Indicative bacterial communities and taxa of disease-suppressing and growth-promoting composts and their associations to the rhizoplane

ABSTRACT Compost applications vary in their plant growth promotion and plant disease suppression, likely due to differences in physico-chemical and biological parameters. Our hypothesis was that bacteria are important for plant growth promotion and disease suppression of composts and, therefore, composts having these traits would contain similar sets of indicative bacterial taxa. Seventeen composts prepared from five different commercial providers and different starting materials were classified accordingly with bioassays using cress plants and the pathogen Pythium ultimum. Using a metabarcoding approach, bacterial communities were assessed in bulk composts and cress rhizoplanes. Six and nine composts showed significant disease suppression or growth promotion, respectively, but these traits did not correlate. Growth promotion correlated positively with nitrate content of composts, whereas disease suppression correlated negatively with factors representing compost age. Growth promotion and disease suppression explained significant portions of variation in bacterial community structures, i.e. 11.5% and 14.7%, respectively. Among the sequence variants (SVs) associated with growth promotion, Microvirga, Acinetobacter, Streptomyces, Bradyrhizobium and Bacillus were highly promising, while in suppressive composts, Ureibacillus,Thermogutta and Sphingopyxis were most promising. Associated SVs represent the basis for developing prediction tools for growth promotion and disease suppression, a highly desired goal for targeted compost production and application.

: Experimental design of the bioassay. Two batches were tested with 9 composts each where compost H was included in both batches. Three replicates for metabarcoding (red circles) and three replicates of cress biomass (green triangles) were sampled. In total, 174 samples were obtained for metabarcoding including 54 compost samples (2 batches x 9 composts x 3 replicates), 3 matrix samples, 108 rhizoplane samples of cress grown in compost substrate (2 batches x 2 inoculation levels x 9 composts x 3 replicates) and 12 rhizoplane samples of cress grown in control matrix (2 batches x 2 inoculation levels x 3 replicates). In total, 240 cress biomass samples were obtained including 182 samples from plants grown in compost substrates (2 batches x 9 composts x 4 inoculation levels x 3 replicates) and 24 samples from plants grown in the control matrix (2 batches x 4 inoculation levels x 3 replicates).
Supplementary Figure 2: Correlations of mean percent growth promotion and mean compost characteristics such as nitrate [mg N (kg dry weight) -1 ], inorganic nitrogen [mg N (kg dry weight) -1 ], salinity [g KCleq (kg dry weight) -1 ], age [d], ammonia [mg N (kg dry weight) -1 ], dry matter [%], content of soluble humic substances (OD 550) and pH. Plots were ordered according to the absolute value of the strength of the Spearman correlation (rho) and a linear trendline was added only for significant correlations (p < 0.05). Correlations were based on 17 composts, except for age only 12 were available. Statistics of correlations are also presented in Supplementary Table 4. Mean growth promotion was defined as the percent increase in shoot weight of plants that were grown in compost substrates compared to the control matrix both without inoculation with the pathogen.
Supplementary Figure 3: Disease suppression among compost substrates (A-Q) and the control matrix (m) inoculated with either 0.25, 0.5 or 1 g of P. ultimum-millet-mix (L of substrate) -1 of the first (A) and the second (B) batch of composts. Percent disease suppression was defined as the relative shoot weight of P. ultimum-inoculated and uninoculated cress plants for each compost treatment and the control matrix. Letters show significant differences among substrates (TukeyHSD-tests, p < 0.05). ANOVA was performed separately for each experiment and inoculation-level and overall statistics for 0.25, 0.5 and 1 g L -1 were F = 1.2 and p = 0.358, F = 2.6 and p = 0.038 as well as F = 19.8 and p = 4 x 10 -8 in batch one (A) and F = 3.0 and p = 0.019, F = 7.2 and p = 0.0001 as well as F = 8.0 and p = 6 x 10 -5 in batch 2 (B). Compost H was tested in both experiments.  Supplementary  Table 4. Percent disease suppression was defined as the relative shoot weight of P. ultimuminoculated and uninoculated cress plants for each compost treatment.
Supplementary Figure 5: Partitioning of robustly detected SVs among substrates and rhizoplanes. (A) Venn diagram displaying the occurrence of SVs in each compartment, i.e., compost (brown), matrix (black), rhizoplanes of cress plants grown in compost substrate ("rhizo c", pink dashed line) and rhizoplanes of cress plants grown in the matrix ("rhizo m", blue dashed line). Rhizoplane data derived from plants inoculated with P. ultimum are presented. The numbers outside of the circles of the Venn diagram show the number of SVs and the mean percentage of SVs in each compartment. The numbers in the different areas of the Venn diagram correspond to mean percentages of SVs from the 17 composts, the matrix, the corresponding rhizoplanes and their intersections. Standard deviations are presented in parenthesis. Data were obtained using a subsampling procedure to 18,743 sequences with 100 iterations. (B) Partitioning of SVs in rhizoplanes of plants grown in compost substrates ("rhizo c"). Percentages correspond to SVs that were also found in compost ("c"), in compost and the matrix ("c+m"), the matrix ("m") and in rhizoplanes of plants grown in the control matrix ("rhizo m"), or that were only detected in rhizoplanes of plants grown in the compost substrates ("rhizo c").
Supplementary Figure 6: Partitioning of robustly detected SVs among substrates and rhizoplanes of the most and the least growth promoting (A) and disease suppressing (B) composts. The numbers in the different areas of the Venn diagram correspond to mean percentages of SVs. The first number is the mean percentage of SVs in the 5 most active and the second number the 5 least active compost for growth promotion (A) and disease suppression (B). Colours represent different compartments, i.e., compost (brown), matrix (black), rhizoplanes of cress plants grown in compost substrate ("rhizopl. compost", pink dashed line) and rhizoplanes of cress plants grown in the matrix ("rhizopl matrix", blue dashed line). Rhizoplane data derived from uninoculated plants (A) and plants inoculated with P. ultimum (B) are presented. Data were obtained using a subsampling procedure to 18,743 (A) and 13,173 (B) sequences with 100 iterations. The Asterisk represents a significant difference of percentage of SVs of the five most active and least active composts (p < 0.05). Statistics of all Student t-tests are presented in Supplementary Table 9.   Table 3: Cress shoot weight (mean and standard deviation) and disease symptoms (median) of plants grown in compost substrate (A-Q) or the control matrix (m). Bioassays were conducted in two batches including nine compost substrates each. Compost A-I were tested in the first batch, compost J-Q were tested in the second batch and compost H (H1 and H2) and the control matrix (m1 and m2) were included in both batches. Plants were inoculated with 1 g of P. ultimum-millet-mix (L substrate) -1 . The severity of disease symptoms of damping-off was classified in five categories according to the severity of disease, i.e., plants showing yellowing of leaves and growth reduction.  Table 4: Spearman correlations of mean disease suppression or mean growth promotion with compost characteristics of the 17 composts. Percent disease suppression was defined as the relative shoot weight of P. ultimum-inoculated and un-inoculated cress plants for each compost treatment. Mean growth promotion was defined as the percent increase in shoot weight of plants that were grown in compost substrates compared to the control matrix both without inoculation with the pathogen. Results from a correlation of mean disease suppression with mean growth promotion was also included.       1 point biserial correlation coefficient 2 Spearman correlation of the relative abundance of an SV with mean growth promotion including all 17 composts; with Benjamini Hochberg adjusted p-value. 3 letter preceding the lowest possible taxonomic classification corresponds to the taxonomic rank (g for genus, f for family, o for order, c for class) NA not available Supplementary Table 11: Relative abundance, occurrence and taxonomy of 75 SVs associated to strongly suppressive composts, i.e., H, I, K, L and Q. Potentially disease suppressive SVs were selected by comparing robustly detected SVs of the five most and the five least suppressive composts using indicator species analysis with point biserial correlation coefficient larger than 0.7 as a selection criterion. Mean relative sequence abundance of each SV was correlated with mean disease suppression including all 17 composts. For each SV, which was assigned to a genus, evidence from literature about its occurrence in composts and involvement in suppression against soil borne fungal and oomycete diseases was assessed (end October 2020).