Higher-order interaction inhibits bacterial invasion of a phototroph-predator microbial community

Abstract

9In nature, the composition of an ecosystem is thought to be important for determining its 10 resistance to invasion by new species. Studies of invasions in natural ecosystems, from plant 11 to microbial communities, have found that more diverse communities are more resistant to 12 invasion. It is thought that more diverse communities resist invasion by more completely 13 Recently, theoretical work using consumer-resource models has extended this intuition and 52 suggested that the emergent resource consumption and exchange in cross-feeding communities 53 can be understood as a community-level fitness which provides cohesiveness and therefore in-54 vasion resistance 14 . Experimental efforts suggest that this picture can capture some features of 55 experimental invasions in bacterial communities 15 . Collectively, this work shows that substantial 56 insight into invasion dynamics can come from understanding resource dynamics during an invasion 57 process. 58 However, in nearly all microbial communities there exist interactions that are not directly 59 3 mediated by resources: for example, antagonistic interactions such as the excretion of antibiotics 16 60 and predation by protists 17 or phage 18 . A handful of studies have examined the role these interac-61 tions play in determining the fate of invading species 19 , but, as recently pointed out by Mallon et 62 al. 20 , it remains an outstanding question how antagonistic interactions affect community invasion 63 dynamics. 64Here we use a model microbial community to study invasion dynamics in the presence of 65 antagonistic interactions. Microbial communities in freshwater lakes and nearby saturated soils 66 are occupied by primary producers who fix inorganic carbon, metabolically flexible heterotrophic 67 bacteria who decompose organic matter, and predators who unlock nutrients held in biomass 21 . To 68 study this canonical natural community, we use a three-species model microbial ecosystem com-69 prised of the alga Chlamydomonas reinhardtii which acts as a primary producer and is an endemic 70 phototroph in soils and freshwater 22 , the bacterium Escherichia coli which acts as a decomposer 71 and is common in soils 23 , and the ciliate Tetrahymena thermophila which dwells in freshwater and 72 preys on E. coli. We refer to this model ecosystem as the 'ABC' community for Algae, Bacteria 73 and Ciliates. The ABC community has been studied previously as a model self-sustaining closed 74 microbial ecosystem [24][25][26] . Recent work has shown that long-term abundance dynamics in closed 75 ABC ecosystems are complex and deterministic on timescales of months, exhibiting rich spatio-76 temporal and phenotypic dynamics 25 . The fact that the composition of the ABC community reflects 77 the structure of some natural communities and that quantitative measurements are feasible make 78 this a compelling model ecosystem for quantitative ecology 27 . 79 4Here we show that when E. coli (B) is introduced into communities of C. reinhardtii (A) 80 and T. thermophila (C...

Species richness increases fitness differences, but does not affect niche differences

“…(3) Third, higher-order interactions, through which a third species changes the interaction between a species pair, are by definition absent from two-species communities (but see Letten & Stouffer (2019); Levine et al (2017)). Such higherorder interactions have been found empirically, for example, in communities composed of phytoplankton, bacteria, and ciliates (Mickalide & Kuehn, 2019). In that study, bacteria coexisted with phytoplankton and ciliates, but all three functional groups did not coexist.…”

Species richness increases fitness differences, but does not affect niche differences

“…(3) Higher‐order interactions, through which a third species changes the interaction between two species, are usually considered absent from two‐species communities (but see Letten and Stouffer (2019); Levine et al (2017) who include higher‐order interactions into two‐species communities). Such higher‐order interactions have been found empirically, for example in communities composed of phytoplankton, bacteria and ciliates (Mickalide & Kuehn, 2019). In that study, bacteria coexisted with phytoplankton and ciliates, but all three functional groups did not coexist.…”

Uncovering the rules of microbial community invasions

“…Such an addition of traits to the model would make it possible to incorporate trait-dependent ecological interactions; both antagonistic and cooperative 27,95,96 . A more advanced model still could explicitly capture environmentally mediated interactions, such as facilitation and toxin mediated antagonism that have been shown to promote and suppress invasion across a wide range of systems and can result in both high-order and intransitive interactions [97][98][99][100] . Finally, our model was spatially implicit with invader and resident communities each being well mixed individually with invasion being modelled as a single event; the same dynamics could be simulated on a network of patches or a fully spatially explicit grid of locations.…”