What does colonisation tell us? Revisiting the functional outcomes of root colonisation by arbuscular mycorrhizal fungi
Adam Frew
Abstract
Arbuscular mycorrhiza refers to the symbiotic association between plants and arbuscular mycorrhizal (AM) fungi, a ubiquitous and ecologically significant interaction across terrestrial ecosystems (Powell & Rillig, 2018). In this partnership, AM fungi colonise plant roots and surrounding soil, exchanging mineral nutrients such as phosphorus for photosynthetically derived carbon from their host plants (Smith & Read, 2008). While nutrient uptake, particularly phosphorus, is often regarded as the central functional outcome of the symbiosis, AM fungi also influence a broader suite of plant traits, including phenology, drought tolerance and defence against herbivores and pathogens (Jung et al., 2012; Delavaux et al., 2017; Frew et al., 2022). The physical interface of this exchange, the proportion of root length colonised by fungi, is frequently measured and used as a proxy for mycorrhizal function (McGonigle et al., 1990; Smith & Smith, 2011). This is logical; after all, it is within the roots that the symbiosis is established. Yet while it is broadly acknowledged that colonisation–benefit relationships are context dependent (Hoeksema et al., 2010), this nuance often fades in practice. Perhaps because assessment of colonisation has been such a foundational and longstanding focus in mycorrhizal research, do we now interpret it with implicit assumptions of function – even when empirical support is weak or absent? While some studies report clear positive linear correlations between colonisation metrics and plant growth (Fioroni et al., 2024) or phosphorus uptake (Püschel et al., 2016; Ryan et al., 2016), others show non-linear responses (Gange & Ayres, 1999; Garrido et al., 2010; Claassens et al., 2018), or none at all (Ryan & Angus, 2003; Ryan & Kirkegaard, 2012; Leiser et al., 2016). This raises the question of whether we continue to rely on colonisation measures out of convention, without sufficiently interrogating what they actually mean in different contexts. Interpreting measures of colonisation is further complicated by variation among AM fungal species in their foraging strategies (although see Camenzind et al., 2024) and carbon demands (Hart & Reader, 2002a,b), and also by methodological limitations. The proportion of root length colonised by AM fungi (McGonigle et al., 1990) is often measured at a single time point on a sub-sample of roots (understandably the assessment of an entire root system is typically not feasible), which may fail to capture the dynamic turnover of fungal structures or the spatial heterogeneity in colonisation across root types or developmental zones. Arbuscules, for example, are ephemeral structures that undergo regular turnover during the symbiosis (Toth & Miller, 1984) meaning that their abundance at any single time point may not reflect cumulative exchange activity. Additionally, Kokkoris et al.'s (2019) valuable comparison of methods underscores that standard approaches can miss variation that may exist in the ‘intensity’ of colonisation – which incorporates the abundance of fungal structures in addition to their presence (Trouvelot, 1986) – potentially obscuring meaningful links to function. Despite these limitations, such measures of colonisation remain widely used as a proxy in both research and applied contexts, likely because it is relatively easy to quantify and intuitively connected to the symbiosis. However, could it also risk narrowing our expectations of what AM fungi do? Focusing exclusively on nutrient-related benefits risks underestimating other important functions of the AM symbiosis (Delavaux et al., 2017). One such function is the modulation of plant defence chemistry. AM fungi are known to prime host plants for enhanced resistance to herbivores and pathogens, sometimes through the accumulation of metabolites such as phenolics (Cameron et al., 2013; Pozo et al., 2013; Frew, 2020). These responses may occur with little or no corresponding change in growth or plant phosphorus (Pozo de la Hoz et al., 2021; Weinberger et al., 2024). Are all benefits of the symbiosis for plants mediated through exchange? Are all functions equally traceable to arbuscules? Vannette & Hunter (2011) proposed a resource exchange model of defence induction, predicting that benefits such as enhanced resistance increase with the proportion of root length colonised to a point, but decline at higher colonisation levels due to carbon costs exceeding nutrient gains. Yet, it is also plausible that defence-related outcomes are initiated early in colonisation of the host by the fungi, perhaps independent of arbuscule formation or nutrient exchange (Cameron et al., 2013). Defence activation from hyphal colonisation has been observed even in host species that are not arbuscular mycorrhizal (Anthony et al., 2020), likely triggered by fungal surface molecules (e.g. chitin) that activate immune responses upon contact or attempted penetration (Zhang & Zhou, 2010). Aspects of fungal colonisation can serve as cues rather than costs; if that is the case, we may be conflating structural presence and functional meaning, making it unclear what we are actually measuring. Plant identity is another key factor shaping relationships between the proportion of root length colonised and functional outcomes. Species differ in their growth strategies, root traits and nutrient foraging behaviours, all of which can affect these colonisation-function associations (Lekberg & Koide, 2005; Smith & Smith, 2011; Bergmann et al., 2020). C4 grasses, in particular, are often considered more responsive to AM fungi, especially in terms of phosphorus uptake (Treseder, 2013). These differences offer an opportunity to test when and where the root length colonised by the fungi serves as a meaningful proxy. In this study, I revisit the function of colonisation metrics using a simple but structured experiment. I assessed colonisation using the McGonigle et al. (1990) intersect method and report it as the percentage of total root length colonised by fungal structures. Throughout the manuscript, I refer to the proportion of root length containing AM fungal structures (hyphae, arbuscules or vesicles) as ‘total colonisation’, and the proportion of root length colonised specifically by arbuscules as ‘arbuscular colonisation’. I measured root length colonised by AM fungi in four globally important crops (two C3 and two C4) and tested their relationships to plant biomass, phosphorus and phenolics. I hypothesised that: (1) the total colonisation and arbuscular colonisation would be positively associated with plant biomass, phosphorus and phenolic responses (responses being the change in each trait in plants with AM fungi relative to controls without AM fungi), particularly in the C4 crops; and (2) any relationships would be predominantly non-linear, reflecting threshold or saturation dynamics. While measures of AM fungal colonisation remain a cornerstone metric in mycorrhizal research, their interpretation is often assumed rather than explored. This study offers empirical data across different traits and species to revisit what the root length colonised can, and cannot, tell us about symbiotic outcomes. I hope this serves as a timely prompt for deeper reflection on how we use colonisation metrics, and what we expect them to mean. I conducted a full factorial pot experiment with four plant species that are some of the most significant crops globally: wheat (Triticum aestivum L. cv ‘Yitpi’), barley (Hordeum vulgare L. cv ‘Hindmarsh’), sorghum (Sorghum bicolor L. Moench cv ‘MR Taurus’) and maize (Zea mays L. cv ‘Amadeus’). The plants were grown in 8-l pots filled with gamma irradiated soil/sand mix (Table 1), which contained moderate to high levels of available phosphorus (46 mg kg−1 Colwell P) within or above the critical phosphorus levels for these crops (see M. J. Bell et al., 2013; R. Bell et al., 2013), consistent with high-input crop systems. Each plant species was grown either with no AM fungi (n = 12 per species), or were inoculated (n = 24 per species) with a commercial inoculant (MicrobeSmart, Melrose Park SA) containing four species of AM fungi (Entrophospora etunicatum (Błaszk., B.T. Goto, Magurno, Niezgoda & Cabello) C. Walker & A. Schüßler, Funneliformis coronatum (Giovann.) C. Walker & A. Schüßler, F. mosseae (T. H. Nicolson & Gerd.) C. Walker & A. Schüßler and Rhizophagus irregularis (Błaszk., Wubet, Renker & Buscot) C. Walker & A. Schüßler). To encourage a broad range of fungal colonisation in roots, the AM fungal inoculant was applied at three rates – low (0.05 g kg−1 soil), medium (0.5 g kg−1) and high (2 g kg−1) – with each rate given to eight plants per species, totalling 24 AM fungal-inoculated plants per species. These inoculation rates are roughly equivalent to 200, 2000 and 8000 spores per kg of soil, respectively, based on the average of 4000 spores per gram of inoculant. The inoculant was thoroughly mixed into the soil : sand substrate before potting to ensure even distribution. To account for any potential effects of non-AM fungal microbes, non-AM fungal pots were supplemented with 200 ml of microbial liquid filtrate derived from washing the AM fungal inoculum mixed with soil : sand mix filtered through sieves down to 20 μm to standardise the non-AM fungal microbial community across all pots. Plants were grown in a glasshouse with day/night temperatures of 27 and 18°C, respectively, with a 12 h photoperiod. Watering was initially 100 ml every day for the first 3 wk, followed by 400–500 ml every 3 d. Volumes were adjusted as needed to maintain consistent soil moisture across pots, monitored every 3 d using a handheld moisture meter (PMS-714; Lutron Electronic Enterprise; Taipei, Taiwan). Plants were harvested 55 d from germination, where all hosts were toward the latter vegetative growth stages, before booting or flowering. For each plant, from six locations distributed across the root system, 10–15 root fragments were collected from each location. These were cut into 5-cm segments, composited and stored in 50% ethanol before mycorrhizal fungal colonisation assessment, and the remaining aboveground and belowground plant tissues were all oven-dried at 60°C. Total biomass was recorded, and all foliar tissue was ground and homogenised before any chemical analyses. Phosphorus (P) concentrations in plants were assessed via inductively coupled plasma spectroscopy following digestions with nitric acid (Zarcinas et al., 1987). Total foliar phenolics were determined as described in Salminen & Karonen (2011), in technical triplicates, using a Folin–Ciocalteu assay with gallic acid monohydrate (Sigma-Aldrich, St Louis, MO, USA) as the quantification standard. For assessing root colonisation by AM fungi, and to confirm the absence of any root colonisation by AM fungi in the control plants, ethanol-stored root samples were placed into histology cassettes, then cleared with 10% potassium hydroxide (KOH) at 90°C for 10 min, stained with 5% ink-vinegar (using Quink ink; Parker Nantes, France) at 90°C for 15 min (Vierheilig et al., 1998). Root fragments of 5 cm each were mounted on glass slides with glycerin under a coverslip. For each plant, 30 root fragments were assessed in total, representing 150 cm of root length. The percentage of root length colonised by any AM fungal structure (total colonisation), as well as the percentage of root length colonised by arbuscules (arbuscular colonisation), and vesicles (vesicular colonisation) were examined microscopically using the intersect method at ×200 magnification for at least 100 intersections per plant (McGonigle et al., 1990). The McGonigle et al. (1990) intersect method was used as it remains the most widely applied approach (Füzy et al., 2015), allowing comparability across the literature, despite the availability of alternative methods that may offer greater sensitivity (e.g. Trouvelot, 1986; Kokkoris et al., 2019). To examine the relationship between root length colonised (total colonisation and arbuscular colonisation) and plant outcomes, plant mycorrhizal responses were calculated as ((plant response − mean plant responses with no AM fungi)/mean plant responses with no AM fungi) × 100, where the plant response was either the total biomass, P concentration, or phenolic concentration. These provided the mycorrhizal growth responses (MGR), mycorrhizal phosphorus responses (MPR) and the mycorrhizal phenolic responses (MPhenR). Due to very low vesicular colonisation across samples, vesicles were not included in the main analyses; however, these data are provided in the Supporting Information (Fig. S1c). Generalised additive models (GAMs) were fitted for each plant species separately. These models were initially fitted for each response variable using the mgcv package in R (Wood, 2003). Inoculation rates were used to generate variation in colonisation and were not included in GAMs as plant responses were analysed in relation to root length colonised, the biologically relevant predictor for the hypotheses. GAMs were chosen for their flexibility in capturing both linear and nonlinear trends without assuming a predefined functional form. Smooth terms were fitted using the default basis dimension (k = 10) in the mgcv package, the suitability of the k value was confirmed using diagnostic checks with the gam.check function. Model smoothness was determined using generalised cross-validation, and the effective degrees of freedom (edf) of the smooth term were examined to assess the shape of the relationship. Where the GAM indicated a significant relationship and edf was close to 1 (suggesting an approximately linear trend), a linear model (LM) was fitted and model selection was confirmed using Akaike Information Criterion (AIC). If a GAM indicated a significant relationship that was nonlinear (edf > 1.8), a nonlinear asymptotic model was tested using nonlinear least squares The asymptotic model followed the asymptotic function in R. Model were tested for and were used to confirm the model Where a GAM no significant relationship > for the smooth no further was and the response variable was considered to be independent of root length GAM were used to relationships and the of the models or were on (Fig. were using the function an of colonisation were from the GAM or and as in all Where GAMs no or were in the Model assumptions were via (using the gam.check function for for in and checks for Model selection was confirmed based on and of fitted were in R using the mgcv (Wood, and & the proportion of root length colonised by AM fungi offers a proxy for the functional outcomes of the symbiosis. Yet this study to others that that the relationship between colonisation and is from across traits, species and colonisation types – not all functional outcomes to with the proportion of root length colonised in the all plant responses the and most consistent in relation to plant defence chemistry. In all four plant species, the total colonisation was and associated with in mycorrhizal phenolic responses with each an asymptotic relationship (Table In all species, GAMs were significant with effective degrees of freedom (edf) from in maize to in and high from to These were described by models that relationships would often be non-linear, and which positive associations between total colonisation and phenolics. total colonisation plant phenolic responses at low to moderate before a This may that phenolic defence responses are triggered by the plant associated with the presence of fungi within the root (Jung et al., rather than arbuscule and that these responses do not continue to in proportion to the percentage of root length this across both C3 and C4 crops not support first that C4 hosts would be more no significant relationships were observed between arbuscular colonisation and in any species P > edf While some models that nutrient exchange defence by greater resource (Pozo & Vannette & others that defence responses can be upon fungal independent of arbuscule formation or (Cameron et al., 2013; et al., Pozo de la Hoz et al., This that phenolic responses are not mediated by the arbuscules – despite their as the interface for carbon and nutrient Arbuscular colonisation can sometimes weak or associations with plant traits due to the ephemeral of the consistent absence of any relationship across all hosts was particularly when to the consistent relationship plant phenolic responses with the total root length than reflecting methodological due to of on of the ephemeral this may point to a phenolic responses could be triggered by fungal presence more not by arbuscular function per This is important when in of the resource exchange model of plant defence from Vannette & Hunter (2011), which that defence benefits from AM fungi initially increase with colonisation but may at higher fungal due to carbon costs and nutrient the observed – consistent defence responses even in the absence of growth or phosphorus benefits – that the phenolic responses are not to exchange per as they occur even in the absence of growth or phosphorus responses and are not by arbuscule than a the with the of a resistance which can be through fungal such as root penetration or other independent of resource exchange (Cameron et al., 2013; et al., Weinberger et al., 2024). The saturation of the plant phenolic responses may reflect an threshold of defence which either host further or fungal of host defence further The absence of a corresponding relationship with arbuscular colonisation the that nutrient exchange is for defence it a broader for colonisation as a a which may then these that AM fungi can plant in that are independent of resource Smith & Smith (2011) that P uptake via the AM fungal can occur without growth particularly in species or under the that AM fungi may consistent defence-related benefits even when of growth and P show little In to the phenolic and first no significant associations were observed between colonisation (total or and across any crop species P > The plants in study were grown in soil with high P availability mg kg−1 Colwell which the of such this P within the critical range for these crops mg kg−1 on soil and and the soil P would reflect of high-input in which these crops are and Additionally, while plant root growth was root was observed in a sorghum with particularly high root biomass – an of pot that may influence root or function. such could to the absence of average were positive across all species, with sorghum the Phosphorus responses (MPR) particularly in the C4 In was and associated with total colonisation edf = = P and by an In the relationship was linear = P = These are consistent with first and and with that C4 species often P benefits from the symbiosis in more to the proportion of root length colonised (Treseder, 2013). The asymptotic shape of the response in sorghum may early saturation of P while linear could reflect a more uptake These responses could assumptions that P may at high barley a significant linear relationship between arbuscular colonisation and = P = a not by first or in that I the root length colonised in C4 plants to relationship with than in C3 plants, and that this relationship would be the of this study would that phenolic responses may be among the most consistent outcomes of AM fungal colonisation in high-input and that these may occur of growth or I not test whether the phenolics resistance to pathogens or This would given a basis for a functional interpretation of the across the plant of arbuscular are well with context and the are no While this study on data from four globally important the relationships observed are likely to with crop AM fungal and than these as I offer them as a prompt to revisit how we assess AM symbiosis function. This may be particularly valuable in where high nutrient availability can growth and P mycorrhizal (Ryan & Kirkegaard, While this study to the functional meaning of colonisation metrics as they are it is clear that methodological remain a to deeper including the used do not between and fungal structures and could functional colonisation with fungal colonisation was assessed at a single into the of the symbiosis. including such as et al., or hosts that report arbuscule formation through et al., offer to colonisation approaches et al., or could be and perhaps could to quantify fungal while could more of including potentially important in fungal across host species. These methods offer valuable to colonisation measures and to how we interpret and the and functional of AM fungal while the proportion of colonised and measures remain and metrics, I hope this the positive associations observed between total colonisation and plant phenolics all four species), and between arbuscular colonisation and foliar phosphorus in three species, as an of symbiotic outcomes. colonisation remains to the study of the AM the presence of AM fungi the root is a for the symbiosis Yet, as we continue to rely on colonisation metrics to symbiotic we not what it but what assumptions we with If we are to the of AM fungi in shaping plant traits and we also how we interpret the most measures of the symbiosis. This was by an to by as of the via the of that support this study are available from the at the following Plant responses and root length colonised across plant species and from generalised additive is not for the or of any Supporting Information by the than be to the The is not for the or of any by the than be to the corresponding for the The remains with to in and in any