UC Berkeley

Microbial residents of soil interact with one another to form communities that serve critical roles in carbon sequestration, agriculture, and nutrient cycling. Soil microbes are abundant – a gram of soil can house billions of cells – and diverse – cells represent thousands of species of archaea and bacteria. However, as new technologies have enabled cataloging a greater portion of which organisms persist and coexist in soil, mapping the network of microbial interactions is the next hurdle for appreciating the critical role of the soil microbiome in terrestrial biogeochemical cycling. From microbial genomes we can predict how organisms live and from culturing microorganisms in laboratories we can see how they behave, but illuminating how microbes act in situ remains the canonical paradox: in order to observe soil microbes we must perturb them. In this dissertation, multiple genome resolved ‘omic approaches are combined to study how microbes may interact in soil. Each chapter focuses on a distinct type of interaction and the cascading effects that interaction has on soil microbial communities at-large.

The first chapter reviews the varied microbial interactions that exist in soil: from contact-dependent molecular transfers to interactions at a distance that depend on molecules moving through soil. Here, I introduce the framework that a subset of interactions can be inferred in an organism’s genome. Reading genomes provides the basis for appreciating an organism’s nutritional requirements which can be used to predict how and with which organisms act in community.

In the second chapter, I turn to often overlooked ultrasmall microbes in soil from the Candidate Phyla Radiation (CPR) bacteria and DPANN archaea (an acronym of the first included phyla: Diapherotrites, Parvarchaeota, Aenigmarchaeota, Nanoarchaeota and Nanohaloarchaeota) lineages. We affirm that, in part, these organisms are overlooked because they comprise the soil rare biosphere and that we were able to capture these genomes by optimizing previous methods used to establish viral-enriched metagenomes. We found that, as in other systems, these organisms must acquire much of their essential building blocks from neighboring cells, suggesting they require living in close association with other organisms perhaps as epibionts or parasites. We also determined that soil CPR appear to harbor a likely local adaptation to the oxic environment of top soils – the capacity for some form of aerobic respiration.

Moving into the third chapter, I turn attention to understanding the role of viruses in the soil microbiome. Viruses, like other soil microbes, are ubiquitous and recently have become better cataloged, but a comprehensive understanding of the ecological role of these specific and diverse predators has not yet emerged. I examined a highly dynamic moment in soil where carbon is mineralized at a disproportionately high rate compared to the rest of the year – the first rainfall in a seasonally dry annual California grassland. This system has long been fruitful for relating microbes directly to the turnover of soil carbon. Thus, we probed the role of viruses in soil and how viruses may contribute to soil carbon turnover via the characteristic resuscitation and mortality of specific microbial lineages immediately following wet-up. We combined metagenomics with stable isotope probing (SIP) using H218O and found that dry soil is a sparse yet diverse reservoir of putative virions, of which only a subset thrives following wet-up. These active viruses appeared to closely follow host population dynamics.

In the final chapter of this dissertation, I focus on a model nutrient and its rippling effects on the microbiome, shifting focus from a specific type of soil microbe. Turning to corrinoids – cobalt-containing enzymatic cofactors used in a variety of carbon and nitrogen metabolisms by most organisms, typified by Vitamin B12 – we analyzed how varying types of this non-assimilated nutrient could differentially shift composition and function of soil microbiomes both in soil and in soil-derived enrichment cultures. We observed in soil that the addition of distinct corrinoid types led to differences in community composition, but that this effect was more muted in soil-enrichment cultures. Using metagenomics and metatranscriptomics on soil enrichment cultures, we found that some organisms showed specific differences in the number of genes differentially expressed for certain corrinoid types over others, while other organisms showed a non-specific response to distinct corrinoid types, but these expression changes only corresponded to subsequent changes in an organism’s abundance in a few instances.

Overall, the results presented in this dissertation use a variety of sequencing methods and lenses to tease apart the types of interactions that cohere the soil microbiome. These interactions illustrate both the highly physically connected nature of the soil microbiome and the simultaneous diffuseness of these connections. CPR bacteria and DPANN archaea often live physically associated with other cells. Similarly, viruses are only alive inside host cells, but may also persist in soil awaiting a physical association with a host – perhaps facilitated by rain. Finally, soils maintain a stock of corrinoids, but changes in the soil corrinoid profile may preferentially target growth of certain bacteria. Through unraveling how microbes interact we will advance our understanding of critical soil processes.