Latent Pathogenic Fungi in the Medicinal Plant Houttuynia cordata Thunb. Are Modulated by Secondary Metabolites and Colonizing Microbiota Originating from Soil

Microbes play a prominent and important role in the plant-microbe ecosystem (Hodge and Fitter 2010; Clemmensen et al. 2013), but there is a need to understand better the linkages between the colonizing latent pathogens in host plants and microbiota from the soli colonizing host plant roots. To date, although the number of studies on interactions within the fungus-host plant symbiotic system is multiplying rapidly (Etalo et al. 2018; Thiergart et al. 2020), the number of fungal species, as well as the number of plant species investigated, represent only the tip of the iceberg. The mutualistic relationship between the colonizing latent pathogens and their host plant is mediated needs further research. In a long time, fungi can colonize the rhizosphere or grow intracellularly in the root epidermis and exodermis. Thus, they may establish stable colonization with the host and not cause plant diseases or be eliminated by the defense system of the host plant (Álvarez et al. 2020). Fungi colonizing the host plant are directly and/or indirectly affected by abiotic factors and functional traits of the plant; in turn, colonizing fungi can also affect host plant traits. Colonizing fungi forming mutualistic relationships with the host plants may promote host plant growth and increase stress tolerance and/or pest/pathogen resistance (Chitnis et al. 2020). Some latent plant pathogens may initiate disease processes in response to unknown factors, such as stress (Aly et al. 2011). LPFs can cause plant diseases, but they can maintain symbiotic relationships with the host and cause disease only under certain conditions. For example, grapevine trunk diseases fungi can transform from the endophytic to the pathogenic under the influence of abiotic and/or biotic stress factors (Úrbez-Torres et al. 2020). Many renowned pathogens were found in asymptomatic hosts. There is no immediate explanation, and whether these LPFs can provide benefits to the host like other endophytes needs further research (Précigout et al. 2020; Salvatore et al. 2020). LPFs can invade the interior of plants asymptomatically and accompany the growth of host plants, and live together for a long time. However, the factors that drive the mutualistic colonization in host plants are still poorly understood. Although the relationships of the LPFs with the host plants are adequately documented (Arnold and Lutzoni 2007; Bamisile et al. 2018), the driving factors for maintaining the stable colonization of the LPFs in the host plants have been not sufficiently elaborated.

Many studies on the relationships between colonizing fungi and host plants have been described (Arnold and Lutzoni 2007; Delaye et al. 2013; Brader et al. 2017; Bamisile et al. 2018). It has been reported that LPFs were isolated from many healthy plants (Hyde and Soytong 2008; Fernandes et al. 2018; Chen et al. 2020). LPFs are considered to have colonized plants at the asymptomatic infection stage and might subsequently have initiated morbidity or spread under suitable conditions (Carroll 1988; Malcolm et al. 2018; Ridout and Newcombe 2018). Interestingly, there is a dynamic switching between a necrotrophic and an endophytic lifestyle of LPFs (Delaye et al. 2013; Brader et al. 2017). Some pathogens can invade a plant species and not cause disease symptoms but may infect other plants and trigger disease (Wheeler et al, 2019). Some pathogens lose pathogenicity because of a gene mutation and might exhibit colonization in plant tissues (Lofgren et al. 2018). However, it is of great interest to define how LPFs colonize but did not cause a severe injury or a complete plant disease outbreak.

The microorganisms colonizing the plant tissues (such as roots or rhizomes) are selected from the soil microorganisms. Healthy plants host diverse but taxonomically structured communities of microorganisms, the plant microbiota (Trivedi et al. 2020). The microbiota plays an important ecological and physiological role in the symbiotic relationships as they confer fitness advantages to the plant host, including growth promotion, nutrient uptake, stress tolerance, and resistance to pathogens (Stringlis et al. 2018; Huang et al. 2019; Trivedi et al. 2020). However, whether the colonizing microbiota of plants regulates the stable maintenance of the LPFs is poorly understood.

The ability to produce secondary metabolites of herbal plants is a significant functional trait, which has played a crucial role in human health and disease treatment for thousands of years and has critical functions in plant physiology and ecology (Huang et al. 2019; Wetzel et al. 2020). However, the accumulation of such physiologically active compounds is affected by colonizing microbes (including the fungi and bacteria) and soil, climatic, and geographic factors (Luo et al. 2019). Houttuynia cordata Thunb. is a perennial herb, which is widely distributed in eastern and Southwest Asia (including most parts of China) and is widely planted in Southwest China. It is not only a traditional Chinese herbal medicine with a long history; this plant is also regarded as an exceptional food in Southwest China because of its unique aroma (Wang et al. 2020). H. cordata is believed to have the effects of “clearing heat” (such as treating fever, headaches, and sweating), detoxifying and purifying the body, could effectively inhibit cancer cell growth and reduce the risk of atherosclerosis (Lai et al. 2010; Li et al. 2013; Park 2015). H. cordata secondary metabolites (including volatiles and phenolics) are affected by abiotic factors. Several studies focused on the effects of soils on H. cordata quality, including phytochemical composition (Fu et al. 2013; Aramsirirujiwet et al. 2016; Řebíčková et al. 2020). However, whether the secondary metabolites of H. cordata can regulate the LPFs is unknown.

H. cordata has been demonstrated to contain a large number of colonizing microorganisms in its rhizomes. However, current research has mainly focused on the microbiota diversities and the accumulation of secondary metabolites (Talukdar et al. 2020). The endophytic fungi isolated from H. cordata have been shown to exhibit broad-spectrum antifungal activity (Li et al. 2013; Pan et al. 2016). Trichocladium sp. isolated from H. cordata rhizomes can metabolize the antifungal metabolites (Nakashima et al. 2017; Tran-Cong et al. 2019). Endophytic fungi isolated from H. cordata can antagonize pathogenic fungi and induce indole-3-acetic acid metabolism (e.g., Colletotrichum, Lasiodiplodia, and Fusarium) (Aramsirirujiwet et al. 2016). However, studies on the effects of secondary metabolites on the LPFs, after being isolated from the complex microbial communities, are still scarce (Aramsirirujiwet et al. 2016). Secondary metabolites can actively respond to pathogenic fungal infections; for instance, phenolics, terpenes, sulfur (S)- and nitrogen (N)-containing compounds can be synthesized by oil palm to resist pathogenic microorganisms (Sahebi et al. 2017). Simultaneously, secondary metabolites can regulate the composition and structure of rhizosphere microbiota; for example, flavonoids are the bridge between microbiota and host plant e.g., white clover and Glomus sp., they represent a potential source of carbon and impact the rhizosphere function (Singh and Singla 2020; Shaw et al. 2006). Whether and/or how the identity of the plant-LPFs is related to H. cordata metabolites or whether they maintain a relatively stable dynamic equilibrium modulated by antifungal metabolites in H. cordata rhizomes is also poorly understood.

In this current study, we hypothesize that maintaining the stable colonization of the plant-LPFs in H. cordata is regulated by the colonizing microbiota, which originates from the soil microorganisms, and is mediated by H. cordata rhizome metabolites. Specifically, we hypothesized: (1) the plant-LPFs are inhibited by the colonizing microbiota in H. cordata; (2) there is a mutually inhibitory relationship between the LPFs; and (3) the LPFs are controlled by H. cordata rhizome metabolites. To test these hypotheses, three LPFs were isolated from healthy rhizomes of H. cordata and identified as Ilyonectria liriodendri, an unidentified fungal sp., and Penicillium citrinum. We investigated the direct antagonism between three representative plant-LPFs, explored whether they could infect sterile plants growing in sterile or non-sterile soil and whether the extracts from H. cordata rhizomes controlled them.

We want to assess the factors driving the maintenance of stable colonization of the plant-LPFs in H. cordata in this study. More importantly, we look forward to providing an insight into the interactions between the LPFs, the secondary metabolites of H. cordata, and the origin of the colonizing microbiota.

Isolation and identification of latent pathogens from H. cordata rhizomes. Our aim was to investigate how the latent pathogens were present but constrained stably in the healthy H. cordata rhizomes. We found that the rhizomes were heavily colonized when the healthy rhizomes, following surface sterilization, were cultured on a half-strength MS medium (Fig. S1). We isolated the LPFs, which emerged from inside the surface-sterilized H. cordata rhizomes on PDA. Fig. 2 and Table SI showed that three fungal species isolated were identified as I. liriodendra (IL), an unidentified fungal sp. (UFS), and P. citrinum (PC), using ITS1-5.8S-ITS2. They were numbered NSF-1, NSF-2, and NSF-3, respectively (Fig. 1 and Table SII). IL proliferated rapidly on the PDA medium. The surface mycelium was pure white with edges, with abundant aerial mycelium, shaped like cotton wool. The hyphae had few branches; colonies were light yellow to dark red on the reverse side of the plate three-septate conidia were long and cylindrical (Fig. 3A-C). UFS colonies had a gray woolly appearance, with white hyphae on the outside and edges, gray to dark black on the reverse side, and obvious concentric annular bands. Hyphae were slender and loose, with few or no branches; two forms of hyphae coexisted, one with a smooth surface and the other with dense cup-shaped protrusions, and no spores were observed (Fig. 3D-F). Colonies of PC had a gray green felt-like appearance with white edges. After about two weeks of growth, the colony gradually stopped extending and formed white, velvety, aerial hyphae on the original colony, white to yellow on the reverse. The hyphae were dense and separate, the sporangia were broom-like, the conidia were light green and round, and were easy to spread (Fig. 3G-I).

Fig. 1.

Fig. 2.

Fig. 3.

Growth properties of H. cordata-colonized LPFs in sterile soils and non-sterile soils. To investigate the interactions between the colonizing microbiota and the LPFs within the rhizomes, we investigated any possible interactions between the colonizing microbiota (originating from nonsterile soils) and the LPFs. To achieve this goal, we investigated whether the growth of the sterile H. cordata seedlings in the nonsterile soils (containing soil microorganisms) and the sterile soils were affected by IL, UFS, and PC (Fig. 4 and 5). When the sterile H. cordata seedlings growing in sterile soils were inoculated with IL, the growth was strongly negatively affected (Fig. 4B). Compared with CK (Fig. 4A) (the sterile water was added to the sterile H. cordata seedlings growing in sterile soils), the control seedling significantly grew faster than the inoculated seedlings. The color of the young leaves of the seedlings began to fade on the 3^rd day, after which the whole plant quickly wilted, and, by 15 days, the plants were completely dead, and dead tissue showed obvious white mycelium.

Fig. 4.

Fig. 5.

When the sterile H. cordata seedlings growing in sterile soil were inoculated with UFS, the effects on H. cordata seedling growth were not significantly adverse. However, the average of dry weight, seedling height, number of live leaves, and total leaf area declined when compared with CK (Fig. 4C and 5). When the sterile H. cordata seedlings growing in sterile soil were inoculated with PC (Fig. 4D), the old leaves began to fade after five days. The plant growth slowed down, compared with CK (Fig. 4A), and after 15 d, the old leaves had died, but the young leaves remained green, the dry weight, seedling height, and total leaf area declined by 50.5%, 48.6, and 69.8%, respectively (p < 0.05) (Fig. 5). When the sterile H. cordata seedlings grown in non-sterile soil were inoculated with IL, UFS, or UFS, and compared to CK (the sterile H. cordata seedlings grown in non-sterile soils and treated with sterile water) (Fig. 4E), the growth of H. cordata seedlings was barely affected (Fig. 4F-H). Furthermore, the growth of sterile H. cordata seedlings in sterile soils (Fig. 4A) was significantly slower than the growth in non-sterile soils (Fig. 4E); the total live leaf area increased 47.9% (p < 0.05) (Fig. 5).

Antagonism between LPFs. The LPFs isolated from the rhizomes colonized the healthy H. cordata rhizome (Fig. S1) with no symptoms, suggesting that they formed a stable system in which the pathogenic factors of LPFs were suppressed. The results from our in vitro studies showed that the colony diameters of the three strains (IL, UFS and PC) were significantly inhibited when three strains were grown in pairs for seven days in the antagonism assay on the PDA medium (Fig. 6). It was clear that, after seven days, the mycelia of the pairs of LPFs in the antagonism trials were not in obvious contact. IL significantly inhibited the colony growth of UFS, with an inhibition rate (IR) of 85.7% (Fig. 6A and Table I), and PC (IR = 52.1%, Fig. 6A and Table I). UFS significantly inhibited colony growth of IL (IR = 26.1%, Fig. 6B and Table I), and PC (IR = 18.5%, Fig. 6C and Table I). PC significantly inhibited colony growth of IL (IR = 52.6%, Fig. 6A and Table I), and UFS (IR = 62.2%, Fig. 6C and Table I). These findings confirm that there is a competitive relationship between any pair of the LPFs.

Fig. 6.

Antifungal activities of H. cordata rhizome extracts toward the LPFs. We investigated the interactions between the LPFs and the H. cordata rhizome extracts in terms of the antifungal activities of H. cordata rhizome extracts toward the LPFs (IL, UFS, or PC). The antifungal activities of H. cordata rhizome extracts on each of the three LPFs were tested in vitro (%Table II). The antifungal activities (expressed as inhibition rate, IR = 11.97–23.08%) of H. cordata rhizome extracts toward IL increased with an increasing concentration (from 1 to 8 mg/ml), ranging from 11.97 to 23.08%. IR of the different concentrations of H. cordata rhizome extracts on UFS was in the range 3.37–12.36% (%Table II), whereas H. cordata rhizome extracts did not significantly inhibit PC (%Table II). Furthermore, the average colony diameter of IL in response to different concentrations of H. cordata rhizome extracts (from 0 to 8 mg/ml) was significantly lower than that of CK (p < 0.05). The average colony diameter of UFS on PDA with H. cordata rhizome extracts, except for 8 mg/ml (p < 0.05) (Fig. 7A), was not significantly different from that in the CK (p > 0.05) (Fig. 7B). The average colony diameter of PC on PDA concentrations (from 0 to 8 mg/ml) of H. cordata rhizome extracts was not significantly different from that in the CK (p > 0.05) (Fig. 7C).

Fig. 7.

Changes in H. cordata rhizome phenolic profile infected with the LPFs from non-sterile soil. The antifungal activities of H. cordata rhizome extract toward the LPFs hinted that the accumulation of antifungal secondary metabolites in the H. cordata rhizome crude extract might change in response to infection by the LPFs. The changes in the H. cordata rhizome phenolic profile (specifically, chlorogenic acid, rutin, afzelin, isoquercitrin, quercitrin, and quercetin) were analyzed to test whether the LPFs (IL, UFS and PC) affected the secondary plant metabolites in the host’s rhizomes (Fig. S3). It might play a particular role in H. cordata defense when the sterile H. cordata seedlings cultured in non-sterile soil were challenged by the LPFs. The changes in the concentrations of chlorogenic acid, rutin, afzelin, isoquercitrin, quercitrin, and quercetin in H. cordata rhizomes, colonized by IL, UFS, and PC, are shown in Fig. 8 and Table SIII. The results showed that quercetin was not detected in any of the H. cordata rhizome samples, CK, or challenged. Furthermore, the concentrations of chlorogenic acid in the H. cordata rhizomes (19.12 mg/kg, IL; 24.62 mg/kg, UFS; 20.60 mg/kg, PC) were significantly higher than in the CK (7.67 mg/kg) (Fig. 8A and Table SIII). Except for the rutin concentration (3.94 mg/kg) in the H. cordata rhizomes inoculated with PC, the rutin concentrations (2.18 mg/kg, IL; 2.13 mg/kg, UFS) in the H. cordata rhizomes inoculated with LPFs were not significantly different from those in the CK (2.16 mg/kg) (Fig. 8B and Table SIII). Similarly, the concentrations of afzelin (188.78 mg/kg) and isoquercitrin (9.57 mg/kg) in the H. cordata rhizomes inoculated with PC were significantly higher than in the CK (afzelin, 85.64 mg/kg; isoquercitrin, 4.69 mg/kg), whereas the concentrations of afzelin and isoquercitrin in the H. cordata rhizomes inoculated with IL and UFS were not significantly different from those in the CK (Fig. 8C-D, and Table SIII). The concentrations of quercitrin in the H. cordata rhizomes (1.10 mg/kg) inoculated with IL were significantly lower than in the CK (2.65 mg/kg), whereas the concentrations of afzelin (4.89 mg/kg, UFS; 3.30 mg/kg, PC) in the H. cordata rhizomes inoculated with UFS and PC were significantly higher than in the CK (2.65 mg/kg) (Fig. 8E and Table SIII).

Fig. 8.

We identified three endophytic fungi I. liriodendra, an unidentified fungal sp., and P. citrinum, which had colonized H. cordata rhizomes in the wild but not affected the growth or caused symptoms in H. cordata plants, forming stable relationships with H. cordata and other colonizing microbiota (including its rhizomes). We verified the pathogenicity of these three fungi and defined them as LPFs. The relationships between LPFs and host plants have previously been reported (Arnold and Lutzoni 2007; Delaye et al. 2013; Brader et al. 2017; Bamisile et al. 2018; Etalo et al. 2018). The various LPFs isolated from many healthy plants have been shown to convert between endophytic and pathogenic stages in response to different environmental factors (Hyde and Soytong 2008; Fernandes et al. 2018; Chen et al. 2020). Ilyonectria spp. were also reported to cause rusty root of Panax ginseng (Lu et al. 2015; Martínez-Diz et al. 2018), critical microbiota involved in rusty roots of P. ginseng was reported to have substantial inhibiting effects on Ilyonectria (Liu et al. 2019). P. citrinum was also confirmed as LPFs of Cannabis (Jerushalmi et al. 2020). According to Fig. 2, UFS may come from Pleosporale. Pleosporale spp. were also reported as pathogens of wheat (Rybak et al. 2017). However, it has not been reported that these fungi caused diseases of H. cordata in the wild. In the current study, three species of LPFs were isolated from healthy H. cordata rhizomes. This may indicate that the stable colonization of these three LPFs in H. cordata had been formed, which influenced and restricted each other. For example, IL could infect and significantly influence the growth of sterile H. cordata seedlings grown in sterile soil and inoculated with each of the three fungi (Fig. 4A-D), (Fig. 4B), proving that this fungus has a powerful potential to cause host disease. Similarly, UFS (Fig. 4C) and PC (Fig. 4D) also negatively affected the growth and vigor of the sterile H. cordata seedlings. At the same time, the sterile H. cordata seedlings growing in the non-sterile soil (Fig. 4E-H) were not influenced by subsequent inoculation with any of the three LPFs (Fig. 4F-H). The only difference between these two infection experiments on sterile seedlings were microorganisms in the non-sterile soils. When the sterile H. cordata seedlings were cultured in non-sterile soils, some of the microorganisms from the soil had presumably colonized the rhizomes of the sterile H. cordata seedlings and formed colonizing microbiota. These results obtained prove that the influence of IL, UFS, and PC, which stably colonize in H. cordata, was regulated by the colonizing microbiota of H. cordata, originating from the soil microorganisms. Interestingly, the growth of branches and leaves of sterile H. cordata seedlings in the non-sterile soils was more luxuriant than that in the sterile soils (Fig. 4A and 4E), suggesting that the colonizing microbiota of H. cordata rhizomes could play an important role in improving the growth of H. cordata plants.

All colonizing microbes stably live in the same tissues of plants (Kia et al. 2019), a situation which would be expected to inevitably lead to competitive or synergistic relationships among the various microbes. In this study, the confrontation (“antagonism”) assay revealed interspecific competition among IL, UFS and PC, the mycelial extension of each LPF being significantly suppressed by antagonistic action, although none of the strains died under these circumstances (Fig. 6A-C). The mutual restriction of hyphen growth may help maintain the balance of each LPF and may be one of the important ways by which the diversity of fungi is maintained (Cline and Zak 2015; Luo et al. 2019). The mycelial extension between IL and PC (Fig. 6A), IL and UFS (Fig. 6B), and PC and UFS (Fig. 6C) were inhibited and was less affected by growing on the same plate by another colony of the same strain. This suggested that antagonism between species was one of the factors responsible for maintaining the balance among IL, UFS, and PC in H. cordata rhizomes. In addition, no contact or minimal contact between the mycelium of the different fungi suggested that the inhibition of LPFs might be achieved by producing water-soluble or volatile inhibitory chemicals (Rajani et al. 2021).

Previous studies have reported that H. cordata contains a large number of secondary metabolites, such as essential oils, flavonoids, water-soluble polysaccharides, alkaloids, organic acids, fatty acids, and sterols (Yang and Jiang 2009; Fu et al. 2013), with some of the secondary metabolites being proven to have therapeutic activities including anaphylactic inhibitory, antiinflammatory, antiviral, antibacterial, and antiallergic effects (Kumar et al. 2014). Antifungal assays of plant extracts always focus on pathogenic fungi (Huang et al. 2012). The colonizing microbes live stably together within the plants (Kia et al. 2019), and the metabolic network is modulated selectively by microbiota in the plant roots (Huang et al. 2019). We considered that the stable relationship between the LPFs and H. cordata might be related to the accumulation of the secondary metabolites of H. cordata. However, the factors driving stable colonization are poorly understood. The secondary metabolites biosynthesized in H. cordata during natural growth must have physiological and ecological functions (including inhibiting the growth of fungi and other microorganisms) (Etalo et al. 2018). Recent studies showed that the root microbiota of Arabidopsis selectively modulated a metabolic network reported in microbiome research (Huang et al. 2019), with beneficial colonizing root microbiota modulating host plant chemistry (Etalo et al. 2018). There is reason to believe that if H. cordata rhizome extracts can inhibit the LPFs (such as IL, UFS, and PC), some secondary metabolites may play an important role in regulating these LPFs. The results showed that, except for PC (%Table II and Fig. 7A), there were significant dosage-dependent antifungal activities in H. cordata rhizome extracts on IL (IR = 11.97–23.08%) and UFS (IR = 3.37–12.36%; %Table II), and that the average diameters of colonies of IL and UFS on PDA with different concentrations of H. cordata rhizome extracts were significantly lower than in the CK (p < 0.05; Fig. 7A–B). It indicated that some of the components in the H. cordata crude rhizome extract, presumably secondary metabolites, might regulate the growth of IL and UFS. On the other hand, PC appeared not to be regulated by H. cordata rhizome secondary metabolites but were certainly controlled by other factors (including other colonizing microorganisms).

Colonizing beneficial microbiota of plant roots can modulate the secondary metabolites reported (Etalo et al. 2018), and the metabolic network modulated plant root microbiota (Huang et al. 2019). In light of this, we decided to investigate whether LPFs influence the secondary metabolites of H. cordata. The experimental results showed that IL, UFS, and PC changed the accumulation of phenolics when colonized H. cordata, and the phenolics had antifungal activities (Oliveira et al. 2016; Martínez et al. 2017). IL significantly increased chlorogenic acid accumulation in H. cordata, relative to the uninoculated CK (p < 0.05, Fig. 8A). UFS significantly increased the accumulation of chlorogenic acid and quercitrin in H. cordata (p < 0.05, Fig. 8A, E), while PC significantly increased the accumulation of chlorogenic acid, rutin, afzelin, isoquercitrin, and quercitrin in H. cordata (p < 0.05, Fig. 8B–D). The increase in concentrations of H. cordata phenolics may be a stress response of the plant to the challenge by the IL, UFS, and PC. For example, when IL infected the sterile H. cordata seedlings, the seedlings must respond to inoculation with IL, and provide a “countermeasure”, apparently by increasing the concentration of chlorogenic acid in H. cordata to inhibit the growth of IL in H. cordata. Similarly, increasing the concentrations of chlorogenic acid and quercitrin in H. cordata inhibited the growth of UFS, whereas increasing the concentration of chlorogenic acid, rutin, afzelin, isoquercitrin, and quercitrin in H. cordata inhibited the growth of PC. These findings suggested that colonization of H. cordata by LPFs may be controlled by the secondary metabolites, particularly flavonoids, in H. cordata, which appears to be among the factors driving to form the symbiotic relationships with the host.

Usually, colonizing microbes are considered to have “friendly” relationships with host plants (Jia et al. 2016), so that long-term colonizing microbiota will not only not cause disease in their host plants but will even provide various benefits, improving host plant growth (Yan et al. 2019). In the absence of colonizing microbiota, IL, UFS, and PC, isolated from healthy H. cordata rhizomes, inhibited H. cordata growth (Fig. 4B-D), which indicates that IL, UFS, and PC are LPFs, and could cause infections to H. cordata. In the presence of colonizing and/or symbiotic microorganisms, however, the LPFs did not influence H. cordata growth (Fig. 4A) and could thrive in H. cordata rhizomes without causing overt symptoms of H. cordata disease. It suggests that colonizing and/or symbiotic microorganisms may modulate the balance of the LPFs in H. cordata rhizomes. In addition, our antagonism assays show that IL, UFS, and PC control one another (Fig. 6) and that there are significant antifungal activities toward IL and UFS in H. cordata rhizome extracts on IL and UFS (Fig. 7). It also indicated that these were factors controlling the stable colonization of IL, UFS, and PC in H. cordata rhizomes. IL, as an LPF, could be recognized by the plant immune system and provide a defense response (Nishad et al. 2020). Penicillium spp. and other LPFs also have antifungal activity (Pan et al. 2016; Garrigues et al. 2018). Therefore, when it colonized the host plant tissue, IL suffered at least two types of stresses, from other colonization fungi and the host plant. It did not cause disease and cannot be eliminated. Sensitivities of the three LPFs to H. cordata extracts were different, which meant that H. cordata regulated LPF diversity selectively. Our findings may explain why the LPFs thrive within H. cordata rhizomes without causing overt symptoms.

In summary, this study results showed that I. liriodendri, an unidentified fungal species, and P. citrinum were LPFs colonized in H. cordata rhizomes. Maintaining stable colonization of the fungi in H. cordata rhizome was controlled by the colonizing microbiota of H. cordata, originating from the soil microorganisms, by mutual inhibition and competition among the fungi. LPFs colonized the H. cordata rhizome tissues without causing disease, and their behavior was modulated by the secondary metabolites of H. cordata induced, in turn, by challenge by the LPFs. Our future research will focus on characterizing the microbiota that modulates pathogenic fungi in H. cordata rhizome tissues, elucidating the microbiota composition, and identifying how regulation of the secondary metabolites responds to colonizing microbiota.