Volume 237, Issue 3 p. 987-998
Full paper
Free Access

Dirt cheap: an experimental test of controls on resource exchange in an ectomycorrhizal symbiosis

Amber L. Horning

Corresponding Author

Amber L. Horning

Department of Biology, University of Mississippi, PO Box 1848, University, MS, 38677 USA

Authors for correspondence:

Amber L. Horning

Email: [email protected]

Jason D. Hoeksema

Email: [email protected]

Search for more papers by this author
Stephanie S. Koury

Stephanie S. Koury

School of Biological, Environmental and Earth Sciences, The University of Southern Mississippi, 118 College Drive #5018, Hattiesburg, MS, 39406-0001 USA

Search for more papers by this author
Mariah Meachum

Mariah Meachum

Department of Biology, University of Mississippi, PO Box 1848, University, MS, 38677 USA

Search for more papers by this author
Kevin A. Kuehn

Kevin A. Kuehn

School of Biological, Environmental and Earth Sciences, The University of Southern Mississippi, 118 College Drive #5018, Hattiesburg, MS, 39406-0001 USA

Search for more papers by this author
Jason D. Hoeksema

Corresponding Author

Jason D. Hoeksema

Department of Biology, University of Mississippi, PO Box 1848, University, MS, 38677 USA

Authors for correspondence:

Amber L. Horning

Email: [email protected]

Jason D. Hoeksema

Email: [email protected]

Search for more papers by this author
First published: 08 November 2022
Citations: 3

Summary

  • To distinguish among hypotheses on the importance of resource-exchange ratios in outcomes of mutualisms, we measured resource (carbon (C), nitrogen (N), and phosphorus (P)) transfers and their ratios, between Pinus taeda seedlings and two ectomycorrhizal (EM) fungal species, Rhizopogon roseolus and Pisolithus arhizus in a laboratory experiment.
  • We evaluated how ambient light affected those resource fluxes and ratios over three time periods (10, 20, and 30 wk) and the consequences for plant and fungal biomass accrual, in environmental chambers.
  • Our results suggest that light availability is an important factor driving absolute fluxes of N, P, and C, but not exchange ratios, although its effects vary among EM fungal species. Declines in N : C and P : C exchange ratios over time, as soil nutrient availability likely declined, were consistent with predictions of biological market models. Absolute transfer of P was an important predictor of both plant and fungal biomass, consistent with the excess resource-exchange hypothesis, and N transfer to plants was positively associated with fungal biomass.
  • Altogether, light effects on resource fluxes indicated mixed support for various theoretical frameworks, while results on biomass accrual better supported the excess resource-exchange hypothesis, although among-species variability is in need of further characterization.

Introduction

Mutualisms are interspecific interactions ubiquitous to all ecosystems (Bronstein, 2015). Mutualists benefit one another through trade of services or resources, such as in plant–pollinator mutualisms where pollination services are exchanged for nectar resources. In resource-exchange mutualisms, such as mycorrhizae and rhizobia, soil microbes provide mineral nutrients to plants in exchange for photosynthates. Resource trade can be costly for organisms in a mutualism, but traded resources may not always be costly to the species trading them, with by-product benefits commonly exchanged in some major types of mutualisms (Connor, 1995). Moreover, the costs and benefits of traded resources may vary among environments (Johnson et al., 1997, 2013), and this context dependency can have a variety of important consequences for the ecology and evolution of mutualisms (Hoeksema & Bruna, 2015).

Various models have utilized economic theory to predict the outcomes of resource-exchange mutualisms (Schwartz & Hoeksema, 1998; Hoeksema & Schwartz, 2003; Kummel & Salant, 2006; Akçay & Roughgarden, 2007; Franklin et al., 2014; Akçay, 2015). Although each economic model differs in approach, the singular common variable is the resource ‘exchange price’, defined as the ratio of units of one resource that are traded for another. These economic models generally assume that the resources being exchanged are costly to each species, and thus the benefits and outcomes of resource-exchange mutualisms are driven by the ratio of resources being exchanged, that is the exchange price. The comparative advantage model makes predictions for how variable environmental factors (e.g. light) should affect resource exchange and the outcomes of the mutualism (Schwartz & Hoeksema, 1998; Hoeksema & Schwartz, 2001, 2003). Specifically, those models predict that an increase in the availability to one species of the resource they are trading away will lower the price (i.e. exchange ratio) offered by that species, increase the overall volume of trade, and increase the growth benefits (of trade) to both species. These predictions are consistent with the phenotypic plasticity component of the sun-worshipper hypothesis (Veresoglou et al., 2019). Grman et al. (2012) proposed a population dynamics model based on comparative advantage principles, which predicts how light affects resource exchange and resource-exchange ratios between plant and arbuscular mycorrhizal (AM) fungal mutualists, depending on nutrient availability. Alternatively, resource-exchange prices may be fixed for particular species pairs, potentially varying among different species (Kummel & Salant, 2006) but not varying with environmental conditions or availability of traded resources (Kiers et al., 2011). Tit-for-Tat models similarly assume that exchanged resources are consistently costly (regardless of environment), with one partner modifying investment in the symbiosis in response to changing investment by the other partner (Axelrod & Hamilton, 1981; Connor, 1995).

Despite the importance of resource-exchange ratios in models of symbiosis and mutualism, they have rarely been quantified in naturally occurring resource-exchange symbioses. Previous studies that measured resource fluxes have typically only provided a short-term snapshot of the relationship or lacked a distinction in the fates (whether to roots or microbial symbionts) of carbon (C) allocated belowground, not accounting for both root and fungal respiration vs assimilation (Douds et al., 1988; Jones et al., 1991, 1998; Colpaert et al., 1996; Qu et al., 2004; Hortal et al., 2017). As a result, exchange ratios have not been explicitly linked to environmental factors or host/symbiont performance.

In contrast to economic models, Corrêa et al. (2008, 2011, 2012) hypothesized that fitness benefits to plants of ectomycorrhizal (EM) mutualisms are not driven by the ratio of resources being exchanged, but rather by the absolute flux of whatever resource is most limiting to plant growth at a given point in time. Specifically, Corrêa et al. (2012) suggested that C is only a limiting resource to plant growth when plants are severely light-limited (< 9% of full sun exposure) and that light intensity has little effect on plant growth responses to mycorrhizal fungi under most normal conditions. Their hypothesis predicts that the absolute fluxes of limiting resources (e.g. nitrogen (N) or phosphorus (P) for plants) exchanged between symbionts better predict outcomes of resource-exchange mutualisms, rather than exchange ratios.

Mycorrhizal mutualisms are ideal systems to test hypotheses on how resource fluxes are affected by environmental conditions and how they influence outcomes of resource-exchange mutualisms. Mycorrhizae are mutualisms between plant roots and fungal hyphae whereby plants photosynthetically fix C compounds into simple sugars and trade them, or other compounds such as lipids, to their fungal symbionts for soil nutrients, such as N and P (Smith & Read, 2008; Luginbuehl et al., 2017). Mycorrhizal resource exchange is discrete, quantifiable, and occurs over relatively short timescales, which allows feasible nutrient-flux measurements. In the present study, we measured the cumulative resource exchange between loblolly pine (Pinus taeda) and two EM fungal symbionts (Rhizopogon roseolus and Pisolithus arhizus) in order to determine how resource exchange (total fluxes and ratios) differed between fungal species and between high- and low-light availability, which we hypothesized would influence C availability to the plant. Although ambient and elevated CO2 was not manipulated, high- and low-light treatments may elicit a parallel response as CO2 is predicted to affect N : C resource-exchange ratios (Franklin et al., 2014). Here, we utilized a modified root-mycocosm approach (Rygiewicz et al., 1988; Rygiewicz & Andersen, 1994) that allowed us to measure cumulative amounts of exchanged resources over time and to partition the fate of C allocated belowground, that is into root biomass, fungal biomass, root respiration, or fungal respiration.

Overall, we sought to address two key questions regarding P. taeda EM mutualisms: (Q1) How does light availability affect resource exchange? (Q2) Do resource-exchange ratios or absolute resource fluxes better predict EM fungal and pine seedling biomass accumulation? For Question 1, the comparative advantage hypothesis from biological market models (Schwartz & Hoeksema, 1998; Hoeksema & Schwartz, 2003) predicts that N : C (N per unit C) and P : C (P per unit C) exchange prices would be lower in a high-light environment than in a low-light environment, due to pine seedlings in high light having an excess of C to offer their fungal symbionts. It also predicts an overall increased volume of trade under increased light. Alternatively, the Tit-for-Tat hypothesis predicts that N : C and P : C exchange prices should be constant among environments (Axelrod & Hamilton, 1981; Connor, 1995), although they may vary depending on the partner species involved (Kummel & Salant, 2006). For Question 2, the comparative advantage hypothesis predicts that across experimental conditions, pine seedling growth would be positively correlated, and EM fungal biomass negatively correlated, with P : C and/or N : C exchange prices (i.e. the ratio of P or N received by the plant from the fungus, relative to the C transferred from the plant to the fungus). By contrast, the excess resource-exchange hypothesis (Corrêa et al., 2008, 2011, 2012) predicts that N or P would be most limiting to plant growth and that the absolute amount of N or P received from EM fungi would be a better predictor of plant growth than N : C or N : P ratios. That hypothesis makes no predictions on how resource exchange should affect fungal biomass accumulation, but a logical extension of it would predict that EM fungal biomass would be better predicted by the absolute amount of C received from the plant than by N : C or N : P exchange prices.

Materials and Methods

Pinus taeda L., loblolly pine, is a coniferous tree species native to the southeastern United States, chosen for this study due to the ease to propagate it, to find fruiting bodies of associated EM fungi and to inoculate seedlings with fungal spores. Ectomycorrhizal mutualisms are particularly important in facilitating pine seedling establishment in acidic, nutrient-poor soils, which were utilized in this experiment (Brundrett, 2009). Pinus taeda seeds were obtained from two open-pollinated families from lines selected previously for Leptographium pathogen resistance (Singh et al., 2014; Piculell et al., 2018). Pine seeds were sterilized in a 3% H2O2 solution for 24 h and then rinsed with running water for 2 min. Seeds were cold-stratified at 4°C for 40 d in moist conditions and agitated daily to deter mold growth. To further prevent contamination after stratification, seeds were soaked in 10% bleach for 5 min, 70% alcohol for 1 min, and 10% bleach again for 1 min, followed by a sterile water rinse for an additional minute. Seeds were germinated in a Conviron Model ATC40 environmental chamber (Controlled Environments Ltd, Winnipeg, MB, Canada) in groups of four on 10-inch water agar plates tilted at 70°C, and with the lower half covered in foil, in order to orient the direction of shoot and root growth. Seeds were germinated on a 16-h photoperiod (400 μmol m−2 s−1) with a consistent temperature of 18°C until seedlings were 2 to 3 inches in length, c. 3 to 4 wk.

Ectomycorrhizal inoculation of seedlings

Pine seedlings were dip-inoculated in spore slurries of fungal sporocarps from two target fungi (R. roseolus (Corda) Th. Fr. and P. arhizus (Scop.) Rauschert) collected from under P. taeda trees in Oxford, MS, in 2014 and 2016, respectively. Rhizopogon roseolus and P. arhizus (hereafter ‘Rhizopogon’ and ‘Pisolithus’) are common in pine forests of the southeastern United States, important for seedling establishment, and are early and thorough colonizers of pines, making them ideal for this seedling study. Identities of the fungal isolates used for inoculation, and of EM root tips from harvested seedlings, were confirmed through Sanger DNA sequencing and comparison of sequences with public databases (see Rúa et al., 2015; Craig et al., 2016; Rasmussen et al., 2017; Hoeksema et al., 2018).

To make the spore slurries for inoculation, sporocarps were blended with DI water and spore concentrations were adjusted to c. 107 spores ml−1. Pine seedling root systems were dipped in the slurry and planted in seedling cones (21 cm × 4.5 cm) filled with a sterile 1 : 20 soil : sand mixture (to be described later) and incubated in full light (400 μmol m−2 s−1) on a 16-h photoperiod in an environmental chamber (Conviron ATC40; Controlled Environments Ltd) at a constant temperature of 25°C. Seedlings were watered to saturation on a weekly basis and allowed to develop for 5 months. After 5 months, 10 ml of background soil microbe slurry (created by filtering 6 l of deionized water through 1 l of fresh soil on a 44-micron sieve) and 10 ml of a 50% diluted MMN media (as a dilute nutrient source) without C source were added to each cone, after which mycorrhizal development was allowed to continue for another 4 wk before transplanting into experimental root-mycocosms.

Root-mycocosm assembly

Root-mycocosms, modified by M. Booth from the original design of Rygiewicz & Andersen (1994) and Rygiewicz et al. (1988), were constructed of two clear polycarbonate plates (23 cm tall by 38 cm wide) separated on the sides and bottom by three sections of PVC 2.5-cm wide, adhered with wing nut bolts and general-purpose silicone sealant. The volume of the root-mycocosm was separated into halves by a PVC spacer (5 mm thick) routed to 90% openness, filled with a mix of fine and coarse sand substrate, and covered on both sides with a nylon mesh (44 μm) to allow for the passage of fungal hyphae, but block Pinus root growth between sections, thus making it possible to measure soil CO2 on one side of the barrier in the absence of the plant root system (Fig. S1). Each root-mycocosm was sterilized in a 10% bleach solution for 30 min, rinsed with DI water, and stored in a room protected with a HEPA air filter to reduce the likelihood of contaminants from nontarget fungi before being filled with a growth substrate. The growth substrate was composed of a 1 : 20 soil : sand mixture, where the sand was a 1 : 1 mixture of commercial play sand and natural sand sourced from northern Mississippi, and the soil was a loamy field soil collected from beneath P. taeda trees in Oxford, Mississippi. The resulting soil mixture was low in total C (0.088%), N (0.005%), and P (0.002%), and contained no detectable fungal biomass (ergosterol). The substrate was sieved to 1 mm to remove coarse particles and autoclaved twice at 121°C for 1 h, with a 24-h waiting period between sterilizations. Each half of all root-mycocosms was filled with c. 800 ml of the substrate and then covered with 50-μm-thick black plastic bag material to reduce algal growth and entrance of airborne fungal spores. One liter of fresh homogenized field soil was suspended in 6 l of DI water and filtered to 5 μm to create a microbial wash. Ten milliliters of microbial filtrate was added to each half of the root-mycocosms before planting.

Experiment setup

The experiment was a 2 × 2 × 3 factorial design: Two EM fungal species crossed with two light levels and three harvest times (10, 20, and 30 wk). Each combination of fungal species, light level, and harvest time was replicated six (Rhizopogon) or eight (Pisolithus) times. Three tree seedlings (one each for high and low light in the Rhizopogon third harvest group, and one for high light in the Pisolithus third harvest group) died during the experiment leaving a total of 81 treatment root-mycocosms. The two light levels tested were high light (400 μmol m−2 s−1) and low light (135 μmol m−2 s−1) on a 13-h light cycle. Both light treatments were below typical photosynthetic saturation levels for pine seedlings (Teskey et al., 1994). An additional 36 control root-mycocosms (six per light level × harvest time combination) containing seedlings with microbial wash, but without mycorrhizal inoculation, were also included for the purposes of estimating background CO2 respiration throughout the experiment. Thirteen of these control seedlings died throughout the experiment, leaving a total of 104 root-mycocosms (81 treatment, 23 control). Two environmental chambers were utilized in this experiment (both Conviron Models ATC40; Controlled Environments Ltd). Each chamber contained two shelves of growing space, one of which was set at the low-light level and the other at the high-light level, creating four blocks (two low-light, two high-light). Each of these contained at least three root-mycocosm replicates of all combinations of EM fungal species and harvest times, whose placement was randomized. Due to environmental chamber mechanical failure, the third harvest root-mycocosms were moved to a nearby grow room, 4 wk before the final harvest, with light exposure levels matching those in the environmental chambers for both high and low-light treatments.

Seedling planting

Mycorrhizal colonization of root tips was verified on each seedling before transplanting a single seedling into each root-mycocosm. Five additional seedlings colonized by each target fungus and five nonmycorrhizal control seedlings were rinsed with DI water and immediately frozen for the analysis of initial ergosterol, C, P, and N contents. In addition, subsamples of the root-mycocosm soil substrate were collected for the analysis of initial nutrient content. After planting, root-mycocosms were fertilized (on both sides) with 10 ml of Hoagland's No. 2 (Sigma-Aldrich, St Louis, MO, USA) solution (for a total of 20 ml per root-mycocosm) at the start of the experiment and immediately after the first two harvests. Root-mycocosms were watered to saturation on a weekly basis. When harvested, seedlings were verified to have 100% root tip colonization.

Overview of data collection and synthesis

Data collection and analysis methods were adapted from Meachum (2016). The total C flux from the seedling to the EM fungal symbiont was estimated as the sum of C respired (as CO2) by the fungi, plus C accumulated in EM fungal biomass on the roots and in the soil (Fig. 1); accumulation in fungal biomass was estimated by measuring ergosterol, on the whole root system and in homogenized soil from each half of the root-mycocosm. Ergosterol is the main sterol found in fungal cell membranes of the Ascomycota and Basidiomycota (Weete et al., 2010) and is indicative of living and recently dead fungal biomass (Grant & West, 1986; Newell et al., 1987). Ergosterol in collected soil samples was extracted in alcoholic KOH (0.8% KOH in HPLC-grade methanol) for 30 min at 80°C in tightly capped thick-walled digestion tubes (total extraction volume 10 ml). The resultant extract was partitioned into n-pentane and evaporated to dryness under a stream of N gas using a N-Evap 111. Ergosterol in dried samples was redissolved by sonication in 1 ml of HPLC-grade methanol, transferred to HPLC screw-capped autosampler vials, and stored at −20°C in the dark until quantified by HPLC (Hendricks et al., 2016). Total fungal biomass was calculated using ergosterol : biomass conversion factors (5.455 μg ergosterol mg−1 R. roseolus fungi (dry weight); 1.534 μg ergosterol mg−1 P. arhizus fungi (dry weight)), which had been determined previously (Koury, 2019). Here, we assumed that all C transferred and incorporated into fungal biomass originated from the pine seedlings, due to the rarity of EM fungi obtaining C compounds directly from soil (Zak et al., 2019).

Details are in the caption following the image
Fates of carbon (C), nitrogen (N), and phosphorus (P) transferred between tree and fungus. Carbon transferred to the fungus will be incorporated into fungal biomass (CFungal Biomass) or respired by the fungus (CFungal Respiration). Nitrogen and P transferred to the plant will be incorporated into shoot (Nshoot, Pshoot) or root biomass (Nroot, Pshoot).

Cumulative respired C was estimated using instantaneous measurements just before harvest of each experimental unit using a LI-6400XT infrared gas analyzer (IRGA; Li-Cor Bioscience, Lincoln, NE, USA), equipped with a custom chamber placed over the root-mycocosms (to be described later). These measurements and harvests took place over the course of several days (5, 3, and 4 d, respectively, for growth periods 1, 2, and 3). The accumulated mass of N and P in plants was estimated from N and P analyses of dry plant biomass from the above and belowground plant parts; root material analyzed for N and P concentration did not include mycorrhizal root tips. Cumulative values for C, N, and P were calculated for each time period; resulting values for each harvested cohort of seedlings represented totals accumulated from the beginning of the first growth period, not just since the previous harvest. Total resource transfers were used to calculate N : C and P : C exchange ratios by dividing the total amount of N and P transferred to the seedling by the total amount of C transferred to the fungus. Full details on resource-exchange calculations can be found in Methods S1 (detailed description of resource-exchange calculations) and Figs S1, S2.

Respiration measurements

Soil CO2 efflux was estimated using a LI-6400XT infrared gas analyzer (IRGA; Li-Cor Biosciences) mounted to a custom polycarbonate box (Fig. S1). All CO2 measurements were taken during the same time period each day (between 11:00 h and 15:00 h) using the ‘closed’ system method as in Meachum (2016). To test for diurnal fluctuations in soil CO2 flux from root-mycocosms, a subset of root-mycocosms (three from each light × fungal species treatment combination) was measured at five time points (09:00 h, 14:00 h, 20:00 h, 01:00 h, and 05:00 h) over a 24-h period. This test was conducted between the second and third harvest (26 wk after the start of experiment). We found no consistent effect of time of day on CO2 flux rates (Fig. S3).

Consideration of background values of saprobic C

Despite very low content of organic matter and nutrients in the experimental soil designed to minimize populations of saprobic microbes, we considered that small numbers of saprobic bacteria and fungi (introduced with the background microbial slurry) might contribute to soil respiration and (in the case of saprobic fungi) fungal biomass in the experimental root-mycocosms. One approach we considered to account for this potential saprobic activity was to assume it was the same in the nonmycorrhizal control root-mycocosms as in the treatment root-mycocosms. If so, C in ergosterol and respired CO2 in nonmycorrhizal controls would represent an estimate of background levels of saprobic fungi to be subtracted from experimental root-mycocosms. In preliminary calculations, subtracting control averages (from nonmycorrhizal controls) from experimental treatment values (from mycorrhizal experimental root-mycocosms) frequently resulted in values near or below zero, even when seedlings in experimental treatment root-mycocosms had abundant EM fungal colonization and mycelium, suggesting that saprobic activity was likely inhibited by mycorrhizal fungi in the experimental root-mycocosms, which is especially likely given the low nutrient levels in our experimental soils (Orwin et al., 2011; Fernandez & Kennedy, 2015; Averill & Hawkes, 2016; Sterkenburg et al., 2018). Ultimately, for these reasons, we chose not to subtract control values of C in ergosterol and respired CO2 from the experimental values (see Table S1 for a full data summary of mycorrhizal and nonmycorrhizal seedlings).

Data analysis

To test effects of light, fungal species, and time on absolute amounts and ratios of transferred resources (Question 1), data on C, N, and P fluxes, and N : C and P : C exchange prices were analyzed as response variables in separate univariate analyses using linear mixed-effects models using the lmertest package in R v.3.5.2, with growth period (1, 2, and 3), light level (high and low), EM fungal species, and their interactions as fixed effects (repeated-measures modeling was not required, since separate replicates were destructively harvested at each sampling point). Because the light treatment was applied to whole growth chamber shelves, and to account for variation among P. taeda genetic families, we included ‘shelf’ and ‘family’, respectively, as random effects. Highly nonsignificant three-way interactions (P > 0.2) were removed from models. In the case of significant effects of harvest, light treatment, EM fungal species, or their interactions, means were separated using orthogonal simple-effect contrasts to test for the effects of light and fungi when possible, or using Tukey HSD tests in other cases, using the emmeans package. The flux of C was also partitioned out into separate variables of C in fungal biomass and C respired by fungi, as well as carbon use efficiency (CUE), calculated as C in fungal biomass divided by total C transferred to fungi. These variables were also analyzed as above, to further understand how experimental factors may have affected C partitioning in the system.

To test for effects of resource transfer on accumulated fungal and plant biomass (Question 2), we conducted model selection (Burnham & Anderson, 2002) among all possible linear mixed-effect models for which the candidate variables were the main effects of the five resource flux variables representing effects of total resource transfer and exchange ratios (C, P, N, N : C, and N : P). As such, there were 32 candidate models in each of the fungal and plant analyses. Shelf and family were included as random effects in all models. This analysis was conducted using Akaike's information criterion corrected for small sample sizes (AICc), using the dredge() function in the mumin package. Model weights were used to calculate relative variable importance (RVI) for each candidate predictor across the entire set of candidate models, and predictors with RVI > 0.5 were considered important for inference. Models were initially fitted using the ML (maximum likelihood) approach to select the top models; then, REML (restricted maximum likelihood) fitting was used to obtain model-averaged parameter estimates from a 95% confidence set of the best models. Normality of residuals was confirmed using inspection of histograms for the global model containing all predictors.

In addition, we analyzed the effects of the light, fungal species, and time treatments on accumulated fungal and plant biomass, using linear mixed-effect models as described for the analyses of resource fluxes in Question 1.

Results

How does light availability affect resource exchange?

Total N : C exchange ratios (‘prices’) for the two EM fungal species changed differently over time (Harvest × Fungus interaction: F2,70.1 = 6.80, P = 0.0020; Fig. 2; Table S2), but were not affected by the light treatment. N : C transfer ratios were > 4× higher for seedlings inoculated with Pisolithus than for seedlings inoculated with Rhizopogon during the first growth period (P < 0.001), but reduced to values similar to seedlings inoculated with Rhizopogon through the second and third harvest (P = 0.32 and P = 0.43, respectively).

Details are in the caption following the image
Relationship between ectomycorrhizal fungal species (charcoal bars indicate Pisolithus fungi, and light gray bars indicate Rhizopogon roseolus fungi) and harvest (10, 20, 30 wk) on N : C exchange ratio (mean μmol, ±SE). Asterisks denote significant pairwise differences (P < 0.001) between the two fungi, for a particular growth period.

The P : C exchange ratio between seedlings and fungi was not affected by light treatment, nor did it differ between the two fungi (Table S3); however, cumulative P : C exchange ratios were > 2× higher during the first growth period than the subsequent growth periods (main effect of Harvest: F2,70 = 3.65, P = 0.031; Fig. S4), although those differences were not significant in pairwise Tukey HSD tests (0.05 < P < 0.1).

The total amount of C transferred from plants to fungi was affected by an interaction between light level, EM fungal species, and growth period (Light × Harvest × Fungus interaction: F2,66.9 = 53.5, P < 0.001; Fig. 3; Table S4). Specifically, cumulative C transferred to Rhizopogon was substantially higher under high light than under low light during the second and third growth periods, while cumulative C transferred to Pisolithus increased by the harvest period but was not affected by light (Fig. 3). A large majority of the total volume of C transferred to the fungi during the experiment was respired as CO2 (Pisolithus: average proportion = 0.996, SE = 0.0005; Rhizopogon: average proportion = 0.958, SE = 0.0056), and as a result, this pool of C exhibited a nearly identical response to the experimental treatments as total C (data not shown). Carbon transferred and incorporated into fungal biomass exhibited a different pattern of change over time for the two fungi (Fungi × Harvest interaction: F2,69.4 = 3.54, P = 0.035; Table S5), with biomass C of Rhizopogon being substantially lower compared with Pisolithus but steadily increasing through all three growth periods, while in Pisolithus reaching a maximum after the second growth period and not increasing during the third growth period (Fig. S5). Carbon use efficiency was influenced by an interaction between harvest date and EM fungal species (Fungi × Harvest interaction: F1,70.1 = 112.2, P < 0.001), generally declining over time, starting dramatically higher for Pisolithus than for Rhizopogon during the first growth period, declining but still significantly higher for Pisolithus than for Rhizopogon in the second growth period, and not different by the end of the third period (Fig. S6; Table S6).

Details are in the caption following the image
Relationship between ectomycorrhizal (EM) fungal species (black bars indicate Pisolithus arhizus EM fungi, and gray bars indicate Rhizopogon roseolus EM fungi), light treatment (orange frames indicate high-light environment, and blue frames indicate low-light environment), and harvest (10, 20, 30 wk) on total C transfer from seedlings to EM fungi (mean μmol, ±SE). Asterisks denote significant pairwise differences (P < 0.001) in cumulative C transfer between high and low light.

Total N transferred to seedlings was significantly affected by an interaction between light and EM fungal species (Light × Fungus interaction: F1,70 = 4.85, P = 0.031; Fig. 4; Table S7). Total N transfer was more than double in the high-light treatment compared with the low-light treatment for seedlings inoculated with Rhizopogon (P = 0.033) but did not differ between light treatments for seedlings inoculated with Pisolithus (P = 0.68). Cumulative total N transfer also increased steadily across growth periods, irrespective of EM fungi, approximately doubling between each subsequent growth period (main effect of Harvest: F2,70 = 23.7, P < 0.001; Fig. S7; Table S7).

Details are in the caption following the image
Relationship between ectomycorrhizal (EM) fungal species (Pisolithus arhizus and Rhizopogon roseolus), light treatment (orange bars indicate high-light treatment, and blue bars indicate low-light treatment), and total N transfer from fungi to seedling (mean μmol, ±SE). Asterisks denote a significant difference (P = 0.033) between high and low-light treatment means within an EM fungal species.

Total P transfer from both fungi to seedlings was significantly higher in the high-light treatment (main effect of Light: F1,70 = 5.5, P = 0.022; High Light: marginal mean = 3.40 μmol, SE = 0.96 μmol; Low Light: marginal mean = 2.02 μmol, SE = 0.95 μmol; Table S8), was significantly higher for seedlings inoculated with Rhizopogon (main effect of Fungus: F1,70 = 7.06, P = 0.010; Pisolithus: marginal mean = 1.91 μmol, SE = 0.93 μmol; Rhizopogon: marginal mean = 3.50 μmol, SE = 0.97 μmol) and increased significantly by the 30-wk harvest (main effect of Harvest: F2,70 = 12.4, P < 0.001; Fig. S8; Table S8).

Do resource-exchange ratios or absolute fluxes better predict EM fungal and pine seedling biomass accumulation?

For EM fungal biomass accumulation, the important predictors were total N transferred (RVI = 0.95), total P transferred (RVI = 0.93), and N : C transfer ratio (RVI = 0.55) (Table S9). The AICc-best model also contained exactly those three predictors, while P : C and total C had RVI scores of 0.34 and 0.25, respectively. Total N and N : C transfer ratio were positively associated with fungal biomass (model-averaged coefficients: 305.5 and 7846000.0, respectively), while total P was negatively associated with fungal biomass (coefficient: −6418.0). For plant biomass accumulation, the only important predictor was total P transferred (RVI = 1.0) (Table S9), and it was positively associated with plant biomass (model-averaged coefficient: 0.0305). The other four factors, total N, P : C, N : C, and total C, had RVI scores of 0.04, 0.32, 0.28, and 0.24, respectively.

In the analysis of how experimental factors of light, fungal species, and time affected accumulated fungal biomass, results were nearly identical to those for C in fungal biomass (see Question 1), as expected, with fungal biomass in Rhizopogon being substantially lower than that in Pisolithus but steadily increasing through all three growth periods, while in Pisolithus reaching a maximum after the second growth period and not increasing during the third growth period (Fig. S9; Table S10). Minor differences in fungal biomass compared with C in fungal biomass were due to slightly different C-to-biomass conversion factors used for Rhizopogon and Pisolithus. Accumulated plant biomass was significantly affected by an interaction between light and EM fungal species (Light × Fungus interaction: F1,67.3 = 5.94, P = 0.017; Fig. S10; Table S11). Specifically, plant biomass was significantly larger in the high-light treatment than in the low-light treatment for seedlings inoculated with Rhizopogon (P = 0.048) but did not differ between light treatments for seedlings inoculated with Pisolithus (P = 0.49). Accumulated plant biomass also increased from the first harvest to the second harvest, but did not increase from the second harvest to the third harvest (main effect of Harvest: F2,68.9 = 9.71, P < 0.001; Fig. S11; Table S11).

Discussion

We found that resource exchange between P. taeda seedlings and two different EM fungal species was strongly influenced by ambient light availability and that plant and fungal performance in the symbiosis was predicted by magnitudes of absolute resource exchange. Some of these relationships were consistent with predictions of the comparative advantage and excess resource-exchange hypotheses, but often varied strongly between the two different EM fungi and changed over the time course of the development of the symbioses.

How does light availability affect resource exchange?

Absolute fluxes of C, N, and P, but not N : C or P : C flux ratios, were affected by light levels, although the effects of light depended on both time and fungal species. Specifically, Rhizopogon transferred more total N and P under high light than under low light (Fig. 4; Table S8) and through the second and third growth periods had received more total C from seedlings under high light than under low light (Fig. 3). Pisolithus also transferred more P under high light (Table S8), but N and C transfer with Pisolithus was not influenced by light levels (Figs 3, 4). This result for Rhizopogon supports the Grman et al. (2012) model for AMF mutualisms, which predicts increases in total fluxes of nutrients under high light. It is also consistent with the results of Kiers et al. (2011), who found that AM fungi can discriminate among hosts with varying C supply and allocate more nutrient transfer to those root tips with higher C supply. Previous experiments with EM symbiosis have found that plants may allocate more C to fungi providing more N (Bogar et al., 2019), but here, the resource exchange responded to C supply being manipulated, suggesting plant control. The fact that absolute C transfer to Rhizopogon, but not N : C or P : C transfer ratios, was stimulated by high light suggests that pine seedlings in symbiosis with Rhizopogon may have been C-limited under low light, although they did not pay a lower price of C (relative to N or P) when C was less available, as predicted by economic models of resource exchange (Schwartz & Hoeksema, 1998; Hoeksema & Schwartz, 2003; Kummel & Salant, 2006; Akçay & Roughgarden, 2007; Franklin et al., 2014; Akçay, 2015). Altogether, our results support the idea that increased light availability may stimulate overall trade between plants and mycorrhizal fungi, but that the phenomenon may be fungal species-specific.

Previous analyses of the effects of light on mycorrhizal resource exchange are rare. However, elevated CO2 studies have commonly found an increase in C allocation to mycorrhizal fungal biomass with increased C availability (Treseder, 2004; Näsholm et al., 2013), which is consistent with our results here showing larger transfer of C to Rhizopogon under high light. In addition, multiple studies have found a shift in the composition of EM (Pena & Polle, 2014) and AM (Knegt et al., 2016) fungal communities with reduced light (Johnson et al., 1997; Shi et al., 2014). Those results may imply shifts in resource exchange with light. For example, if nutrient-flux responses consistently vary among mycorrhizal fungal species, as we observed here, those differences could mediate changes in fungal communities in response to partner choice by plants, as predicted by the economic model of Kummel & Salant (2006).

Both EM fungi tested (Rhizopogon and Pisolithus) are considered high biomass exploration types (Agerer, 2001, 2006), but showed a distinct difference in resource-exchange responses to light treatments (Figs 3, 4). It is therefore not apparent that these results are applicable to high biomass exploration biomass types in general. Some have suggested that high biomass EM fungal types may require more C for long-distance growth (Godbold et al., 1997). However, higher biomass exploration types may provide more efficient nutrient uptake and long-distance transport (Koide et al., 2014), and there is also evidence for a higher C demand from low exploration biomass types (Bidartondo et al., 2001), perhaps resulting from higher respiration costs. Thus, our results imply that fungal traits other than exploration type may be important for driving variation in nutrient fluxes between different EM fungi.

Although the cumulative total transfer of C to Pisolithus was lower than to Rhizopogon at the end of each growth period (Fig. 3), Pisolithus exhibited a higher CUE than Rhizopogon, especially during the first two growth periods, converting a higher proportion of transferred C to increased biomass (relative to respiration) (Fig. S6). This increased allocation to biomass in turn apparently resulted in an increase in the initial N transferred and a spike in the Pisolithus N : C exchange ratio (Fig. 2). However, as nutrient availability in the soil system became more limited over the course of the experiment, the ratio of N : C exchange dropped in subsequent harvests (Fig. 2). Pisolithus fungi are known to vary widely (among isolates) in capabilities to access organic soil N (Chambers & Cairney, 1999). The Pisolithus isolate used in the present experiment seemingly used its available C to build biomass to explore soil for easily accessible nutrients. Rhizopogon, by contrast, exhibited consistently lower CUE (Fig. S6) and lower total biomass accumulation, although its biomass did increase during each growth period, including the third period when Pisolithus biomass did not (Fig. S9). This low growth rate of Rhizopogon may be expected, as other Rhizopogon isolates have been found to grow poorly under low C : N conditions, as in our soil system (Hatakeyama & Ohmasa, 2004). Moreover, Rhizopogon species are relatively specialized for accessing organic N via extracellular enzymes and therefore may respire a higher proportion of C transferred from the plant (Fransson et al., 2007). This high respiration rate was especially true later in the experiment under high light, as C investment from the plant was highest and soil nutrient availability was likely at its lowest (Fig. 3). These results are consistent with a model by Franklin et al. (2014), and with the comparative advantage biological market models, which predict that as soil N availability decreases, mycorrhizal N export to plants decreases, resulting in a lower N : C exchange rate.

Do resource-exchange ratios or absolute fluxes better predict EM fungal and pine seedling biomass accumulation?

Consistent with the excess resource-exchange hypothesis (Corrêa et al., 2008, 2011, 2012), and in contrast to economic models (Schwartz & Hoeksema, 1998; Hoeksema & Schwartz, 2003), total P transferred from fungi was the only important resource predictor of seedling biomass, with total N, total C, N : C, and N : P being unimportant predictors of seedling biomass. Moreover, absolute P transfer was positively associated with plant mass, implying that P was limiting to plant growth in this system. The low-light treatment in this experiment was designed to produce C limitation on the symbiosis, and these results suggest that such limitation may not have been substantial enough (despite a positive influence of light on total C transfer to Rhizopogon) to influence the outcome of the EM symbiosis from the plant perspective.

Similarly, biomass accumulation in EM fungi was best predicted by the absolute fluxes of N and P from fungi to plants, again in contrast to predictions of economic models. Fungal biomass was positively associated with flux of N to plants and negatively associated with flux of P to plants, implying P limitation of fungal growth in the system. This P limitation of fungal growth may help explain the low observed CUE of the fungi in this study. Especially in the later growth stages, plants continued to transfer substantial C to fungi, and C transfer to Rhizopogon was stimulated by light (Fig. 3), but the ability for EM fungi to utilize this additional C is dependent on nutrient availability in the soil and C may be quickly respired by the fungi if mineral nutrients are low in availability (Fransson et al., 2007; Albarracín et al., 2013). Indeed, fungal biomass was not influenced by light availability, although the respiration of Rhizopogon and the biomass of plants growing with Rhizopogon were both stimulated by high light, supporting the idea that Rhizopogon was investing C in costly mechanisms to scavenge for soil nutrients, including P to alleviate its own P limitation, and N for trade to plants in exchange for C. The N : C transfer ratio was also positively associated with fungal biomass, which could be explained by a dependence of CUE on N and C flux; for example, it is possible that under some conditions, fungi took up more N, transferred more N to plants, receiving less C from plants but building more biomass through higher CUE. Alternatively, this result could also further indicate the lack of limitation on fungal growth by either N or C, and a complex interaction of N : C with P limitation of fungal growth.

Changes in resource exchange over time

We observed that some nutrient fluxes changed over time in different ways for the two different EM fungal species. Although P : C exchange ratios declined after the first growth period for both fungi (Fig. S4), the two fungi differed in their temporal patterns of N : C nutrient exchange. Pisolithus inoculated seedlings had sharply decreasing N : C transfer ratios after the first harvest (Fig. 2), while both N (Fig. S7) and C (Fig. 3) total fluxes increased with time, apparently because C transfer to Pisolithus was increasing over time at a faster rate than were N transfers to the plant. This pattern, as well as the decline over time in P : C exchange ratios, may have occurred due to the overall low N and P availability in the soil, leading seedlings to invest increasingly more C in EM fungi in an attempt to gain more soil nutrients (Treseder & Allen, 2002; Hobbie, 2006; Corrêa et al., 2008; Hasselquist et al., 2016), and/or due to increased total photosynthetic capacity of seedlings over time, consistent with predictions of comparative advantage biological market models. By contrast, Rhizopogon inoculated seedlings had stable N : C transfer ratios throughout the experiment (Fig. 2) even though fluxes of both N (Fig. S7) and C (Fig. 3) increased throughout the experiment, suggesting a consistent price of N : C trade of resources between mutualists (Kiers et al., 2011).

Methodological considerations

Several aspects of the methods and assumptions used here are worth noting for their potential effects on our conclusions. The artificial light intensity and photoperiod of this environmental chamber study may have affected resource exchange. However, our light treatments were similar to those used in other studies in environmental chamber systems (Corrêa et al., 2012), and we wanted to use low enough light levels so that light limitation on photosynthesis would be a possibility. Also, the soil substrate used here was, by design, poorer than would be found in many natural pine-EM systems, with low nutrient and organic matter content, and we supplemented the system with inputs of mineral nutrients periodically. It is possible that supplying more nutrients overall, alleviating potential nutrient limitation of the fungi, would shift the system towards C limitation, under which predictions of the comparative advantage economic models might make more sense. Augmenting nutrients in an organic form might change resource-exchange fluxes for any particular combination of plants and EM fungi, since EM fungal species vary in their abilities to use mineral vs organic forms of nutrients (Koide et al., 2014).

Nutrient fluxes in pairings of a single seedling with a single isolate of a single EM fungal symbiont species, as utilized here, may not be indicative of how those same fungi would interact with plant hosts when in competition with other fungal species or isolates on the same root system (Nowak & May, 1994; Kummel & Salant, 2006; Franklin et al., 2014; Hortal et al., 2017; Bogar et al., 2019), or with how interactions with saprotrophs may affect resource exchange (Lindahl et al., 2001; Leake et al., 2002). However, measuring resource exchange between individually paired plants and mycorrhizal fungal species is an important first step in understanding discrete resource exchange between mutualists. Future studies should test how resource exchange is altered by co-colonization and or cohabitation of multiple host seedlings and/or fungal species. Also, future studies would ideally directly partition biomass and respiration of saprobic organisms (Bol et al., 2003), and would make more frequent measurements of respiration to account for potential trends in respiration rates during particular growth periods; in the current study, there is some uncertainty in the estimates of respiration, due to the lack of such additional measurements.

Because all root tips were colonized by EM fungi, we have for simplicity made the assumption that 100% of N and P found in plants was transferred by fungi; however, it is likely that the true value is something < 100%, that is some proportion of N and P in plant tissue was taken up directly by pine seedling roots, bypassing the EM fungi. To whatever degree the latter is true, the presented resource flux values including N, P, N : C, and P : C would be overestimates or should be interpreted as flux from both roots and fungi. This would imply a degree of decoupling of plant and fungal roles in nutrient exchange in the symbiosis, although we would not expect it to qualitatively change our conclusions here. In future studies, it would be ideal to include estimates of the proportion of nutrient uptake via symbionts vs directly through roots.

Future directions and conclusions

Future studies of resource-exchange fluxes in EM systems could benefit from evaluating a larger variety of EM fungal mutualists to compare groups of EM fungal species differing in traits, such as their exploration morphotypes, successional status, ability to utilize organic vs mineral sources of nutrients, and propensity to form common mycorrhizal networks (CMNs). The latter could help to clarify the interpretation of results from field experiments manipulating the presence and absence of EM CMNs, which tend to also change the community composition of the EM fungi, favoring early-stage EM fungi in non-CMN treatments (Hoeksema & Bruna, 2015), and making it difficult to know whether changes in plant performance are due to CMNs per se or to changes in EM fungal functional traits.

Evaluating central assumptions and predictions of economic and other models, in addition to evaluating key effects of contextual variables, is crucial to fully understanding resource-exchange mutualisms. Our study suggests that, at least under the experimental conditions used here, light availability can be important for nutrient fluxes and exchange ratios between EM mutualists, but that limiting nutrient fluxes, not light and exchange ratios, are most important in predicting plant and fungal growth. These results especially support predictions of models other than economic models, such as the tit-for-tat and the excess resource-exchange hypothesis.

Acknowledgements

We thank Dr Ami Lokhandwala and student research assistants including, Michelle Ha, Thomas Moorman, Richard Easterling, Branden Jones, Pearl Reed, Mary Moss, Alexis Richardson, Madison Woodruff, and Meghan Van, for their assistance in accomplishing this research, and Dr Michael G. Booth for his early contributions and for building the first root-mycocosms and custom respiration chamber. This research was supported by an award from the University of Mississippi Graduate Student Council to ALH and by grants from the National Science Foundation to JDH (DEB no. 1119865) and KAK (DBI no. 0923063).

    Competing interests

    None declared.

    Author contributions

    ALH, JDH, KAK and MM planned and designed the research. ALH, KAK and SSK performed experiments. ALH and JDH analyzed data and drafted the manuscript. KAK, MM and SSK edited the manuscript.

    Data availability

    The data and R code that support the findings of this study are openly available in Data Dryad at doi: 10.5061/dryad.zkh1893d4.