Plants and animals house microbes that provide critical nutrients, but little is known about host control over microbial cooperation when resources are also accessed from the environment. Changes in nutrient access can challenge the host’s ability to detect and selectively reward beneficial partners, destabilizing symbiosis. Legumes acquire nitrogen from soil and from symbiosis with rhizobia, but it is unclear if extrinsic sources of nitrogen interfere with host control systems. We inoculated the legume, Lotus japonicus, with rhizobia bearing nitrogen fixation or nitrogen metabolism knockouts, and factorially varied molecular sources of nitrogen fertilizer. Lotus hosts selectively rewarded beneficial rhizobia and sanctioned non-fixing strains when extrinsic nitrogen was unavailable. Host benefits were undiminished when inoculated with rhizobia bearing nitrogen metabolism knockouts, suggesting redundancies in nitrogen provisioning systems. However, under nitrogen fertilization hosts did not discriminate between fixing and non-fixing rhizobia. Fertilized hosts formed miniaturized nodules housing limited rhizobia, divesting from symbiosis. Thus, sanctioning mechanisms rely on detection of nitrogen fixation differences among rhizobia strains and can break down in nitrogen rich environments. Nonetheless, divestment from symbiosis offers legumes robust host control, minimizing investment into rhizobia strains, irrespective of their capacity to provide benefit, when symbiosis services are not needed.

Sole nitrogen source fertilization experiment 

The L. japonicus ecotype Miyakojima MG-20 was used, inbred at the University of California Riverside to generate seeds following published protocols. The M. loti genotype MAFF303099 (hereafter MAFF) is a highly effective nitrogen fixing strain on L. japonicus. MAFF was modified to express dsRed, a red fluorescent protein visible under natural light. The near-isogenic strain STM6 (strain ID 17T02d02) was employed as a non-nitrogen fixing mutant. STM6 was generated by a signature tagged transposon inserted in the nifD gene, and was subsequently modified to express GFP. Both modified strains efficiently form nodules on L. japonicus, with MAFF providing substantial benefits to host growth while STM6 provides none. MAFF and STM6 were grown on agar plates with a modified arabinose gluconate medium (MAG). In mixed cultures, MAFF and STM6 can be easily differentiated by colony color. Other than differences in nitrogen fixation and colony color, previous work uncovered no differences between the strains in terms of their invitro growth rate or capacity to induce nodule formation under clonal conditions.

L. japonicus plants thrive under organic and inorganic sources of extrinsic nitrogen (i.e., alanine, aspartic acid, ammonium sulfate, potassium nitrate) and plants fertilized with these nitrogen sources reduce investment in nodules infected with MAFF. In this experiment, L. japonicus was exposed to fertilization treatments and rhizobia that vary in nitrogen fixation to quantify effects on rhizobia fitness and examine how sanctions traits respond under fertilization. L. japonicus seedlings were exposed to clonal treatments of nitrogen fixing MAFF and non-fixing STM6, and mixed infections of both strains. Rhizobia cell proliferation, nodule biomass, and nitrogen fixation were quantified to examine rhizobia fitness, and fitness effects in planta under different treatments.

L. japonicus seeds were surface sterilized in bleach (5% sodium hypochlorite), rinsed in sterile water, nick scarified, and planted under axenic conditions in bleach sterilized Ray-Leach SC10 pots (Stuewe & Sons, Corvallis, OR, USA) using an equal mix of coarse sand on top of a layer of fine calcined clay that was autoclave-sterilized (Pro League, Quickdry; Turface Athletics, Buffalo Grove, Illinois, USA). In a factorial experiment, plants were inoculated (i.e., on 12/21/2021) with one of four treatments including MAFF, STM6, an equal mix of both strains, and sterile ddH₂O as a control. Inoculations were 5 ml and included 5 x 10⁸ cells, except for the water controls. High cell concentrations are needed to assure consistent nodulation in sterile coarse soils, likely because of low bacterial survival during inoculation. For mixed inoculation treatments 2.5 x 10⁸ of each strain were combined, and the mixed inoculum was plated to empirically confirm the ratio of MAFF to STM6. All plants were supplemented weekly with 5 ml of Broughton & Dilworth solution, including key nutrients but lacking nitrogen. Hosts were fertilized with one of four sources of nitrogen dissolved in the Broughton & Dilworth solution, including alanine (C₃H₇NO₂, 2.64 g/L), ammonium sulfate ((NH₄)₂SO₄, 3.92 g/L), aspartic acid (C₄H₇NO₄, 3.95 g/L), or potassium nitrate (KNO₃, 3 g/L). Concentrations were set to equalize the nitrogen moiety at levels previously shown to optimize host growth in the absence of rhizobia inoculation, thus attempting to isolate the growth effect of nitrogen. Fertilization occurred weekly, starting one day before inoculation (i.e., 12/20/2021) with 2 mL of fertilizer per plant and increasing by 1 mL per week until a maximum of 5 mL was reached 3 weeks later. This final level of fertilizer was maintained until the plants were harvested.

Fourteen experimental blocks each contained a single plant replicate per each of the twenty treatments, including four inoculation treatments (i.e., MAFF, STM6, MAFF+STM6, water) crossed with five fertilization treatments (i.e., alanine, ammonium, aspartate, nitrate, no added nitrogen), totaling 280 plants. To minimize within-block variance, seedlings were assigned to blocks based on size at time of inoculation. Treatments within blocks were assigned randomly. Two blocks were harvested at three weeks post inoculation, the time when nodules first become visible on L. japonicus roots. Two additional blocks were harvested five weeks post inoculation, when most nodules have typically established. Data from early harvests were used to quantify rhizobia nodulation and growth effects over time in relation to fertilization treatments. The remaining ten blocks were harvested at week seven post inoculation to quantify root and shoot biomass, nodule count and nodule biomass, to quantify rhizobia fitness in planta (i.e., rhizobia population size in nodules), and to test for evidence of host sanctions (i.e., overrepresentation of the nitrogen fixing strain MAFF in the nodules of coinoculated plants). The mixed inoculum was plated to estimate the initial ratio of MAFF:STM6 by distinguishing colony color.

During harvests, plants were removed from pots, washed free of soil, and dissected into root, shoot, and nodule portions. Nodules were counted and photographed. Within each inoculated treatment, four plants were randomly selected for quantitative culturing, wherein two nodules each were randomly sampled, surface sterilized in bleach for 30 seconds followed by 3 rinses in autoclave-sterilized ddH₂O. Nodules were then crushed in 200µl of sterile ddH₂O and the slurry was serially diluted to 0.5 x 10^-6 and 0.5 x 10^-8 on MAG-agar plates. Colony counts were averaged across at least two plate replicates per nodule to estimate rhizobia population size within each nodule. The ratio of MAFF:STM6 colonies was estimated from nodules plated from coinoculated plants and compared to the MAFF:STM6 ratio in the initial inoculum. Roots, shoots, and remaining nodules were weighed for biomass after being oven dried (> 3 days, 60°C). Ten plants were removed from the analysis, 5 due to mortality, 5 due to contamination.

Leaf ∂¹⁵N ‘atom percent difference’ was estimated as the percentage of ¹⁵N atoms over total nitrogen in each sample and the values used as a measure of nitrogen fixation. Only plants harvested at seven weeks post inoculation were used. Individual leaves from each plant were oven dried, weighed, and placed in tin capsules for isotopic analysis at the UC Davis Stable Isotope Facility.

Nitrogen metabolism knockout experiment

Rhizobia strains with individual gene knockouts were inoculated onto L. japonicus MG-20 plants to test for changes in nodulation or host benefits. M. loti strains were selected from a large collection of single gene mutants, focusing on knockouts of genes whose loss is predicted to disrupt the symbiotic transfer of nitrogen from rhizobia to the host. No nitrogen fertilizer was used so we could examine effects of these mutations when hosts have no other nitrogen source. Five M. loti strains were tested, including MAFF as a control, and four near isogenic, signature-tagged mutants. M. loti 03T03g02 has a knockout for a periplasmic amino acid ABC transporter (mll3861), a protein similar to one in Rhizobium that transfers alanine or aspartate to the host. M. loti 27T02f13, has a knockout for L-2,4-diaminobutyric acid transaminase (mlr5943), an enzyme involved in alanine synthesis that is upregulated in nodules compared to growth on media, suggesting importance for nitrogen transfer to the host during symbiosis. The last two mutants carry knockouts for paralogs of aspartate aminotransferase, which produces aspartate from oxaloacetate, predicted to be a mechanism to generate nitrogen usable to the host. M. loti 8T11d03 carries a knockout for mlr5883, encoded on the symbiosis island, and whose expression is upregulated in planta during symbiosis, whereas M. loti 20T04d01 has a knockout for the other copy, mlr2541, which is encoded on the bacterial chromosome, and is constitutively expressed. None of the mutants have been previously tested on L. japonicus.

Fifteen replicate plants per treatment were singly inoculated with one of the mutant strains, with MAFF, or with sterile water, to examine nodulation traits and growth effects, totaling 90 plants. Plants were prepared, inoculated, and maintained as outlined in the fertilization experiment, with the modification of only receiving the nitrogen-free Broughton & Dilworth (micronutrient solution). Plants were harvested at seven weeks after inoculation to quantify root and shoot biomass, nodule count, and nodule biomass.

Symbiosis trait measurements

Host biomass was used to estimate host growth in different treatments. Nodulation was used to estimate host investment into symbiosis, quantified as the number of nodules formed, their mean biomass, or total nodule biomass per plant. Host investment into symbiosis was also calculated by dividing the dry nodule biomass of each plant over the total plant biomass. Rhizobia population size in nodules, a proxy of rhizobia fitness in planta, was calculated by quantitative culturing of nodules of individual plants. The ratio of MAFF:STM6 colonies cultured from coinoculated plants was used to analyze relative fitness of the beneficial strain and to test for sanctions. Nitrogen fixation was measured using the abundance of ¹⁵N in the plant shoots from different treatments.

Data analysis

Statistical analyses were conducted using linear mixed models in JMP (Ver. 16.0.0). All data distributions were tested for normality using the Shapiro-Wilk test, as this is a prerequisite for parametric statistics. Data were transformed when necessary to achieve normal distributions to satisfy this requirement. In some cases, log or square root transformations were used. Full models, transformations, and random versus fixed factors are all listed (Data Supplement). Tukey’s HSD tests were used to test for significant differences among groups in a post hoc manner, after observing a significant effect. The ∂¹⁵N of each sample was calculated by comparing ¹⁵N abundance expressed as parts per thousand to atmospheric N₂. The values were then used to compare values of ∂¹⁵N between treatments by subtracting the mean sample atom % ¹⁵N of uninoculated plants from the sample atom % ¹⁵N of inoculated plants. When plants incorporate fixed nitrogen, leaf tissues exhibit a decrease in ¹⁵N (compared to uninfected plants) due to isotopic fractionation by rhizobia. To test for sanctions, a chi-square test was used to test whether MAFF was overrepresented in nodules of coinoculated plants, relative to the proportion of MAFF in the inoculum.