Session F23: Physics of Microbiomes and Bacterial Communities

A statistical-mechanics approach to microbiome data analysis

Next-generation sequencing, high-throughput metabolomics and other measurement technologies have opened vast new horizons for collecting data on the structure and function of microbial communities. But it remains unclear how to leverage this data for effective intervention in medical and agricultural applications. We do not know which quantities can be reliably predicted, which are hopelessly contingent, and what the predictors are for the former. In this talk, I will draw on conceptual tools from statistical physics, which were designed to answer precisely these sorts of questions. In particular, I will argue that the key features of community structure are encoded in a susceptibility matrix, which contain the response of species population sizes to small changes in growth rates. I will show how to estimate this matrix in different scenarios from existing data sets, and then explain how it can be used to cluster species into functionally redundant groups for enhanced predictability of community composition.   

Quantifying multi-species bacterial interactions in larval zebrafish

The microbial communities resident in animal intestines are composed of many species and play an important role in host development, health and disease. The complex nature of these diverse communities makes it challenging to determine the driving forces behind microbial composition. Further, it is unclear for many multi-species consortia whether species-level makeup can be predicted based solely on pairwise inter-species interactions, or whether higher-order interactions are needed to explain the composition of these communities.
To address this, we consider commensal intestinal microbes in larval zebrafish, initially raised germ-free to allow introduction of controlled combinations of bacterial species. Using dissection and plating assays we demonstrate the construction of communities of 1 to 5 bacterial species and show that the outcomes from the 2-species competitions do not contain enough information to predict the abundances in more complex communities. Furthermore, we observe a dampening of strong negative interactions as the microbial composition of the gut becomes more diverse. Our data suggests higher order interactions are important in the zebrafish gut, possibly due to changes in spatial organization that ongoing live imaging experiments are illuminating.   

Species covariance in C. elegans microbiome demonstrates existence of bacteria-bacteria and host-bacteria interactions

The structure and internal dynamics of complex microbial communities in the guts of organisms is poorly understood. Here we analyze a simplified community of bacteria in the gut of Caenorhabditis elegans, a roundworm. Initially germ-free adult C. elegans are introduced into an environment with equal concentrations of eight bacterial species from a native worm microbiome. After four days, we examine individual worm gut communities and calculate the covariance structure of the bacterial abundances. We find statistically significant off-diagonal covariances. We demonstrate that a simple model only considering migration, birth, death, and competition for space among the bacteria can capture the mean values of bacterial abundances and their variances. However, it is incapable of explaining the off-diagonal covariances. We also show that the positive off-diagonal covariances can be partially explained by variation of the birth rate and other bacterial parameters among the worm hosts. However, to explain the off-diagonal negative covariances observed in the experiment requires interactions between bacteria beyond competition for space. We thus show that the structure of the microbiome is affected by both bacteria-bacteria and bacteria-host interactions.   

Evidence for a multi-level trophic organization of the human gut microbiome

The human gut microbiome is a complex ecosystem, in which hundreds of microbial
species and metabolites coexist, in part due to an extensive network of cross-feeding
interactions. However, both the large-scale trophic organization of this ecosystem and
its effects on the underlying metabolic flow, remain unexplored. Here, using a simplified
model, we provide quantitative support for a multi-level trophic organization of the
human gut microbiome, where microbes consume and secrete metabolites in multiple
iterative steps. Using a manually-curated set of metabolic interactions between
microbes, our model suggests about four trophic levels, each characterized by a high
level-to-level metabolic transfer of byproducts. It also quantitatively predicts the typical
metabolic environment of the gut (fecal metabolome) in approximate agreement with
the real data. To understand the consequences of this trophic organization, we quantify
the metabolic flow and biomass distribution and explore patterns of microbial and
metabolic diversity at different levels. The hierarchical trophic organization suggested
by our model can help mechanistically establish causal links between the abundances
of microbes and metabolites in the human gut.   

Regime shifts in a phage-bacterial ecosystem and strategies for its control

We show [1] that phage-bacterial ecosystems can have several alternative species compositions separated by abrupt regime shifts. In one of these states the fast-growing bacteria competitively exclude the slow-growing species by depleting common nutrient. In the other state the slow-growing bacteria with a large burst size make so many phages that the other species cannot survive. This scenario can be realized e.g. if two bacterial strains are protected from the same phage by, respectively, abortive infection or CRISPR, and partial resistance defence mechanisms. Alternative stable states and regime shifts greatly complicate the task of manipulation of microbial communities. We propose and study a successful control strategy via short population pulses aimed at inducing the desired regime shifts. In particular, we predict that a fast-growing pathogen could be eliminated by a combination of its phage and a slower-growing susceptible host.

[1] S Maslov, K Sneppen, mSystems (in press), https://doi.org/10.1101/797456   

Biophysical consequences of sublethal antibiotics on gut bacterial persistence and transmission

Antibiotics induce large changes in the composition of the gut microbiome even at sublethal concentrations, by mechanisms that have remained elusive. Using larval zebrafish, which allow controlled studies of microbial dynamics, we found that sublethal doses of the common antibiotic ciprofloxacin cause severe drops in bacterial abundance. Contrary to conventional expectations, disruption was more pronounced for slow-growing, aggregated bacteria than for fast-growing, motile species. Three-dimensional live imaging revealed that antibiotic treatment promoted physical aggregation of bacteria in both planktonic and cohesive species and increased susceptibility to intestinal expulsion. Intestinal mechanics therefore amplify antibiotics effects on resident bacteria. A biophysical model, reminiscent of models of polymer growth and gelation, describes microbial dynamics and makes testable predictions of aggregation properties. The antibiotic-induced expulsion of live bacteria from the host influences the transmission of microbes to new hosts, which we suggest may contribute to the spread of antibiotic resistance.   

Delayed antibiotic exposure induces population collapse in enterococcal communities with drug-resistant subpopulations

The molecular and genetic causes of bacterial antibiotic resistance are increasingly understood, while less is known how these molecular events influence population dynamics. In this work, we show the dynamics of E. faecalis communities exposed to antibiotics can be surprisingly rich, as increasing population size or delaying drug exposure can promote population collapse. We combine experiments in computer-controlled bioreactors with simple mathematical models to reveal density-dependent feedback loops, coupling growth and antibiotic efficacy of populations with drug-resistant (β-lactamase) subpopulations. With a wide range of behavior- population survival, collapse, or one of two qualitatively distinct bistable behaviors where either small or large populations survive- competing density-dependent effects arise: drug-sensitive cell growth increases while drug-resistant cell growth decreases drug efficacy. We experimentally show how populations receiving immediate drug influx may thrive, while identical populations exposed to delayed drug influx (and lower average drug concentrations) collapse. These results illustrate the spread of drug resistant determinants—even in single-species communities—may be governed by counterintuitive dynamics driven by population-level interactions.   

Predicting microbial community metabolic function from genomic structure

The genes and organisms present in microbial communities determine metabolic flows that drive global nutrient cycles. A primary objective of microbial ecology is therefore to predict the metabolic function of a community from its genomic structure. We approach this prediction problem using denitrification as a model metabolic process. Denitrification is mediated by bacterial consortia that convert nitrate to dinitrogen gas. Using metabolite measurements and sequencing of denitrifying bacteria isolated from local soils, we develop a statistical-empirical approach to predicting community function from genomic structure. We show that for each strain in monoculture, the dynamics of denitrification are parameterized by a consumer-resource model. With some well-defined exceptions, we then find that the metabolite dynamics of simple communities are predictable from monoculture dynamics. This means we need only predict single-strain metabolite flows from genomes to successfully predict community-level metabolism. We solve this by predicting consumer-resource parameters via regression onto the presence/absence of each strain's denitrification genes, thereby providing a complete map from community genomic structure to metabolic function.   

Cross-feeding is not necessarily evolutionarily stable

Cross-feeding is widely observed in microbial communities. Intuitively, cross-feeding provides mutual benefits: one species exports metabolic waste which can be taken up and utilized by another species. However, it remains unclear whether mutually beneficial cross-feeding is evolutionarily stable in the face of mutations, invasion by other species, and the cost of transporting metabolites. To address this question, we constructed a minimum metabolic model of cross-feeding, with trade-offs in enzyme allocation and reversible energy-producing reactions. The population dynamics of this system in a chemostat reveal that cross-feeding is evolutionarily unstable. Instead, we find stable coexistence between a species that exports an intermediate and a species that consumes the same nutrient but processes the intermediate. We show why cross-feeding is unstable for this system, and generalize our results for the coexistence of different consumption strategies to multi-step reactions. Further, we identify a scaling relation for the population ratios of coexisting species at large nutrient supply. In summary, our model reveals that cross-feeding is not necessarily evolutionarily stable, and identifies a mechanism for coexistence based on species polluting their own metabolic environment.   

Percolation transition of pusher-type microswimmers

In this talk I will present the presence of a continuum percolation transition in model suspensions of pusher-type microswimmers. The clusters dynamically aggregate and disaggregate resulting from a competition of attractive and repulsive hydrodynamic and steric interactions. As the microswimmers' filling fraction increases, the cluster size distribution approaches a scale-free form and there emerge large clusters spanning the entire system. We characterize this microswimmer percolation transition via the critical exponents, analytical arguments, and scaling relations known from percolation theory. This finding opens new vistas on microswimmers' congregative processes.   

Biphasic Chemotaxis of E. coli to the Microbiota Metabolite Indole

Bacterial chemotaxis to microbiota metabolites in the GI tract is important in the development of microbial communities. A prominent metabolite is indole, which has received wide attention for its role in regulating a broad range of bacterial phenotypes including virulence. The basis of chemotaxis to indole however, remains poorly understood. With a combination of flagellar motor and FRET assays, we have discovered a time-dependent inversion from a chemorepellent to chemoattractant response to indole. Such an inversion caused a bipartite response – wild-type cells were attracted to regions of high indole concentration if they had previously adapted to indole but were otherwise repelled. Interestingly, the flagellar motor itself mediated a repellent response independent of the chemoreceptors. I will discuss analytical models to explain these and other physiological responses to indole. The spatial filtering of cells by indole is likely important in repelling invaders while recruiting beneficial resident bacteria to growing microbial communities within the GI tract.   

Dynamic motility selection drives population segregation in a bacterial swarm

Ecological models usually take growth rate fitness as the essential driver of population dynamics. However, as a widespread natural phenomenon, population expansion in space (or range expansion) is often governed by both motility and growth. Microbial communities offer unique systems to study how individual’s motility contributes to space competition among heterogeneous microbial populations during range expansion. Here we show that motility heterogeneity can promote the spatial segregation of sub-populations in structured microbial communities via a dynamic motility selection mechanism. Our findings are relevant to microbial stress response and microbial ecology. The results may also provide new insight to range expansion in other biological systems, such as tumor invasion and collective stress tolerance of cancer cells in densely packed environments.   

Continuum modeling of bacterial biofilm development

Biofilms are surface-associated bacterial communities embedded in an extracellular matrix. Vibrio cholerae, a rod-shaped bacterium, forms biofilms starting from a single cell and growing into a three-dimensional structure. Recent advances in single-cell imaging reveal the kinematics of this developmental process with single-cell resolution. To understand the role of mechanics in shaping the biofilm, we develop a multiphase continuum model of biofilm development. We find that the kinetic friction between the growing biofilm and the surface leads to a universal “fountain-like” flow of material within the biofilm, in good agreement with experimental data. Using a phase-field approach, we study how cell-cell and cell-matrix mechanical interactions affect the internal organization of cells in the growing biofilm.   

E. coli Bacteria near "Black Hole"

In microfluidic environment, we create hydrodynamic horizon from which no E. coli bacteria can escape ("black hole") to study the collective behaviors of those organisms under the influence of such background, probing for their strategy to avoid potentially harmful region where part of the population disappears.   

Microbial communities governed by interplay of bacterial interaction and biofilm mechanics

Biofilms are highly structured, densely packed bacterial consortia. Their structures are often explained as the result of social interactions between bacteria, e.g. cooperation and competition. Others concentrate on the role of local mechanics in biofilm formation. These two lines of argumentation are typically treated separately. Here, we show that mechanics and social interactions can be strongly interrelated and their combination can crucially impact biofilm formation and dynamics. Using Vibrio cholerae strains that kill each other on contact, we examine how bacterial antagonism impacts biofilm mechanics, and vice versa. We find that this interplay leads to counterintuitive results. For example, killing produces dead cells, which can prevent subsequent killing, and, thus, stabilizes coexistence.