- Top of page
- Materials and Methods
- Supporting Information
Arbuscular mycorrhizal (AM) fungi are an important component of the soil microbial community. They form a symbiotic relationship with c. 90% of terrestrial plant species, supplying nutrients, particularly phosphorus (P), to the host plant in exchange for carbon (Smith & Read, 2008). This symbiotic relationship can also result in increased resistance to drought, soilborne fungal pathogens and heavy metals in the host plant. There is also evidence to indicate that AM fungi improve soil structure (Rillig & Mummey, 2006).
The diversity of AM fungi is low relative to that of host plants, with c. 230 species described morphologically (http://schuessler.userweb.mwn.de/amphylo/), compared with the estimated 300 000 terrestrial plant species (Mora et al., 2011), although evidence from molecular methods suggests that the true diversity of AM fungi is somewhat higher (Opik et al., 2009; Lumini et al., 2010). Despite the relatively low diversity of AM fungi, there are evident differences between AM communities in different habitats and between AM fungi colonizing different host species in the same habitat (Helgason et al., 1998; Scheublin et al., 2004; Uibopuu et al., 2009; Hazard et al., 2012). The mechanisms driving differences in AM fungi colonizing different host species are unclear, but seem likely to be attributable, at least in part, to functional differences between AM fungi (Maherali & Klironomos, 2007). Distinctions between AM fungal communities at different sites or in different habitats may be partially explained by variation in host plant occurrence, but are also a result of environmental factors. For instance, agricultural practices have been repeatedly shown to impact AM symbioses (Gosling et al., 2006), with more intensive production systems having both reduced AM populations and AM diversity when compared with more extensive or pastoral production systems (Hijri et al., 2006; Gosling et al., 2010; Van der Gast et al., 2011).
One environmental factor shown to have a negative impact on the AM symbiosis is a high concentration of extractable soil P. Assimilation of P by AM colonized plants reflects the sum of uptake directly via plant cells, and indirectly via AM fungi, with the importance of the AM pathway declining when P availability is high, which is usually associated with a decline in AM colonization (Smith & Smith, 2011). These effects are mediated via the impact on host nutrition and, although the exact mechanisms involved are uncertain, changes in expression of host genes related to the different pathways of P uptake, with consequent changes to host signalling responsible for regulating formation of the symbiosis, are likely to be involved (Balzergue et al., 2011; Smith & Smith, 2011).
Recently, using in vitro tissue culture systems, Kiers et al. (2011) suggested that the trade in host C for fungal P could control the precise fungal communities colonizing roots, the plant selecting fungi which provide P at the lowest C cost by preferentially allocating them C. In turn, AM fungi preferentially transfer P to plants with the highest C supply. The extent to which the plant specifically rewards other benefits provided by individual AM species, such as uptake of N and water, and pathogen suppression is unknown.
The way in which the interplay of trade in fungal P for host C is impacted by P availability is uncertain. While colonization of host roots by AM fungi and the AM P uptake pathway is suppressed and may even be eliminated at high soil P concentrations (Jensen & Jakobsen, 1980; Hicks & Loynachan, 1987; Thingstrup et al., 1998; Ryan & Ash, 1999; Khaliq & Sanders, 2000; Liu et al., 2000; Kahiluoto et al., 2001), suggesting reduced C allocation to AM fungi, the relationship between soil P and AM diversity is unclear.
Plant species have varying requirements for P, reflecting differences in biomass and cellular concentration, and this requirement can change between developmental stages (Raghothama, 1999). The value of AM-derived P, in terms of its C cost, is therefore likely to vary between plant species and with developmental stage. It is also likely to depend on the supply of P available via direct and mycorrhizal pathways.
Counts of AM spores in the soil suggest that high concentrations of available soil P can reduce diversity, or at least populations of some species of AM fungi in the soil (Johnson, 1993; Hamel et al., 1994; Kurle & Pfleger, 1996), but results based on counts of AM spores in the soil are unlikely to reflect activity within host roots. Hijri et al. (2006) suggested that lower diversity of AM fungi colonizing roots in some agricultural fields compared with others was attributable to soil P, but there were significant confounding influences, not the least of which were differences in the host plants examined.
We therefore sought to assess the following. Does a high concentration of available soil P have a negative impact on colonization of a range of host plants by AM fungi? Does a high concentration of available soil P alter the diversity of AM fungi colonizing a range of host plants? Does the effect of soil P depend on host identity? Is there is an interaction between induced shifts in AM community diversity and growth stage?
- Top of page
- Materials and Methods
- Supporting Information
The relationship between AM fungi and their plant host is strongly controlled by the host P status, which is in turn strongly influenced by soil P availability (Smith & Read, 2008). There is considerable evidence in the literature to support this hypothesis. A reduction in the percentage AM colonization of roots as extractable soil P concentration increases is often evident (Jensen & Jakobsen, 1980; Ryan et al., 1994; Kahiluoto et al., 2001), but the influence of soil P on the diversity of AM fungi colonizing roots is not so clear. Although some authors have suggested a link between reduced AM fungal diversity and high soil P (Hijri et al., 2006), others have suggested there is no connection (Mathimaran et al., 2005; Beauregard et al., 2010). Furthermore, published assessments of the effect of soil P on AM diversity are sometimes confounded by factors such as host species, and the true influence of soil P is uncertain.
Our results showed a complexity of response that may go some way to explaining the contradictory results in the literature. Although there was a clear response to soil P, it was only evident at higher soil P concentrations and was strongly mediated by host plant identity. While we did record a significant decrease in AM colonization in response to increasing soil P in maize, there was no significant decrease, even at the highest soil P concentrations, in soybean and Viola. The soil P concentration range used was very large and thus ought to have been sufficient to produce a response in all species, yet it did not. The lack of response in soybean was particularly unexpected, as large reductions in AM colonization of soybean have been reported in response to P fertilizer application (Hicks & Loynachan, 1987; Isobe et al., 2008), although Fernandez et al. (2011) reported a response of colonization in soybean to soil P only at low soil P concentrations, with no response above c. 15 mg kg−1 P (Bray 1 extraction). The fact that there was no response in Viola and soybean combined with the generally low to moderate degree of colonization in all three host species might suggest that the AM community at this site is degraded and dominated by AM fungal species with weak colonization potential and weak response to soil P. However, the degree of colonization we recorded was within the range reported in the literature for maize and soybean (Galvez et al., 2001; Ryan & Graham, 2002; Liu et al., 2003; Murillo-Williams & Pedersen, 2008; Duan et al., 2010; Fernandez et al., 2011) and the AM diversity analysis does not support this hypothesis.
Examination of clones did not indicate amplification of nontarget DNA (see Notes S1) and the mean number of T-RFs found was similar to that reported elsewhere for AM fungi in arable agricultural systems (Vandenkoornhuyse et al., 2003; Mummey & Rillig, 2008; Bainard et al., 2012), indicating that the background AM community diversity at this site was not unusual. The dominance of the AM community by Glomus spp., evident from the cloned sequences, is also common for tilled agricultural soils (Johnson et al., 1991; Hendrix et al., 1995; Kurle & Pfleger, 1996; Daniell et al., 2001; Oehl et al., 2003; Toljander et al., 2008), also suggesting there was nothing unusual about this site.
More significantly, AM diversity was significantly impacted by soil P in maize and soybean. The effect at P8 in soybean was especially large, with fewer than half the number of T-RFs at P8 than at any other P concentration, at all the sampling times. Although the mean number of T-RFs was not significantly reduced by high soil P in Viola, there was a trend for fewer T-RFs at P8, the probability level for a significant effect was marginal and a t-test comparing P1 and P8 does result in a significant difference (P < 0.05, data not shown).
Further evidence to show that the AM community at the site was P-responsive comes from the T-RF community analysis. Soybean roots at P8 contained a significantly different T-RF community from the other P concentrations, as did the AM community at P8 in Viola (Figs 4 and 5 respectively). In maize, the T-RF community at both P4 and P8 was significantly different from those at P1 and P2. The contrast between the results for colonization and AM community size and structure suggests a degree of compensation, at least in soybean and Viola. Although the AM community diversity was reduced, colonization was impacted to a lesser degree. AM fungi that were eliminated or reduced by high soil P were evidently replaced to some extent by AM that were less sensitive to high soil P. This could reflect the persistence of fungi which are able to provide the plant with an extremely economical supply of P, requiring little receipt of C in return, or fungi which receive C from the host to perform functions unrelated to P supply, such as uptake of water or nutrients such as inorganic nitrogen, or even ‘cheaters’, parasites that receive C without providing useful benefits to the host.
In contrast to the influence of soil P, sampling time had a significant influence on AM colonization in all host species, with increased colonization between the first and second sampling time points across soil P concentrations, but a nonsignificant influence on AM diversity. This is consistent with Vandenkoornhuyse et al. (2002) and Santos-Gonzalez et al. (2007), who showed little change in diversity between different sampling dates, although these results were from perennial grassland species. Daniell et al. (2001) showed strong seasonal changes in AM diversity in annual crop species, but the direction of change was inconsistent and sample numbers were small, making interpretation uncertain. In annual plant species, there must be an initial period of colonization by AM fungi, during which there is an increase in diversity, but our results suggest this period is short, and after the initial colonization, overall diversity remains fairly consistent. Changes in the occurrence of T-RFs between sampling times that did occur included both the appearance and disappearance of T-RFs and were confined to rare T-RFs, resulting in nonsignificant differences between communities at different sampling times in NMDS plots. A similar small but insignificant shift in AM communities was recorded by Beauregard et al. (2010) in alfalfa, between June and September, while Bainard et al. (2012), using a semiquantitative approach to T-RFLP analysis of AM communities in maize, showed a change in abundance of different phylotypes between June and August, but very little change in occurrence. This does not suggest that succession of AM species during the growing season happens to any significant extent once initial establishment has occurred and this hypothesis is supported by results from Mummey et al. (2009), which showed that plants preinoculated with Glomus spp. retained a similar AM community after transplanting into a field situation. The functional significance of changes in colonization is uncertain. Highest host P demand will be during grain/seed filling, which coincided with the highest degree of colonization in all species, but function is not directly related to the degree of colonization (Smith & Smith, 2011).
Although soil P concentration had a significant influence on the T-RF communities in all the three host plants, the greater influence was host plant identity itself. Only at the highest soil P concentration, when diversity was greatly reduced in all host plants, did soil P override host identity in determining the T-RF communities present. Evidence for host specificity or at least host preference in AM fungi has accumulated over recent years (Sykorova et al., 2007a,b; Opik et al., 2009; Uibopuu et al., 2009; Li et al., 2010; Hazard et al., 2012), overturning the previous paradigm that AM fungi were generalists showing little host preference (Smith & Read, 1997). Li et al. (2010) found distinctly different AM communities in three plant species, but in contrast to our results found that habitat was a stronger influence than host plant; however, they gave little information about the two habitats. It is clear that host plant species has a strong influence on the AM community present in roots, but it appears that environmental factors, such as soil P, override this effect if they are large enough, a result also hinted at in Bainard et al. (2012). Host plant neighbour identity can also influence the AM community present in roots (Hausmann & Hawkes, 2009, 2010), which may explain the wider range of AM communities present in Viola compared with maize and soybean (Fig. 2), Viola potentially having a range of neighbour species in the mixed weed flora. This mixed neighbour effect may have been further enhanced or suppressed depending on whether Viola germinated before or after its immediate neighbours (Hausmann & Hawkes, 2010).
The significance of the differences in AM community between host plants and the reduction in diversity at high soil P is uncertain. Functional diversity within the AM fungi is a well established fact, with evidence that phylogenetically distantly related AM species are functionally dissimilar (Maherali & Klironomos, 2007). Evidence from our cloning data hints that differences in the T-RF communities in the host plants have at least some basis in phylogeny, suggesting a functional basis. Opik et al. (2009) reported distinct AM fungal communities in forest specialist plant species vs generalists and tentatively suggested a functional basis to the difference. Our three host plants, a C3 herb (Viola), a C4 grass (Maize) and a legume (Soybean), are ecologically distinct and thus would be expected to host distinct communities of functionally different AM fungi if host preference has a functional basis.
Whatever the immediate impact of the reduction in the diversity of AM fungi caused by high soil P on plant growth or nutrient uptake, a significant impact on the AM community in the soil would occur as propagule numbers of the absent species rapidly declined (Troeh & Loynachan, 2003). There were some particularly interesting patterns when individual T-RFs were considered. For instance, all the T-RFs that were found at lower P concentrations and that were unique to soybean were absent at P8, although this was not a pattern repeated with T-RFs unique to Viola. This strongly suggests that response to soil P may be dependent on AM identity, rather than random elimination of AM in response to high soil P. Helgason et al. (2007) reported a similar effect with a differential impact of disturbance on rare and common AM fungi, in their case in response to application of benomyl, while comparisons between AM communities in agroecosystems managed in different ways often show that AM families, such a Gigasporaceae and Acaulosporaceae, are more easily reduced or eliminated by unsympathetic management than the Glomeraceae, with agricultural ecosystems often dominated by Glomus group A (Franke-Snyder et al., 2001; Oehl et al., 2003; Hijri et al., 2006). This may help to explain some of the contradictory results in the literature regarding the influence of soil P on AM fungi. If some AM fungi have already been suppressed because soil P is elevated for historical reasons, then adding P to the soil will have little impact, as the remaining AM fungi are insensitive. The available soil P concentration at P8 (201–280 mg kg−1) used in this study was unusually high even for agroecosystems; the AM fungal community diversity in soybean and Viola was resistant to moderately high soil P, perhaps because it was very efficient at delivering P to the host, for a small C cost, only being reduced at very high soil P concentrations. The P4 concentration of 46–70 mg kg−1, to which AM diversity in maize responded, is, however, well within the range that could be expected in agroecosystems, also indicating a differential response, dependent on host species, possibly reflecting the different ‘value’ placed on P delivered by the AM pathway by different plant species.
The introduction of molecular methods to the study of AM fungi has revealed a previously unsuspected degree of complexity in the ecology of the fungi and in their relationship with the host plant. Here we have provided further evidence of strong host specificity in the AM fungi and evidence that different AM communities colonizing different hosts respond in a broadly similar way to increased soil P. However, we have also shown that there are subtle differences between the responses of different AM communities that may help explain seemingly contradictory results in the literature.