Cynomorium songaricum Rupr., is a rare medicinal herb plant with pharmacological activities of treating various sexual conditions such as premature ejaculation, low sexual function, impotence, and spermatorrhea colic (Cui et al., 2013; Lee et al., 2013). It is often found parasitizing the roots of several plant families such as Nitrariaceae, Tamaricaceae, and Chenopodiaceous shrubs located in dry, rocky, or sandy soils of extreme desert environments (Cui et al., 2013). In China, Nitraria tangutorum and Nitraria sibirica of the Nitrariaceae are the preferred hosts of C. songaricum, and the flowering-fruiting season for C. songaricum has been described to be from May to August (Wan and Chen, 2000). C. songaricum parasitizing N. tangutorum forms a successfully evolving sustainable ecosystem in which mature seeds are dispersed with sand under natural desert conditions. Occasionally, a few seeds encounter the roots of N. tangutorum, invading plants using host root-secreted hormone, strigolactones (Kruh et al., 2017). The seed then begins to develop special intrusive organs such as haustorium which directly connects to the vascular system of host under suitable spatial and temporal conditions. Because C. songaricum plant has no chlorophyll and is unable to photosynthesize, it lives as a holoparasite, i.e., totally dependent on its host for nutrients and water. The parasitic seed germinates to form tubercle (rhizome) on host plant root in autumn and winter. Under suitable temperature and water conditions in March to early April, the tubercle begins to develop quickly, and emerges out of the ground as a low-growing inflorescence in late April to early May. After about 1 month, the tissue above ground becomes senile and forms numerous mature seeds upon pollination by flies, then drop and disperse in the desert sand. The rhizome underground, however, stays connected to the host root staying dormant for several years until the next development cycle begins. As an ethno-pharmacologically important and endangered medicinal species, the production of C. songaricum is far from meeting the demand, one tough problem is that the seed is very difficult to germinate due to the secretion of germination inhibitors such as abscisic acid and thicker wax coat which inhibits absorption of water and air that are essential for seed germination (Dong et al., 2010; Chen et al., 2011). For successful seed germination and development of C. songaricum, 5 years or even longer time is needed under natural conditions. Further, the studies on the mechanism of seed germination and development under natural parasitizing environment are largely undocumented to date. Compounding the matter further is the over-exploitation of stems and inflorescences (before seed maturation) for their popularity has proven detrimental to the preservation and conservation of C. songaricum on an ecological scale. Hence, for conservation of these medicinal plants a sustainable harvesting and manual cultivation technique is warranted (Cui et al., 2013).

Addressing this key issue, several tests were performed on improving seed germination rate (Dong et al., 2010). This includes breaking seed cover with ultrasonic treatment, soaking and then removing seed cover with a series of chemicals such as sodium hydroxide and potassium permanganate, or seed treatment with hormones such as gibberellins (GA) and naphthylacetic acid (NAA), or treatment with extracts of N. tangutorum root, and/or a combination of the above methods (Luo et al., 2013). But these methods exhibit challenges such as low germination rate, environment pollution, etc. Further, to date no one method has successfully been applied in sustainable production on a large scale. Other alternative methods yielding long-term reproducible solutions to sustainable production needs further investigation taking into consideration their development under natural parasitic habitats.

Endophytes are an integral part of plant biology and are perceived as host organs in and of themselves serving as important components of plant micro-ecosystems (Christian et al., 2016; Jia et al., 2016). Endophytes, especially fungal endophytes play important beneficial roles in host plant development and physiology including increased stress tolerance, enhanced root growth and provision for special nutrition and water (Wani et al., 2015). It is hypothesized that over long periods of co-evolution the endophytes not only establish special relationship with one another but can also significantly influence the formation and accumulation of certain key metabolic products altering the quality and quantity of crude drugs derived from medicinal plants (Faeth and Fagan, 2002; Firáková et al., 2007; Jia et al., 2016). Recently, a comprehensive review of published data on medicinal plants and endophyte relationships obtained in the last 30 years suggested that medicinal compounds could be effectively produced on a biotechnological scale if the relationships and the conditions that promote such production are clearly understood and cultivable endophytic fungi associated with medicinal plants be successfully established (Jia et al., 2016). The factors affecting the host-endophytic relationships and endophyte assembly in plant-parasitic interactions are limited, but previous results have suggested an important role for fungal endophytes and the diversity of fungal assemblages in key trophic interactions (Chen et al., 2005; Zhang et al., 2006; Behie et al., 2012).

Nearly 18 years ago, research showed no overlap of endophytic fungal assemblages of parasitic Cuscuta reflexa and it host Cucurbita maxima (Suryanarayanan et al., 2000). However, recent studies have highlighted the rich and diverse endophyte fungal communities in the hemi-parasitic plant Macrosolen tricolor and its host plant Camellia oleifera, and that they were always distributed in nearly the same orders of the four classes of Ascomycota and one class of Basidiomycota (Zhou et al., 2014). Surprisingly, the endophytic fungal assembly are shown to exhibit clear shifts between the parasitic plant and its host, for example, Arceuthobium americanum and its host Pinus contorta (Martin et al., 2012), and Phoradendron perrottettii and its host Tapirira guianensis (Teixeira-Costa and Ceccantini, 2016). In other studies, it has been demonstrated that endophytic fungi from parasitic Viscum coloratum, secrete cellulolytic enzymes which can degrade cellulose assisting the haustorium of mistletoe to break through the cell wall as well as intracellular tissue spaces within the host Pterocarya stenoptera (Ding et al., 2008). Besides the above roles of endophytes in parasitic development, they also exhibit profound influence on seed germination, and it is a well-established fact that Mycena osmundicola and Mycena dendrobii are essential for seed germination of Gastrodia elata by secreting indoleacetic acid (Xu and Guo, 1989; Guo and Wang, 2001). Recent investigations are also providing new and exciting information on the potential of cultivable endophytic fungi (associated with medicinal plants) as reliable sources of plant secondary metabolite accumulation and bioactivity (Nisa et al., 2015; Chagas et al., 2017; Cui et al., 2017; Ran et al., 2017). While the importance of endophytic fungi to the growth and development of host plants is increasingly being substantiated, exploiting these endophytic relationships for production of medicinal plant-based drugs and mechanistic understanding of the benefits of these exact relationships is still very limited.

To gain insights into the fungal endophytes and the diversity of fungal assembly in C. songaricum and its host N. tangutorum, here we (1) investigated the genetic relationship, community assembly and distribution characteristics of endophytic fungi across the different developmental stages of C. songaricum parasitizing N. tangutorum. (2) Evaluated the potential of the identified endophytic fungal assembly for functional interaction, providing evidence for the application of a target endophytic fungal spp. for promoting C. songaricum seed germination under in vitro axenic culture conditions. Our results demonstrate that parasitic interaction is a significant factor in shaping endophytic fungal assemblages. We provide evidence for the involvement of few endophytes in plant development and physiology, enabling promotion of seed germination of C. songaricum. The information obtained from this study provides new insights into the tripartite interactions of N. tangutorum hosts, root-parasitic plant C. songaricum, and its fungal endophytes. Our study highlights the potential for exploiting alternative bioengineering approaches to sustainable production of C. songaricum.

With the rapid development of high-throughput sequencing methods and bioinformatics tools, a huge number of cryptic fungi are being discovered (Yang et al., 2016). Previous estimates based on high-throughput sequencing suggest as many as 5.1 million fungal species and a significant proportion of which is estimated to be found in plant–endophyte relations (Blackwell, 2011). Recently, the use of “meta-omic” technologies such as metagenomics has helped to realize the potential of endophytes in plant-fungal interaction studies (Kaul et al., 2016). While many aspects of endophytic biology are still unknown, metagenomic approaches in identifying sequences of interacting partners is unraveling the nuances of plant–endophyte relationship. The identified sequences help exploring endophytic species distribution, community structure, lifestyles and nutrition modes, and provide objective insights into endophytic fungal richness, ecological function and fungal-mediated processes in host plant (Ottosson et al., 2015). While, an urgency to establish sustainable programs for conservation and resource utilization of medicinal plants has long been proposed (Cui et al., 2013), further studies on mechanisms and exploitation of co-existing endophytes is still largely limited. Here, in our study we aimed at understanding the diversity of fungal endophytes in one such ethnopharmacological relevant spp. of the genus Cynomorium, that forms root-parasitic associations with N. tangutorum. Previously, a pioneering study on characterization of endophytic fungal communities in 29 traditional Chinese medicinal plants had suggested a predominant occurrence of fungal endophyte species of the taxa; Alternaria, Colletotrichum, Phoma, Phomopsis, Xylariales, and mycelia sterilia (Huang et al., 2008). Further, quite interestingly to date most other fungal endophytes discovered have been shown to be the fungal species that could be obtained easily and isolated commonly in the environment, such as Gibberella spp., Fusarium spp., Aspergillus spp., Candida spp., and Penicillium spp. (Strobel and Daisy, 2003). In this study, using high-throughput sequencing we detected dominant genera/order of endophytic fungi at different distribution levels among our CSR, SNT, and NT samples (Figures 1, 2A,C and Supplementary Figure S2). Our results concur with previously reported findings that under traditional methods used in laboratory conditions only 0.1–1.0% of endophytic fungi could be obtained because of artificial selection pressure (Amann et al., 1995). However, considering the medicinal value of Cynomorium spp. in the clinic as well as in folk-medicines information on the endophytes found in the root-parasitic interactions of C. songaricum and its host N. tangutorum could be deemed invaluable.

Plant endophytic fungi are shown to originate mainly from seeds through vertical transmission or from the environment through horizontal transmission (Christian et al., 2016; Guo et al., 2017). However, studies on endophytic fungal distribution in parasitic plants and its host are more complicated, for fungal colonization and movement within the two plants can either be facilitated via xylem tubes (Compant et al., 2010) or through apoplasts, intercellular space (Maheshwari et al., 2013). For example, it was found that the mistletoe Phoradendron perrottettii and its host tree Tapirira guianensis shared some common endophyte species while still maintaining distinct communities (de Abreu et al., 2010). Similarly, in our studies (Figure 2A) we found that CSR and its host SNT shared much more OTUs (76 in total) for endophytic fungi, compared to CSR and NT (4 in total). The reason for this observed discrepancy may be that C. songaricum is succulent while N. tangutorum is a ligneous plant, which may lead to the damage of C. songaricum tissue by sand more easily as compared to N. tangutorum. Thus, it can be hypothesized that the fungal invasion and colonization to C. songaricum is easier than N. tangutorum. This may probably explain the observed differences in the number of unique fungi in CSR (419) being more than those in SNT (290) (Figures 2A,C). Further, we also found that NT and SNT together shared the maximum numbers (1120) of endophytic fungi, which we believe supports the notion of species specificity and selective specificity existing between endophytes and their hosts (Rai and Agarkar, 2016). On the other hand, the parasitic connection in CSR and SNT would lead to chemical exchange, and further influence selection and/or fungal inhibition. Thus, the genetic distance of fungal communities from SNT intermediated between CSR and NT, however, results suggest that the fungal assemblages are more associative toward CSR than NT (Figure 2B).

In one life cycle, C. songaricum grows originating from a tubercle (about several millimeters) to an exuberant organism (about 0.5 m) spanning the T to A developmental stage, then they shrink gradually toward the end of the year. With the growth of root tips or other tissue in sand, more and more environmental fungi would invade and colonize plants via the tissue that is injured by sand particles. Thus, it can be speculated that the amount of endophytic fungi increases from T to U developmental stage with the rising desert temperatures (Figure 3). After the peak (U stage), the fungal number decreased until to A stage (Kruh et al., 2017). The invaded fungi may be subjected to selection pressure from the host plants (Nikolov et al., 2014; Rai and Agarkar, 2016), and only part of them capable of adapting to the host internal environment, may survive and multiply in the host plants (Huang et al., 2017). Hence, the longer the plants grow, the more is the shared endophytic fungi among CSR, SNT, and NT (Figure 3). On the other hand, because of exchange of fungi through parasitic connections (Kruh et al., 2017), the shared fungi between CSR and SNT were more than that between CSR and NT in each developmental stage. This suggests a most probable co-evolution and natural selection among host species (Ambrose et al., 2014).

The role and mechanism of endophytes in parasitic interaction between C. songaricum and N. tangutorum remains largely uninvestigated, and our results provide key cues that could further be used for functional evaluation of fungal assemblages in isolation or in unison. Previous evidence indicates that products from endophytic fungi affect physiology and development of host plants (Rodriguez and Redman, 2008). For example, strigolactones produced almost exclusively from parasitized host plant, serve as essential signal molecules stimulating the germination of parasitic plant. Recent evidence that strigolactone biosynthesis can be regulated by gibberellins which were produced by endophytic Gibberella spp. and Fusarium spp. provides further opportunities for analyzing the roles of endophytic fungal assemblages (Lale and Gadre, 2010; Troncoso et al., 2010; Ito et al., 2017). Interestingly, even in our study both Gibberella spp. and Fusarium spp. were identified as the dominant fungi in parasitic C. songaricum and its host N. tangutorum. And, further analysis of relevance and function of the Fusarium spp. on the growth and development of C. songaricum was tested using in vitro axenic conditions, which suggested that the Fusarium redolens (KY379544) can serve as a target endophytic fungal species in promoting C. songaricum seed germination (Supplementary Figure S3). However, it remains to be investigated whether they are involved in the tripartite interaction of fungal endophytes with the root-parasitic plant and/or its host. While, some endophytic bacteria were also shown to possess the ability of promoting host seed germination (Shittu et al., 2009; Xu et al., 2014), studies with respect to the use of fungal endophytes as sources of medicinal host similar bioactive component (biosimilar) is only now beginning to be addressed to its full potential (Jia et al., 2016). In this study, five isolates (F. redolens KY379544, F. nematophilum KY379572, A. terreus KY379684, C. rosea KY379716, and F. nematophilum KY379764) showed the ability of promoting seed germination, which suggests that endophytic fungi may play important roles in root-parasitic plant interactions and parasitic host growth and development.

This study firstly demonstrates that parasitic behavior of C. songaricum on N. tangutorum could significantly affect the inhabiting distribution, growth and decline, and other plant fungal endophyte dynamics. We show that the endophytic fungal community shifts between the parasitic plant and its host, which helps one to understand the significance of such ecological relationships of parasite-endophyte-host organism(s) in tripartite associations. Some of the endophytes we isolated in this study exhibit the ability to promote C. songaricum seed germination, which concurs with previously suggested studies that endophytes could significantly influence the production and accumulation of medicinal host metabolites. Our results in the ethnopharmacological relevant Cynomorium spp., highlights the important and integral part fungal endophytes occupy in the C. songaricum’s desert ecosystem, thus playing a vital role in parasitic development and physiology. Our results provide evidence for future investigations on novel ways for improving the parasitic survival rate and production of C. songaricum, which in the long-run may aid maintenance and sustainable development of desert plant ecology.

J-LC and GZ designed the research, performed the experiments, and analyzed the data. J-LC collected the samples. J-LC, VV, and GZ co-wrote the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.