Evaluation of foliar fungus‐mediated interactions with below and aboveground enemies of the invasive plant Ageratina adenophora

Abstract Plant–fungal associations are frequently key drivers of plant invasion success. Foliar fungi can benefit their invasive hosts by enhancing growth promotion, disease resistance and environmental stress tolerance. However, the roles of foliar fungi may vary when a given invasive plant faces different stresses. In this study, we designed three independent experiments to evaluate the effects of a foliar fungus, Colletotrichum sp., on the growth performance of the invasive plant Ageratina adenophora under different soil conditions, as well as the responses of A. adenophora to the foliar fungal pathogen Diaporthe helianthi and to herbivory. We found that the soil type was the most influential factor for the growth of A. adenophora. The role of the foliar fungus Colletotrichum sp. varied in the different soil types but generally adversely affected leaf development in A. adenophora. Colletotrichum sp. may be a weak latent foliar pathogen that can enhance the pathogenicity of D. helianthi on leaves of A. adenophora and marginally reduce signs of herbivory by natural insects in the wild on A. adenophora seedlings. In general, the benefits of the foliar fungus Colletotrichum to the fitness of A. adenophora are not significant in the context of this experimental design. However, our data highlight the need to consider both aboveground and belowground biota in different soil habitats when evaluating the effects of foliar fungi.


| INTRODUC TI ON
The rapid expansion of exotic plants has caused serious damage to the structure and function of invaded ecosystems and has resulted in enormous social and economic losses (Richardson & Ricciardi, 2013;Vilà et al., 2010). Determining why invasive plants succeed in their introduced range has been a major goal of invasion ecology. Plant-fungal associations are frequently key drivers of plant invasion success; a fungal species can act as a mutualist by enhancing plant defense, growth and stress tolerance or as a pathogen to cause establishment failure (Dickie et al., 2017;Flory & Clay, 2013). Foliar fungi are diverse in terrestrial ecosystems, including endophytes and epiphytes, and are often assumed to be mutualists, latent saprotrophs or pathogens of plants (Arnold, 2007;Osono et al., 2004;Schulz et al., 1999). Foliar fungi have also been observed to be capable of altering plant disease severity (Arnold et al., 2003;Busby et al., 2016). For exotic plants, foliar fungi can assist the host to grow, resist disease, and change herbivory.
For example, Aschehoug et al. (2012) indicated that the inoculation of endophytic Alternaria (CID120) on the invasive forb Centaurea stoebe could directly increase its growth. Currie et al. (2019) demonstrated that foliar fungi (Colletotrichum acutatum, Alternaria alternata and Cladosporium oxysporum) appeared to mitigate the effect of the pathogen Puccinia komarovii on the invasive weed Impatiens glandulifera. Rudgers and Clay (2008) showed that the foliar fungus Neotyphodium coenophialum could significantly reduce the abundance and diversity of herbivorous arthropods on the invasive plant Lolium arundinaceum.
In addition to aboveground biota, invasive plants also interact with belowground biota. Escape from soil-specific host pathogens has been suggested to play a key role in determining the success of exotic plant invasion into local ecosystems, and this phenomenon is referred to as positive plant-soil feedback (PSF) (Callaway et al., 2004;Hawkes et al., 2005;Klironomos, 2002;Kulmatiski et al., 2008).
Nonetheless, exotic plants have not experienced uniformly positive PSF, for example, Andonian et al. (2011) indicated that an invasive weed experienced negative PSF; Crawford and Knight (2017) found that Lespedeza cuneata experienced positive PSF in invaded soils only in the absence of competition; and for some plant species, a positive PSF was observed in a glasshouse but not in the field ((Schittko et al., 2016). These discrepancies suggest that the PSF experienced by the exotic invaders is context dependent and may involve the soil habitat type in which the plants grow (Pizano et al., 2019).
More interestingly, interactions of belowground and aboveground biota may exist when performing PSF studies. For example, the invasive plant I. glandulifera experienced a positive PSF, and the effect of the PSF even extended beyond the soil microbial community to affect foliar fungi, which in turn enhanced resistance to herbivory in I. glandulifera and thus accentuated the invasive properties of this species (Pattison et al., 2016). In reality, invasive hosts expand into different habitats across a large geographical scale (Vilà et al., 2010) and potentially interact with diverse belowground biota (e.g., soilborne pathogens) and aboveground biota (e.g., foliar fungi). However, it is unclear whether the invasive host experiences distinct soil biota effects in different habitats and how the roles of the soil biota change after a foliar fungus is introduced; in addition, it is also unknown whether a foliar fungus can modify both aboveground diseases and herbivory in the invasive host.
The invasive plant Ageratina adenophora (Sprengel) R. M. King and H. Robinson is a perennial herb of the Compositae family that is native to Central America. The spread of A. adenophora is considered to be a severe problem in more than 30 countries in Asia, Africa, Oceania, Europe and North America (Datta et al., 2019). In China, this weed has been reported to reduce the diversity of native plant species, crop production in agricultural land and forage production in pastures; moreover, this weed is poisonous to domestic animals (Poudel et al., 2019). Therefore, understanding the mechanisms of successful weed invasion is ecologically and economically important. Previous results have indicated that the roles of soil biota in the establishment of A. adenophora are habitat dependent, for example, Niu et al. (2007) indicated that soil biota collected from mixed evergreen-broadleaf-deciduous forests had strong inhibitory effects on A. adenophora growth; however, Xiao et al. (2014) indicated that sterilization had no significant effect on A. adenophora biomass growing in soils collected from a tropical botanical garden. Nevertheless, once established, A. adenophora can change the soil biota to promote its own growth (Niu et al., 2007), for example, by increasing the abundance of nitrogen-fixing bacteria (Xu et al., 2012) and root endophytic Enterobacter bacteria . Moreover, A. adenophora has also been reported to host various foliar fungi, including pathogens (Poudel et al., 2019;Zhou et al., 2010) and endophytes (Mei et al., 2014. In addition, foliar herbivory was frequently observed in our routine investigations. These data suggest the occurrence of potential aboveground and belowground biotic interactions with A. adenophora, which remain to be characterized. In this study, we designed three independent experiments to determine the effect of a foliar fungus, Colletotrichum, on the growth performance of A. adenophora in different soil conditions, as well as the functional responses of A. adenophora to aboveground pathogenic fungi and herbivores.

| Cultivating seedlings of A. adenophora
Seeds were collected from wild populations of A. adenophora in Kunming city (24°58′22″ N, 102°27′49″ E, 1980 m), Yunnan Province, Southwest China. Completely filled seeds were selected and surface sterilized by submerging them in 95% ethanol for 30 s and 2% sodium hypochlorite for 20 min and rinsing them with sterile water three times. After surface sterilization, the seeds were submerged in sterile water for 24 hr and then planted in pots containing sterile soils. The soils used to grow the plants were sterilized three times at 24-hr intervals by autoclaving (121°C, 0.135 MPa, 2 hr). Pots were disinfected by submerging them in 0.2 g/L potassium permanganate solution for 30 min. Plants were cultivated in an RXZ-380D growth chamber (Ningbo Southeast Instrument Co., Ltd, Ningbo, China) with a temperature of 25 ± 1°C, light intensity of 80-100 µmolm −2 s −1 , 12-hr photoperiod and humidity of 80 ± 5%. The germinated seedlings were supplemented with the same amount of sterile water in each pot every day as needed and watered with Hoagland nutrient solution (detailed in Table 1) once a month. The seedlings were grown for 5 months, and then those of similar sizes (approximately 20 cm shoot length) were randomly selected for the subsequent experiments.

| Inoculation and detection of foliar fungi
The fungus JK99 was previously isolated from healthy leaves of A. adenophora, making up nearly 30% of the total isolated fungi (Mei et al., 2014). It has been identified as Colletotrichum sp. (see colony morphology in Figure S1. Its ITS gene was sequenced and was completely identical to the sequence of a strain of Colletotrichum sp. A285 that we previously submitted to GenBank with an accession number of MK247540). Calcium carbonate medium (1,000 ml of distilled water containing 30 g of CaCO 3 , 20 g of glucose and 18 g of agar) was used for sporulation at 28°C. The preparation of the spore suspension and inoculation were performed as previously reported (Arnold et al., 2003). Briefly, the spores produced on the plate were washed with 0.5% sterilized gel solution (100 ml of distilled water containing 0.5 g of gel). The mycelium was removed with sterile filter paper, and the spore suspension was diluted to 10 6 CFU/ml using a haemocytometer. Then, the A. adenophora leaves were misted to saturation on both the upper and lower surfaces using this spore suspension with a sterile sprayer, and the control group was sprayed with sterile 0.5% gel solution. The whole plant was wrapped in a sterile black plastic bag for 48 hr to promote fungal infection, and the fungal infection was detected after 15 days, according to Arnold et al. (2003). The leaves inoculated and not inoculated (control plants) with Colletotrichum sp.
(named Col+ and Col−, respectively) were surface sterilized by submerging in 0.5% sodium hypochlorite for 2 min and 70% ethanol for 2 min and rinsing with sterile water 3 times, and then the surface-sterilized leaves were cut into fragments of approximately 6 mm 2 . Sixteen fragments were randomly selected from each leaf sample, placed on a malt extract agar (MEA, 1,000 ml of distilled water containing 30 g of malt extract, 3 g of soy peptone and 18 g of agar; pH 5.6 ± 0.2) plate, sealed with parafilm and cultured at room temperature for several days ( Figure S2). The number of leaf fragments with fungi and the colony morphology were observed and recorded. The fungal infection rate of the leaves (Col+ and Col−) was calculated by dividing the number of fragments with fungi by the total number of fragments.

| Growth experiments with A. adenophora in different soil types
The soils used in this experiment were collected from agricultural land and a forest in Kunming city, Yunnan Province, Southwest China, and are referred to as agricultural soil (AS) and forest soil (FS).
The agricultural soil collection site commonly rotates maize, pea and cabbage crops. The forest soil collection site is dominated by Pinus yunnanensis and Cyclobalanopsis glaucoides. Both sites have red soil.
The soil properties are shown in Table 2.
A total of 8 treatments were used to evaluate the effects of the different soil conditions on the growth of A. adenophora seedlings, that is, 2 soil types (AS, FS) × 2 treatments (soil sterilization, nonsterilization) × 2 inoculations (with, without Colletotrichum sp.) ( Figure S3).
Because we focused on the role of soil microbes in the growth of the host plant, the soil as an inoculation source was mixed with a sterilized matrix at a volume ratio of 1:9 to lessen the nutrient effects of the different soil types (Whitaker et al., 2017). The sterilized matrix was a mixture of equal proportions of agricultural soil and forest soil, and high-pressure steam sterilization was carried out as described above. Then, 5-month-old A. adenophora seedlings (Col+ and Col−) were transplanted into sterilized pots with the different soil treatments. The plants were cultured in an RXZ-380D growth chamber (Ningbo Southeast Instrument Co., Ltd, Ningbo, China) at a temperature of 25 ± 1°C, light intensity of 80-100 µmolm −2 s −1 , 12-hr photoperiod and humidity of 80 ± 5% and watered with the same amount of sterile water as needed every day. After 2 months, the leaf dry biomass (LDB) per unit area, aboveground dry biomass (ADB), belowground dry biomass (BDB) and shoot length (SL) were measured. Six seedlings (replicates) were planted in each treated soil, and three of the seedlings were randomly selected for fungal infection rate (FIR) detection. The FIR was significantly different between the seedlings with and without Colletotrichum sp. inoculation (Figure 1, FIR(Col−) = 1.46%, FIR(Col+) = 19.17%). Since half of the seedlings were used to measure the fungal infection rate, the biomass data (ADB, BDB) included only the remaining three replicates in each treatment, and the SL measurements had 6 replicates. The LDB per unit area was measured by punching the first five pairs of mature leaves with a 5-mm

| Inoculation experiment with leaf pathogenic fungi
The pathogenic fungus used in this experiment was previously isolated from spots on A. adenophora leaves, and a phylogenetic analysis of the ITS gene indicated that it was close to Diaporthe helianthi (JK58) (see colony morphology in Figure S4, with GenBank accession No. JN854227 for the ITS gene). Its pathogenicity ( Figure S5) was verified by the method previously described by Gilbert and Webb (2007). Briefly, the pathogenic fungi cultured on potato dextrose agar (PDA, 1,000 ml of distilled water containing 200 g of potato, 10 g of dextrose and 18 g of agar) medium were made into agar disks with a 3-mm internal diameter punch. A sterile needle was used to lightly wound the underside of the leaf. The agar disk was then at-

| Herbivory experiment
To date, no report has characterized the aboveground insect en-  Province. After a week in the field, the seedlings were brought back to the laboratory to record feeding by a guild of natural herbivores in the wild. In most A. adenophora individuals, only rare herbivory occurred on very few leaves during the period of our experiment; thus, we only evaluated signs of herbivory from insects attracted to A. adenophora. We used "1" to represent the occurrence of feeding when one individual of A. adenophora was observed to show herbivory, regardless of the amount of herbivory damage, and "0" to represent that no feeding occurred. The signs of herbivory were calculated and compared between Col+ and Col− seedlings. Each treatment contained 14 individuals, for a total of 28 seedlings.

| Data analysis
Nonparametric Mann-Whitney U tests were used to identify the differences in growth between inoculated and noninoculated seedlings, sterilized and nonsterilized soil and agricultural and forest soil, as well as differences in the fungal infection rate, leaf spot area and seedling feeding frequency in the wild between Col+ and Col− leaves. Linear regression was used to analyze the relationships between leaf spot area and leaf age as well as infection time.
All analyses were performed using SPSS, version 22.0 (SPSS Inc., Chicago, IL, USA). Visualizations of the fungal infection rate, growth performance, leaf spot area and feeding frequency were generated using GraphPad Prism 7 (GraphPad Software Inc., San Diego, CA, USA).

| Effects of the foliar fungus Colletotrichum sp. on the growth of A. adenophora depending on the soil type
The

| Effect of the foliar fungus Colletotrichum sp. on herbivory on A. adenophora leaves
Inoculation with Colletotrichum sp. marginally reduced the signs of herbivory on seedlings in the wild (p = 0.063) ( Figure 6).

F I G U R E 3
The infection rate of foliar fungus Colletotrichum sp. in herbivory experiment. Col+ and Col− represent seedlings with and without Colletotrichum sp., respectively. Nonparametric Mann-Whitney U tests were used to identify the difference in fungal infection rate between leaves with and without Colletotrichum sp. (* represents p < 0.05). The error bar represents the standard error

| D ISCUSS I ON
Local soil biotic and abiotic factors ultimately determine the positive or negative effects of PSF on plants (Bennett & Klironomos, 2019).
Our report characterized the growth response of the invasive plant A. adenophora when grown in agricultural and forest soils upon interaction with a foliar fungus. We found that the soil source played the most important role in the growth of A. adenophora. Although the forest soil showed significantly higher levels of macronutrients, such as N, P, and K, than the agricultural soil (Table 2), on average, all measurements for A. adenophora growth were lower in forest than in agricultural soil (Figure 4). Because the soil, as an inoculation source, was mixed with a sterilized matrix at a volume ratio of 1:9 and watered with Hoagland nutrient solution during the growth of A. adenophora (see Methods), the effects of the macronutrients in the soils on A. adenophora were likely lessened.
These data thus reflected that the unmeasured micronutrients in the agricultural soil still contributed to the growth of A. adenophora. For example, Zhang et al. (2004) found that A. adenophora tends to enrich soil selenium.
Regarding the soil biota effects, we found that sterilization caused declines in most measures of A. adenophora growth in forest soil, including LDB per unit area, ADB and SL, and conversely, improved the LDB per unit area in agricultural soil (Figure 4). These findings suggested that A. adenophora experienced positive soil biota effects in forest soil but negative or neutral effects in agricultural soil, supporting previous results that soil biota from different habitats have distinct inhibitory effects on A. adenophora growth F I G U R E 4 Effects of the foliar fungus Colletotrichum sp. on the growth of A. adenophora in different soil conditions. Col+ and Col− represent seedlings with and without Colletotrichum sp., respectively; AS and FS represent agricultural and forest soil, respectively; S and NS represent sterilization and nonsterilization, respectively. Nonparametric analysis was used to compare the differences in growth performance between Col+ and Col− seedlings, sterilized and nonsterilized soils and agricultural and forest soils (n.s. represents p > 0.10, + represents p < 0.10, * represents p < 0.05, ** represents p < 0.01). The error bar represents the standard error. Since one Col+ seedling died in the S-FS treatment, statistical analysis between the Col+ and Col− seedlings was not performed for the biomass measurements (b, c) (Niu et al., 2007;Xiao et al., 2014). The negative feedback in agricultural soil indicates that heavy loads of pathogens exist in these soils (Ashizawa et al., 2010;Etebu & Osborn, 2010;Kohn, 1995). In contrast, the positive effect of forest soils on A. adenophora growth may be related to the high abundance and/or diversity of beneficial microbes, for example, AM fungi, in intact forest soils (Bordoloi et al., 2015). The invasiveness of a given plant and the invasibility of the ecosystem together determine the successful colonization of exotic plants in non-native habitats (Milbau & Nijs, 2004). Therefore, the positive feedback in forest soil does not indicate that A. adenophora invades this habitat more easily than agricultural soils, which may contain unknown nutrients that support A. adenophora growth better than forest soils. Great variation can be observed even among agricultural soils (Franklin & Mills, 2003) based on many factors, and the same is true for forest soils (Yang et al., 2018). Therefore, a variety of factors need be ruled out before multiple soil types in developed faster with increasing leaf age ( Figure 5b). These results suggest the existence of a synergism between these two fungi that causes them to be more virulent to A. adenophora. Coinfection is common in nature and usually increases the pathogenicity of pathogens to the host (Laine, 2011). For example, coinfection with Verticillium dahliae and Colletotrichum coccodes causes more severe foliar disease symptoms and crown rot in potato (Nicola) than inoculation with either of the two pathogens separately (Tsror & Hazanovsky, 2001).
In addition, although Colletotrichum sp. JK99 has been isolated from healthy leaves and reported as an endophyte of A. adenophora (Mei et al., 2014), we found that healthy Col+ leaves developed a larger leaf spot area than healthy Col− leaves only through puncturing (Figure 5d), and symptoms worsened with leaf age (Figure 5e).  (Viret & Petrini, 1994). Colletotrichum has various life habits and can be broadly categorized as necrotrophic, hemibiotrophic, latent or quiescent and endophytic (De Silva et al., 2017). Some endophytic fungi, such as latent pathogens and quiescent pathogens, may induce disease symptoms later when the plant tissues age or become weakened (Kogel et al., 2006;Viret & Petrini, 1994). Indeed, very recently, our group indicated that many members of Colletotrichum sp. were asymptomatic latent pathogens of both A. adenophora and co-occurring native plants (Chen et al., 2020). Nonetheless, the disease symptoms elicited by Colletotrichum sp. were very weak, and disease development was halted within two weeks after wounding ( Figure 5f).
Foliar fungi, as a secondary metabolite source for plants, can provide their hosts with antiherbivore defenses (Nisa et al., 2015).
A recent review indicates that fungal endophytes commonly have a negative effect on insect herbivores . Our data also showed that inoculation with Colletotrichum sp. marginally reduced the signs of herbivory on intact A. adenophora seedlings by natural insect enemies in the wild ( Figure 6). The plant pathogen Colletotrichum sp. has previously been reported to cause extensive mortality in an insect herbivore by directly colonizing insect tissue (Graziosi & Rieske, 2015). Similarly, the members of Colletotrichum have been observed to infect scale insects (Marcelino et al., 2008). However, it is not known whether Colletotrichum species produce defense compounds to directly kill insects. Here, it was also unclear how Colletotrichum sp. JK99 modifies the herbivory of A. adenophora.
In conclusion, our preliminary study characterized the interactions among an invasive plant, a foliar fungus and natural enemies aboveground and belowground. We found that the soil type played the most important role in the growth of A. adenophora, and the role of the foliar fungus, Colletotrichum sp., varied according to the habitat but primarily adversely impacted leaf development in A. adenophora. The foliar fungus Colletotrichum sp. might be a latent pathogen, and it enhanced the pathogenicity of D. helianthi on A. adenophora leaves; this fungus also marginally reduced the signs of herbivory on seedlings from natural insect enemies in the wild. Previously, invasive hosts have been shown to benefit from the presence of foliar fungi Currie et al., 2019;Evans, 2008;Xiao et al., 2012).
The adverse impacts in this study may be related to the latent pathogenic effects of this foliar fungus but also reflect the fact that the ecological functions of foliar fungi are usually context dependent (Hawkes & Connor, 2017). In particular, our study highlights the need to consider both aboveground and belowground biota in different habitats when evaluating the effects of foliar fungi.

Funding by Major Science and Technology Project in Yunnan
Province, China (No. K264202011020) and the National Natural Science Foundation of China, No. 30960077.

F I G U R E 6
Herbivores on the leaves of A. adenophora with and without the foliar fungus Colletotrichum sp. Col+ and Col− represent leaves with and without Colletotrichum sp. inoculation. Nonparametric Mann-Whitney U tests were used to identify differences in herbivory between Col+ and Col− leaves (+ represents p < 0.1). The error bar represents the standard error

CO N FLI C T O F I NTE R E S T
The authors declare no competing financial interests.