Understanding the impact of the peaola microbiome on soil fertility, crop yield, and plant nitrogen content

Results

Land Equivalent Ratio and Yield Analysis

For our first objective, we found that the cropping system had a significant impact on the yield of pea and canola, and on the LER. Across the trials grown at the WSU research farm in 2021 and the trials grown near Colfax, WA in 2020, it was found that the average LER was 1.63 for peaola compared to the normalized value of 1 for the monocultures. The application of N fertilizers had no significant impact on the yield of peas and canola, or the LER for both of these locations. For the strip trials grown in Colfax, WA, the LER for peaola, at 1.37, was not significantly different from the monocultures. It was found that the yield of canola was not significantly different from when it was grown in monoculture, but the yield of pea was significantly reduced compared to when it was grown in monoculture at the Colfax, WA location. At the WSU research farm there was a significant increase in the LER of peaola from monoculture. However, there was a significant reduction in pea and canola yields from monoculture. At the Greene farm in 2021, no yield data was collected due to poor performance of the peas, so it was considered a loss. LERs were not calculated for the crops planted in 2022 at the WSU research farm and at the Greene farm as crops at the WSU research farm location were damaged significantly by deer and yield data was not collected at the Greene farm due to Dr. Isaac Madsen's move to Africa. The LERs that were recorded do match the range that has been recorded for peaola in the literature at low N application rates (Dowling et al., 2021).

Microbial Community Analysis

To accomplish our second, third, and fourth objectives, we have collected 264 soil samples, 459 root and rhizosphere samples, and 333 plant shoots across 3 different locations from 2020 to 2022. Substantial delays were experienced in regard to accomplishing our second and third objectives due to there being initial issues with our DNA extraction and quality control pipeline. As a result, we are slightly behind schedule in regard to our sequencing timeline, but we are quickly getting back on schedule. To date, we have extracted DNA from all of the soil and rhizosphere samples we collected in 2020, 2021, and 2022. Extractions have begun on the root samples DNA with sequencing expected to happen by August. Sequencing has been done on all of the soil and rhizosphere samples collected in 2020, 2021, and 2022. Some of these samples are still actively being sequenced by the Michigan State University Genomics Core with results expected to be back by late August. Due to delays in our optimization of our lab's protocol of the SmartChip Real-Time PCR system, we have not performed qPCR for the nifH gene on all of the samples. Currently, work is being done to test a small batch of samples within the system. Once this test has been completed and is successful, all extracted samples will be put through the protocol. Network analysis has been successfully done on the sequences for our 2020 samples. Stable isotope analysis for C and N percentages has been performed on the leaf tissue of the shoot samples collected in 2020 and 2022. Analysis of these results has been completed for both years. All results can be found bellow.

Data analysis has been done on the 2020, 2021, and 2022 soil and rhizosphere sequences. Through the inclusion of a microbial community standard, we determined we were able to detect bacteria at the appropriate abundance down to a relative abundance of 0.089%. When analyzing the 2020 soil using Shannon diversity index, Observed Features, and Evenness (measures of α-diversity) we did not find any significant differences (P < 0.05) or trends (0.05 ≤ P ≤ 0.1) between pea monoculture, canola monoculture, and peaola. We did find a trend towards the monoculture pea soil being richer than the monoculture canola soil using Faith's Phylogenetic Diversity index (Kruskal–Wallis Test, n1 = 12, n2 = 12, H = 5.60, p = 0.0537) (Figure 1). No other significant differences or trends were found using Faith's Phylogenetic Diversity index. For the 2020 rhizosphere, we found that the bacterial community in pea is significantly more even than canola (P < 0.05) and that there was a trend towards it being more even than intercropped canola (0.05 ≤ P ≤ 0.1) (Figure 2).

When looking at the β-diversity of the 2020 soil using Jaccard distance, Bray-Curtis distance, and unweighted UniFrac distance, we did not find any significant differences (P < 0.05) or trends (0.05 ≤ P ≤ 0.1) between pea monoculture, canola monoculture, and peaola. Under weighted UniFrac distance, we found that between pea and canola monoculture soils, and between peaola and canola monoculture soils, there was a trend toward there being a difference in the community composition (PERMANOVA, F = 2.06, P = 0.0705; PERMANOVA, F = 1.72, P = 0.0705) (Figure 3). This was not observed between pea monoculture and peaola soils. When looking at the 2020 rhizosphere, we found a significant difference in the β-diversity between canola and pea rhizosphere regardless of cropping system (PERMANOVA, P<0.05), and trends between like intercropped and monoculture crops (PERMANOVA, 0.05≤P≤0.1) based on Bray-Curtis distance (Figure 3). Similar observations were found using other measurements of β-diversity.

Analysis of the 2021 soil samples showed that there are no significant differences across all measure of α-diversity and β-diversity between the different cropping systems and treatment groups (Figure 5). For the 2021 rhizosphere samples, it was found that pea with a fertilizer treatment of 0 kg N ha^-1 and 33 kg P ha^-1 is significantly (P < 0.05) more diverse than intercropped canola with the same fertilizer treatment and a fertilizer treatment of 67 kg N ha^-1 and 0 kg P ha^-1 (Figure 4). Pea with a fertilizer treatment of 0 kg N ha^-1 and 33 kg P ha^-1 trended (0.05 ≤ P ≤ 0.1) towards being more diverse than canola with a fertilizer treatment of 67 kg N ha^-1 and 33 kg P ha^-1 (Figure 4). The same pattern was seen for intercropped pea with a fertilizer treatment of 67 kg N ha^-1 and 0 kg P ha^-1, but it was significantly (P < 0.05) more diverse than canola with a fertilizer treatment of 67 kg N ha^-1 and 33 kg P ha^-1 (Figure 4).

Using Bray-Curtis distance, it was found that there was a significant difference (P < 0.05) between most pea and canola samples regardless of treatment or cropping system (Figure 5). However, intercropped canola with a fertilizer treatment of 33 kg N ha^-1 and 0 kg P ha^-1 is not significantly different (P > 0.05) from pea with no fertilizer and intercropped pea with fertilizer treatments of 33 kg N ha^-1 and 0 kg P ha^-1 and 67 kg N ha^-1 and 33 kg P ha^-1 (Figure 5). There was also a trend (0.05 ≤ P ≤ 0.1) towards it being significantly different from intercropped pea with a fertilizer treatment of 33 kg N ha^-1 and 33 kg P ha^-1 and pea with a fertilizer treatment of 0 kg N ha^-1 and 33 kg P ha^-1 (Figure 5). Intercropped canola with a fertilizer treatment of 67 kg N ha^-1 and 33 kg P ha^-1 was also found to not be significantly (P > 0.05) different from pea with no fertilizer and intercropped pea with a fertilizer treatment of 67 kg N ha^-1 and 33 kg P ha^-1 (Figure 5). Some significant differences (P > 0.05) were also observed between like crops. These were seen between intercropped canola with a fertilizer treatment of 67 kg N ha^-1 and 33 kg P ha^-1 and canola with the same fertilizer treatment and intercropped canola with the fertilizer treatments of 0 kg N ha^-1 and 33 kg P ha^-1 and 67 kg N ha^-1 and 0 kg P ha^-1 (Figure 5). A trend (0.05 ≤ P ≤ 0.1) towards intercropped pea with a fertilizer treatment of 0 kg N ha^-1 and 33 kg P ha^-1 and 67 kg N ha^-1 and 0 kg P ha^-1 was found (Figure 5).

Similar to the results we saw in 2021, we saw no significant differences in the measures of α-diversity and β-diversity for the 2022 soil samples across the different cropping systems and treatments (Figure 7). When looking at the results for α-diversity in the 2022 rhizosphere samples there were significant differences (P < 0.05) and trends (0.05 ≤ P ≤ 0.1) across different cropping systems and seeding treatments using Faith's Phylogenetic Diversity index (Table 2 and Figure 6). Significant differences (P < 0.05) also existed between canola 2 and 8 and canola 8 and null Figure 6. Trends (0.05 ≤ P ≤ 0.1) existed between canola treatments 1 and 8, canola 2 and 3, canola 2 and 6, canola 4 and 8, canola 5 and 8, and pea 1 and 4 (Figure 6).

Table 2. Results for Faith's Phylogenetic Diversity index, with yes indicating a significant difference (P < 0.05) and trend indicating a trend (0.05 ≤ P ≤ 0.1)

The results for β-diversity for the 2022 rhizosphere samples using Bray-Curtis distance showed a similar pattern as was observed for the previous 2 years of data. Significant differences (P < 0.05) existed between pea and canola samples across treatments groups (Figure 7). Interesting, significant differences (P < 0.05) were observed between canola 6 and canola 1 and canola 6 and canola 4 (Figure 7). Trends (0.05 ≤ P ≤ 0.1) were observed between canola 6 and canola 9, canola 7 and null, canola 11 and canola 1, canola 1 and canola 3, canola 5 and canola 6, canola 3 and null, canola 1 and canola 4, canola 2 and canola 6, canola 4 and null, and pea 1 and 5 (Figure 7).

Analyses of the bacterial core microbiomes of our 2020 soil and rhizosphere samples were performed. Analysis of the 2021 and 2022 samples will be done. For soil, we analyzed the strict bacterial core microbiome, consisting of bacteria found in 100% of our samples. We found that the peaola core microbiome consisted of 34 members, the pea core microbiome consisted of 23 members, and the canola core microbiome consisted of 29 members. When comparing those microbiomes, we found that 13 bacteria were shared across them. From the canola core microbiome, 8 bacteria were shared with peaola. Out of the pea core microbiome, 3 bacteria were shared with peaola. Despite the peaola microbiome having shared members with monoculture pea and canola, it does have 10 core members that were unique to its own microbiome (Figure 8). When analyzing the rhizosphere, we did not analyze the strict core microbiome because it was not found across all 3 cropping systems. Instead, we analyzed core microbiomes consisting of bacteria found across 90% of our samples since a core microbiome was found across all 3 cropping systems at this level. At this level, we found that the peaola core microbiome consisted of 23 members, the pea core microbiome consisted of 14 members, and the canola core microbiome consisted of 18 members. When comparing those microbiomes, we found that 8 bacteria were shared across them. From the canola core microbiome, 4 were shared with peaola, and from the pea core microbiome, 1 was shared with peaola. Again, we found that despite peaola having shared core members with the microbiomes of pea and canola monoculture, we did find 10 unique members of the peaola core microbiome (Figure 8).

Network Analysis

Network analysis using SPIEC-EASI has been performed successfully at the genus level on the 2020 soil samples. It should be noted, however, that identification to the genus level is rare when performing analyses on Illumina amplicon sequences due to the small size of the sequences. The peaola co-occurence network consists of 1016 nodes and 10,058 edges (Figure 9). The pea monoculture network consisted of 984 nodes and 8674 edges and the canola monoculture network consists of 965 nodes and 8330 edges (Figures 10 and 11).

This shows that there is a change in the structure of the microbial community from monoculture pea and canola to peaola, which can also be seen visually (Figures 9, 10, and 11). Although there are nodes that are more highly connected than others, there does not appear to be any one node that is more highly connected than the others. Instead, there are clusters of highly connected nodes found throughout all 3 networks.

Stable Isotope Analysis

After performing statistical analyses on the nitrogen percentages for the 2020 samples collected at the Davenport location, we determined that there were no significant differences (P < 0.05) or trends (0.05 ≤ P ≤ 0.1) between the intercropped and monoculture pea and canola, and between pea and canola regardless of cropping system (Figure 12). For the 2022 samples, we found monoculture pea had significant more (P < 0.05) nitrogen in its leaf tissue than canola monoculture. Additionally, we found that pea in peaola treatments 1, 6, and 8 had significantly more (P < 0.05) nitrogen in their leaf tissue than monoculture canola. For canola in peaola treatments 1 and 6 we found that it had significantly less (P < 0.05) nitrogen in its leaf tissue than monoculture pea. For the samples collected from the Greene Farm, we also found no significant differences (P < 005) or trends (0.05 ≤ P ≤ 0.1) between the pea and canola in the intercropping system (Figure 13). Comparisons between intercropped and monoculture pea and canola were not able to be made at the Greene Farm as monoculture trials were not planted for both crops.

Discussion

Land Equivalent Ratio and Yield Analysis

It is not surprising that the peaola system did not appear to benefit from increasing the rates of synthetic N fertilizer, as previous work done on legume-oilseed intercropping has found similar results (Porter et al., 2020). Despite these results, more work needs to be done to determine if the peaola cropping system can be grown independently of N fertilizer application. Largely, this is due to there being a potential that N was not a limiting factor in the crops grown in this study. Therefore, future research will be addressing if N is being transferred from peas to canola via plant-plant-microbe interactions. This research will help us determine if these interactions are responsible for the increased LER with decreased synthetic N input of peaola. The relative yield of peas to canola, calculated based on monoculture pea and canola, showed that winter peaola did not strongly favor either pea or canola. Although some studies have found that pea is favored over canola in overyielding peaola, our data align with the majority of studies on peaola with an overall LER of 1.63 with neither pea nor canola being strongly favored (Fletcher et al., 2016). The crop grown in 2021 at the Greene farm likely experienced failure of its pea crop due to extreme weather conditions around the time of flowering. This is because heat is known to decrease yield in both pea and canola, especially around flowering (Mohapatra et al., 2020).

Microbial Community Analysis

Through our analyses of the α-diversity and β-diversity of the microbial communities in the soil and rhizosphere, we did not find that intercropping pea and canola had a large impact on the diversity of the microbial community alone. Despite this finding, we did find that peaola did have a unique core microbiome compared to pea and canola when grown in monoculture. Taken with our findings for α-diversity and β-diversity, it does show why we did not see a significant difference here as in the core microbiome there does appear to be a combining of the pea and canola microbiomes despite peaola's unique members. Therefore, it would make sense that the diversity would not be significantly different from pea and canola if there were not many unique members added to the peaola microbiome overall. Despite this, our findings that peaola does have unique core members does suggest to us that it is possible that peaola is creating a unique soil environment from pea and canola, potentially due to changes it is making to the soil chemistry. Further work will need to be done in order to confirm this. In addition, since changes were observed in the core microbiome, it is possible that some members of the microbial community are impacted by peaola in terms of their overall abundance. Therefore, we will be performing differential abundance analysis on our sequencing data in order to determine if there are any bacteria that are changing significantly in their overall abundance between cropping systems. This will be done in R version 4.2.1 in RStudio version 2022.07.0+548. Additionally, data on the abundance of the nifH gene across cropping systems from qPCR will also be insightful into how the microbial community is changing across cropping systems, as it will show us if there is a change in the abundance of bacteria with the ability to fix nitrogen.

The results from the sample taken in 2021 and 2022 do show that crop management practices do have a significant impact on microbial communities, especially when different seeding rates were used in peaola. From looking at the 2022 data, it appears that treatments 3, 6, and 8 had the most significant impact on the canola bacterial communities overall, as these communities were significantly richer than almost all of the pea bacterial communities across treatments. Interestingly, treatments 3 and 6 had the higher pea seeding rate tested. Treatment 8 had the mid-range pea seeding treatment, but the highest canola seeding rate. Analysis of the core microbiome will be helpful in determining what is happening in these systems, as it is possible that bacteria from the pea core microbiome could be entering the canola core microbiome in these treatments. Differential abundance analysis will also be useful here in determining which bacterial taxa are shifting. Knowing which taxa are shifting in the microbiome will allow us to determine if these bacteria are playing a role in nutrient cycling and therefore are potentially contributing to overyielding in peaola.

Network Analysis

Based on the changes in the number of nodes and edges found in the networks from pea and canola to peaola, it is clear that there is a change in the structure of the microbial community. There is also the potential that this is demonstrating the combining of the pea and canola microbial communities in peaola as we saw through our microbial community analysis. Although there is useful information that can be extracted from these networks, we have decided to remake our networks with bacteria at the phylum level instead. This will allow patterns to become clearer in the networks, as this reduces the number of nodes. Further investigation will then be done into phyla that are highly connected in the network to determine the microbial members present in each phylum and their predicted ecological role.

Stable Isotope Analysis

From our analyses of our stable isotope data, it appears that the seeding rate of pea and canola in peaola does have a significant impact on the N content of pea and canola. Although no difference was seen between pea and canola N content in 2020, in 2022 pea was found to have significantly more N than canola. In peaola, this trend was no longer observed. This suggests that there could be an exchange of nutrients between pea and canola, and in some treatment groups there could be better exchange of N from pea to canola. For example, in seeding treatments 1 and 6 canola was found to have significantly less N content than monoculture pea, but pea was found to have significantly more N than monoculture P. In the other seeding treatments, canola and pea N content was not significantly different from monoculture pea in either direction. With that said, it is possible that in seeding treatments 1 and 6 that pea could be withhold more N for itself and even competing with canola, which is why we see an increase in N content from monoculture pea, but a maintained pattern of canola having significantly less N content than monoculture pea. With that said, greenhouse experiments will be necessary to determine if there is transfer of N from pea to canola. Preliminary results from a greenhouse experiment we performed measuring the dry shoot mass of pea and canola across cropping systems suggests that pea is providing N to canola while experiencing no negative consequences to its own success.

This is because we saw that there was no significant difference in the dry shoot mass of pea across cropping systems both with and without a N treatment (Figure 10). When looking at canola, we did see that the dry shoot mass of canola when grown with pea was significantly greater than canola grown in monoculture with no N treatment (Kruskal–Wallis Test, n1 = 11, n2 = 20, P = 0.0366), but was not significantly different from canola grown in monoculture with a N treatment (Kruskal–Wallis Test, n1 = 11, n2 = 20, P = 0.445) (Figure 11). Therefore, further investigation will need to be done in the greenhouse to determine how pea and canola are interacting in order to confirm these results were the results of N nutrient transfer between plants.

Sources

Dowling A, Victor SO, Penny R, Ashlea D, Yi Z, Matthew DD. Legume-oilseed intercropping in mechanised broadacre agriculture – a review. Field Crops Research. (2021) 260:107980. doi: https://doi.org/10.1016/j.fcr.2020.107980

Fletcher AL, Kirkegaard JA, Peoples MB, Robertson MJ, Whish J, Swan AD. Prospects to utilise intercrops and crop variety mixtures in mechanised, rain-fed, temperate cropping systems. In: Crop and Pasture Science, Vol. 67. CSIRO (2016). p. 1252–67. doi: 10.1071/CP16211

Mohapatra C, Chand R, Tiwari JK, Singh AK. Effect of heat stress during flowering and pod formation in pea (Pisum sativum physiology L). Mol Biol Plants. (2020) 26:1119–25. doi: 10.1007/s12298-020-00803-4

Porter MJ, Pan WL, Schillinger WF, Madsen IJ, Sowers KE, Tao H. Winter canola response to soil and fertilizer nitrogen in semiarid mediterranean conditions. Agron J. (2020) 112:801–14. doi: 10.1002/agj2.20119