Metallic micronutrients are associated with the structure and function of the soil microbiome

The relationship between metallic micronutrients and soil microorganisms, and thereby soil functioning, has been little explored. Here, we investigate the relationship between metallic micronutrients (Fe, Mn, Cu, Zn, Mo and Ni) and the abundance, diversity and function of soil microbiomes. In a survey across 180 sites in China, covering a wide range of soil conditions the structure and function of the soil microbiome are highly correlated with metallic micronutrients, especially Fe, followed by Mn, Cu and Zn. These results are robust to controlling for soil pH, which is often reported as the most important predictor of the soil microbiome. An incubation experiment with Fe and Zn additions for five different soil types also shows that increased micronutrient concentration affects microbial community composition and functional genes. In addition, structural equation models indicate that micronutrients positively contribute to the ecosystem productivity, both directly (micronutrient availability to plants) and, to a lesser extent, indirectly (via affecting the microbiome). Our findings highlight the importance of micronutrients in explaining soil microbiome structure and ecosystem functioning.

on soil microbial abundances.The relationships between micronutrients and bacterial abundance (a and b), and between fungal abundance (c and d) were conducted by the partial correlation.The total PC1 and total PC2 were obtained by using the principal component analysis of the total micronutrients, and the available PC1 and available PC2 were obtained using available micronutrients.The total PC1, total PC2, available PC1 and available PC2 represented the integrated effects of Fe, Mn, Cu, Zn, Mo and Ni.The solid dots represent the positive correlation coefficients and the empty dots represent the negative correlation coefficients in radar plots.Fig. S6.The effects of available micronutrients (Fe, Mn, Cu, Zn, Mo, Ni) on the abundances of bacterial and fungal taxa at genus level.The identified genera had the relative abundances that were significantly higher or lower in relative low micronutrients as compared with high micronutrient group.The low group included the 60 samples with relative lower micronutrient concentrations in the total of 180 samples and the high group included the 60 samples with relative higher micronutrient concentrations in the 180 samples.The blue and red dots represent the identified genera whose abundance decreased and increased as available micronutrient concentrations increased respectively.Fig. S7.The effects of total micronutrients (Fe, Mn, Cu, Zn, Mo, Ni) on the abundances of bacterial and fungal taxa at genus level.The identified genera had the relative abundances that were significantly higher or lower in relative low micronutrients as compared with high micronutrient group.The low group included the 60 samples with relative lower micronutrient concentrations in the total of 180 samples and the high group included the 60 samples with relative higher micronutrient concentrations in the 180 samples.The blue and red dots represent the identified genera whose abundance decreased and increased as total micronutrient concentrations increased respectively.

Fig. S3 .
Fig. S3.The range of total (a and b) and available (c and d) micronutrient concentrations across 180 soil samples.

Fig. S4 .
Fig. S4.Effects of total and available micronutrients(Fe, Mn, Cu, Zn, Mo, Ni)  on soil microbial abundances.The relationships between micronutrients and bacterial abundance (a and b), and between fungal abundance (c and d) were conducted by the partial correlation.The total PC1 and total PC2 were obtained by using the principal component analysis of the total micronutrients, and the available PC1 and available PC2 were obtained using available micronutrients.The total PC1, total PC2, available PC1 and available PC2 represented the integrated effects of Fe, Mn, Cu, Zn, Mo and Ni.The solid dots represent the positive correlation coefficients and the empty dots represent the negative correlation coefficients in radar plots.

Fig. S5 .
Fig. S5.Partial correlation for the relationship between soil pH, macronutrients and bacterial (a) and fungal (b) abundances.The solid dots represent the positive correlation coefficients and the empty dots represent the negative correlation coefficients in radar plots.

Fig. S8 .
Fig. S8.The differences in soil bacterial (a and b) and fungal (c and d) communities as revealed by community dissimilarity between different climate zones and land uses, performed by two-way ANOVA.

Fig. S9 .
Fig. S9.Partial correlation coefficients (avoiding the pH effects) for the relationship between soil macronutrients and the alpha, beta diversity, the abundance of dominant phyla and network connectivity for bacteria (a) and fungi (b).

Fig. S10 .
Fig. S10.Score coefficients of principal component 1 (PC1) and principal component 2 (PC2) for the total (a) and available (b) micronutrients across 180 soil samples, shown by the principal components analysis.

Fig. S11 .
Fig. S11.The original paths of SEM hypothesizing how environmental factors contribute to ecosystem production.

Table S1 .
The correlation coefficients of the correlations between the abundances of identified genera and micronutrient available concentrations with and without data normalization.The "*" represents that the correlation is significant at adjusted p < 0.05 (two-sided).

Table S2 .
The correlation coefficients of the correlations between the abundances of identified genera and micronutrient total concentrations with and without data normalization.The "*" represents that the correlation is significant at adjusted p < 0.05 (two-sided).

Table S3 .
The correlation coefficients and p-values (two-sided) of the correlations between micronutrient concentrations and the relative/absolute abundances of functional genes that are shown in the C, N, P and S pathways in Fig.3.

Table S4 .
Microorganisms containing the genes involved in C, N, P and S cycling at genus level.The genera with the top 5 highest abundances across the samples and the genera identified by Deseq2 (marked with "*") are presented.

Table S5 .
The name and annotation of genes involved in C, N, P and S cycling measured by high-throughput qPCR based chip.

Table S6 .
Pearson correlation between soil total micronutrients, available micronutrients and soil macronutrients.The "*" represents that the correlation is significant at p < 0.05 (two-sided).The number in the table represents the correlation coefficient (r).

Table S7 .
The differences in the overall microbial community revealed by PCoA between the soils with low, medium and high concentrations of available micronutrients.The low and high groups included the 60 samples with relatively lowest and highest micronutrient concentrations among 180 samples respectively.The other 60 samples were included in medium groups.

Table S8 .
The differences in the overall microbial community revealed by PCoA between the soils with low, medium and high concentrations of total micronutrients.The low and high groups included the 60 samples with relatively lowest and highest micronutrient concentrations among 180 samples respectively.The other 60 samples were included in medium groups.