American Journal of Botany 97(8): 1313–1327. 2010. A NARROW GROUP OF MONOPHYLETIC TULASNELLA (TULASNELLACEAE) SYMBIONT LINEAGES ARE ASSOCIATED WITH MULTIPLE SPECIES OF C HILOGLOTTIS (ORCHIDACEAE): IMPLICATIONS FOR ORCHID DIVERSITY 1 Sean A. Roche, Richard J. Carter, Rod Peakall, Leon M. Smith, Michael R. Whitehead, and Celeste C. Linde 2 Evolution, Ecology and Genetics, Research School of Biology, 116 Daley Road, The Australian National University, Canberra, ACT 0200, Australia • Premise of the study: The Orchidaceae is characterized by exceptional species diversity. Obligate orchid mycorrhizae are pre- dicted to determine orchid distributions, and highly specifi c relationships between orchids and fungi may drive orchid diversi- fi cation. In this study, mycorrhizal diversity was examined in the terrestrial, photosynthetic orchid genus Chiloglottis to test the hypothesis of mycorrhizal-mediated diversifi cation in the genus Chiloglottis. This orchid genus secures pollination by sexual deception, an obligate and highly specifi c pollination strategy. Here we asked whether the obligate orchid – fungal interactions are also specifi c. • Methods: Two sequenced loci, the internal transcribed spacer region (ITS) and mitochondrial large subunit (mtLSU), were used to identify fungal isolates and assess fungal species diversity. Symbiotic germination of two species Chiloglottis aff. jeanesii and C. valida were used to assess germination potential of isolates and confi rm mycorrhizal association. • Key results: Phylogenetic analyses revealed that six representative Chiloglottis species spanning a broad survey of the genus were all associated with a narrow group of monophyletic Tulasnella fungal lineages. • Conclusions: The Chiloglottis – Tulasnella interaction appears to be the fi rst known case of such a narrow symbiont association across a broadly surveyed orchid genus. It appears that the specifi c pollination system of Chiloglottis , rather than specifi c or- chid – fungal interactions has been the key driving force in the diversifi cation of the genus. These fi ndings also indicate that plant groups with highly specifi c mycorrhizal partners can have a widespread distribution. Key words: ectomycorrhizae; Chiloglottis ; ITS; mtLSU; molecular phylogeny; orchid mycorrhizal fungi; Orchidaceae; Tulasnellaceae. Coevolution involving intimate interactions between two dif- interactions is therefore critical for a more complete under- ferent groups of organisms is proposed to be a key driving force standing of the evolution of orchid diversity. in speciation, especially when relationships display a high level The minute, wind-dispersed seeds of orchids are undifferen- of specifi city such as host – parasite relationships and specifi c tiated and do not store suffi cient nutrition for unaided develop- pollination systems ( Barrett, 1986 ; Thompson, 1986 ; Weiblen, ment of the embryo ( Arditti, 1992 ). Thus, all orchids require 2004 ). The Orchidaceae represents one of the most diverse mycorrhizal fungi, at least for germination and early seedling plant families, and it has been proposed that this diversity could, phases ( Warcup, 1990 ; Arditti, 1992 ). During the early seedling in part, be driven by the often-specifi c interactions with animal phase, they may be considered parasitic on their fungal symbi- pollinators ( Peakall, 2007 ; Scopece et al., 2007 ; Peakall et al., onts because nutrients are directed exclusively to the plant 2010 ). It has been predicted that this diversity has been ( Hadley, 1970 ; Waterman and Bidartondo, 2008 ; Rasmussen strongly infl uenced by the unique features of orchid reproduc- and Rasmussen, 2009 ). Even as photosynthetic adults, many tion and the associated specifi c interactions with insect pollina- terrestrial species may still rely on fungal-mediated transfer of tors ( Cozzolino and Widmer, 2005 ). The below-ground orchid nutrients and carbon, especially in forests where light can be interactions involving symbiotic fungi, which in some cases limited ( Bidartondo, 2005 ; Cameron et al., 2006 ). Whether this demonstrate high levels of specifi city, may have also played relationship remains parasitic in later growth stages may vary a key role in speciation within certain orchid groups ( Taylor according to the photosynthetic ability of the orchid ( Cameron et al., 2003 ; Otero and Flanagan, 2006 ; Waterman and Bidartondo, et al., 2006 , 2009 ). 2008 ). Documenting the underlying patterns of orchid-fungal Most orchids are photosynthetic, and there is no evidence that they obtain a majority of their energy from the fungi, ex- cept at the early stages in development following germination 1 ( Taylor, 2004 ; Rasmussen and Rasmussen, 2009 ). However, Manuscript received 8 February 2010; revision accepted 19 May 2010. mutualistic orchid – fungal associations have been confi rmed by The authors thank Prof. Kingsley Dixon (Director of Science, Kings Park and Botanic Gardens, Perth) for advice on seed germination radioactively labeled carbon and phosphorous studies, where classifi cation. bidirectional fl ow of carbon has been observed between fungal 2 Author for correspondence (e-mail: [email protected]) symbionts and the orchid Goodyera repens ( Cameron et al., 2006 , 2007 ). More species need to be investigated to determine doi:10.3732/ajb.1000049 whether this bidirectional nutrient fl ow is a general pattern in American Journal of Botany 97(8): 1313–1327, 2010; http://www.amjbot.org/ © 2010 Botanical Society of America 1313 1314 American Journal of Botany [Vol. 97 the trophic interactions between green orchids and fungi. If so, As is generally the case for sexually deceptive orchids, polli- an association of mycorrhizae with orchids may provide a long- nator interactions with Chiloglottis are highly specifi c; a one to term benefi t to both partners, and the exchange of limiting com- one relationship between orchid and pollinator is typical for a pounds may broaden the ecological range of either partner and given site ( Bower, 1996 , 2006 ; Bower and Brown, 2009 ), al- overcome physiological stresses ( Arditti, 1992 ; Bidartondo, though some pollinator sharing the edge of a species range is 2005 ). Interactions between photosynthetic orchids and their known ( Peakall et al., 2002 ). The discovery of the specifi c chem- mycorrhizae may therefore represent a complex and dynamic icals representing a new class of compounds used by Chiloglottis relationship between parasitism and mutualism, which can orchids to lure their male pollinators ( Schiestl et al., 2003 ; Franke change across the life cycle of the host/symbiont and in response et al., 2009 ; Peakall et al., 2010 ) confi rms that pollinator specifi c- to changes in environmental conditions ( Harley and Smith, 1983 ; ity has a strong chemical basis. It has been hypothesized that Arditti, 1992 ; Hynson et al., 2009 ). changes in fl oral chemistry that enable pollinator switching, com- Highly specifi c relationships with obligate fungal partners bined with subsequent pollinator-mediated selection, have played have been proposed as a possible driving force in the diversifi ca- a key role in the evolution and diversifi cation of Chiloglottis tion of some photosynthetic orchids ( Otero and Flanagan, 2006 ; ( Mant et al., 2002 , 2005 ; Peakall et al., 2010 ). Shefferson et al., 2007 ), nonphotosynthetic orchids ( Taylor et al., Despite this extensive work on pollinator specifi city, it is im- 2003 ), and members of the Monotropoideae ( Bidartondo and portant that the potential role of below-ground interactions — the Bruns, 2001 ). However, the occurrence of species-specifi c or- obligate interaction between orchid and fungal mycorrhizal as- chid – fungal relationships is unlikely to be a consequence of sociates ( Waterman and Bidartondo, 2008 ) — is not neglected in cospeciation (reciprocal speciation as a consequence of the in- studies of orchid diversifi cation. Therefore, the overarching teraction) ( Thompson, 1986 ; Taylor et al., 2003 ) because orchid question we ask in this study is what role, if any, might specifi c mycorrhizae are not dependent on orchids for reproduction and fungal relationships have played in the diversifi cation of the or- dispersal and can survive as either saprophytes (e.g., Tulasnella chid genus Chiloglottis? We attempted to answer this by ad- species [ Roberts, 1999] ) or parasites (e.g., Armillaria mellea dressing four specifi c questions: (1) Which fungi and how many [ Rasmussen, 2002] ). One proposed mechanism for fungal- fungal species are involved in partnership with six Chiloglottis driven speciation is analogous to “ mixed-process coevolution ” species strategically chosen to represent the phylogenetic ( Thompson, 1986 ), by which changes in specifi city of orchid breadth across the genus? (2) What is the germination potential species for compatible fungi are predicted to result in reproduc- of the fungal isolates associated with Chiloglottis species? (3) tive isolation among plant hosts with different specifi c fungi. What are the phylogenetic relationships of the fungal symbi- However, in their recent review, Waterman and Bidartondo onts? (4) Finally, what are the implications of our fi ndings for (2008) stated that a direct link between mycorrhizal specifi city the current hypotheses on the diversifi cation of Chiloglottis and and reproductive isolation of the plant host is hard to envision. for other orchids more generally? A second mechanism for fungal driven speciation has been suggested in which mycorrhizal specialization has indirect con- sequences