Partial mycoheterotrophy is more widespread among orchids than previously assumed

With > 22 000 species in 880 genera occurring on all continents in almost every habitat, the Orchidaceae is considered the most diverse and most widely distributed plant family on Earth (Merckx et al., 2013). All orchids produce tiny, dust-like seeds (Arditti & Ghani, 2000) that require carbon (C) and nutrient supply by mycorrhizal fungi for germination and in the early seedling development phase (Smith & Read, 2008). This kind of nutrition is called initial mycoheterotrophy (Merckx, 2013). Later the majority of orchids develop green leaves and photosynthesize, yet remain mycorrhizal. There are, however, also > 200 achlorophyllous orchid species. These orchids require complete C and presumably mineral nutrient supply from mycorrhizal fungi and are known as full mycoheterotrophs (Leake, 1994; Merckx, 2013). Owing to their use of a 13C- (Gleixner et al., 1993) and 15N-enriched (Gebauer & Dietrich, 1993) fungal source for nutrition, mycoheterotrophic orchids are significantly enriched in 13C and 15N, when compared with autotrophic plants from the same microhabitats (Gebauer & Meyer, 2003; Trudell et al., 2003; Hynson et al., 2013). Based on their isotopic positioning between autotrophic nonorchids and achlorophyllous orchids, during the last decade a couple of chlorophyllous orchids have been found to acquire C from both autotrophic photosynthates and associated fungi (Gebauer & Meyer, 2003; Hynson et al., 2013). This mode of mixotrophic C gain is called partial mycoheterotrophy (Gebauer & Meyer, 2003; Merckx, 2013). In the past, the use of C stable isotope natural abundance played an important role in elucidating C gain by higher plants from two origins. This approach was highly successful when the two C sources of interest were sufficiently distinguished in their isotope composition, for example, in the case of hemiparasites with C3 photosynthesis simultaneously utilizing host plants with C4 or CAM photosynthesis (Press et al., 1987; Schulze et al., 1991; Ziegler, 1996). Limitations of this approach emerged in cases with two potential C sources not sufficiently distinguished in their C isotope composition, for example in the case of C3 hemiparasites utilizing C3 host plants (Schulze et al., 1991; Ziegler, 1994). Since early investigations by Bernard (1909), the majority of chlorophyllous orchids are considered as forming mycorrhizas with fungi that belong to the phylogenetically heterogeneous assemblage of rhizoctonias (Dearnaley et al., 2012). Interestingly, all known partially mycoheterotrophic orchids are additionally or exclusively associated with fungi that simultaneously form ectomycorrhizas with neighbouring trees (Selosse et al., 2002; Bidartondo et al., 2004; Hynson et al., 2013). Regarding the C and nitrogen (N) isotope abundance pattern, rhizoctonia-associated orchids do not provide a uniform picture. Hynson et al. (2013) speculated that a partially mycoheterotrophic C gain may be hidden behind findings of almost uniform 15N enrichment and inconsistent 13C positioning (Bidartondo et al., 2004; Liebel et al., 2010; Hynson et al., 2013) and called these orchids ‘cryptic partial mycoheterotrophs’. In order to elucidate this phenomenon of cryptic C gain from fungal sources, Stöckel et al. (2014) investigated C and N isotope abundances in initially mycoheterotrophic orchid seedlings and in mycoheterotrophic, partially mycoheterotrophic and putatively autotrophic orchid adults. It emerged that mycoheterotrophic seedlings of ectomycorrhiza-associated orchids were enriched in 13C and 15N in the same manner as mycoheterotrophic adults. However, mycoheterotrophic seedlings of rhizoctonia-associated orchids were by far less enriched in 13C and 15N than were mycoheterotrophic seedlings of orchids with ectomycorrhizal fungi. Stöckel et al. (2014) concluded that C and N stable isotope abundance alone is not suited to unequivocally identify partially mycoheterotrophic species among adults of rhizoctonia-mycorrhizal chlorophyllous orchids. Here we provide a new approach to detect partial mycoheterotrophy among rhizoctonia-mycorrhizal orchids. In addition to C and N stable isotope abundance, we analysed hydrogen (H) stable isotope natural abundance in dry tissues of an achlorophyllous orchid, chlorophyllous orchids associated either with ectomycorrhizal fungi or with rhizoctonias, autotrophic reference plants and sporocarps of ectomycorrhizal and saprotrophic fungi growing together in the understorey of a beech forest. We hypothesized that H isotope abundance can be used as a substitute for C isotope abundance in cases of limited information available from C isotope composition, because all organic compounds exchanged between fungi and plants are composed of C and H atoms. We furthermore hypothesized that the relative abundance of the heavy isotope 2H increases proportionally with the heterotrophic C gain of the investigated species. Thus, partially mycoheterotrophic orchids should be positioned in their H isotope abundance between autotrophic and fully mycoheterotrophic plants irrespective of their fungal partner. The second hypothesis is based on the conclusion of Yakir (1992) that secondary heterotrophic organic compounds, that is in our case organic compounds of fungal origin, are enriched in 2H compared with primary photosynthetic organic compounds. In order to determine whether transpiration (Ziegler, 1988) is responsible for differences in H isotope abundance of orchids and reference plants, we analysed oxygen (O) isotope abundances. Transpiration affects H and O isotope abundances in the same direction (Sternberg, 1988). Plant samples were collected in June and fungi in September in a temperate Fagus sylvatica-dominated broadleaf forest in northeastern Bavaria, Germany (Supporting Information Table S1; Zimmer et al., 2008). Leaves of target orchids and reference plants were sampled on 29 plots (each sized 1 m2) with one or two target orchid species and two to four autotrophic reference plant species per plot (Gebauer & Meyer, 2003). The following eight target orchid species belonging to three groups with respect to their type of C gain and fungal association were collected in five to six replicates each: Neottia nidus-avis, an achlorophyllous orchid mycorrhizal with one specific strain of ectomycorrhizal fungi (Selosse et al., 2002); Cephalanthera damasonium, Cephalanthera rubra and Epipactis atrorubens, chlorophyllous orchids mycorrhizal with ectomycorrhizal fungi and considered as partially mycoheterotrophic based on their C and N isotope composition (Bidartondo et al., 2004); and Cypripedium calceolus, Neottia ovata, Ophrys insectifera and Platanthera bifolia, chlorophyllous orchids mycorrhizal with fungi of the rhizoctonia group. A total of 92 leaf samples obtained from six autotrophic understorey species served as reference plant collection (Table S2). The fungal sporocarp collection comprised 54 samples from 11 species of ectomycorrhizal fungi and four species of saprotrophic fungi (Table S2). The classification according to ectomycorrhizal and saprotrophic fungi follows Tedersoo et al. (2010) and, in the case of Ramaria, Agerer et al. (2012). Although fungal fruit bodies of this investigation grew in the same forest, their spatial distribution did not follow the plot wise plant sampling scheme. Carbon/nitrogen and hydrogen/oxygen isotope abundances of plant and fungi samples were measured with an elemental analysis–isotope ratio mass spectrometry or a thermal conversion/isotope ratio mass spectrometry coupling, respectively (Table S3; Gehre & Strauch, 2003; Bidartondo et al., 2004). For a reliable H isotope abundance determination, each sample was analysed three times consecutively. The first two samples were biased by memories of the previous sample and were thus neglected. In order to minimize a bias of post-sampling H atom exchange between organically bound hydroxyl groups in our samples and H2O in ambient air (Yakir, 1992), we analysed target orchid samples and respective reference plant samples together in identical sample batches. Relative isotope abundances are denoted as δ values. For details on standard gas calibrations and tests used for statistical data evaluation, see Tables S3 and S4. Carbon and H isotope abundances of all investigated plants and fungi covered wide ranges from −22.1 ± 1.1‰ to −34.0‰ for δ13C and from −19.6 ± 4.4 to −110.4 ± 6.3‰ for δ2H (Fig. 1a). Chlorophyllous plants were significantly more depleted in 13C (U = 8, P < 0.001) and 2H (U = 126, P < 0.001) than fungi. By contrast, the achlorophyllous orchid was significantly more enriched in 2H than fungi (U = 13, P < 0.001) and not distinguished in 13C from fungi (U = 115, P = 0.596). Ectomycorrhiza-associated chlorophyllous orchids and the achlorophyllous orchid were significantly distinguished in 13C and 2H abundance from autotrophic reference plants (Table 1). Also 15N abundance distinguished all ectomycorrhiza-associated chlorophyllous orchids and the achlorophyllous orchid from the reference plants (Tables 1, S2). The 13C and 15N abundance of the rhizoctonia-mycorrhizal chlorophyllous orchid species, however, showed no uniform picture (Table 1). Only O. insectifera and P. bifolia were distinguished in both 13C and 15N abundance from the reference plants. The 2H abundance of all rhizoctonia-associated orchids, by contrast, showed a uniform pattern. Like ectomycorrhiza-associated orchids, all chlorophyllous rhizoctonia-mycorrhizal orchids were significantly enriched in 2H compared with reference plants (Table 1). In contrast to H, C (and N) isotope abundances, O isotope abundances of the plants were significantly enriched compared with fungi samples (U = 127, P < 0.001) and varied little (23.6 ± 1.5 to 29.2 ± 1. 2‰; fungi, 8.9 ± 4.3 to 23.8 ± 0.4‰) (Fig. 1b,c). With one exception (C. rubra), orchids and reference plants were not distinguished in their O isotope abundances (Table 1). Correlation analyses revealed a strong positive and linear relationship between δ13C and δ2H (Fig. 1). By contrast, δ13C/δ18O and δ18O/δ2H were weakly negatively correlated (Fig. 1). Simultaneous C gain from two sources, photosynthesis and heterotrophic use of an alternative C source, has previously been reported for diverse groups of protists (Pringsheim, 1958; Flynn et al., 2013; Wilken et al., 2014), but only in a very few cases for higher plants utilizing either the xylem of host plants as hemiparasites (Press et al., 1987; Marshall & Ehleringer, 1990; Schulze et al., 1991; Ziegler, 1996; Tĕšitel et al., 2011) or ectomycorrhizal fungi as partial mycoheterotrophs (Gebauer & Meyer, 2003; Hynson et al., 2013). Our investigation provides evidence that substituting analysis of C isotope abundance by H stable isotope abundance can overcome the limitation of insufficiently distinguished C source isotope abundance. Early indications about unique H isotope abundances in parasitic plants and in fungi were given by Ziegler (1994, 1996), however, without explanation for potential reasons. In the following we discuss the physiological background potentially responsible for the observed 2H enrichment. Hydrogen in organic tissue of autotrophic terrestrial plants originates from precipitation water that is used in photosynthesis to convert CO2 into glucose: 6CO2 + 6H2O → C6H12O6 + 6O2. Thus, like C atoms from atmospheric CO2, H atoms from precipitation H2O ultimately drive the isotopic composition of tissues in autotrophic terrestrial plants (Ziegler, 1988). In analogy to the discrimination against 13C in photosynthesis of terrestrial C3 plants by c. 20‰ (Farquhar et al., 1989), photosynthesis also discriminates against the heavy H isotope 2H. Owing to the higher relative mass difference between 1H and 2H in comparison to 12C and 13C, the fractionation against 2H in photosynthesis is even larger and ranges in the order of 170‰ (Yakir, 1992). This huge fractionation against 2H in photosynthesis is postphotosynthetically reduced by the exchange of H atoms from hydroxyl groups in organic molecules with H atoms from surrounding water molecules and by fractionations in metabolic reactions (Schmidt et al., 2003). The type of photosynthetic pathway is the major factor simultaneously shifting C and H isotope abundances in a direction as observed in our study. In principle, chlorophyllous orchids can follow C3 or CAM photosynthesis (Ziegler, 1995). Many orchids living as epiphytes in tropical forests are known as CAM plants (Osmond, 1978) and are thus enriched in heavy C and H isotopes compared with C3 plants (Ziegler, 1988; Farquhar et al., 1989). The orchids of our investigation, however, grow in a temperate or boreal climate on dark, humid forest floors – a habitat unsuited for CAM photosynthesis. Moreover, they are C3 plants (Ziegler, 1995) or nonphotosynthetic. Thus, CAM as a driver for the C and H isotope abundances observed here can be ruled out. Transpiration is another factor simultaneously affecting C (Farquhar et al., 1989), H (Ziegler, 1988) and O (Barbour, 2007) isotope abundance. Leaf tissues of hemiparasites that transpire at higher rates than their host plants were shown to be depleted in 13C and 18O (Cernusak et al., 2004) and should be also depleted in 2H (but see Ziegler, 1996). Thus, higher transpiration by green orchids potentially capable of using fungi as an alternative C source analogous to hemiparasitic plants should end up with depletion in 13C, 18O and 2H. Simultaneously, positive correlations among δ13C/δ2H, δ18O/δ2H and δ13C/δ18O should be found. The studied orchids were not depleted in 18O and were enriched in 2H and in many cases enriched in 13C, and δ18O/δ2H and δ13C/δ18O were negatively correlated. Thus, transpiration is unlikely to be driving the 2H enrichment of the orchids. A third factor potentially responsible for the significant 2H enrichment found for orchids in comparison to neighbouring autotrophic plants is a selective organic matter gain from their fungal hosts through mycoheterotrophy. Yakir (1992) proposed a distinction between primary photosynthetic organic compounds characteristically depleted in 2H and secondary postphotosynthetic, and thus heterotrophic, organic compounds, which are relatively enriched in 2H. This concept is exactly mirrored by the H isotope compositions of the functional groups of plants and fungi in our investigation. Autotrophic reference plants were most depleted in 2H and thus most influenced by primary photosynthetic processes. Fungi as heterotrophic organisms were enriched in 2H when compared with autotrophic plants, and the achlorophyllous orchid was even more enriched in 2H as a result of the exclusive use of heterotrophic ectomycorrhizal fungi as an organic matter source. All chlorophyllous orchids of this investigation, whether associated with ectomycorrhizal or rhizoctonia fungi and enriched in 13C or not, were positioned in 2H isotope composition between autotrophic plants and the achlorophyllous orchid. We conclude that all chlorophyllous orchids investigated here have to be considered as gaining H together with C from two sources, namely primary photosynthesis and secondary heterotrophic fungal origin. Thus, they are partially mycoheterotrophic. This is the first evidence that organic matter gain from a fungal source by adult chlorophyllous orchids is not limited to the few species known as ectomycorrhiza-associated (Bidartondo et al., 2004; Hynson et al., 2013), but also plays an important role for adults of rhizoctonia-mycorrhizal orchids, the vast majority of the > 22 000 orchid species worldwide, as proposed by Selosse & Martos (2014). Further investigation has to elucidate to what extent an 2H enrichment is distributed among the Orchidaceae. The discovery of H isotope abundance as a substitute for C isotope abundance provides perspectives for elucidating organic matter gain by heterotrophic means in all those cases where the two potential C sources, autotrophic C from photosynthesis and heterotrophic C from a host, are not sufficiently distinguished in their C isotope composition. Thus, the use of this approach is not limited to orchids, but can be also applied to a wider spectrum of plant groups, such as hemiparasites and plant families producing dust seeds (Merckx, 2013). Furthermore, this investigation provides, to our knowledge for the first time, evidence that fungi are not only distinguished in their C, N and H stable isotope abundance from neighbouring autotrophic plants (Gebauer & Dietrich, 1993; Gleixner et al., 1993; Ziegler, 1994), but also in O stable isotope abundance. Further investigation has to elucidate whether different functional and taxonomic fungal groups are systematically distinguished in their O isotope abundance patterns. Factors underlying the 18O depletion of fungi compared with neighbouring plants have to be identified. We thank Iris Schmiedinger for technical assistance in establishing the facilities for H/O isotope abundance analysis at the BayCEER – Laboratory of Isotope Biogeochemistry, Christine Tiroch for assistance with C/N isotope abundance measurements, and the Regierung von Oberfranken for authorization to collect the orchid samples. Furthermore, we thank Maarja Öpik for her efficient handling of our manuscript as second Editor. This work was supported by the German Research Foundation (DFG, project GE565/7-1). G.G. had the idea for this investigation, supervised the sample analysis, performed the data analysis and drafted the manuscript. K.P. collected and prepared the plant samples for further analysis and contributed to the manuscript. A.C.G. collected and prepared the fungi samples for further analysis and contributed to the data analysis and to the manuscript. Please note: Wiley Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Table S1 Sampling site characteristics Table S3 Equipment and conditions as used for stable isotope abundance analysis Table S4 Statistical tests and software as used for data evaluation Table S2 Single and mean values of δ13C, δ15N, δ2H and δ18O of plant and fungi samples Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.