1. Plants face a constant struggle to acquire nutrients and defend themselves against herbivores. Mycorrhizae are fungal mutualists that provide nutrients that can increase plant growth and alter resistance to herbivores. The beneficial effects of mycorrhizae for nutrient acquisition can depend on the quantity and type of soil nutrients available, with plants usually benefiting more in terms of growth from mycorrhizae when nutrients are limited. However, it is unclear how the addition of different nutrients might shift mycorrhizal conferred resistance to herbivores by changing defensive secondary chemistry and nutrient availability.

2. We conducted two concurrent greenhouse experiments to test how three levels of fertilizer (low, medium, and high) and three types of fertilizer (organic, organically derived, and inorganic) altered mycorrhizae-conferred resistance to herbivores in tomato plants. In addition, we looked at whether mycorrhizae-conferred resistance was driven by plant secondary metabolites or the nutrient content of the leaves.

3. Association with mycorrhizae was associated with an increase in biomass at low levels of fertilization and decreased biomass at high levels of fertilization. Interestingly, mycorrhizae increased resistance to herbivores at medium levels of fertilization, but had no effect at low and high levels of fertilization. Mycorrhizae improved resistance most strongly when plants were fertilized with a high phosphorus, organically derived fertilizer. In both experiments, increased resistance was correlated with changes in the plant’s foliar nitrogen content.

4. Synthesis and applications: Our study supports the potential for mycorrhizae to improve either crop growth or pest resistance under lower fertilizer conditions. However, mycorrhizae did not provide both growth and resistance benefits under any treatment. While mycorrhizae have the potential to benefit crops in in lower input systems, it may be challenging to maximize both growth and resistance benefits.

We conducted a 2x5 factorial greenhouse experiment to test the interacting effects of mycorrhizae and five fertilizer treatments on tomato plant resistance to herbivores.

Study system

We conducted these experiments using tomato plants var. Castlemart (Solanum lycopersicum). Tomatoes are a valuable field and greenhouse crop that associate with mycorrhizal fungi and have a range of chemical defenses against herbivores. The defensive chemistry of tomatoes has been well characterized, and protease inhibitors in particular have been identified as important for defense (Felton, Broadway, & Duffey, 1989; Shrivastava et al., 2015).

We used first instar cabbage looper (Trichoplusia ni) caterpillars in a bioassay to measure plant resistance. Cabbage loopers are generalist noctuid caterpillars that feed on a wide variety of crop plants including solanaceous and cruciferous vegetables. They were chosen because they are sensitive to changes in host plant quality. We used them in their first instar because this stage is most sensitive to plant defenses (Thaler unpublished data), and because later instar caterpillars may have developed resistance to plant defenses if they were fed on leaves (Lee & Berenbaum, 1989). These larvae were obtained from a colony maintained on artificial cabbage looper diet (Southland Products Inc.) at Cornell University for many years.

In both experiments, we used mycorrhizae extracted from soil collected at the Dilmun Hill student organic farm at Cornell University (Ithaca, NY). Diverse, field collected mixtures have been found to be more beneficial to plants than monocultures (Rowe, Brown, & Claassen, 2007; Rúa et al., 2016), and are more representative of the conditions crop plants experience in the field. To isolate mycorrhizal spores, the soil was wet sieved to remove large rocks and debris and then blended for 20 seconds using a Cuisinart™ immersion blender. The resulting liquid was passed through a series of sieves with the smallest having a pore size of 600 µm. We then used a 20 µm nylon mesh to remove excess water. The soil slurry was divided into 5 mL aliquots and resuspended in 40 mL of a 30% sucrose solution. This mixture was centrifuged in a bucket attachment at 2200 rpm for 2 minutes. The supernatant was decanted through a 30 µm mesh set over a funnel. The spores on the mesh were washed with 20 mL of DI water into a beaker.

Spores were filtered through 30 µm mesh, surface sterilized with a solution of 4% chloramine T, 0.05% Tween^® 20, 0.02% Gentamicin and 0.01% Streptomycin using the methods outlined by Mukerji et al., 2002 (page 305). We used microscopy to determine that extracted spores were clean and contained a diverse range of morphologies. The spores were resuspended in DI water such that 10 µL of water contained between 10-12 viable spores. The solution was kept suspended using a vortex during application. Two weeks after the plants germinated, half were inoculated using 100 mL of the spore solution pipetted at the base of the plant and watered down. The control plants were treated with 100 mL of DI water and also watered down. To recover soil microbes, which could be have an impact on both plant health (Berendsen et al. 2012 and references there in), mycorrhizal fungi (Desirò et al 2014), and the interaction between the two (Artursson, Finlay, & Jansson, 2006), we filtered a mixture of Lambert LM-AP potting soil and water through a 1 µm sieve, and added 20 mL of the resulting solution to each pot. We used potting soil to reduce the risk of introducing pathogenic species.

Experimental conditions

We grew 220 Castlemart tomato plants in individual 10 cm pots filled with a 1:1 sand and calcined clay media. The media was autoclaved for one hour 3 times, 24 hours apart at 121°C to sterilize it before use. The tomato seeds were surface sterilized for 15 minutes in a 15% household bleach solution and then rinsed under running water for 1 minute. The plants were grown at 34 °C and watered with 60 mL of water every 2-4 days.

We divided the plants into two concurrent experiments. In the first, we tested the effects of mycorrhizae and different amounts of inorganic fertilizer. Plants were treated with a low, medium, or high dose of inorganic 21-5-20 NPK fertilizer (Table 1). We chose this low phosphorus fertilizer to encourage association with mycorrhizal fungi. Guaranteed Analysis for all fertilizers is available in Supporting Information 1. Sixty plants were given a low dose (20 mL) of a 21-5-20 fertilizer diluted to 150 mg/L. Another 60 plants were given a medium dose (30 mL) of the same fertilizer. A third set of 60 plants were given a high dose (40 mL) of the same fertilizer. Each plant was given supplemental water such that each plant received an equal amount of water.

To test the effect of fertilizer type, we compared three types of fertilizer: an organic (n=20), organically derived (n=20), and inorganic commercially available fertilizer (n=60). Guaranteed Analysis for all fertilizers is available in Supporting Information 1. We chose three commonly used fertilizers and applied them as recommended on the label. We chose not to adjust quantity of fertilizer to equalize the N or P level because differences in macronutrient accessibility and micronutrient levels would limit our ability to link effects to a single macronutrient.

In the organically derived treatment, 20 plants were fertilized using a higher phosphorus fertilizer: Foxfarm Grow Big liquid plant food 6:4:4 diluted to 4 mL/L of fertilizer. In the organic treatment, 20 plants were fertilized with the organic, carbon-rich Alaska brand fish fertilizer 5:1:1 diluted to 14.3 mL/L water as recommended. For the inorganic treatment, we used the same plants that were given a high dose (40 mL) of the 21-5-20 fertilizer from experiment 1. We fertilized the plants grown in the 21-5-20 fertilizer once every two weeks, while the other two fertilizers were applied once every 4 weeks, to maintain a more comparable total nutrient addition.

Measurements

Mycorrhizal Colonization

To confirm mycorrhizal colonization, we bleached the roots using potassium hydroxide and stained the roots using Schiffer black ink (Vierheilig et al. 1998). Using microscopy, we assessed the roots to confirm that plants in the mycorrhizal treatment were colonized and those in the control were not. Previous work by Rutkowski et al. (unpublished) found that variation in colonization had no effect on resistance to herbivores or resistance traits. We had no accidental colonization in the control treatment.

Plant Growth

We harvested the plants two months after germination. We excised the terminal leaflet from the most recently fully extended leaf for protease inhibitor analysis and from the second most recently fully extended leaf for the bioassay. We harvested and dried the remaining leaf and stem tissue for 1 week to measure dry biomass and to analyze for carbon:nitrogen ratio (C:N ratio), a measure of the nutritional content of plants.

Resistance to herbivores

To measure herbivore performance, we excised the terminal leaflet from the third most recent fully extended leaf and placed it in a 9 cm petri dish lined with damp filter paper. We placed 2 neonate Trichoplusia ni caterpillars on each leaflet, closed the petri dish, and sealed the petri dish with parafilm. After 6 days, we measured the mass of each caterpillar. We observed that many caterpillars left the leaf and died. We recorded the number of such caterpillars as a metric of repellence. Caterpillars that died without moving or feeding were removed from analyses as we presumed that they were killed during transportation to the leaf. We also measured levels of herbivory (mm²) using a 4mm² grid to assess the quantity of leaf consumed (Coley 1982).

Using excised leaves to measure herbivory and resistance is a common approach (Kumar, Ortiz, Garrido, Poveda, & Jander, 2016; J S Thaler, Stout, Karban, & Duffey, 1996; Jennifer S. Thaler, Agrawal, & Halitschke, 2010) that prevents confounding temporal variation (Karban, 2011). However, removing the leaves may differentially induce resistance to herbivores in the mycorrhizal and non-mycorrhizal plants (Pozo and Azcón-Aguilar, 2007 and citations therein). Induced defenses are resistance traits that are deployed only once the plant has been damaged, while constitutive defenses exist regardless of damage. We are unable to differentiate whether the effects of mycorrhizae on resistance were due to changes in constitutive or induced resistance. If mycorrhizae are affecting induced, but not constitutive resistance, it may mean that herbivores will be able to damage plants in the field significantly before the defenses are deployed.

Resistance Traits

We measured plant nutritive quality using C:N ratio. The C:N ratio is indicative of both the attractiveness of a plant to herbivores and the health of a plant, with a low ratio correlated to healthier, more fertilized plants. Most herbivores are N limited (White, 1984), so plants with a low C:N ratio can be more attractive and nutritious (Behmer, 2009). To test the role of mycorrhizae and fertilizer on leaf nutrient quality and the effect of leaf nutrient quality on herbivory, we analyzed the C:N ratio of one leaf per plant. Each leaf was ground into a powder using 2.3mm an Mp Biomedical Fastprep 24. Then 5 ± 0.1 mg tissue from each leaf was balled into 4x6 mm tin capsules (Costech Analytical Technologies Inc) and analyzed using a Costech 4010 CHNS-O Analytical Combustion System.

We measured plant chemical defense by measuring protease inhibitor activity. Protease inhibitors are a class of chemical defenses that reduce the digestibility of leaf tissue by breaking down the herbivore’s digestive enzymes (Chen, Wilkerson, Kuchar, Phinney, & Howe, 2005). In tomatoes, they play a strong role in the resistance to herbivores including T. ni (Scott, Thaler, & Scott, 2010). Protease inhibitors are produced through the jasmonic acid pathway and can be used to measure expression of this pathway (Koiwa, Bressan, & Hasegawa, 1997). Mycorrhizae have been shown to alter protease inhibitor levels under different conditions (Barazani 2004). We excised the terminal leaflet from the most recent fully extended leaf and immediately froze it on dry ice. We analyzed 100mg of tissue using a colorimetric assay to calculate the activity of defensive trypsin protease inhibitors using a method adapted from Hegedus et al. (2003) (Supporting Information 2). Data points with erroneous values (>100%) were removed.