. hokurikuensis JH-11M Keta (AB597649)

Mycobiont of L. japonica var. hokurikuensis JH-10M Kanazawa (AB597648)

Mycobiont of L. japonica var. kiiensis JK-02M Tajimi (AB597671)

Mycobiont of L. flavicans var. flavicans FA-01M Mugi Oshima (AB597630)

Mycobiont of L. trachycaula TR-06M Setouchi (AB597710)

Mycobiont of L. trachycaula TR-06M Shishikui (AB597705)

Mycobiont of L. trachycaula TR-09M Amami (AB597713)

Mycobiont of L. nigricans NI-02M Shizuoka (AB597695)

Mycobiont of L. nigricans NI-10M Kiyosumi (AB597703)




Mycobiont of L. thalassica (KP238187)

Mycobiont of L. kiusiana var. kiusiana KK-07M Mugi (AB597681)

0·1

–/0·75

79/1·096/1·0

68/0·99

83/0·98

56/0·99

100/1·0

98/1·0

95/1·0–/0·99

–/0·81

–/0·77

62/0·77

60/0·89

61/0·96

71/0·91

64/0·97

–/0·98

–/0·54

–/0·66

–/0·96

–/0·68

–/0·67

–/1·0

–/0·53

–/0·93

71/0·67

53/0·99

74/0·99

56/0·61

76/0·97

61/0·51

96/1·0

93/1·0

100/1·0

100/1·0

100/1·0

100/1·0

100/0·52

100/1·0

100/1·0100/1·0

100/1.0

100/1·0

86/1·0

89/1·0

–/0·82

–/0·52

51/0·98

94/1·0

100/1·0

100/1·0

100/1·0

100/1·0

100/1·0

99/1·0

86/1·0

100/1·0

91/0·88

63/0·89

98/1·0

80/0·99

97/1·0

–/0·59

FIG. 6. Phylogenetic relationships of the mycorrhizal fungi of Lecanorchis thalassica based on the Bayesian analysis of partial ITS ribosomal DNA sequences ofRussulaceae available in GenBank (with the position of mycorrhizal fungi found in Lecanorchis by Okayama et al., 2012). GenBank accession numbers are shownin parentheses. The values above branches are bootstrap percentages and Bayesian posterior probabilities (>50 %), respectively. ‘–’ indicates that the node was not

supported in MP analysis.


Mycena polygramma (DQ071707)Mycena polygramma (AY207243)Mycena leaiana (AF261411)Mycena galericulata (AF042636)Mycena crocata (AY207241)Mycena aurantiomarginata (AY207246)Mycena niveipes (AY207242)Mycena leptocephala (AY207253)Mycena renati (AY207256)Mycena citrinomarginata (AY207248)Mycena olivaceomarginata (AY207255)Mycena galopus (AY207250)Mycena sanguinolenta (AY207257)Mycena haematopus (AY207252)Mycena maculata (AY207254)Mycena insignis (AF261413)Mycena calavicularis (AF042637)Mycena tintinnabulum (AY207258)Mycena epipterygia (AY207249)

Mycena plumbea (DQ470813)

Mycenoporella griseipora (AF261428)

Mycena galericulata (AY647216)Mycena galericulata (AY207251)Mycena rubromarginata (AY207245)Mycena zephirus (AY207259)

Mycena aff. pura (AY261409)Mycena aff. pura (AF261410)

Mycena aff. pura (AY207244)

Mycena aff. pura (DQ457688)Mycena aff. pura (DQ457689)Mycena aff. pura (DQ457690)Cotobrusia calostomoides (AF261424)

Mycobiont of Gastrodia confusa (AB454414)


Mycobiont of Wullschlaegelia aphylla (FJ179476)

Mycobiont of Wullschlaegelia aphylla (FJ179475)Marasmius androsaceus (AF261585)

Marasmius scorodonius (AF261332)Marasmius alliaceus (AY635776)Marasmius alliaceus (AY639436)Marasmius copelandii (AY639438)

Marasmius oreades (DQ156126)Marasmius rotula (DQ457686)Henningsomyces candidus (AF287864)Nia vibrissa (AF334750)Cyphella digitalis (AY635771)

Marasmiellus opacus (AF261330)

Marasmiellus palmivorus (AY639434)

Marasmiellus synodicus (AY639435)

Mycobiont of Gastrodia similis (FJ179468)

Mycobiont of Gastrodia appendiculata (KP238191)Mycobiont of Gastrodia nantornsis (KP238188)Mycobiont of Gastrodia fontinalis (KP238190)

Mycobiont of Gastrodia flabilabella (KP238192)

Mycobiont of Gastrodia fontinalis (KP238189)

Mycena rutilanthiformis (AF042606)

Mycena amicta (DQ457692)

Mycena capillaripes (AY207247)Tricholoma matsutake (U62964)

Entoloma prunuloides (AY700180)Catathelasma ventricosum (DQ089012)Lyophyllum decastes (AF042583)

Hydropus cf scabripes (DQ411536)Hydropus marginellus (DQ457674)Megacollybia platyphylla (AY635778)Gymnopus aff moseri (AY639409)

Gymnopus aff menehune (AY639408)

Gymnopus nonnullus var attenuatus (AY639426)

Gymnopus melanopus (AY639422)

Gymnopus brunneigracilis (AY639412)Gymnopus gibbosus (AY639415)Gymnopus termiticola (AY639430)

Gymnopus fusioes (AY256710)Gymnopus bicolor (AY639411)Gymnopus sepiiconicus (AY639427)Gymnopus spissus (AY639428)Rhodocollybia maculata (AY639880)Rhodocollybia butyracea (EU486454)

Gymnopus indoctus (AY639418)

Cruentomycena viscidocruenta (AF261414)Dictyopanus pusillus (AF261425)Dictyopanus sp. (AF261425)

Poromycena sp. (AF261429)

Poromycena sp. (AF261421)

Poromycena manipularis (AF261423)

Fiboboletus gracilis (AF261422)


Panellus stypticus (AF261427)Resinomycena acadiensis (AF042638)Resinomycena rhododendri (AF261415)

0·1

100/0·74

100/0·58

100/0·89

100/0·99

100/0·99

74/0·8194/0·68

52/0·96

88/0·78

64/0·9561/0·97

58/1·0

99/1·0

100/1·0

81/1·0

86/1·0

99/1·0

93/1·0

75/0·92

64/0·9755/0·59

100/0·97

100/0·76

100/1·0

94/1·0

100/1·0

100/0·95

100/0·87

63/0·55

–/0·57

–/0·52

–/0·56

–/0·57

–/0·9

–/0·72

–/0·62

–/0·99

–/0·92

–/0·99

–/0·51

–/0·97–/0·67

–/0·98

–/0·97

–/0·6–/0·91

–/0·7753/1·0

91/0·65

99/1·0

82/1·0

80/0·88

99/1·0

55/1·0

56/0·99

66/0·98

60/0·74

71/1

–/0·67

–/0·51–/0·89

51/0·51

95/1·0

75/1·0

91/1·0

84/1

FIG. 7. Phylogenetic relationships of the mycorrhizal fungi of Gastrodia appendiculata, G. fontinalis, G. flabilabella and G. nantoensis based on the Bayesian analy-sis of LSU ribosomal DNA sequences of marasmioid available in GenBank (Moncalvo et al., 2000, 2002; Wilson and Desjardin, 2005; Matheny et al., 2006; Martoset al., 2009; Ogura-Tsujita et al., 2009). The values above branches are bootstrap percentages and Bayesian posterior probabilities (>50 %), respectively. ‘–’ indi-

cates that the node was not supported in MP analysis.


Despite the fairly high number of fully MH species in subfam-ily Vanilloideae, no partially MH species have been reported sofar to occur among green species of Vanilloideae.

The four Gastrodia spp. investigated here also fall into twogroups with respect to their mycorrhizal partners. Mycena dom-inates as fungal host in the three species growing sympatricallyin a bamboo forest (G. appendiculata, G. fontinalis and G. nan-toensis). Furthermore, Gymnopus was detected in the mycorrhi-zal roots of G. fontinalis. It is worth mentioning that the threeGastrodia species produce their fruit bodies at different timesin the year. Gastrodia fontinalis produces fruit bodies in springand G. nantoensis and G. appendiculata appear successively inOctober. Note that G. fontinalis associates with type I Mycenaand Gymnopus, G. nantoensis only associates with type IIIMycena, and G. appendiculata recruits Mycena types II and III.Thus, the results demonstrate a preference towards different lin-eages of Mycena by three sympatric Gastrodia spp. In G. con-fusa from Japan, an association with various types of Mycenawas found in the same and in different populations, suggestingan additional geographical mosaic of mycorrhizal specificity(Ogura-Tsujita et al., 2009).

The fourth Gastrodia sp., growing in a coniferous forest(G. flabilabella), has Hydropus as a mycorrhizal partner.Unlike the other three Gastrodia spp., G. flabilabella forms itsown clade (Hsu, 2008) and mainly occurs in coniferous forests.The change of mycorrhizal fungal associations may reflect notonly the independent evolution of Gastrodia lineages, but alsoa switch in forest types as their habitats. Previous studies haveshown that Gastrodia spp. associate with litter- and wood-decomposing SAP basidiomycetes of the genera Armillaria,Mycena, Resinicium and Campanella or Marasmius (Kusano,

1911; Martos et al., 2009; Ogura-Tsujita et al., 2009; Dearnaleyand Bougoure, 2010). Here we provide molecular evidence thatGymnopus and Hydropus are additional mycorrhizal partners ofGastrodia spp., and this broadens the diversity of SAP basidio-mycetes associating with Gastrodia. Gastrodia is one of thelargest MH genera, consisting of >50 fully MH species, distrib-uted in Asia, Australia and Africa (Hsu, 2008), and the recruit-ment of various basidiomycetes as mycorrhizal partners needsfurther investigation. All fungal hosts so far known of allGastrodia spp. so far investigated are either litter- or wood-decaying non-Rhizoctonia SAP fungi. A better knowledgeabout the preferred substrates (litter or wood) of the respectivefungal hosts would be desirable for the future in order to allowan even more comprehensive use of information available fromstable isotope natural abundance (see below).

Carbon and nitrogen sources and total nitrogen concentrations

All seven of the MH study orchids turned out to be signifi-cantly enriched in the heavy isotopes 13C and 15N in compari-son with autotrophic plants growing in identical habitats. Thisfinding confirms that fungi also known to be enriched in 13Cand 15N (Gebauer and Dietrich, 1993; Gleixner et al., 1993)serve as their C and N source. Though leafless, some putativeMH orchids have green flowering stems and thus a low photo-synthetic activity. One leafless orchid with frequently greenflowering stems and green seed capsules is Corallorhiza trifida.This species is associated with an ECM fungus and is less en-riched in 13C and 15N than fully MH orchids also associatedwith ECM fungi (Zimmer et al., 2008). Thus, C. trifida hasbeen classified as being partially MH. All study orchids of thisinvestigation provided no indications for chlorophyllous flower-ing stems and had isotope signatures typical of various types offully MH plants. According to their pattern in 13C and 15N en-richment, the MH orchids fall into three groups. (1) The ECM-associated L. thalassica is the species most enriched in 15N.This pattern is in agreement with current knowledge (Hynsonet al., 2013) and can be traced back to the fact that ECM fungiare more enriched in 15N than sympatrically growing wood- orlitter-decaying SAP fungi (Gebauer and Taylor, 1999). (2) Thetwo orchid species associated with wood-decayingMeripilaceae, Galeola falconeri and Cyrtosia javanica, belongto the group most enriched in 13C. The specifically high 13C en-richment of this group of MH orchids is expected due to thefact that wood is more enriched in 13C than leaf tissue (Gebauerand Schulze, 1991; Cernusak et al., 2009) and confirms previ-ous findings of similar patterns in MH orchids associated withwood-decaying fungi (Ogura-Tsujita et al., 2009; Martos et al.,2009; Hynson et al., 2013). The third MH orchid belonging tothis group is the Hydropus-associated Gastrodia flabilabella,suggesting that this Hydropus sp. should be a wood-decayingfungus. (3) The third MH orchid cluster characterized by a sig-nificantly lower 15N enrichment than (1) and a significantlylower 13C enrichment than (2) is composed of the threeGastrodia spp. growing sympatrically in a bamboo forest,G. appendiculata, G. fontinalis and G. nantoensis. The isotopicpattern of this cluster of MH orchids provides strong evidencethat their mycorrhizal associates, Mycena of three differenttypes and Gymnopus, are all litter-decaying fungi. This

12

12

10

10

8

8

6

6

4

4

2

2

0

0

Enr

ichm

ent f

acto

r e

15N

(%

o)

Enrichment factor e 13C (%o)

–2

–2

Ref

Lecanorchis thalassica

G. nantoensis

Galeola falconeri

G. fontinalis

G. flabilabella

Cyrtosia javanica

G. appendiculata

Marasmius sp.

FIG. 8. Mean enrichment factors e13C and e15N 6 1 s.d. as calculated for six MHorchid species associated with SAP non-Rhizoctonia fungi (triangles), onemycoheterotrophic orchid species associated with a fungus forming ECM (filledsquare), one litter-decaying SAP fungus (open circle) (n¼ 5 each) and of photo-synthetic reference plants (Ref, n¼ 90, see Supplementary Data Table S2) col-lected together with each of the orchids at four sites in the Xitou ExperimentalForest, Nantou County, Taiwan. Please note that the orchids associated withSAP non-Rhizoctonia fungi fall into two significantly distinct groups, Galeolafalconeri, Cyrtosia javanica, Gastrodia flabilabella (filled triangles) andGastrodia fontinalis, G. nantoensis and G. appendiculata (open triangles). Bothof these groups are significantly different from Lecanorchis thalassica, the or-chid associated with an ECM fungus. Mean e values of reference plants are zero

by definition. The box represents 6 1 s.d. for the reference plants.



evidence is further strengthened by the rather similar isoto-pic pattern found for the fruit bodies of the litter-decayingfungus Marasmius sp. This is the first report we are awareof that indicates a significantly different isotopic patternfor MH orchids associated with either wood-decaying orlitter-decaying fungi. The only hint towards an isotopicdistinction between MH orchids associated with wood- orlitter-decaying fungi was previously given by Martos et al.(2009), who investigated the isotopic compositions ofthe MH orchids Gastrodia similis associated with wood-decaying Resinicium from La Reunion andWullschlaegelia aphylla mycorrhizal orchids with litter-decaying Gymnopus and Mycena from Guadeloupe.Unfortunately, comparisons of the data from the study ofMartos et al. with each other and with our investigation basedon normalized enrichment factors is not possible due to a lackof stable isotope data for forest ground plants growing in mi-cro-environments identical to the MH orchids.

Total N concentrations in MH orchids have been reported tobe significantly higher than in the majority of co-occurring au-totrophic plants (Gebauer and Meyer, 2003; Liebel et al., 2010;Liebel and Gebauer, 2011; Sommer et al., 2012). The mostlikely reason for these unusually high total N concentrations inMH orchid tissues is their N gain through the fungal source.Fungi are known to have considerably higher total N concentra-tions in their tissue than autotrophic plants growing in the sameenvironment (Gebauer and Dietrich, 1993; Gebauer and Taylor,1999). Total N concentrations found in the flower stems of theMH orchid species investigated here (2�88 6 0�50 mmol gd. wt–1) are rather similar to the total N concentrations reportedfor other fully or partially MH orchids (Gebauer and Meyer,2003; Liebel et al., 2010; Liebel and Gebauer, 2011; Sommeret al., 2012). The reference plant leaf total N concentrations in-vestigated here (2�67 6 0�67 mmol g d. wt–1) were only slightlylower than N concentrations in the flower stems of the MH or-chids and statistically not distinguishable. Total N concentra-tions in the leaves of the understorey plants investigated herefrom four different forests in Taiwan are about twice as high asleaf total N concentrations in autotrophic forest ground plantsfrom temperate Central Europe (Gebauer et al., 1988; Gebauerand Meyer, 2003) and Mediterranean Italy (Liebel et al., 2010),and three times higher than leaf total N concentrations in auto-trophic forest ground plants from severely N-limited regions,such as SW Australia (Sommer et al., 2012) or boreal Norway(Liebel and Gebauer, 2011). The most likely reason for thehigh total N concentrations in the leaves of autotrophic forestground plants in Taiwan is a high mineral N availability due tohigh soil N mineralization rates. Thus, N limitation as a driverfor the switch from autotrophy towards mycoheterotrophy isunlikely for the MH orchids investigated here. More likely as adriving factor for the development of mycoheterotrophy is lightlimitation for photosynthesis on the forest ground. As reportedfor other habitats of MH plants (Bidartondo et al., 2004;Zimmer et al., 2007; Preiss et al., 2010) only 2–6 % of incom-ing light reached the ground of our study forests and thus se-verely limited the photosynthetic capacity of forest groundvegetation.

In conclusion, our data provide further evidence for the im-portance of associations with non-Rhizoctonia SAP fungiamong fully MH orchids from sub-tropical Asia. Abundant

fallen litter and wood in the warm and humid forests of Taiwanobviously supplies ideal substrates for the continuous growth ofSAP fungi. For the MH orchids studied here, as seeds germi-nate, they can construct efficient mycorrhizal interactions pref-erably with wood- or litter-decaying SAP fungal partners, butsympatrically also with ECM fungi, in order to thrive in deeplyshaded forest understoreys with low light conditions.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxford-journals.org and consist of the following. Table S1: voucherand sampling sites of seven mycoheterotrophic orchids in XitouExperimental Forest used in molecular identification of mycor-rhizal fungi. Table S2: mean d15N and d13C values, mean en-richment factors e15N and e13C, and mean total Nconcentrations of all plant and fungal samples in this study.Table S3: putative taxonomic identity of the fungi detected inthis study. Table S4: ITS sequence types in Mycena from my-corrhizal roots of three sympatric Gastrodia spp. in a bambooforest. Figure S1: morphology of rhizomes and roots of myco-heterotrophic orchids in this study.

ACKNOWLEDGEMENTS

The authors thank Christine Tiroch (BayCEER–Laboratory ofIsotope Biogeochemistry, University of Bayreuth) for skilfultechnical assistance with stable isotope abundance measure-ments. The authors also thank Yu-Hsiu Cho (NationalMuseum of Natural Science) for molecular identifications offungal partners, and Sheng-Kun Yu for providing the photo ofLecanorchis thalassica. This work is a contribution to theGerman Research Foundation Project GE 565/7-2. This workwas also supported by the funding from National Museum ofNatural Science, Taiwan to Y.-I.L.

LITERATURE CITED

Akaike H. 1974. A new look at the statistical model identification. IEEETransactions on Automatic Control 19: 716–723.

Bidartondo MI, Burghardt B, Gebauer G, Bruns TD, Read DJ. 2004.

Changing partners in the dark: isotopic and molecular evidence of ectomy-corrhizal liaisons between forest orchids and trees. Proceedings of the RoyalSociety B: Biological Sciences 271: 1799–1806.

Binder M, Justo A, Riley R, et al. 2013. Phylogenetic and phylogenomic over-view of the Polyporales. Mycologia 105: 1350–1373.

Cameron KM. 2009. On the value of nuclear and mitochondrial gene sequencesfor reconstructing the phylogeny of vanilloid orchids (Vanilloideae,Orchidaceae). Annals of Botany 104: 377–385.

Cernusak LA, Tcherkez G, Keitel C, et al. 2009. Why are non-photosynthetictissues generally 13C enriched compared with leaves in C3 plants? Reviewand synthesis of current hypotheses. Functional Plant Biology 36: 199–213.

Cha JY, Igarashi T. 1996. Armillaria jezoensis, a new symbiont of Galeola sep-tentrionalis (Orchidaceae) in Hokkaido. Mycoscience 37: 21–24.

Dearnaley JDW. 2007. Further advances in orchid mycorrhizal research.Mycorrhiza 17: 475–486.

Dearnaley JDW, Bougoure JJ. 2010. Isotopic and molecular evidence for sap-rotrophic Marasmiaceae mycobionts in rhizomes of Gastrodia sesamoides.Fungal Ecology 3: 288–294.

Dearnaley JDW, Martos F, Selosse M-A. 2012. Orchid mycorrhizas:molecular ecology, physiology, evolution and conservation aspects. In: BHock, ed. Fungal associations. The mycota IX, 2nd edn. Berlin: SpringerVerlag, 207–230.



www.aob.oxfordjournals.org

www.aob.oxfordjournals.org

Esnault AL, Masuhara G, McGee PA. 1994. Involvement of exodermal pas-sage cells in mycorrhizal infection of some orchids. Mycological Research98: 672–676.

Felsenstein J. 1985. Confidence limits on phylogenies: an approach using thebootstrap. Evolution 39: 783–791.

Fry B. 2006. Stable isotope ecology. New York: Springer.Gardes M, Bruns TD. 1993. ITS primers with enhanced specificity for basidio-

mycetes: application to the identification of mycorrhizae and rusts.Molecular Ecology 2: 113–118.

Gebauer G, Dietrich P. 1993. Nitrogen isotope ratios in different compartmentsof a mixed stand of spruce, larch and beech trees and of understorey vegeta-tion including fungi. Isotopenpraxis 29: 35–44.

Gebauer G, Meyer M. 2003. 15N and 13C natural abundance of autotrophic andmyco-heterotrophic orchids provides insight into nitrogen and carbon gainfrom fungal association. New Phytologist 160: 209–223.

Gebauer G, Schulze ED. 1991. Carbon and nitrogen isotope ratios in differentcompartments of a healthy and a declining Picea abies forest in theFichtelgebirge, NE Bavaria. Oecologia 87: 198–207.

Gebauer G, Taylor AFS. 1999. 15N natural abundance in fruit bodies of differ-ent functional groups of fungi in relation to substrate utilization. NewPhytologist 142: 93–101.

Gebauer G, Rehder H, Wollenweber B. 1988. Nitrate, nitrate reduction and or-ganic nitrogen in plants from different ecological and taxonomic groups ofCentral Europe. Oecologia 75: 371–385.

Gleixner G, Danier HJ, Werner RA, Schmidt HL. 1993. Correlations betweenthe 13C content of primary and secondary plant products in different cellcompartments and that in decomposing basidiomycetes. Plant Physiology102: 1287–1290.

Hamada M. 1939. Studien uber die Mykorrhiza von Galeola septentrionalisReichb. F. – ein neuer Fall der Mykorrhiza-Bildung durch intraradicaleRhizomorpha. Japanese Journal of Botany 10: 151–211.

Holm S. 1979. A simple sequentially rejective multiple test procedure.Scandinavian Journal of Statistics 6: 65–70.

Hsu TC. 2008. Taxonomy of Gastrodia (Orchidaceae) in Taiwan. Master thesis,National Taiwan University, Taipei.

Hynson NA, Madsen TP, Selosse M-A, et al. 2013. The physiological ecologyof mycoheterotrophy. In: VSFT Merckx, ed. Mycoheterotrophy: the biologyof plants living on fungi. Berlin: Springer Verlag, 297–342.

Justo A, Hibbett DS. 2011. Phylogenetic classification of Trametes(Basidiomycota, Polyporales) based on a five-marker dataset. Taxon 60:1567–1583.

Kikuchi G, Higuchi M, Morota T, Nagasawa E, Suzuki A. 2008. Fungal sym-biont and cultivation test of Gastrodia elata Blume (Orchidaceae). Journalof Japanese Botany 83: 88–95.

Kusano S. 1911. Gastrodia elata and its symbiotic association with Armillariamellea. Journal of the College of Agriculture Imperial University of Tokyo4: 1–65.

Leake JR. 1994. The biology of myco-heterotrophic (‘saprophytic’) plants. NewPhytologist 127: 171–216.

Leake JR. 2004. Myco-heterotroph/epiparasitic plant interactions with ectomy-corrhizal and arbuscular mycorrhizal fungi. Current Opinions in PlantBiology 7: 422–428.

Liebel HT, Gebauer G. 2011. Stable isotope signatures confirm carbon and ni-trogen gain through ectomycorrhizas in the ghost orchid Epipogium aphyl-lum Swartz. Plant Biology 13: 270–275.

Liebel HT, Bidartondo MI, Preiss K, et al. 2010. C and N stable isotopesignatures reveal constraints to nutritional modes in orchidsfrom the Mediterranean and Macaronesia. American Journal of Botany 97:903–912

Liebel HT, Bidartondo MI, Gebauer G. 2015. Are carbon and nitrogen ex-change between fungi and the orchid Goodyera repens affected by irradi-ance? Annals of Botany 115: 251–261.

Martos F, Dulormne M, Pailler T, et al. 2009. Independent recruitment of sap-rotrophic fungi as mycorrhizal partners by tropical achlorophyllous orchids.New Phytologist 184: 668–681.

Matheny PB, Curtis JM, Hofstetter V, et al. 2006. Major clades of Agaricales:a multilocus phylogenetic overview. Mycologia 98: 982–995.

Merckx VSFT. 2013. Mycoheterotrophy: an introduction. In: VSFT Merckx, ed.Mycoheterotrophy: the biology of plants living on fungi. Berlin: SpringerVerlag, 1–17.

Moncalvo JM, Lutzoni FM, Rehner SA, Johnson J, Vilgalys R. 2000.

Phylogenetic relationships of agaric fungi based on nuclear large subunitribosomal DNA sequences. Systematic Biology 49: 278–305.

Moncalvo JM, Vilgalys R, Redhead SA, et al. 2002. One hundred andseventeen clades of euagarics. Molecular Phylogenetics and Evolution 23:357–400.

Nylander JAA. 2004. MrModeltest 2.2. Program distributed by the author.Evolutionary Biology Centre, Uppsala University.

Ogura-Tsujita Y, Yukawa T. 2008. High mycorrhizal specificity in a wide-spread mycoheterotrophic plant, Eulophia zollingeri (Orchidaceae).American Journal of Botany 95: 93–97.

Ogura-Tsujita Y, Gebauer G, Hashimoto T, Umata H, Yukawa T. 2009.

Evidence for novel and specialized mycorrhizal parasitism: the orchidGastrodia confusa gains carbon from saprotrophic Mycena. Proceedings ofthe Royal Society B: Biological Sciences 276: 761–767.

Okayama M, Yamato M, Yagame T, Iwase K. 2012. Mycorrhizal diversityand specificity in Lecanorchis (Orchidaceae). Mycorrhiza 22: 545–553.

Pate JS, Gunning BES. 1972. Transfer cells. Annual Review of PlantPhysiology 23: 173–196.

Porras-Alfaro A, Bayman P. 2007. Mycorrhizal fungi of Vanilla: diversity, spe-cificity and effects on seed germination and plant growth. Mycologia 99:510–525.

Preiss K, Gebauer G. 2008. A methodological approach to improve estimatesof nutrient gains by partially myco-heterotrophic plants. Isotopes inEnvironmental and Health Studies 44: 393–401.

Preiss K, Adam IKU, Gebauer G. 2010. Irradiance governs exploitationof fungi: fine-tuning of carbon gain by two partially myco-heterotrophicorchids. Proceedings of the Royal Society B: Biological Sciences 277:1333–1336.

Rasmussen HN. 1995. Terrestrial orchids – from seed to mycotrophic plant.Cambridge: Cambridge University Press.

Ronquist F, Huelsenbeck JP. 2003. MRBAYES 3: Bayesian phylogenetic in-ference under mixed models. Bioinformatics 19: 1572–1574.

Roy M, Whatthana S, Stier A, Richard F, Vessabutr S, Selosse M-A. 2009.

Two mycoheterotrophic orchids from Thailand tropical dipterocarpaceanforests associate with a broad diversity of ectomycorrhizal fungi. BMCBiology 7: 51.

Sommer J, Pausch J, Brundrett MC, Dixon KW, Bidartondo MI, Gebauer

G. 2012. Limited carbon and mineral nutrient gain from mycorrhizal fungiby adult Australian orchids. American Journal of Botany 99: 1133–1145.

Stockel M, Tesitelova T, Jersakova J, Bidartondo MI, Gebauer G. 2014.

Carbon and nitrogen gain during the growth of orchid seedlings in nature.New Phytologist 202: 606–615.

Stover BC, Muller KF. 2010. TreeGraph 2: combining and visualizing evidencefrom different phylogenetic analyses. BMC Bioinformatics 11: 7.

Su HJ. 2000. Orchidaceae. In TC Huang, ed. Flora of Taiwan, 2nd edn, Vol. 5.Editorial Committee of the Flora of Taiwan, Department of Botany,National Taiwan University, Taipei, Taiwan, 729–1086.

Swofford DL. 2002. PAUP*: phylogenetic analysis using parsimony (*and othermethods), version 4.0b10 for Macintosh. Sunderland, MA: SinauerAssociates.

Taylor DL, Bruns TD. 1997. Independent, specialized invasions of ectomycor-rhizal mutualism by two non-photosynthetic orchids. Proceedings of theNational Academy of Sciences of the USA 94: 4510–4515.

Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. 1997.

The Clustal X windows interface. Flexible strategies for multiple se-quence alignment aided by quality analysis tools. Nucleic Acids Research24: 4876–4882.

Trudell SA, Rygiewicz PT, Edmonds RL. 2003. Nitrogen and carbon stableisotope abundances support the myco-heterotrophic nature and host-specif-icity of certain achlorophyllous plants. New Phytologist 160: 391–401.

Umata H. 1995. Seed germination of Galeola altissima, an achlorophyllous or-chid, with Aphyllophorales fungi. Mycoscience 36: 369–372.

Umata H. 1997a. Formation of endomycorrhizas by an achlorophyllous orchid,Erythrorchis ochobiensis, and Auricularia polytricha. Mycoscience 38:335–339.

Umata H. 1997b. In vitro germination of Erythrorchis ochobiensis(Orchidaceae) in the presence of Lyophyllum shimeji, an ectomycorrhizalfungus. Mycoscience 38: 355–357.

Vilgalys R, Hester M. 1990. Rapid genetic identification and mapping of enzy-matically amplified ribosomal DNA from several Cryptococcus species.Journal of Bacteriology 172: 4238–4246.

White TJ, Bruns TD, Lee S, Taylor JW. 1990. Amplification and direct se-quencing of fungal ribosomal RNA genes for phylogenetics. In: MA Innis,DH Gelfand, JJ Sninsky, TJ White, eds. PCR protocols: a guide to methodsand application. San Diego: Academic Press, 315–322.


Wilson AW, Desjardin DE. 2005. Phylogenetic relationships in the gymnopoidand marasmioid fungi (Basidiomycetes, euagarics clade). Mycologia 97:667–679.

Yamato M, Yagame T, Suzuki A, Iwase K. 2005. Isolation and identificationof mycorrhizal fungi associating with an achlorophyllous plant, Epipogiumroseum (Orchidaceae). Mycoscience 46: 73–77.

Yang CK, Lee YI, Chen YF, Chung NJ, Wang YN, Chung KF. 2010. Studyon species diversity and conservation of wild orchids at Xitou TractExperimental Forest of National Taiwan University. Journal of theExperimental Forest of National Taiwan University 24: 123–136. (in Chinese)

Yang Z, Rannala B. 1997. Bayesian phylogenetic inference using DNA se-quences: a Markov chain Monte Carlo method. Molecular Biology andEvolution 14: 717–724.

Yeung EC. 1984. Histological and histochemical staining procedures. In: IKVasil, ed. Cell culture and somatic cell genetics of plants. Orlando:Academic Press, 689–697.

Zimmer K, Hynson NA, Gebauer G, Allen EB, Allen MF, Read DJ. 2007.

Wide geographical and ecological distribution of nitrogen and carbon gainsfrom fungi in pyroloids and monotropoids (Ericaceae) and in orchids. NewPhytologist 175: 166–175.

Zimmer K, Meyer C, Gebauer G. 2008. The ectomycorrhizal specialist orchidCorallorhiza trifida is a partial myco-heterotroph. New Phytologist 178:395–400.


part of a highlight on orchid biology...in temperate forests, mh orchids usually associate with...

Documents