Plant-microbe networks in soil are weakened by century-long use of inorganic fertilizers

Abstract

Understanding the changes in plant-microbe interactions is critically important for predicting ecosystem functioning in response to human-induced environmental changes such as nitrogen (N) addition. In this study, the effects of a century-long fertilization treatment (> 150 years) on the networks between plants and soil microbial functional communities, detected by GeoChip, in grassland were determined in the Park Grass Experiment at Rothamsted Research, UK. Our results showed that plants and soil microbes have a consistent response to long-term fertilization-both richness and diversity of plants and soil microbes are significantly decreased, as well as microbial functional genes involved in soil carbon (C), nitrogen (N) and phosphorus (P) cycling. The network-based analyses showed that long-term fertilization decreased the complexity of networks between plant and microbial functional communities in terms of node numbers, connectivity, network density and the clustering coefficient. Similarly, within the soil microbial community, the strength of microbial associations was also weakened in response to long-term fertilization. Mantel path analysis showed that soil C and N contents were the main factors affecting the network between plants and microbes. Our results indicate that century-long fertilization weakens the plant-microbe networks, which is important in improving our understanding of grassland ecosystem functions and stability under long-term agriculture management.

Figure 1

The relative frequency of plant species in (A) control and (B) long‐term fertilized plots of Park Grass.

Figure 2

The normalized signal intensity of the detected key genes families involved (A) C degradation, (B) C fixation, (C) N cycling and (D) P cycling. The normalized signal intensities showed by cylindrical diagrams of (A), (B) and (D) were the sum of the average of individual gene sequences among 21 soil samples, divided by the number of detected sequences and then divided by 100. All data are presented as mean ± SD.*P < 0.05, **P < 0.01. The relative changes of the detected genes involved in N cycling (C) in fertilized plot were calculated, where the percentage of a functional gene in a bracket was the sum of the signal intensity of all detected sequences of this gene divided by the grand sum of the signal intensity of the detected N cycling genes, and weighted by the fold change of the signal intensity of this gene in the fertilized plot compared to that in control plot. For each functional N cycling gene, red colour means that this gene had a lower signal intensity in the fertilized plot than the control plot with significance at P < 0.05(*). Grey‐coloured genes were not targeted by this GeoChip, or not detected in those samples.

Figure 3

Network interactions between microbial functional communities in (A) control and (B) long‐term fertilized plots. The node size represents the relative abundance of the functional gene. Dotted grey dashed lines represent positive correlations, and orange solid lines represent negative correlations. The relationships between microbes at the phylum level were also calculated and showed below the network charts. The hollow circle represents no significant association; grey and orange in the circles represent positive and negative correlations, and numbers in the grey and orange circles represent the number of significant correlation edges.

Figure 4

Network interactions between plants and soil microbes in (A) control plot and (B) long‐term fertilized plot. Circles and diamonds represent functional genes and plants respectively.

Figure 5

Mantel path analysis links associations between plants and microbes to soil geochemical attributes and community structure of soil microbes and plants with long‐term fertilization. Solid and dashed lines indicate significant and insignificant relations respectively; lines width is proportional to r the correlation coefficient. The symbols *, ** and *^** represent P < 0.05, P < 0.01 and P < 0.001 respectively.