Cropping practices manipulate abundance patterns of root and soil microbiome members paving the way to smart farming

Background Harnessing beneficial microbes presents a promising strategy to optimize plant growth and agricultural sustainability. Little is known to which extent and how specifically soil and plant microbiomes can be manipulated through different cropping practices. Here, we investigated soil and wheat root microbial communities in a cropping system experiment consisting of conventional and organic managements, both with different tillage intensities. Results While microbial richness was marginally affected, we found pronounced cropping effects on community composition, which were specific for the respective microbiomes. Soil bacterial communities were primarily structured by tillage, whereas soil fungal communities responded mainly to management type with additional effects by tillage. In roots, management type was also the driving factor for bacteria but not for fungi, which were generally determined by changes in tillage intensity. To quantify an “effect size” for microbiota manipulation, we found that about 10% of variation in microbial communities was explained by the tested cropping practices. Cropping sensitive microbes were taxonomically diverse, and they responded in guilds of taxa to the specific practices. These microbes also included frequent community members or members co-occurring with many other microbes in the community, suggesting that cropping practices may allow manipulation of influential community members. Conclusions Understanding the abundance patterns of cropping sensitive microbes presents the basis towards developing microbiota management strategies for smart farming. For future targeted microbiota management—e.g., to foster certain microbes with specific agricultural practices—a next step will be to identify the functional traits of the cropping sensitive microbes. Electronic supplementary material The online version of this article (10.1186/s40168-017-0389-9) contains supplementary material, which is available to authorized users.


130
Across the four cropping systems, OTUs from the Actinobacteria were equally well 131 represented. OTUs from the Proteobacteria and Bacteroidetes, were more abundant in 132 reduced and no-tillage plots. Like in the soil bacterial community, the Firmicutes were 133 generally more abundant in root samples from organically managed plots. This 134 appeared to be driven by the increased abundance of several OTUs from the family 135 Peptostreptococcaceae (bOTU23, bOTU119), Clostridiaceae (bOTU341), 136 Erysipelotrichaceae (bOTU36), and Lachnospiraceae (bOTU1403), a family that was 137 exclusive to organically managed plots. (Figs. S8, S11). 138 The 36 csOTUs (Fig. S6) in root fungal communities were classified into at least 139 three phyla. Most sequences belonged to the Ascomycota (75.9%), followed by 140 unassigned phyla (14.1%), and Basidiomycota (9.8%; Fig. S7). We noted that OTUs 141 from the Ascomycota favored the C-NT system and, to a lesser extent, the organically 142 managed plots. The O-RT system supported a higher abundance of OTUs belonging to 143 unassigned phyla and the Basidomycota. Many of the cropping sensitive OTUs were 144 unassigned at lower taxonomic levels (Fig. S8). However, in the Ascomycota, fOTU63 145 (Pleosporaceae) and fOTU97 (Phaeosphaeriaceae) were abundant in the C-NT system, 146 while the Psathyrellaceae fOTU86 was abundant in the O-RT system (Figs. S8, S12). 147 We also noted a number of OTUs from the family Lasiosphaeriaceae with higher mean 148 abundances in the O-IT treatment. 149

151
Cropping system effects on soil microbial communities 152 We found significant effects of cropping system on soil microbial communities, 153 explaining approximately 30% of the total variation in both bacteria and fungi (Fig. 2). 154 More specifically, bacterial communities were more strongly separated by the different 155 tillage regimes rather than by management type, with the biggest differences between 156 intensive tillage samples and those receiving less intensive tillage (Table S4). This 157 finding is somewhat unexpected given that earlier work has shown that the addition of 158 manure, as is the case in the organically managed plots, can result in substantial shifts 159 in soil bacterial community [8][9][10][11]. Moreover, bacteria are generally thought to be 160 relatively unaffected by tillage practices, given their small cell size and constrained 161 dispersal and are therefore, less likely to be affected by the homogenization of soil 162 microsites [12,13]. 163 It has also been suggested that bacteria introduced into soils from manure 164 amendments do not become prominent [9] and that any bacterial community 165 compositional shifts as a result of manure additions tend to diminish over time [8][9][10]. 166 However, these results would seemingly conflict with a number of recent studies that 167 have profiled microbial communities in soils receiving inorganic and organic fertilizer 168 and found substantial differences between the two fertilizer regimes [14][15][16][17] In contrast to soil bacteria, constrained ordinations of soil fungal communities 184 revealed that differences between conventional and organic management types 185 explained most of the variation (Fig. 2). Despite the relatively short term management of 186 the FAST site, our results are more in accordance with previous studies on long-term 187 (>20 years) agricultural trials that reported significant effects of organic management 188 with manure fertilization on soil fungal community composition [15,19]. Studies on soil 189 communities subjected to organic management with manure additions over the short 190 term (typically less than 10 years) have tended to report no significant differences in 191 fungal community structure between manure amended and non-amended soils [20,21].

192
However, these shorter-term studies relied on older molecular tools, which may be less 193 precise in capturing subtle community shifts compared to amplicon sequencing [19]. 194 Nevertheless, there is evidence that the addition of manure to soils represents an 195 input of external microbes that could affect strong changes in the diversity and 196 composition of both bacterial and fungal communities over the course of a growing 197 season [19,22]. With this in mind, our results highlight the need for future studies to 198 assess the temporal variability in soil communities receiving external microbial inputs, 199 such as manure. Sampling at multiple time points, including before manure application, 200 would shed light on the dynamics of the bacterial and fungal communities during the 201 course of the growing season. This could help to improve estimates of microbial α-202 diversity, which have been shown to exhibit greater temporal variability than across 203 different land use types [23]. Furthermore, future studies would benefit from the 204 inclusion of manure samples in high-throughput sequencing runs for the direct 205 identification of manure-derived bacteria and fungi OTUs based on sequence similarity.

206
We found that an increase in tillage intensity from reduced tillage to intensive 207 tillage resulted in significantly different soil fungal communities in organically managed 208 plots; whereas the same was not observed between no-till and intensive tillage samples 209 in conventional plots ( site due to very low abundances of AMF sequences (Fig. S3). It is generally thought 215 that tillage affects soil AMF communities through physical destruction of dense hyphal 216 networks [27]. Such mechanisms of physical disturbance are also thought to influence 217 communities of general soil fungi, and therefore less soil disturbance and more 218 heterogeneous resource distribution, common of no till and reduced tillage systems, 219 may promote fungal communities [28]. Many hypotheses about the effects of tillage on 220 fungal communities also focus on indirect effects, namely that tillage influences edaphic 221 factors like soil organic carbon content [29,30] and soil nutrient pools like extractable P 222 [31], which have been shown to influence soil fungal community composition. Similarly, 223 our unconstrained ordination analyses revealed that differences in pH explained 224 approximately 24% and 27% of community variation in the soil bacterial and fungal 225 communities, respectively (Fig. S4). These results are generally consistent with 226 previous findings showing soil pH as a significant driver of primarily bacterial community 227 composition [32,33], but also of fungi [34]. However, it is important to stress that our 228 findings were less the result of a true pH gradient across multiple samples and more the 229 result of a low pH value in one subplot. 230 231 Cropping system effects on microbial α-diversity 232 We have assessed the effects of cropping systems on observed bacteria and fungi OTU 233 richness in both soil and root samples, confirming that soils were more diverse than root 234 microbial communities [35,36]. With respect to the effects of cropping system, we found 235 the soil bacteria and fungi tended to be richest in the O-IT system (Fig. S5, Table S3).

236
These observations are in accordance with previous studies reporting higher soil 237 microbial richness in organically managed compared to conventionally managed soils 238 (bacteria: 29, 50, 51; fungi: 29, 52). However, there are also studies reporting no 239 differences between conventional and organic managements [40,41]. We speculate that 240 timing differences between application and sampling might explain conflicting results, in 241 that any enhanced diversity effects might disappear in the time span between manure 242 application and sampling. 243 The effects of differential soil managements on the root microbes appear to vary 244 depending on the root compartment analyzed. Edwards et al., [42] found differences in 245 bacteria α-diversity in the rhizosphere but not rhizoplane and endosphere compartments 246 when comparing samples from conventional and organically managed cropping 247 systems. Also Seghers et al., [43] found no difference in maize root endophyte richness 248 (bacteria and fungi) in samples taken from conventionally and organically managed 249 plots. Soil management seems to affect microbial communities to lesser extents the 250 more intimate they associate with their host plant. We think that our root sampling 251 method without physical (no sonication) or chemical (no detergent or bleach) separation 252 from the rhizosphere compartment yields a rather low-intimacy type of compartment and 253 we expected to find impacts by soil management. Indeed, we found effects of cropping 254 practices on observed root OTU richness. We found significantly higher richness in in 255 O-IT plots compared to conventionally managed plots for the bacteria ( Fig. S5; Table  256 S3).

257
Taken together, we find enhanced richness in root and soil microbiota in O-IT 258 systems. We think that the application of animal manure as fertilizer coupled with 259 structural disturbance presents a likely explanation for the enhanced diversity in organic 260 intensive tillage systems. 261