In vitro selection of a microbial consortium predictive of synergistic functioning along multiple ecosystem scales

Soil microbes form complex interactive networks throughout the soil and plant rhizosphere. These interactions can result in emergent properties for consortia that are not predictable from the phenotypes of constituents in isolation. We used a four-species consortium to assess the capacity of individual microbial species versus different consortia permutations of the four species to contribute to increased P-solubilization using soil incubations and plant growth experiments. We found that as different combinations of bacterial species were assembled into differing consortia, they demonstrated differing abilities to stimulate soil P cycling and plant growth. The combination of all four microbes in the consortia were much more effective at solubilizing P and stimulating plant growth than any of the individual bacterial species alone. This suggests that in vivo functionally synergistic soil microbial consortia can be adept at performing specific ecosystem functions in situ. Improving our understanding of the mechanisms that facilitate synergistic functioning examined in this study is important for maximizing future food production and agroecosystem sustainability.


Introduction 25
Microbial inoculants are increasingly promoted and adopted to enhance crop productivity, soil health, and overall nutrient use efficiency [1].

26
Given the key role of the phytobiome for soil ecosystem functioning and plant health, microbiome enhancement and engineering seems like a 27 logical intervention in crop production systems. Yet, many challenges lie between microbial discovery in the laboratory and successful 28 implementation in the field [2]. The functional attributes of isolated bacteria and fungi can be routinely assayed in the laboratory. Microbial 29 genomics identifies the potential for specific functions and traits, and can be predictive of phenotype. For some conserved traits, microbial 30 phylogeny alone predicts function [3]. However, the expressed phenotypes of individual isolates often vary with environmental conditions,

31
including pH, nutrient availability, and in response to stress [4]. As a result, single isolates that show promise in the laboratory often prove 32 ineffective when inoculated into field soils [5,6].

34
Predictions of in situ functioning of isolates are further challenged by the strong effects of interactions among microbial taxa. Deep learning 35 approaches applied to constructed synthetic communities have revealed causality between microbiome composition and host phenotypes in 36 complex systems under nutrient limited or stressed conditions [7]. Herrera Paredes, Gao (7) showed how different synthetic microbial consortia 37 can affect plant gene expression associated with phosphorus starvation. They suggested that the microbe-microbe interactions were likely 38 responsible for microbial community assembly, which, in turn, allowed for emergent interactions with the host plant to occur. This recent work 39 suggests a framework explaining why broadly efficacious microbial inoculants based on isolates are rare and often provide only a limited solution 40 for improving plant productivity and agronomic efficiency in real-world agriculture applications [8].

41
Functional microbial consortia may be robust alternatives to single isolates for use as beneficial inoculants in agricultural systems. We define a 42 true functional consortium as two or more taxa that demonstrate enhanced function when interacting relative to any individual constituent alone. all native bacteria) was added to allow for the determination of synergistic effects independent of native soil communities. Immediately after 91 planting the seeds, 5 mL of the different cultures outlined in Table 1 were added to each of the tray slots which resulted in previously determined 92 saturation of the root zone. Plant were allowed to germinate for 3-5 days under a mist bench after which they were inoculated a second time with 5 93 mL and kept moist by daily watering. After two weeks, the plants were moved out of the mist bench and were watered daily in a greenhouse set at 94 22 ± 2.5°C with a day length of 16 hours. To assess plant growth, we manually measured the number of rosette leaves, the maximum rosette 95 diameter and plant height twice weekly according to Boyes,Zayed (17). Soil available P was determined according to Olsen (18). Plants unable to 96 bolt (emerge) in 32 days which were equally distributed among treatments, were excluded from the analysis.

98
Statistics 99 The differences among treatments in process rates were tested using an analysis of variance approach (ANOVA) and Tukey pairwise comparisons.

100
The relationship between the number of bacterial species and quantified processes was conducted using Spearman's correlations. All data was 101 checked for normality and, if needed, log or square root transformed to acquire normal distribution. Synergistic or antagonistic effects were 102 determined by subtracting the summed isolate activity from the combination activity and dividing by the summed isolate activity, thus, providing a 103 percentage of the relative non-additive effect. This represents a conservative estimate of non-additive effects since it does not account for 104 reductions in bacterial species-specific abundance when in a consortium. P-mobilization sensitivity was defined as the increase in P-mobilization 105 with the number of bacterial species and abiotic P-mobilization indicates the rate in sterile soils as indicated by the intercept of that relationship.

106
Treatment effect on plant growth metrics were also assessed using repeated measures ANOVA analyses in combination with contrast analyses to 107 determine differences among treatments. All statistics were conducted using JMP 11.

110
Microbial synergy in culture 111 We found individual bacterial species incubated on low-orthophosphate selective media mobilized P at significantly slower rates than the full 112 consortium (Figure 1; Table S1), indicating strong synergistic effects ( Table 4). The most successful individual species was C. testosteroni 113 followed by C. freundii, both of which were significantly greater than E. cloacae but not than P. putida. The simpler consortia with similar mobilization rates to the full consortia (PCo, PE, PCiE. PCi and CiE) were not significantly different from the full consortia but also not from any 115 other treatments except for the control and the E and P single strain inoculation. P-mobilization rates within consortia of 2-3 bacterial species were 116 up to 4-fold greater than those of individual bacterial species.

118
Microbial synergy in soil 119 We tested the potential for microbial P-solubilization in soils across an estimated gradient of phosphate saturation of the Fe and Al oxides 120 (Fe equivalent :PO 4 ). The mobilization of P in soils varied with both inoculum composition and soil characteristics (Figure 2; Table S2). In the soil 121 with high a Fe equivalent :PO 4 , we found the full consortia, PE. PCo, PE and PCiCo to result in greater P-mobilization rates than the control. In 122 contrast, the soil with low Fe equivalent :PO 4 showed all treatments except for P, Ci and CiCo to result in significantly lower P-mobilization rates. In 123 the soil of medium Fe equivalent :PO 4 no inoculation type resulted in greater P-mobilization than in the control. PCiCo and ECiCo were even 124 significantly lower in their soil mobilization rate than the control. We found that resource availability, as indicated by the ratio between total P and 125 the total P-binding capacity of Al and Fe (Fe equivalent ), influenced whether consortium members acted synergistically or antagonistically (Table 4).

126
In soils with a greater concentration of potentially bound P (narrow ratio of Fe equivelent to P ratio), the four microbes exhibited an antagonistic 127 relationship, while soils with a wider ratio showed a synergistic response (Figure 2; Table S2). Soil P-mobilization sensitivity, defined by the rate 128 at which a greater number of bacterial species increases P-mobilization, was positively correlated to the total Fe equivalent :P ( suggested that the full four species consortia treated plant treatment was greater than the control treatment (p < 0.05). We could not detect specific 134 treatment effects for any of the treatments in the maximum rosette diameter or the number of rosette leaves using repeated measures analysis. We

135
found that increasing the number of bacterial species in the consortium from one to four species increased plant height by up to 32% (Figure 3).

136
Further, inocula containing 1-3 bacterial species were greater in height than the control. Residual available P in the soil after plant harvest was 137 greatest in the zero-bacterial species control (44 ± 2 mg kg -1 ) and CiE (46 ± 4 mg kg -1 ) treatment and lowest in the C (25 ± 5 mg kg -1 ) and four 138 species consortia (28 ± 4 mg kg -1 ) treatments. Inoculation with all four bacterial species reduced the available P by 36% in comparison with the 139 non-inoculated treatment. exhibited positive effects of the tested trait relative to the control (i.e. P-solubilization, P-mobilization or plant growth). However, the relationships 151 in culture clearly showed that with increasing consortia complexity, the cultures shifted from P-immobilization to P-solubilization.

152
In synergistic consortia, individual constituents can enhance overall performance through interactive rather than direct effects. Previous research 153 has identified Comamonas spp. to be important for P-solubilization [28], yet, the Comamonas testosteroni in our study proved important for 154 consortium performance in culture but showed no P-solubilization when inoculated as a single strain. We found the communities with two species 155 to be very inconsistent in P-solubilization rates while communities of 3-4 species consistently solubilized P in culture. These results contradict a 156 theoretical model predicting that synergistic interactions should emerge in consortia greater than three bacterial species [29], but it is not 157 uncommon for synergistic effects to occur within simple ecosystems akin to our laboratory conditions [20,[23][24][25]. Overall, these findings suggest 158 that although variance was substantial, consortia are often superior to individual strains of the studied four bacteria and have emergent properties 159 with regards to P-solubilization in culture.

161
The effects of different constructed consortia on P-mobilization were inconsistent across different soils. Pseudomonas putida consistently mobilize 162 P independent of soil type, while the direction and magnitude of the synergistic response to increasing consortia size depended on soil type.
163 Surprisingly, even though one of our source soils was alkaline (dominated by calcium oxides), the ratio between Fe equivalent to P appeared to control 164 the sensitivity of P-mobilization, not the combination of Ca, Fe and Al as might be expected since these three elements are part of the dominant 165 three oxides in soil responsible for P-immobilization [30]. This suggests that solubilization of phosphate sorbed to Fe and Al oxides are the main 166 mechanisms used by the selected bacteria to increase available phosphorus. The patterns of antagonistic and synergistic effects follow the principles of stoichiometric theory where microbes mobilize P when bound-P is high (low Fe equivalent :P) but immobilize P at low availability (high

168
Fe equivalent :P) [31]. Indeed, we found antagonistic effects under a low Fe equivalent :P ratio and high synergistic effects under a high Fe equivalent :P ratio.

169
This suggests that the effects observed in culture only translate to the soil environment in soils high in occluded phosphate.

170
In contrast to the results of the soil incubations, when plants were grown in a mixture containing autoclaved soil from the site high in total P 171 concentrations we observed strong synergistic effects with inoculation of increasing consortium size. As would be expected, it is highly unlikely 172 the control treatment remained sterile throughout the experiment but it prevented native microbial interaction from immediately outcompeting the 173 inoculum species. Although, these results do not guarantee these synergistic relationships will also manifest with inoculations in more complicated 174 ecosystems, previous research on the same consortium has found positive effects on plant growth when inoculated in native soil [10]. Microbial P-175 solubilization has been linked with plant performance in many previous studies summarized by Richardson and Simpson (32)

Supporting information
268 Table S1: P-mobilization rates in culture (nM/h) 269 Table S2: P-mobilization rates in soil (µM/d) 270 271 272 273     The points represent the mean and the error bars indicate the standard error of the mean. For every treatment n=12.