Mycorrhizal fungi are diverse, with strains differing in the magnitude and types of benefits provided to their hosts. In mixture, mycorrhizal strains that have complementary functions could provide their hosts with greater benefits than any constituent strain in isolation. Conversely, mycorrhizal strains could also interfere with one another, competing for host resources, providing less benefits than the most beneficial strain. The actual, realized effects are likely a composite of these complementary and competitive effects. Spatial structure in the root system may allow plants to benefit from complementary benefits while reducing competition among mycorrhizal strains. I tested the role of spatial structure using a species of plant from the family Podocarpaceae, which host arbuscular mycorrhizal fungi inside both finely separated nodules and longer, contiguous sections of fine root. In root nodules, co-inoculation with both growth-promoting and non-growth-promoting arbuscular mycorrhizal fungi resulted in complementary increases in aggregate root / fungal phosphatase activity. This complementary effect was not present in longer, contiguous sections of fine root or in either root type with P-fertilization. I dub this interaction a “White Album effect”, a reference to the Beatles’ best-selling album, where complementary effects of single artists were only revealed when individual conflicts were avoided by separating band members and synthesizing their contributions in post-production.

Growth experiment

Seeds of Podocarpus gracilior (syn. Afrocarpus gracilior; Afrocarpus falcatus gracilior), an east-African podocarp commonly grown as a hedge in tropical and Mediterranean climates, were purchased from a commercial nursery (The Banana Tree, Oregon House, California, USA, banana-tree.com). Seeds were germinated in Metromix soil (Hummert International, Earth City, Missouri, USA) that had been autoclaved for 8 hours at 121 ºC (in two four treatments on consecutive days).

After 3 months, seedlings were transferred to pots containing a 50:50 (volume:volume) sand:soil mixture. Soil was collected from the Kankakee Sands Nature Preserve (Morocoo, Indiana, USA) and mixed at a 1:1 ratio with Indiana river sand. The soil mixture was autoclaved with the same time/temperature regime as the Metromix used as germination media. The background soil mix had a pH of 8.2, with 0.1% organic matter, 70 nitrate, 19 P, 33 K, and 102 Mg (all elemental quantities in ppm, using Mehlich-3 extraction).

Experimental seedlings were inoculated either singly or in a 50:50 mixture of Claroideoglomus candidum and/or Gigaspora margarita (deposited as NC172 and NC175, in INVAM, respectively). Single-strain inoculum had been maintained at Indiana University (Bloomington, IN, USA) in association with Sorghum bicolor. In each case, inoculate was prepared as follows. Prior to the experimental setup, single species cultures of C. candidum and G. margarita consisting of spores, mycelium, and fine root segments were air dried, chopped, and thoroughly mixed with the potting soil in a 1:5 (volume:volume) ratio for use as fungal inoculum. The total inoculate volume was the same for each treatment, such that in the mixed inoculate treatment, each species was present at half the volume as in the single-strain treatment.

Seedlings were grown over a period of 6-months in a greenhouse with daily watering. Photosynthetic active radiation was not measured on the greenhouse bench where experimental seedlings were grown; however, published figures from an experiment conducted during the same period in the same greenhouse give an average value of 1339 +/- 71 μmol m^-2 s^-1 (Zheng et al., 2015).  Seedlings were fertilized twice (2 and 12 weeks after planting) with either a no-phosphorus fertilizer (P-,Scott’s 15-0-15) or a complete mixed fertilizer (P+, Scott’s 15-15-15). Both fertilizers contained, in units % mass, 15% N (from urea) and 15% K (soluble potash), 0.02% Boron, 0.05 % Cu, 3% Fe, 0.05% Mn, and 0.05% Zn. The P+ fertilizer also contained 15% P.  Each mycorrhizal x fertilizer treatment had 8 replicates, for a total of 64 plants. 

Plant performance measurements

Plants were harvested by gently removing them from their pots and washing adhering soil particles from roots under water. Root segments for enzyme assays, and root length colonized by mycorrhizal fungi measurements were excised and weighed. The rest of the plant was dried at 80°C for one week and leaves, stems, and roots weighed separately. Dried leaf samples were sent to the Cornell Nutrient Analysis Laboratory (cnal.cals.cornell.edu) and foliar P concentration analysed via dry ash elemental analysis using inductively coupled plasma atomic emission spectrometry (ICP-AES, Spectro, Analytical Instruments). Additionally, dry combustion at 1100°C was used to measure %C and %N of foliar dry mass using non dispersive infrared and thermal conductivity detection analysis, respectively (SNC-100 Carbon and Nitrogen Analyzer, Skalar Primacs).

Mycorrhizal colonization

The % root length colonized was measured using the standard clearing and staining protocol of (McGonigle et al., 1990). Eight 2 cm nodulated root segments (technical replicates) were cleared in KOH and stained with Trypan blue. Mycorrhizal colonization was scored as “present” if it appeared in at least one technical replicate, based on 6 complete passes along the root axis for the cortex of both the fine roots and the adhering nodules.

Phosphatase enzyme assays

Immediately following plant harvest, four root tips that contained both root nodules and long, un-nodulated sections of fine root were excised from each experimental plant for use in phosphomono- and diesterase enzyme assays. Root samples contained either a 2 cm section of roots with all nodules removed (fine roots) or 10 individual root nodules. For each replicate plant and root structure, potential enzyme activity was assayed in 8 subsamples (technical replicates). Statistical analyses were conducted on the average value of up to 8-technical replicates, with exclusions made for outliers > 2 standard deviations from the mean of each experimental plant x root-type.

Root samples were added to fluorescence assay well plates along with 200 μL of acetate buffer (pH 5.0). Potential P-mono and -diesterase activity was measured fluorometrically by adding MUB-phosphate substrates to root samples. Additionally, blank wells (buffer), sample controls (buffer and root sample), reference standards (MUB standard and buffer), quench standard wells (MUB standard and root sample), and negative control wells (substrate and buffer) were run to control for quenching by the substrate and buffer alone. Samples were incubated in the dark at 23°C for 2 hrs.  At the end of incubation, 20 μL of 1 M NaOH was added to each well, raising the pH to improve fluorescence (German et al., 2011). Fluorescence was measured on a plate reader using 365 nm excitation and 450 nm emission filters. Potential enzyme activity rates for each sample were calculated as μmol MUB g⁻¹ dry root h⁻¹. Dry root mass was determined by multiplying fresh root mass to the ratio of dry to fresh root mass of one additional replicate sample that was not assayed for enzymes.

Statistical Analyses 

For all dependent variables (plant growth, biomass allocation, foliar nutrient concentrations, and nodule and fine root P-mono and di-esterase activity), ANOVA was used to partition variance among the main effects of mycorrhizal treatment (df=3), fertilization (df=1) and the interaction between mycorrhiza x fertilization (df=3) (table 1).

Post-hoc tests were used to assess differences in treatment means among mycorrhizal treatments within each fertilization treatments (least significant difference with Bonferroni correction for multiple contrasts, using the ‘agricolae’ package in R).

 Chi-square tests were used to assess the probability that mycorrhizal colonization frequency was distributed according to chance between different groups (rejecting the null-hypothesis when p<0.05). Differences in the frequency of mycorrhizal fungal colonization were tested between different root types (nodules vs. fine-roots, df=1) and mycorrhizal and fertilization treatments (df=9).

  Linear regression was used to assess the amount of variation in nodule enzyme activity that could be explained by fine root enzyme activity and, within a root group, the amount of P-diesterase activity that could be explained by variability in P-monoesterase activity. Additionally, principal component analyses were used to assess the relationship between enzyme activities, growth, and foliar nutrient concentrations.