Elevated CO2 spurs reciprocal positive effects between a plant virus and an arbuscular mycorrhizal fungus

Authors

  • Megan A. Rúa,

    Corresponding author
    1. Department of Biology, University of Mississippi, Oxford, MS, USA
    • Curriculum for the Environment and Ecology, University of North Carolina, Chapel Hill, NC, USA
    Search for more papers by this author
  • James Umbanhowar,

    1. Curriculum for the Environment and Ecology, University of North Carolina, Chapel Hill, NC, USA
    2. Department of Biology, University of North Carolina, Chapel Hill, NC, USA
    Search for more papers by this author
  • Shuijin Hu,

    1. Department of Plant Pathology, North Carolina State University, Raleigh, NC, USA
    Search for more papers by this author
  • Kent O. Burkey,

    1. USDA – ARS Plant Science Research Unit, and Department of Crop Science, North Carolina State University, Raleigh, NC, USA
    Search for more papers by this author
  • Charles E. Mitchell

    1. Curriculum for the Environment and Ecology, University of North Carolina, Chapel Hill, NC, USA
    2. Department of Biology, University of North Carolina, Chapel Hill, NC, USA
    Search for more papers by this author

Errata

This article is corrected by:

  1. Errata: Corrigendum Volume 199, Issue 4, 1109, Article first published online: 10 July 2013

Author for correspondence:

Megan A. Rúa

Tel: +1 662 915 7203

Email: marua@olemiss.edu

Summary

  • Plants form ubiquitous associations with diverse microbes. These interactions range from parasitism to mutualism, depending partly on resource supplies that are being altered by global change. While many studies have considered the separate effects of pathogens and mutualists on their hosts, few studies have investigated interactions among microbial mutualists and pathogens in the context of global change.
  • Using two wild grass species as model hosts, we grew individual plants under ambient or elevated CO2, and ambient or increased soil phosphorus (P) supply. Additionally, individuals were grown with or without arbuscular mycorrhizal inoculum, and after 2 wk, plants were inoculated or mock-inoculated with a phloem-restricted virus.
  • Under elevated CO2, mycorrhizal association increased the titer of virus infections, and virus infection reciprocally increased the colonization of roots by mycorrhizal fungi. Additionally, virus infection decreased plant allocation to root biomass, increased leaf P, and modulated effects of CO2 and P addition on mycorrhizal root colonization.
  • These results indicate that plant mutualists and pathogens can alter each other's success, and predict that these interactions will respond to increased resource availability and elevated CO2. Together, our findings highlight the importance of interactions among multiple microorganisms for plant performance under global change.

Introduction

Effects of increased atmospheric CO2 on plant growth and productivity are expected to occur both directly via plant physiological responses (Lee et al., 2001) and indirectly via impacts on microbes that associate with plants (Malmstrom & Field, 1997; Johnson et al., 2005). Plant pathogens and arbuscular mycorrhizal (AM) fungi are two ubiquitous classes of microorganisms that can directly impact plant allocation of carbon (C), and may in turn indirectly alter each other's success (Bennett et al., 2006; Smith & Read, 2008). Elevated CO2 generally increases the positive impact of AM fungi on plant growth (Treseder, 2004). Additionally, elevated CO2 can increase disease tolerance by reducing the negative impacts of pathogen infection on plant growth (Malmstrom & Field, 1997). Together, these studies suggest the potential for interactive effects of plant pathogens and AM fungi on plant performance under elevated CO2. Yet, there have been no studies considering their joint impact on plant performance under elevated CO2. Thus, the goals of this experiment were to explore the independent and interactive effects of viral plant pathogens, fungal mutualists and changing resource levels on plant performance.

Plants often simultaneously support mutualists and are attacked by natural enemies such as herbivores and pathogens, creating the potential for interactions that impact plant performance. A meta-analysis of plant–enemy–mutualist interactions concluded that, on average, the presence of mutualists lessens the negative effect of both herbivores and pathogens on plant performance (Morris et al., 2007). However, the impact of AM fungi on the performance of plants exposed to natural enemies depended upon the identity of the enemy examined (i.e. whether the enemy was a pathogen or herbivore and what type; Borowicz, 2001). Similarly, the effects of herbivore damage on plant performance can also depend on the identity of the plant mutualist (Bennett & Bever, 2007). Furthermore, while AM fungi are generally mutualistic, their overall effect on host plants may vary from beneficial to pathogenic depending on host identity and abiotic factors such as nutrient availability (Johnson et al., 1997). Also, plant viruses, even those considered pathogenic and that cause agricultural losses, can range from pathogenic to beneficial depending on abiotic factors such as water and heat stress – two factors highly relevant to global change (Roossinck, 2011). The context dependency and variability present in these interactions demonstrates that we cannot safely extrapolate from one type of interaction to another, so understanding interactions between mycorrhizal fungi and plant viruses will require the direct study of those systems.

Interactions between plant mutualists and pathogens are likely mediated by their shared host. Theoretical models predict that, by improving plant nutrition and tolerance, AM fungi will also increase enemy populations (Bennett et al., 2006). This prediction may be extended to explain the effects of AM fungi on plant viruses, with plant phosphorus (P) as a major mediating factor. Among natural enemies, viruses may be particularly limited by P availability within hosts because they are comprised mainly of nucleic acids that have a relatively high concentration of P (Clasen & Elser, 2007). AM fungi generally increase plant P concentration under both ambient and elevated CO2 conditions (Smith & Read, 2008). Therefore, host plants associated with mycorrhizal fungi may have higher viral titer due to their higher shoot phosphorus content. There is some evidence in agricultural systems that suggest an increase in viral titer as a result of association with mycorrhizal fungi (Daft & Okusanya, 1973; Schonbeck, 1979), but such reports are limited.

The impact of global change on plant communities may be mediated through indirect effects, including via pathogens (Burdon et al., 2006). Because pathogens do not fix C, such indirect effects must begin with effects of elevated CO2 on plant physiology. Although the effect of CO2 can vary considerably across plant species and environmental gradients (Lee et al., 2001), plants grown under elevated CO2 generally show increased concentrations of photosynthates (Pritchard et al., 1999; Ward et al., 2005), potentially increasing the resources available to pathogens infecting the host plant (Clasen & Elser, 2007; Alexander, 2010). Alternatively, elevated CO2 may alter plant–pathogen interactions by changing plant defense traits, including those traits associated with tolerance and resistance (Burdon et al., 2006). Elevated CO2 can alter traits that are associated with pathogen tolerance, the capacity to vegetatively or reproductively compensate for damage by enemies (Strauss & Agrawal, 1999). Specifically, elevated CO2 can enhance traits associated with tolerance such as photosynthetic capacity, root biomass and C stores (Strauss & Agrawal, 1999; Ainsworth & Long, 2005), leading to an increase in plant tolerance of infection (Malmstrom & Field, 1997). Overall, changes in plant performance and physiology in response to elevated CO2 may change the growth, fecundity and population dynamics of pathogens (Alexander, 2010).

Just as with plant pathogens, alterations of plant physiology due to elevated CO2 may, in turn, impact mycorrhizal fungi (Johnson et al., 2003; Treseder, 2004; Klironomos et al., 2005). The carbon limitation hypothesis suggests that when C is limiting, such as can occur under ambient CO2 or under foliar herbivory, AM fungal growth will be reduced because C will be preferentially allocated to other plant parts or to the soil pool excluding AM fungi (Gehring & Whitham, 2002). Therefore, we expect that elevated CO2 will alter plant physiology to increase the available C to AM fungi, thereby strengthening the mutualism by increasing one currency of the mutualism. A meta-analysis of atmospheric CO2 studies found that mycorrhizal fungi consistently and significantly increased their growth in response to elevated CO2 (Treseder, 2004). However, mycorrhizal fungi have also been reported to reduce their beneficial effects on plant biomass under elevated CO2 (Johnson et al., 2003, 2005).

In total, association with mutualists can alter plant–pathogen interactions, and pathogens can also alter plant–mutualist interactions. In addition, elevated CO2 can alter both plant–pathogen and plant–mutualist interactions. Together, this suggests that elevated CO2 will alter interactions between plant pathogens and mutualists. Yet to date, no experimental studies have examined the effects of elevated CO2 on plants associated with both mutualists and pathogens.

Materials and Methods

Study system

Barley and cereal yellow dwarf viruses (B/CYDVs) are a group of aphid-transmitted generalist viral pathogens that infect over 150 crop and noncrop grasses (D'Arcy, 1995; Halbert & Voegtlin, 1995). Infection is systemic and localized to the phloem where it causes necrosis and disruption of carbohydrate translocation (Irwin & Thresh, 1990; D'Arcy, 1995). B/CYDV infection stunts plant growth (Malmstrom et al., 2005a), reduces root : shoot ratio (Kolb et al., 1991) and reduces longevity. B/CYDVs are obligately transmitted by aphids, including the globally common aphid species Rhopalosiphum padi (L.) (Irwin & Thresh, 1990; D'Arcy & Burnett, 1995).

AM fungi are ubiquitous plant symbionts that provide the host with a wide range of benefits including the acquisition of less mobile mineral nutrients, particularly phosphorus (Smith & Read, 2008). In return, AM fungi receive carbohydrates from the plant. To this end, in addition to altering leaf-level photosynthesis (Smith & Read, 2008), AM fungi can increase plant root growth (Bryla & Eissenstat, 2005).

For this experiment we used two C3 Eurasian annual host plants, Bromus hordeaceus L. and Avena fatua L., known invaders of Western US grasslands (Malmstrom et al., 2005b). These host plants were chosen because they are both colonized by AM fungi (Hu et al., 2005; Rillig, 2006) and are hosts for B/CYDVs (Malmstrom et al., 2005b). To ensure genetically diverse hosts, seeds from c. 100 wild plants per species were hand-collected in Oregon and germinated in experimental pots. When more than one plant germinated, plants were thinned down to one plant per pot. Plants were watered every 3 d.

Experimental conditions

The experiment was conducted in the CO2 exposure facility at the USDA-ARS Air-Quality glasshouse at North Carolina State University in Raleigh, NC, USA. Each continuously stirred tank reactor (CSTR) chamber measured 1.2 m in diameter by 1.4 m tall (Chen et al., 2007). Gases were dispensed and monitored in a laboratory adjacent to the glasshouse. A blower system provided a constant flow of charcoal-filtered air through each CSTR. For those chambers assigned to an elevated CO2 treatment, compressed CO2 was added to the air entering the CSTR. To maintain CO2 at a constant concentration, a rotameter was used to control flow. The potential heating effect of the chambers was alleviated by air which was continuously moved through the CSTR. Monitoring of CO2 concentration was accomplished using computer-activated solenoid valves to direct gas exiting the CSTR into infrared analyzers (model 6252; LiCor Inc., Lincoln, NE, USA).

Experimental design and data collection

We used a fully factorial design in which AM fungi (mycorrhizal and nonmycorrhizal), virus (infected and uninfected) and phosphorus (P; addition and ambient) were manipulated as subplot factors at the individual plant level and atmospheric CO2 concentration (ambient and elevated CO2) was applied as a whole plot factor. There were three chambers per CO2 concentration level for a total of six chambers. Targeted treatments of either ambient or elevated CO2 concentration (ambient + 200 ppm) were randomly assigned to each chamber. Elevated CO2 concentration was within the range of concentrations predicted by the IPCC for the end of this century (IPCC, 2001). Measured values (mean ± SE) for ambient and elevated CO2 treatments during the study were 387 ± 11 and 581 ± 11 ppm, respectively.

There were 32 total factorial treatment combinations, each of which started with 12 replicates (for a total of 384 replicates, with each replicate being an individual plant) but this number changed somewhat over the course of the experiment and with research question. Plants that did not survive, flowered or were not successfully inoculated with the appropriate treatment were eliminated from analyses, resulting in 339 total plants (162 A. fatua and 177 B. hordeaceus). No treatment effects were significantly related to plants that did not survive or flower.

Individual plants were grown in D60 Deepots (Steuwe and Sons Inc., Tangent, OR, USA). We were interested in the effects of P and mycorrhizal fungi on plants growing under nutrient-poor conditions. Each plant was grown in 800 g of steam-sterilized field soil in a mixture of one part sandy loam with two parts of pure sand by mass. Field soil was collected from a site adjacent to the CSTR facility and steam sterilized to remove any existing soil microbes. Before mixing with sand, the soil contained 2.8 g kg−1 total C and 0.22 g kg−1 total N, as well as 7 mg kg−1 available P. The nutrient-poor soil resulted in slow plant growth, and allowed the plants to grow for an extended period without producing enough biomass to become either light-limited or root-bound. To inoculate plants with AM fungi, we added 50 g of active mycorrhizal spore inoculum per pot. We used commercially available inoculum AM120 from Reforestation Technologies International (Salinas, CA, USA) which primarily consists of the AM fungal species Glomus intraradices. Glomus intraradices is a ubiquitously distributed AM fungus and this inoculum has been commonly and effectively used for restorations of grassland habitats on the Pacific coast of the US, where seeds of the host species were collected (Vogelsang & Bever, 2010). Restored populations of perennial grasses in this region are commonly invaded by both experimental host species (Malmstrom et al., 2005a,b) so their response to this mycorrhizal inoculum is relevant to populations in the field. Control plants received 50 g of autoclave sterilized inoculum to control for potential changes in nutrient content due to the inoculum. To ensure that, in addition to the AM fungus, the same soil microbial community was added to all treatments, all pots received 100 ml of microbial filtrate solution filtrated by Whatman No. 1 filter paper. Solution was created from 10.0 g AM inoculum (in which mycorrhizal spores were removed) to correct for possible differences in the microbial community and mineral content between mycorrhizal and mock mycorrhizal treatments. Plants in the P addition treatment received 1.42 g of triple super phosphate (Ca(H2PO4)2) per pot, mixed into the soil before planting while ambient pots received no additional P.

In order to infect plants with virus, we used an isolate of Barley yellow dwarf virus – PAV (hereafter referred to as BYDV for brevity) that has previously been used in inoculation experiments (Cronin et al., 2010). This isolate was obtained in August 2007 from a naturally infected Bromus vulgaris individual in Oregon and has been maintained since its collection (c. three transmission cycles per year) in laboratory plants of the Avena sativa cultivar Coast Black Oats. The virus isolate has not been sequenced yet and is currently not included in GenBank.Virus inoculations occurred c. 2 wk after germination when plants were at the two leaf stage. Uninfected adult aphids of the species R. padi were fed in Petri dishes for 72 h on infected plant tissue. Five infected aphids were then transferred to each experimental plant, at which time a plastic/nylon mesh cap was placed on the plants to prevent the spread of aphids. Aphids were allowed to feed on each experimental plant for 48 h. Plants were then uncapped and sprayed with a horticultural oil solution (SAF-T-SIDE; ClawEl Specialty Products, Pleasant Plains, IL, USA) to kill the aphids. Mock-inoculated plants received the same treatment but uninfected aphids were fed on uninfected tissue before being transferred to experimental plants. To test the plants for BYDV infection and to quantify relative viral titer concentration, a compound indirect double-antibody sandwich Enzyme-linked Immunosorbent Assay (ELISA; Agdia Inc., Elkhart, IN, USA) was used on 0.1–0.3 g wet aboveground tissue collected from each experimental plant at harvest (Cronin et al., 2010).

In order to assess viral responses, we used relative viral titer. Viral titer is the concentration of virus present in plant tissue. ELISAs generate optical density (OD) values that can be used as a measure of relative viral titer (Gray et al., 1991; Cronin et al., 2010). We employed ELISA methods for assessment of BYDV-PAV infection from Gray et al. (1991), as modified by Cronin et al. (2010), with two key modifications being as follows. First, we used monoclonal antibodies (Agdia Inc.). Second, instead of employing dilutions of purified virus, we assessed the potential for compounds of plant origin to reduce the precision or consistency of OD values by serially diluting sap from infected Coast Black Oats (a cultivar of Avena sativa that yields relatively high virus titers) in sap from uninfected individuals of each of our experimental host species (M. G. Dekkers & C. E. Mitchell, unpublished data). If the identity of uninfected plant sap caused substantial variation in OD, then the OD values would need to be standardized to remove this variation. By contrast, these serial dilutions indicated that the variation in OD values among our experimental treatments was several times greater than could be explained by compounds of plant origin. This indicates that most of the variation in optical density of infected plants was caused by titer, and therefore our observed optical density values adequately represent viral titer. The serial dilutions also allowed us to confirm that the OD values reported here were in the range where the response of OD to changes in virus concentration is not saturated (Copeland, 1998). Previous comparisons of a similar ELISA procedure with transmission assays indicated that ELISA was highly sensitive even at low titers (Cronin et al., 2010). To assess treatment effects on viral titer, analyses of OD values were limited to plants that ELISA confirmed to be infected with BYDV, resulting in 161 total plants for this response variable.

Plants were allowed to grow for five and a half months and then harvested. Only five plants produced reproductive structures and we excluded these from our data. At harvest, plants were separated into above- and belowground portions. Both above- and belowground biomass was placed in a drying oven. Plants were dried at 60°C for a minimum of 72 h to obtain dry biomass values. The soil was frozen and stored at −20°C until it could be processed. The belowground fraction was washed to separate roots from soil. A subset of the roots from each individual was collected before drying, stained with trypan blue following the methods outlined in Koske & Gemma (1989) and scored for intraradical AM fungal colonization using the magnified gridline intersect method (McGonigle et al., 1990). Using this method, the proportion of root length colonized by intraradical hyphae was measured using a compound microscope (×200–×400) by examining the intersections between the microscope eyepiece crosshairs and roots for evidence of AM fungal colonization (arbuscules, hyphae and/or vesicles). Plant phosphorus concentration was determined using a colorimetric analyzer (Alpkem Corporation, Clackamas, OR, USA) after dry ash/acid extraction (Stable Isotope/Soil Biology Laboratory of the University of Georgia's Odum School of Ecology; Jones et al., 1990).

Analyses

We used several response variables to assess experimentally induced changes in plant performance. To assess changes in allocation we used root fraction. Root fraction (root biomass divided by total plant biomass) quantifies the portion of the plant's total biomass allocated to roots. We measured root mass and fraction because BYDV is known to have strong negative effects on the root biomass of crop plants (Irwin & Thresh, 1990; D'Arcy & Burnett, 1995) and AM fungi exist within the root portion of the host plant. Total plant biomass was the sum of all above-and below-ground biomass. To account for the portion of aboveground tissue removed for ELISA, we used a wet/dry conversion factor. To calculate this conversion factor, aboveground material minus ELISA tissue was weighed immediately after harvest and divided by the weight of the same aboveground material after drying. The DW for tissue removed from ELISA was then added to complete the total biomass metric. After removing material for ELISA, 81 of the 339 total plants did not have enough plant material for phosphorus analysis and were removed from analyses for this response variable. Thus, analyses considering percent leaf P as a response variable used 258 plants.

In order to evaluate specific microbial responses, subsets of the dataset were further analysed accounting for the split plot design. To assess AM fungal response, analyses were limited to only those plants that received active mycorrhizal inoculum and had > 5% colonization, resulting in 168 total plants. To assess treatment-induced changes in AM fungal response, mycorrhizal hyphal colonization of roots was used as a measure of fungal performance (Smith & Read, 2008). Also, higher proportional colonization values in general indicate a greater proportion of plant resource allocation to the mycorrhizal fungus (Smith, 2009; Smith et al., 2009).

All data were analysed using R (v2.13.1; R Foundation for Statistical Computing, Vienna Austria) with the ‘lme4’ package and the ‘lmer’ function (Bates & Maechler, 2009). Data from the experiment were subjected to analysis of variance using general linear models with error terms appropriate to a split plot design. Response variables were log transformed to fit model assumptions of homogeneity of variances when necessary. Each statistical model analysed included all treatment variables and all their interactions. Differences within in a treatment were determined using Tukey's HSD which is a single-step multiple comparison procedure. To do this we used the ‘glht’ function of the ‘multcomp’ package (Hothorn et al., 2010). When interactions included plant species, Tukey's tests were performed within each species because main effects already indicated differences between species. Full statistical model tables for all response variables can be found in the Supporting Information Tables S1–S5.

Results

Plant biomass and root fraction

Across all treatments and both plant species, virus infection reduced root fraction by 20% (F1,300 = 45.2, < 0.0001; Fig. 1a) and tended to decrease total plant biomass by 8.6% (F1,300 = 2.92, = 0.088; Fig. 1b). Elevated CO2 increased total plant biomass of nonmycorrhizal A. fatua, but not of mycorrhizal A. fatua or B. hordeaceus plants (AM fungi × CO2 × plant species interaction: F1,300 = 4.4, = 0.037, Fig. S1). On average, across all individuals in the experiment, B. hordeaceus individuals had 52% less total biomass (F1,300 = 79.1, < 0.0001) and 43% smaller root fraction than A. fatua (F1,300 = 126.2, < 0.0001).

Figure 1.

Across plant species (Avena fatua and Bromus hordeaceus), virus infection decreased root fraction (a) and tended to decrease total plant biomass (b). Data shown are means ± SEM.

Leaf P concentration

Phosphorus addition to the soil increased P concentration in leaves of both species, but the effect was larger for A. fatua than B. hordeaceus (phosphorus × plant species interaction: F1,257 = 61.9, < 0.0001; Fig. S2). Virus infection also increased leaf P concentration for both species, but more for A. fatua than B. hordeaceus (virus × plant species interaction: F1,257 = 5.15, = 0.0241; Fig. 2). Mycorrhizal fungi did not increase leaf P concentration (F1,257 = 0.21, = 0.644).

Figure 2.

Across mycorrhizal status, virus infection increased leaf phosphorus (P) concentration, and more for Avena fatua than Bromus hordeaceus. Closed bars, +virus; open bars, −virus. Data shown are means ± SEM; letters indicate significant pairwise differences between means (Tukey's HSD; P < 0.05).

Viral titer

As an indicator of relative viral titer (concentration) in leaf tissue, we analysed optical density (OD) values from ELISAs. Plant association with mycorrhizal fungi increased the OD of virus infected plants under elevated CO2, but not under ambient CO2 (AM fungi × CO2 interaction: F1,141 = 4.622, = 0.033; Fig. 3a). Phosphorus addition decreased OD for A. fatua, but not for B. hordeaceus (phosphorus × plant species interaction: F1,141 = 4.26, = 0.0409; Fig. 3b).

Figure 3.

Across plant species and phosphorus treatment, mycorrhizal colonization (+AMF vs −AMF) increased relative viral titer as measured by optical density (OD) value for plants under elevated CO2 but had no effect under ambient CO2 (a). Across mycorrhizal status, phosphorus addition (+P vs −P) decreased relative viral titer as measured by ELISA optical density (OD) value for Avena fatua but not for Bromus hordeaceus (b). Data shown are means ± SEM; letters indicate significant pairwise differences between means (Tukey's HSD; P < 0.05).

AM fungal colonization of plant roots

Mirroring the effect of mycorrhizal fungi on relative viral titer, virus infection increased mycorrhizal colonization of roots by 69% under elevated CO2, but not under ambient CO2 (CO2 × virus interaction: F1,148 = 11.4, = 0.0009; Fig. 4a), Also, elevated CO2 increased hyphal colonization of virus-infected plants more than of virus-uninfected plants. Further, P addition decreased hyphal colonization by 37% under elevated CO2, but not under ambient CO2 (CO2 × phosphorus interaction: F1,148 = 10.5, = 0.0015; Fig. 4b). Phosphorus addition also decreased mycorrhizal colonization of virus-infected B. hordeaceus, but not A. fatua or virus-uninfected B. hordeaceus (virus × phosphorus × plant species interaction: F1,148 = 4.62, = 0.033; Fig. 5). Finally, elevated CO2 increased hyphal colonization of both plant species, but the effect was larger for A. fatua than for B. hordeaceus (CO2 × plant species interaction: F1,148 = 13.6, = 0.0003; Fig. 6).

Figure 4.

Across plant species (Avena fatua and Bromus hordeaceus), virus infection increased root colonization (proportion of sites colonized) by mycorrhizal fungi under elevated CO2 (closed bars) but not under ambient CO2 (open bars; a). Phosphorus addition (+P vs −P) decreased root colonization by mycorrhizal fungi under elevated CO2 (closed bars) but not ambient CO2 (open bars; b). Data shown are means ± SEM; letters indicate significant pairwise differences between means within each figure panel (Tukey's HSD; P < 0.05).

Figure 5.

Phosphorus addition (+P vs −P) did not alter root hyphal colonization (proportion of sites colonized) for (a) Avena fatua or (b) virus-uninfected (−virus) Bromus hordeaceus but decreased hyphal colonization for virus-infected (+virus) B. hordeaceus. Closed bars; +virus; open bars, −virus. Data shown are means ± SEM; letters indicate significant pairwise differences between means within each figure panel (Tukey's HSD;< 0.05).

Figure 6.

Elevated CO2 increased root colonization (proportion of sites colonized) for both Avena fatua and Bromus hordeaceus. Closed bars, elevated CO2; open bars, ambient CO2. Data shown are means ± SEM; letters indicate significant pairwise differences between means (Tukey's HSD; P < 0.05).

Discussion

Anthropogenic changes in abiotic resource supply, such as atmospheric CO2 concentration and soil phosphorus (P) availability, have been hypothesized to alter plant interactions with microbes (Suding et al., 2008; Compant et al., 2010; Kivlin et al., 2011). Our results support this concept, showing that alterations in resource supply can influence performance of both pathogenic and mutualistic plant-associated microbes. In turn, effects on these microbes can influence not only their host but also each other.

General ecological theory has predicted that host associations with mutualists may increase herbivore and/or pathogen populations and thus the severity of enemy damage (Bennett et al., 2006; van Dam & Heil, 2011). The stoichiometric hypothesis for virus production (Clasen & Elser, 2007) leads to a more specific prediction: the association of plants with AM fungi may increase viral concentration, because AM fungi typically increase host P content (Smith et al., 2009). Our experimental results partially supported this prediction and demonstrated that the viral titer was 20% higher in mycorrhizal than in nonmycorrhizal plants (Fig. 3a). However, AM fungi did not significantly increase host tissue P (Table S5), which suggests that the viral response did not result from an improved P supply to mycorrhizal plants. This does not completely rule out a role for P because our P data were collected at the leaf level, rather than at the level most relevant to the virus, which is restricted to the phloem (Irwin & Thresh, 1990; Jensen & D'Arcy, 1995). However, physiological mechanisms other than P transfer may be important. Our finding that AM fungi increased viral titer under elevated CO2, but not under ambient CO2 (Fig. 3a) suggests that the flow of carbon (C) may also be important in viral production (Malmstrom & Field, 1997).

Our results showing that virus infection increased mycorrhizal colonization of plants roots (Fig. 3a) indicate that viral infection increased plant resource allocation to AM fungi, particularly under elevated CO2. By the same token, elevated CO2 increased hyphal colonization of virus-infected plants more than virus-uninfected plants. Also, virus infection interacted with P addition to alter fungal performance for one plant species. Phosphorus addition decreased fungal colonization for virus-infected B. hordeaceus, but not for virus-free B. hordeaceus or for A. fatua (Fig. 5). Together, these results suggest the possibility that the virus derives a fitness benefit under elevated CO2 by stimulating its host to invest more into the mutualism. While the possible selective pressures behind this are unclear, one possible physiological mechanism involves sucrose conductance via the phloem. Typically, B/CYDVs disrupt the flow of carbohydrates, including sucrose flow through the plant (Irwin & Thresh, 1990; Jensen & D'Arcy, 1995; Malmstrom & Field, 1997), which may interfere with or induce the signaling pathways for AM fungi and P transport. However, in Avena sativa grown under elevated CO2, BYDV had the opposite effect on nocturnal reduction of total soluble sugar plus starch in leaves, and in particular virus infection increased the export, respiration, or conversion of sucrose by 30% (Malmstrom & Field, 1997). This may have both provided more carbohydrate to AM fungi, shifting the plant's C : P ratio, and triggering the plant's P starvation response, thereby stimulating greater colonization of roots by AM fungi (Smith et al., 2011).

Within host individuals, pathogen populations can be limited by nutrient supplies (Smith et al., 2005; Smith, 2007). For example, in an algal–viral system post-infection viral production was reduced in low-P host cultures, presumably as a result of insufficient intracellular P for production of P-rich viral particles (Clasen & Elser, 2007). The universally high P concentration of nucleic acids, the main component of viruses, suggests that low P concentration may similarly constrain production and titer of viruses infecting terrestrial plants. This stoichiometric hypothesis predicts that soil P amendments should increase viral titer in experimental plants if P is limiting for the virus. Effects of P amendment on the prevalence of virus infection in a field experiment were consistent with this hypothesis, although relative virus titer and leaf P concentration were not analysed (Borer et al., 2010). In the first experiment to consider the role of BYDV and leaf P concentration in wild grasses, we demonstrated that soil P amendment significantly decreased relative viral titer for A. fatua and had no effect on titer for B. hordeaceus (Fig. 3B). This result indicates that the effects of P supply on viral titer can vary among host species, perhaps depending on their physiological uptake rate or allocation of P, and that P may not be limiting for this viral complex.

In addition to altering microbial performance, changes in resources can also have direct effects on plant biomass and allocation. In a previous study, elevated CO2 increased the biomass of BYDV-infected Avena sativa more than of uninfected plants (Malmstrom & Field, 1997), suggesting that elevated CO2 counterbalances the decrease in plant C uptake caused by BYDV infection. We did not see such a counterbalancing effect in our experiment. This may be because we used two wild host species which have not been selected for agronomic yield, whereas Malmstrom & Field (1997) used A. sativa, an agricultural species which could react differently to changes in CO2 availability due to differences in evolutionary history.

Elevated CO2 and mycorrhizal fungal colonization often jointly stimulate plant growth, but such responses can vary with host-fungal species identity (Johnson et al., 2003; Klironomos et al., 2005). In our experiment, elevated CO2 increased total biomass of non AM-fungal A. fatua, but not of AM-fungal A. fatua, or of B. hordeaceus (Fig. S1). AM fungi had no net impact on total biomass of B. hordeaceus or of A. fatua plants under elevated CO2, even though elevated CO2 stimulated AM fungal colonization of both plant species (Fig. 6). This result suggests that AM fungi did not stimulate plant biomass despite increased activity as measured by hyphal colonization of roots. It also further emphasizes the variability in host species responses to mycorrhizal colonization under elevated CO2.

It is interesting to note that two model species often differed in their responses to our experimental manipulations, although they were similar in life history and growth form, as well as in serving as common hosts for mycorrhizal fungi, aphids and viruses. While such differences may be idiosyncratic, study of a larger number of host species may reveal these to be part of a broader pattern. Further study of the combined effects of abiotic and microbial drivers in such a broader ecological context may be the key to understanding and predicting large-scale changes to ecosystems (Treseder, 2004; Suding et al., 2008).

Acknowledgements

We are thankful to J. Barton at the CSTR facility for technical help. We would like to thank J. Bever and K. Vogelsang for mycorrhizal inoculum. We are grateful to the Borer-Seabloom lab for collecting and providing all plant seed, and to Jack Weiss for statistical advice. We would also like to thank the Mitchell lab for assistance. This research was partially supported by the joint NSF-NIH Ecology of Infectious Disease program through NSF Grants EF-05-25641 and DEB-10-15909 to C.E.M., an NSF GRFP to M.A.R., and the National Science Foundation Postdoctoral Research Fellowship in Biology under Grant no. DBI-12-02676 to M.A.R.

Ancillary