Arbuscular mycorrhizal fungi are a double‐edged sword in plant invasion controlled by phosphorus concentration

The invasion success of some exotic plants depends on associations with arbuscular mycorrhizal (AM) fungi, which range along a continuum from strong mutualism to parasitism frequently affected by soil phosphorus (P) availability. It is unclear how P availability shifts AM associations on native and invasive plants, which in turn influence their competition. In this work, we conducted a three-factor common garden experiment, including manipulations of P availability, AM fungi occurrence and interspecific competition to evaluate how AM fungi influence competition between two pairs of invasive and native plants under different P availabilities. Our results showed that P enrichment reduced positive AM effects on the growth of both native and invasive plants. Competition had no effect on AM colonization on the invasive plants, but reduced AM colonization on the native plants, which led invasive plants to receive greater AM benefits under low-P treatment, but to be exposed to stronger AM detriments under high-P treatment compared to the native competitors. Therefore, the competitive advantage of invasive vs native plants was enhanced by AM fungi under low-P treatment, but weakened under high-P treatment. Our work highlights the necessity to incorporate the interaction between soil microbes and nutrient availability on plant invasion in future studies. Due to relatively low host specificity of AM symbioses, AM fungi can form symbiotic associations with invasive species in nonnative habitats even though they lack a co-evolutionary history with these new hosts (Moora et al., 2011; Dickie et al., 2017). Interactions between plants and AM fungi may enhance species invasiveness by facilitating the establishment, persistence and dominance of invasive plants in new habitats (Pringle et al., 2009; Menzel et al., 2017). Previous studies showed that carbon and mineral nutrients can be transferred from native plants to invasive plants via common mycorrhizal networks (Carey et al., 2004; Awaydul et al., 2019). Thus, greater AM benefits on the invasive than the native species may be an important driver of AM-induced plant invasion. Moreover, plant invasion may magnify the differences of AM benefits on invasive and native plants, which in turn further facilitates invasion. Previous research has shown that invasive and native hosts significantly differ in their preferences for, and dependencies on, different AM taxa (Hawkes et al., 2006; Zhang et al., 2017). On the one hand, invasive plants can selectively accumulate the most beneficial fungal species in favour of their own growth (Reinhart & Callaway, 2006; Zhang et al., 2010). On the other hand, invasive plants can shift the composition of the mycorrhizal community associated with the roots of native species (Stinson et al., 2006; Vogelsang & Bever, 2009) and make their native neighbours suffer from increased relative abundance of low-quality mutualists. A meta-analysis showed that native plants indeed suffer from reduced AM colonization when in competition with invasive plants compared with when they are growing alone, and are often less AM colonized than their invasive counterparts (Bunn et al., 2015). Higher AM colonization on the invasive plants than on the natives may provide competitive advantages to some invaders over the natives (Zhang et al., 2017). However, high AM colonization is not always good for plants, especially when AM effects are negative (Garrido et al., 2010). The effect of AM associations on plants ranges from strong mutualism to parasitism which is strongly influenced by soil nutrient availability (Johnson, 2010). It is uncertain how the nutrient availability alters AM effects on native and invasive plants, which in turn influences their competition. Moreover, P availability is one of the most important factors on the AM fungal symbiont. Where 4–20% of the plant photosynthates are allocated to AM fungi in exchange for nutrients (Garrido et al., 2010; Lendenmann et al., 2011). Mycorrhizal benefits are reduced by P enrichment, because plants can acquire sufficient P without the help of AM fungi under such conditions (Schroeder & Janos, 2005; Elbon & Whalen, 2015). When the benefits AM fungi provide fail to compensate for their demands on plant carbon, AM symbiosis can be regarded as 'parasites' despite the facilitation of P acquisition by AM fungi (Johnson & Graham, 2013). Under high-P conditions, maintaining a high level of AM colonization results in unnecessary carbon output, and leads to growth depression (Smith & Smith, 2011). It remains unclear whether the higher AM colonization on invasive plants than their native neighbours under high-P conditions will result in stronger negative AM effects on them, which in turn results in the loss of competitive advantages against natives. The purpose of this study is to understand how P concentration regulates AM associations with native and invasive plants, and thus influences their competition. We conducted an experiment with a full factorial design to test the competitive ability of invasive plants across three different soil P treatments with or without AM removal. We hypothesized that: (1) AM colonization on native plants would reduce under competition with invasive plants. Because P increase would reduce the positive AM effects on the plant growth, we hypothesized that (2) P increase would weaken the competitive advantage of invasive vs native plants. Two AM plant pairs were used in the experiments, each consisting of a highly invasive plant species and a native competitor: Eupatorium catarium vs E. chinense, and Bidens pilosa vs Sesbania cannabina (Supporting Information Methods S1). The soils from E. catarium and E. chinense rhizosphere were collected and combined at a 1 : 1 (v/v) ratio and then mixed with steam sterilized sand at a 1 : 5 (v/v) ratio; the same procedure was performed for the soils from B. pilosa and S. cannabina rhizosphere. Total nitrogen (N) and P were 0.23 g N kg−1 and 0.10 g P kg−1 in the mixed E. catarium–E. chinense soil, respectively, and 0.17 g N kg−1 and 0.10 g P kg−1 in the mixed B. pilosa–S. cannabina soil. A three-factor common garden experiment was conducted that consisted of combinations of P, AM removal and competition treatments (Fig. S1). All plants were grown in 8-l pots (23 cm in diameter, 21 cm in height). Three levels of P treatments were established by adding sodium dihydrogen phosphate (NaH2PO4; low, no addition; medium, 20 mg P kg−1 soil; high, 100 mg P kg−1 soil). AM removal treatments were realized by applying or not applying fungicide, benomyl (C14H18N4O3, 50 mg kg−1 soil per month) (Marler et al., 1999). Benomyl has been shown to effectively reduce AM colonization in glasshouse soils with minimal direct effects on plants (Hetrick et al., 1990; Smith et al., 2000), and widely used to investigate the effects of AM fungi on the interactions between native and invasive plants (Callaway et al., 2003; Grilli et al., 2014). For the competition treatment, a plant individual was either grown alone or in competition with its counterpart of the same species pair. Eupatorium catarium and E. chinense seedlings were only planted in the mixed E. catarium–E. chinense soil, and B. pilosa and S. cannabina seedlings were only planted in the mixed B. pilosa–S. cannabina soil. Each pot was watered daily and provided with 250 ml of a no-P Hoagland nutrient solution (Hoagland & Arnon, 1950) every month. Three months after transplantation, all seedlings were harvested. We measured the biomass, shoot N and P contents, AM colonization of each plant and calculated the mycorrhizal responsiveness (MR) (Grman, 2012), the tolerance and suppression relative interaction intensity (RII) index (Fletcher et al., 2016) and the correlation between AM colonization and the total biomass of the plants. The MR calculated the differences in the total biomass of plants grown without fungicide relative to those with fungicide under the same P treatment. Tolerance RII represents the ability of invasive plants to maintain growth performance in competition, whereas suppression RII represents the ability of invasive plants to suppress the growth of the native competitors. Higher tolerance RII and suppression RII reflect stronger competitive ability of invasive plants. More detailed descriptions of the methods, MR and RII calculation and statistical analysis are presented in Methods S1. For all studied species, AM colonization was unaffected by P addition, but was significantly reduced by fungicide application (Fig. 1a; Table S1). Competition reduced AM colonization on the two native plants by more than half, but had little effect on that of the two invasive plants (Fig. 1a; Table S1), supporting our first hypothesis, which suggested a competition-induced reduction in the AM colonization exclusively on the native species. Zhang et al. (2018) also found that B. pilosa can reduce the AM colonization of the native plant Setaria viridis in competition. Some invasive plants can alter AM fungal communities and suppress the AM colonization of their native neighbours by secreting allelochemicals (Mummey & Rillig, 2006; Dieng et al., 2015). In fact, both E. catarium and B. pilosa have strong allelopathic effects (Zeng & Luo, 1993; Lin, 2008), and Filho et al. (2012) isolated a flavonoid (5,7,4′-trimethoxyflavone), an allelochemical with antifungal potential from E. catarium. Allelopathic effects on the two invasive plants may be a possible mechanism for the AM colonization reduction on their native counterparts, but further studies are warranted. Without competition, we found that P addition significantly increased the total biomass of all studied species (Table S2; Fig. S2). According to the analyses of MR, all species exhibited positive MR under the low-P treatment, indicating AM facilitation on the growth of all species; all but one species, S. cannabina, exhibited negative MR under high-P treatment, indicating AM inhibition on the growth of all but one species (Table S3), and P addition reduced the positive effects of AM fungi on plant growth. We did not observe any reduction of AM colonization on any species in monocultures despite the occurrence of negative MR. These findings are consistent with those of Olsson et al. (2010), who found that AM colonization on plants was unaffected by P enrichment even when carbon was limited and AM fungi acted parasitically. In addition, the correlation between AM colonization and the total biomass of plants was positive for all species under low-P treatment, whereas it was negative for all species but S. cannabina under high-P treatment (Table S4). These results suggest that higher AM colonization is better for plant growth when MR was positive, but worse when MR was negative. With competition, we calculated RIIs of the two invasive plants to measure their competitive ability under different P treatments. Under low-P treatment, AM fungi removal (i.e. fungicide application) decreased the suppression RII of E. catarium and the tolerance RII of B. pilosa, but had little effect on the tolerance RII of E. catarium and suppression RII of B. pilosa (Fig. 1b; Table S5). However, under the high-P treatment, AM fungal removal resulted in increased tolerance RII and suppression RII of both invaders (Fig. 1b; Table S5). These results showed that the competitive ability of the two invasive plants was strengthened by AM fungi under low-P treatment but weakened by AM fungi under high-P treatment, supporting our second hypothesis. We propose that the change of MR and the correlation between AM colonization and the total biomass of plants under different P treatments may help explain the earlier results. Under low-P treatment, MR and the correlation between AM colonization and the total biomass were both positive (Tables S3, S4). In this case, maintaining a high level of AM colonization and disrupting the AM colonization of the native competitors provided the two invasive plants with greater competitive advantages over native plants. The advantage might come from the improvement in P acquisition by AM fungi suggested by many studies (Thingstrup et al., 2000; Elbon & Whalen, 2015). In our experiment, E. catarium, B. pilosa and S. cannabina had lower shoot P content after AM fungi removal (Table S6; Fig. S3), underlining that AM fungi facilitate P acquisition of the plants. Under low-P treatment, two native plants had lower shoot P content in competition than growing alone, whereas the same pattern was not observed in the two invasive plants (Table S7; Fig. S3). These findings indicate that higher AM colonization rate of the invasive vs native plants allowed the invaders to have stronger P acquisition ability than the natives in low-P conditions. Under high-P treatment, all plants had a low N : P ratio regardless of whether AM fungi had been removed (Table S6; Fig. S3), indicating a less essential role of AM fungi in P acquisition. Thus, maintaining high AM colonization might become a burden for plant growth. Because of the negative MR and the negative correlation between AM colonization and plant growth under high-P treatment (Tables S3, S4), we suggest that the competition-induced reduction in AM colonization reduced the negative AM effects on E. chinense, which indirectly reduced the competitive advantage of its invasive competitor, E. catarium. For the competition of another plant pair, because of the negative MR on B. pilosa and positive MR on S. cannabina under high-P treatment, AM fungi brought about a competitive disadvantage in B. pilosa (the invader) under high-P treatment. Brewer & Cralle (2003) also found that P addition reduced the invasiveness of Imperata cylindrica in the field, which also had a high AM colonization rate (Zhang et al., 2013) and the ability to decrease the AM colonizations on other species (Bajwa, 2005). Here, we propose a hypothetical framework to visualize the differential AM effects on the competition between native and invasive plants and how these effects may be altered by P enrichment and competition (Fig. 2). Increased P shifts AM effects on both native and invasive plants from beneficial to detrimental. Competition reduces the AM colonization on native plants, which in turn significantly weakens the AM effects on the native plants, whereas it has little effect on the AM colonization and AM effects of the invaders. Thus, competition enhances the relatively stronger AM benefits on the invasive vs native plants in low-P conditions, but also enhances the relatively stronger AM detriments on the invasive vs native plants in high-P conditions. Therefore, the AM-induced competitive advantage of the invasive plants is strengthened when soil P is limited but weakened when soil P is high. Our results demonstrated that AM facilitation of plant invasion is nutrient dependent. The competitive advantage of the invasive over native plants was enhanced by AM fungi in low-P conditions, but was weakened in high-P conditions. These highlight the necessity to incorporate the interaction between soil microbes and nutrient availability in future studies on plant invasion. As AM fungi have greater facilitations on plant invasion at low-P levels, P-limited ecosystems may be at higher risk of invasion by the AM plants. Due to the few species used in this study, the generality of our conclusions should be further tested by including more species so as to provide better guidance for the management of invasive plants. In addition, the mechanisms by which invasive species reduce AM colonization of native species are also required to be clarified in further studies. The authors are grateful to Shengsheng Huang for collecting the seeds, harvesting and measuring of the plants. The authors also thank Marc-André Selosse and three anonymous referees for their comments and suggestions on this article. This research was financially supported by the National Natural Science Foundation of China (NSFC 31700450, NSFC 31971556), Special funds of Guangdong Province for Promoting Economic Development (For the Development of Marine Economy, GDME⁃2018E002), Vegegraphy of China (2015FY210200-13), Zhang Hongda Science Foundation and Fu Jia-Rui Scholarship in Sun Yat-sen University. EC and SP conceived the ideas. EC, HL and BC designed the study. EC performed the research and collected the data. EC and HL analysed the output data. EC and SP led the writing of the manuscript. All authors contributed substantially to the drafts and gave final approval for publication. 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. Fig. S1 Common garden experiment design. Fig. S2 Total biomass of the plants in different treatments. Fig. S3 Shoot nitrogen (N) and phosphorus (P) content and N : P ratio of the plants in the different treatments. Methods S1 Precise description of experimental design and statistical analysis. Table S1 Result of three-way analysis of variance (ANOVA) for the effects of phosphorus (P), fungicide and competition on arbuscular mycorrhizal (AM) colonization in the roots of the plants. Table S2 Result of three-way analysis of variance (ANOVA) for the effects of phosphorus (P), fungicide and competition on total biomass of the plants. Table S3 Mycorrhizal responsiveness (MR) of the plants in different phosphorus (P) treatments. Table S4 The correlation between arbuscular mycorrhizal (AM) colonization and total biomass of the plants in different treatments. Table S5 Result of two-way analysis of variance (ANOVA) for tolerance and suppression relative interaction intensity (RII) index of two invasive plants (Eupatorium catarium, Bidens pilosa) under different phosphorus (P) treatments with or without fungicide application. Table S6 The F value of general linear modelling for the effects of phosphorus (P), fungicide and competition on shoot nitrogen (N) and P content and N : P ratio of the plants. Table S7 Result of Student's t-test for the interaction between fungicide and competition on the total biomass of the plants in low-phosphorus (P) conditions. 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.