Symbiosis of soybean with nitrogen fixing bacteria affected by root lesion nematodes in a density-dependent manner

Early maturing varieties of soybean have a high yield potential in Europe, where the main biotic threat to soybean cultivation are root lesion nematodes (Pratylenchus spp.). Nitrogen fixation in root nodules by highly efficient inoculants of Bradyrhizobium japonicum is an incentive to grow soybean in low-input rotation systems. We investigated density-dependent effects of Pratylenchus penetrans on nitrogen fixation by co-inoculated B. japonicum. Less than 130 inoculated nematodes affected the number and weight of nodules, the density of viable bacteroids in nodules, and nitrogen fixation measured as concentration of ureides in leaves. With more inoculated nematodes, the percentage that invaded the roots increased, and adverse effects on the symbiosis accelerated, leading to non-functional nodules at 4,000 and more nematodes. When P. penetrans invaded roots that had fully established nodules, growth of nodules, density of bacteroids, and nitrogen fixation were affected but not the number of nodules. In contrast, nodulation of already infested roots resulted in a high number of small nodules with decreased densities of bacteroids and nitrogen fixation. P. penetrans invaded and damaged the nodules locally, but they also significantly affected the nodule symbiosis by a plant-mediated mechanism, as shown in an experiment with split-root systems.

experiments were performed with high numbers of infective stages of the nematodes that do not reflect normal field densities. In addition, the methods for the measurement of nitrogenase activity by the acetylene reduction assay that were applied in these studies can be highly inaccurate in predicting the rate of nitrogen fixation in nodules 15,16 . Thus, the density-dependent effects of root-lesion nematodes on the symbiosis of soybean plants with inoculants of B. japonicum are still unclear, and whether the nodule formation or the functioning of established nodules is affected by the nematodes. Systemic control of nodulation by the plant 17 suggested that the intimate relationship is affected by systemic changes in the plant that are induced by root invasion of lesion nematodes. However, it was occasionally observed that some plant parasitic nematodes fed and reproduced in nodules 18,19 , which suggested a local mechanism of damage to the nodules.
The objectives of this study were to investigate whether the root lesion nematode P. penetrans affects the symbiosis of soybean plants with the nitrogen-fixing bacteria, and how this depends on the population density of the nematode in soil and in the roots. We further studied the disturbance of the established symbiosis by invading nematodes, and the effect of lesion nematodes in the roots on nodule formation. We determined as a measure for nitrogen fixation rate the concentration of ureides in the leaves. Ureides are synthesized by soybean plants from fixed nitrogen for transport from the nodules to the shoot 20 . As we observed migration of the nematodes into the nodules, we investigated whether the detrimental effect of the nematodes on nodule functioning is caused by the local damage of the nodules, or by a systemic mechanism.

Results
Density dependent effects of P. penetrans on rhizobial nodulation. The roots of soybean were co-inoculated with B. japonicum and various numbers of P. penetrans to investigate density dependent effects of the nematodes on nodulation, symbiotic N 2 fixation, and plant growth. The nematodes invaded the roots in a density dependent manner (Fig. 1). The percentage of inoculated nematodes that entered the roots slightly decreased from 14% with 60 inoculated nematodes to 10% with 375. There was a sharp increase above around 500 inoculated nematodes. With increasing numbers of nematodes in soil, the percentage of invading nematodes rapidly increased to 39%. Nematode counts in roots remained above 31% from 2,000 to 6,000 inoculated nematodes, despite the negative bias of counts at high density in roots. Increasing numbers of lesion nematodes progressively affected the number and weight of nodules per soybean plant, the number of viable bacteroids in the nodules, and the amount of fixed nitrogen measured as concentration of ureides in the leaves (Fig. 2). These negative effects, compared to the control plants without inoculated nematodes, started with about 250 infective nematodes. With 4,000 inoculated nematodes, hardly any viable bacteroids were detected and nodules developed only in small number and size. Ureides in the shoots (allantoin and allantoic acid) were near or even below detection limit, showing that no fixed nitrogen was transported to the leaves at these high infection rates. As a consequence of the Suspensions of infective stages of P. penetrans were equally inoculated around soybean plants into four holes of 5 cm depth, and invaded nematodes were stained and counted microscopically in the roots after two weeks. The numbers and percentages of nematodes that invaded the root system are shown relative to the number of nematodes that were inoculated to the plant two weeks before. nitrogen limitation, the shoot weight declined with increasing numbers of inoculated nematodes (Fig. 3A). The root reacted to low infestation by P. penetrans with slightly increased growth, while 3,000 and more inoculated nematodes affected the root weight (Fig. 3B). To estimate the threshold densities of inoculated P. penetrans above which the symbiosis was affected, the data were fit to the Seinhorst equation (Table 1). The shoot and root weight was affected above 250 or 2,000 nematodes, respectively. However, bacteroids, nodules, or ureides were affected at much lower densities above 0 to 130 inoculated P. penetrans.

Effect of P. penetrans on already nodulated soybean plants. To investigate how an established B.
japonicum-soybean symbiosis will be affected by invading lesion nematodes, soybean plants were first inoculated with B. japonicum and allowed to form nodules. Two weeks later, the nodulated plants were infected with 1,000 P. penetrans each. The effect of the root invasion of P. penetrans on the nodules and N 2 fixation was analyzed two and five weeks after incubation of P. penetrans. The number of nodules did not significantly differ between the treatment with or without P. penetrans, and it did not change significantly over time (Fig. 4A). The mass of the nodules increased over time in both treatments, but was significantly greater in the non-infested control at both time points (Fig. 4B). The density of viable bacteroids in nodules increased over time and was significantly affected by P. penetrans two and five weeks after inoculation (Fig. 4C). Concomitantly, the transport forms of fixed nitrogen, the ureides, increased in concentration in the leaves over time. The concentration of ureides was significantly affected by P. penetrans at the 5-weeks sampling, and this trend was also visible at the 2-weeks sampling (Fig. 4D). In this experiment, the duration of treatments was not long enough to lead to a clear effect of P. penetrans on the growth of nodulated soybean plants (Supplementary Table S1). The shoot fresh weight was significantly decreased two weeks after incubation of P. penetrans compared to non-infected plants, and root weights showed the same trend.
Effect of established P. penetrans on nodulation of soybean roots. To investigate how an established infestation with P. penetrans affects the nodule formation and function by B. japonicum in the soybean roots, the nitrogen fixing bacteria were inoculated to soybean roots that were already colonized by P. penetrans, or non-infected roots. Surprisingly, the infested roots formed a significantly higher number of nodules than the non-infested roots in the two weeks after inoculation of B. japonicum (Fig. 5A). Five weeks after inoculation of B. www.nature.com/scientificreports www.nature.com/scientificreports/ japonicum, the number of nodules even increased to an extreme value of 128 nodules on average per infested root, while the control plants had an average number of 51 nodules per root. The size of the nodules and the distribution across the root hairs clearly differed between the pre-infected and control roots. In the pre-infected soybean roots, where P. penetrans caused distributed lesions within the roots, the size of nodules was very small and clustered abnormally. In contrast, the nodules were large and well distributed across the root hairs in non-infested soybean roots. The higher number of nodules formed on the pre-infected roots was reflected by a slight trend for increased nodule mass, two and five weeks after inoculation of B. japonicum (Fig. 5B). However, the pre-infection of soybean roots with nematodes led to a significant decrease of viable bacteroids in nodules compared to the non-infested roots, two and especially five weeks after inoculation of B. japonicum (Fig. 5C). Consequently, ureide concentrations in leaves of infested plants were significantly lower than in leaves of non-infested plants at both samplings. Plant growth parameters did not significantly differ in this experiment where plants were grown with fertilizer before inoculation of the nodulating bacteria (Supplementary Table S2 Table 1. Estimation of threshold densities (T) of inoculated Pratylenchus penetrans (P i ) above which the measured parameter of soybean plants or nodules was affected, using best fits to the Seinhorst equation a . a y = y m m + y m (1 − m) z (Pi -T) , for P i > T; y = y m , for P i ≤ T; y m = y(P i < T); y m m = y(high P i ); z = slopedetermining parameter with 0.997 ≤ z < 1. (2020) 10:1619 | https://doi.org/10.1038/s41598-020-58546-x www.nature.com/scientificreports www.nature.com/scientificreports/ shows the change in root and nodule morphology caused by P. penetrans. Rhizobial colonization of soybean roots, measured as number of nodules per root mass, was strongly affected by the number of inoculated root lesion nematodes ( Supplementary Fig. S2).
Plant-mediated effect of P. penetrans on the B. japonicum -soybean symbiosis analyzed in split-root systems. Microscopic analysis of nodules and areas near to nodules showed that P. penetrans invaded the nodules and the area close to nodules (Fig. 6A). They laid their eggs in such zones, suggesting that nodulation zones are a favorable habitat for these nematodes (Fig. 6B). The presence of P. penetrans within and close to nodules resulted in brown lesions within nodules and in the cortex and phloem near nodules (Fig. 6C,D). This looked like a direct effect of the nematodes on the functioning of the nodules. Therefore, we tested in a split-root system whether the effect of P. penetrans on the bacteria-plant symbiosis was partially plant-mediated by spatially separating nematodes from nodules. One side of the root system was either inoculated with P. penetrans, or kept as non-inoculated control. Two weeks later, the other half of the root system was inoculated with B. japonicum, and the roots were sampled after three weeks. As in the previous experiment, the number of nodules was significantly increased in roots of infested plants compared to the non-infested control (Fig. 7A). Nematodes were not able to migrate to the nodules in the other root system. The mass of nodules and the density of bacteroids was significantly decreased in infested plants compared to the non-infested control (Fig. 7B,C).

Discussion
Density dependent effects of P. penetrans on rhizobial nodulation. In this study, root invasion by P. penetrans affected the formation and development of nodules and thereby the N 2 fixation by the B. japonicum-soybean symbiosis in a density-dependent manner. The effect on nodulation and N 2 fixation was tested with a range of densities of lesion nematodes that reflected densities observed in the field, as well as with densities of 4,000 and more infective stages per plant that are unusually high at least during the early developmental period of soybean plants 12 . Other studies reporting effects of P. penetrans 14 or Heterodera glycines 21 on soybean nodules only applied very high densities that were largely above damage thresholds. This extreme number of nematodes resulted in severe effects on nodulation, nitrogen fixation and plant growth. In general, these early findings are in line with our results. However, the effect of P. penetrans on nitrogen fixation was not clearly shown in the earlier studies. The applied acetylene-ethylene assay suggested only a 19% reduction in the nitrogen fixing capacity of nodules compared to a control without P. penetrans in a phytotron experiment, and no reduction in a greenhouse experiment 14 . This underestimation of the effects of P. penetrans can be explained by inaccuracy of the assay as it was applied, and because the capacity of nitrogenase activity measured by this assay does not well reflect the amount of fixed nitrogen 15,16 . The ureides allantoin and allantoic acid are synthesized in the nodules of of tropical legumes, such as soybean and Phaseolus vulgaris, while asparagine and glutamine are the final products of nitrogen fixation in temperate legumes, such as pea and Faba bean 20 . In our study, we used a differential-colorimetric  www.nature.com/scientificreports www.nature.com/scientificreports/ method 22 to determine allantoin and allantoic acid that were synthesized with fixed atmospheric nitrogen in the nodules, and transported to the shoot tissues 23,24 .
The nodulation traits like nodule numbers and mass were negatively correlated with the density of P. penetrans. This might be explained by the competition between parasitic nematodes and rhizobial cells for resources provided by the host plant for the establishment of mutualism 25,26 . The soybean roots infected with higher densities of P. penetrans had significantly reduced nitrogen-fixing activity in the nodules as indicated by lower densities of bacteroids and less production of ureides compared to roots infected with lower numbers of nematodes, or uninfected plants. Migration of root lesion nematodes through the roots resulted in a destructive damage of root cells. The basal defense of the plant led to further damage of the root tissue. This might have interrupted the flux and delivery of nutrients that the plant provides to support the bacterial symbiont 27,28 . In addition, pathogen induced defense can negatively affect the mutualistic relationship as shown for both rhizobia and arbuscular mycorrhizal fungi 29 .
Our data demonstrated that high densities of P. penetrans reduced the shoot dry mass. The root reacted to low infestation by additional growth at low infection rates, as the plant probably tries to compensate for damaged tissue to sustain nutrient and water uptake. High numbers of 1,000 and more infective stages invading the root significantly affected the root weight. It is apparent that the relation between density and plant growth is complicated 30 . In some cases, damage by nematodes to the root does not result in the reduction of shoot weight because plants may have more roots than needed to support the shoot, or plants compensate the damage to roots by building more side roots. However, their potential to compensate the damage is determined by the availability of nutrients.

Effect of nematodes on the maturation of nodules.
We showed that the order of establishment and colonization of roots by P. penetrans and B. japonicum largely determined the effects on the nodules. After initial colonization of roots by rhizobial cells and establishment of the nodules, P. penetrans infection had no effect on the number of nodules compared to non infected roots. Probably the acquisition of B. japonicum by the roots had already been completed. However, the further maturation of the nodules was impaired, as indicated by the slower increase in the weight of nodules and density of bacteroids within the nodules. Both, mutualistic and pathogenic partners induce significant changes in phytohormone levels 31 . With rhizobia, this leads to transportation of the fixed nitrogen and in turn enhances the flux of carbon and amino acids to the bacteroids 32 . This nutrient fluxes can be affected by complex changes in the vascular tissues, sugars transport and phytohormones regulation, that in turn affect the growth of nodules and survival of bacteroids 33,34 . In addition, the invasion of nematodes can affect the viability and differentiation of bacteroids by reducing the availability of leghemoglobin, which regulates the supply of oxygen to protect the rhizobial nitrogenase. Accordingly, the concentration of ureides in the leaves was significantly lower in plants infested by P. penetrans.
Interference of lesion nematodes with the regulation of nodulation. Interestingly, P. penetrans that was already established in roots had a very severe impact on nodulation by B. japonicum. A dramatically higher number of nodules developed on infested roots compared to the non-infested roots. This hyper-nodulation resulted in small-sized and aggregated nodules containing low densities of bacteroids. The majority of these clustered small nodules were formed near to the damaged crown of the roots. Nodulation of legume roots is a resource demanding process and therefore tightly regulated 17 . This autoregulation seemed to be disturbed in the infested roots. The nodulation is regulated through systemic mechanisms where the plant coordinates the nodulation and suppresses further formation of nodules by signaling from shoot to root and back again 6,35 . We showed in the split-root experiment that the nematodes had a systemic effect on the nodulation, but local damage to the nodules was also observed. Lesions formed by P. penetrans near the nodules releases extracellular ATP, which acts as a damage-associated molecular patterns but likely plays a role in the regulation of nodulation as well 6 , which might result in a local interference of P. penetrans and the soybean -B. japonicum symbiosis. Another key to this interaction might be cytokinin, because cytokinin signaling is important for the control of nodulation by legumes 7 , and cytokinin was also reported to play a role in the parasitism of root knot and cyst nematodes, which produce cytokinin and manipulate cytokinin signaling of the host plant 36 . A local excess of cytokinin likely leads to hyper-nodulation because it acts as an endogenous inducer of nodule primordia formation 7 . However, cytokinin production was not yet shown for root lesion nematodes 27 . In our study, aggregated hyper-nodulation was observed on locally infested roots but much less in the split-root system where P. penetrans was spatially separated from nodules. A similar effect was achieved by removal of nodules, which was suggested as an indication for a locally generated signal of autoregulation of nodulation 37 .The damage of nodules by the migrating nematodes might result in a new round of nodulation by the same mechanism.
It was previously reported that the number of nodules per plant was not well correlated with the total mass of nodules per plant and the amount of fixed nitrogen 38 . It seems that the plant initiated more nodulation processes but the nodules did not become mature. In a Phaseolus vulgaris -Rhizobium leguminosarum symbiosis, the number of nodules was also systemically affected by a fungal leaf pathogen that was inoculated in the initial phase of nodule formation by the nitrogen fixing symbiont, but in contrast to our study the number of nodules was significantly reduced 29 . The fungal pathogen systemically induced higher activity of polyphenol oxidases in the roots, suggesting that plant defense responses interfered with nodulation. In another study, chemically induced defense pathways in soybean systemically reduced the number of nodules and the nitrogen content in leaves and roots after 51 days 39 . The chemical inducer was applied once as foliar spray 5 days after inoculation of seedlings with B. japonicum, while P. penetrans could continuously affect the soybean -B. japonicum symbiosis. In our study, the small sized nodules from pre-infested roots had very low densities of bacteroids compared to those recovered from healthy roots. Limitation of O 2 supply for the bacteroids in the nodules as a response of the plant Scientific RepoRtS | (2020) 10:1619 | https://doi.org/10.1038/s41598-020-58546-x www.nature.com/scientificreports www.nature.com/scientificreports/ to stress caused by the nematodes could explain the reduced viability of bacteroids 40 . This was also evidenced by a significant decline in the ureides amount in the infected plants.
In sum, results contribute to a better understanding of the interaction between soybean plants, nitrogen fixing bacteria and Pratylenchus that might lead to strategies to improve N 2 fixation under pathogen pressure. Early nodulation before the nematode population builds up, and control of population densities of root lesion nematodes are important for harnessing the positive effects of symbiotic nitrogen fixation. In the future, it will be of interest to expand our investigation to explore the spatial distribution of plant parasitic nematodes within the nodulated roots and to understand the reasons behind the preference of nematodes to migrate and propagate near to nodule formation zones.

Materials and Methods
Soybean cultivation. Seeds from soybean cv. Primus were surface sterilized in 1.5% sodium hypochlorite for 15 min and rinsed with sterile distilled water. Seeds were transferred on moist sterilized filter paper in Petri dishes placed in the dark at 20 °C for 4 days to allow germination. Uniformly germinated seedlings were selected and transplanted in 12 cm diameter plastic pots filled with 500 ml growth substrate. Except for the split-root experiment, the growth substrate consisted of two volumes of sand and one volume of a low nitrogen content field soil (7.2 kg N min per ha, loamy sand, braunerde, pH 6.5, 52°17′57″N/10°26′14″E) 41 . Plants were maintained in the greenhouse at 24 °C and 16 h photoperiod and watered as needed every 2-3 days.
B. japonicum inoculum. An isolate of strain B. japonicum 532C was recovered from HiStick Soy (BASF, Ludwigshafen, Germany) by plating the product on a yeast mannitol (YM) agar supplemented with 1 mg/l vancomycin. To support the axenic recovery of the strain from nodules, a rifampicin resistant mutant was isolated. This was done by plating a high-density culture on YM supplemented with 1 mg/ml vancomycin 42 and 50 mg/l rifampicin, and purifying streaks of a single rifampicin-resistant colony on these agar plates. The B. japonicum strain was cultured for 3 days at 28 °C on YM liquid medium supplemented with rifampicin. The B. japonicum cells were spun down at 4,000 g for 10 minutes, and the pellet was washed twice with sterile tap water to remove antibiotic residues, resuspended in sterile tap water and adjusted to OD 600 = 0.2. In the greenhouse experiments, each pot was inoculated with 4 ml of the B. japonicum cell suspension by adding it to four holes of 5 cm depth around the soybean root. P. penetrans inoculum. The root lesion nematode P. penetrans was extracted from 4 months old carrot disks by the Baermann funnel technique 43 . The densities of nematodes in 1 ml of the suspensions were microscopically determined in a nematode-counting slide. The nematode suspension of mixed infective stages of P. penetrans was equally inoculated around soybean plants into four holes of 5 cm depth. The number of P. penetrans inoculated differed based on the experimental layout.
Greenhouse experiment on density-dependent effects of P. penetrans on nitrogen fixation by co-inoculated B. japonicum. Inocula of P. penetrans were prepared by serial dilution of three independently extracted suspensions of the nematodes. One week after germination, soybean plants were infected with 0, 0, 63, 125, 125, 188, 250, 250, 375, 500, 500, 750, 1000, 1000, 1500, 2000, 2000, 3000, 4000, or 6000 individuals of P. penetrans. Each of the 20 plants was co-inoculated with B. japonicum and randomly distributed in a greenhouse chamber. One month after inoculation, the root weight, shoot weight, concentration of ureides in leaves, number of nematodes in roots, and number and weight of nodules were determined.
Greenhouse experiment on effects of P. penetrans on established nodules. Twelve-day-old soybean seedlings were inoculated with B. japonicum and kept for two weeks until nodules have been established. Sixteen nodulated plants were inoculated with 1,000 infective stages of P. penetrans. Sixteen control plants were not infected with nematodes. Plants were randomly distributed in a greenhouse chamber. Two and five weeks after inoculation, eight plants per treatment and time point were sampled to determine the root weight, shoot weight, ureide concentration in leaves, number of nematodes in roots, and number and weight of nodules.
Greenhouse experiment on effects of established P. penetrans on nodulation. This experiment was implemented to investigate how the process of nodulation and N 2 fixation is affected when the roots are infested by P. penetrans before inoculation of B. japonicum. Sixteen 12-day-old soybean seedlings were infected with 1,000 P. penetrans. Sixteen control plants were not infected with nematodes. Plants were randomly distributed in a greenhouse chamber. Two weeks later, all pots were inoculated with B. japonicum. Eight plants per treatment and time point were sampled two weeks and five weeks after inoculation with B. japonicum to determine the root weight, shoot weight, ureide concentration in leaves, number of nematodes in roots, and number and weight of nodules.
Split-root experiment on plant-mediated effects of P. penetrans on the nodulation by B. japonicum-soybean symbiosis. The root of twelve two-week-old soybean seedlings was split and transplanted into two adjacent square pots (7 cm × 7 cm) filled with two times autoclaved sand. For half of the plants, the split-root system in one pot was inoculated with 500 P. penetrans, the other plants were non-inoculated controls. After two weeks, one-half of each root system without nematodes was inoculated with 2 ml of a B. japonicum suspension (OD 600 = 0.2). The plants were watered and fertilized by a mineral solution 44 supplemented with 0.125 mM NH 4 NO 3 . Three weeks after inoculation of B. japonicum, the plants were sampled to determine the root weight, shoot weight, number of nematodes in roots, and number and weight of nodules.