Earth's great recyclers: Carbon cycling in microbial communities

Earth's great recyclers: Carbon cycling in microbial communities

Three-and-a-half billion years ago, the incessant agitations of organic molecules led to the chemical singularity we now know as life. Since then, life has persisted uninterrupted on Earth. Life's great persistence is partly a result of the constant recycling of nutrients, driven by organisms, to regenerate resources in their local environments.

Nutrient cycling plays a critical role in ecosystems at all scales, from microbial communities to lakes and the entire biosphere. An ecosystem is a region of space that supports life thanks to the interactions among its constituent organisms and the environment. For example, both an entire lake and a single decaying floating tree trunk colonized by insects, bacteria and fungi are ecosystems. These communities are self-sustaining thanks to the complex relationships between its members and with the enclosing environment. Carbon cycling is perhaps the most familiar of nutrient cycles. Most of the carbon is cycled as the result of enormous amounts of carbon compounds being used and converted to CO₂ and similarly large quantities of CO₂ being absorbed and transformed back to carbon compounds. Without carbon cycles, life on this planet would not be possible as we know it. An important challenge in ecology is understanding how carbon cycles emerge from the complex web of ecological processes. However, we lack model ecosystems that can be replicated, manipulated, and quantified in the laboratory, making it challenging to determine how changes in species composition and in the environment impact carbon cycling. 

My doctoral research was guided by two ecological questions related to carbon cycling. First, I wanted to understand how the organisms in an ecosystem self-organize to cycle carbon. To answer this question, I constructed and studied model ecosystems in the laboratory and showed that carbon cycles were sustained. Second, I wanted to understand how the capacity to cycle carbon is influenced by the organismal composition of the ecosystem. I developed a simulation that captures key features of our laboratory ecosystems to help us find a function relating community composition to the carbon cycling rate of the community. 

In a paper published in PNAS, we addressed the first problem by showing that carbon cycling arises in functionally redundant communities. We developed a novel, and cheap method to measure, in real time and with high precision, the number of carbon atoms cycled per day in closed microbial ecosystems (CES) provided only with light. Our model ecosystems were composed of microbes and were materially closed, permitting energy exchange only in the form of light and heat. By imposing closure, we isolated the phenomena of nutrient cycling by necessity, since if the microbes fail to cycle nutrients, the ecosystem ceases to support life. We then assembled CES composed of microbes from nearby prairies to study carbon cycling in a laboratory setting. We also coupled our novel measurement of carbon cycling with existing taxonomic and metabolomic assays. By studying replicate CES that support carbon cycles for at least half a year, we observed variable final community composition, but a conserved set of metabolic capabilities across communities. Furthermore, we found that carbon cycling does not depend strongly on the identity of the microbes present. Therefore, an emergent carbon cycle enforces metabolic, but not taxonomic constraints on ecosystem organization. We thus identified the communities in our self-organized CES as exhibiting functional redundancy. To the best of our knowledge, this is one of the first-reproduced and studied instances of functional redundancy in ecology. 

After demonstrating the formation of carbon cycles in a controlled experiment, we wanted to understand how it depends on the ecosystem from which it emerges. Addressing this second problem required a different approach. The extensive duration of our experiments and the large number of experiments required to survey the effects of changing organismal composition hinted towards adopting a computational approach. We constructed a deterministic model of CES able to capture key features like the emergence of carbon cycling. The model considered the coupled dynamics of resources and consumers in a CES. A library of CES were simulated and regression methods were employed to find a mapping between community composition and carbon cycling rates. Community composition was defined as a binary vector of the presence or absence of the organisms in the CES.

 Our simulations support that functionally redundant communities are mainly composed of organisms that consume resources indiscriminately. We also found that communities mainly composed of organisms with rigid resource-consumption preferences are less amenable to regression methods than the communities with indiscriminate consumers. A venue for future research is understanding how the consumption preference of the organisms impacts the predictive power of the regression methods. Our computational approach serves to inform future experiments in physical closed microbial ecosystems. 

My work establishes CES as reliable models to study persistent carbon cycles and provides one of the few concrete physical instances of functional redundancy in ecology.