Unveiling core ecological interactions in marine microbial communities using omics approaches: INTERACTOMICS.

1.1. The microbial ocean

Microbes are present in all imaginable habitats, being responsible for a myriad of chemical reactions that directly or indirectly shape and maintain our current environment¹. This should not be restricted to a perception of our external environment, microbes are also a crucial component of animals and plants. As the greatest part of the evolutionary process occurred at micro-scales, most of the diversity of life today is found across microbial lineages. Therefore, while a number of metabolic pathways prevail among macro-organisms, microbes present a stunning repertoire of metabolisms. In the oceans, microbes are the dominant organisms², playing key roles in the biogeochemical cycles that move chemical elements through biotic and abiotic compartments contributing significantly to the functioning of the ecosystem. As the oceans are the largest biome on Earth, covering about 70% of its surface, the great importance of oceanic microbes for the functioning of the biosphere is evident. Our collective awareness of such importance has increased steadily during the last 40 years, progressing in concert with the development of new technologies that allowed us to dig deeper into the microbial world. During this period, oceanic microbial ecology gained increasing momentum, becoming today a large, and very dynamic research field. A number of key events could be identified throughout this period and were promoted by a revolution that started about 25 years ago, with the introduction of molecular approaches to marine microbial ecology. This revolution has generated a wealth of new knowledge, and during the last eight years it has gained renovated momentum with the advent of high-throughput sequencing platforms³, and a variety of other high-throughput omics technologies. High-throughput sequencing, especially 454 and Illumina platforms, has been a key addition to the molecular toolbox of microbial ecologists, allowing a deeper exploration of microbial diversity, distributions and metabolic function. Questions that were impenetrable a decade ago are now being addressed and answered to certain extent, such as the magnitude of microbial diversity and its spatiotemporal patterns. In addition, the introduction of High-Throughput omics approaches and the generation of big-data is encouraging microbiologists to switch from a classically reductionist perspective, concentrating in one microbial group, species or metabolism, and embark into challenging studies aiming to understand the functioning of increasingly complex eco-systems and the effects of marine microbes activities (feeding, respiring, growing, excreting) on the physicochemical conditions of their environment.

1.2. Moving towards oceanic eco-systems biology

Despite the overall progress, it is fundamental to start appreciating the global ocean as a large integrated system that regulate various processes in the biosphere. To begin moving into this direction we need at least three types of ecosystem data:

1) list of components (species),

2) connectivity between these components (interactions),

3) spatiotemporal context for the connectivity⁴.

The use of High-Throughput sequencing in genomics, transcriptomics, metagenomics and metatranscriptomics has contributed substantially to point 1, increasing our knowledge about species in oceanic microbial communities as well as their single and combined metabolisms. In addition, most microbes are unculturable, so for most of the known marine species, we do not have access to their genomes, gene-expression patterns or metabolisms using standard techniques. We know even less about point 2: connectivity, which is crucial for understanding how marine ecosystems function. In ecosystems, connectivity may refer to classic ecological interactions (e.g. competition), structural interactions (e.g. one microbe providing physical support for another) as well as metabolic interactions among organisms (e.g. metabolic cooperation)⁴. Ecological interactions in microbes can be negative (antagonistic), positive (synergistic) or neutral for the species involved, and these outcomes can be used to classify interactions^5-7. Competition is the classic example where both species involved are negatively impacted by the interaction, while mutualism is its counterpart, bringing benefits for both. Parasitism and predation represent typical examples of win-loss interactions, where one species gets all the benefits and the other assumes all the costs. Commensalism and amensalism are opposites too, with commensalism bringing benefits to one species at no cost for the other, while in amensalism, one of the interacting species is negatively impacted by another one with no benefits to itself. Furthermore, cooperation between species (mutualism) can be divided into facultative cooperation (protocooperation) or obligate (symbiosis). Cross-feeding (syntrophy) is a good example of protocooperation, with microbes forming transient consortia. Microbial interactions may require physical contact between the cells (direct interactions) or not (indirect interactions). Within microbial communities, indirect interactions can be important, and these would include, for example, microbes releasing to the environment toxic chemicals that inhibit the growth of others (allelopathy, a form of competition), the production of ethanol by yeasts that inhibit the growth of several bacteria in fermenting fruit juices (amensalism), and microbes that release to their surroundings substances that are by-products of their metabolisms which are essential for other members of the community (e.g. vitamins; commensalism). Despite the examples and literature on microbial interactions are significant, we are still very far from a general understanding of microbial interactions in nature, which represents a major knowledge gap in ecology.

In general, most studies are typically focused in single specific relationships, thus providing no insight on the complex network of ecological interactions occurring in microbial communities that altogether sustain ecosystem functioning.

1.3. Understanding oceanic microbial eco-systems as interaction networks

The natural approach to start investigating microbial interactions in the ocean was to look at feeding habits, and it was not complicated to make simple diagrams between functional groups. In a nutshell, a classic microbial food chain in the sunlit ocean is composed by primary producers that generate organic matter through photosynthesis, and a large number of heterotrophic species that feed, directly or indirectly, on the organic carbon produced by them⁸. A key alternative carbon and energy pathway is the so-called microbial loop, in which prokaryotes absorb dissolved organic carbon from the environment and reintroduce it into the classic food chain through heterotrophic flagellates (eukaryotes) that prey on them. Viruses attack and eventually kill several members of this community releasing nutrients, complex molecules and organic matter that is either recycled in the sunlit community or exported to deeper water. All the players in this food web contribute to the pool of dissolved organic carbon.

This simple model has been undoubtedly useful to start understanding the functioning of marine microbial communities, but it should be upgraded into a much more detailed description of reality that integrates species and their individual interactions. During the last 20 years, macro ecologists have started to substitute the classic linear food-chain models by highly-interconnected food webs (Figure 1), which reflects the natural evolution of a field that has been studying food chains for decades¹⁰. A network is much more than just visualization tool; the architecture of the network, or topology, reflects community self-organization processes (e.g. how incoming members interact with pre-existing in a community) and has implications for ecosystem functioning. For example, networks presenting few highly connected species (nodes) instead of an even distribution of connectivity among species are more robust to random removal of species; yet, if highly connected species are successively lost, the network may break into disconnected sub-webs⁹. Thus, without understanding the network of species interactions in oceanic microbes, we cannot estimate realistically the effects of extinctions or invasions, which is pivotal in a context of global change.

Figure 1. Ecosystems include complex networks of interacting species. Here, species are the nodes, while the links represent ecological interactions. Nodes with several links are hubs.

Unlike the progress observed in macro ecology, our understanding of microbial communities as networks of interacting species is still on its infancy. Overall, network research has attracted growing interest among marine microbial ecologists, and several researchers acknowledge its great potential for increasing our understanding of oceanic ecosystems. Yet, the main obstacle for applying the network approach has been the technological limitation for identifying 1) all network components (microbial species) and 2) their interactions (connectivity). Some microbiologists have started using networks based on analysis of taxonomic correlations over time (co-occurrence or co-exclusion) derived from standard community rDNA data. So far, studies are normally restricted to interaction predictions based on individual (cell) abundances and confirmation of just a few of the predicted interactions. This project proposes to go beyond the state of affairs, by aiming to corroborate a larger proportion of the predicted interactions using single-cell genomics (Figure 2) as well as metagenomics and metatranscriptomics. The latter is possible thanks to the recent advance in different fields, such as high-throughput microfluidics, single cell genomics and high-throughput sequencing as well as in several other minor omics approaches. Single cell genomics will be fundamental to this project, providing extra and extremely valuable evidence on certain ecological interactions such as parasitism, symbiosis or predation, following the basic premise that when a single cell is isolated, it will carry with it its symbionts, pathogens, and sometimes prey.

1.4. Microbial interactions can also be investigated from a purely metabolic perspective

A “reverse ecology” approach can be used to infer pairwise ecological interactions between specific microbes based on their cellular metabolisms inferred in silico from whole genome data. For example, complementarity in metabolite requirement and production in different microbes pointed to host-parasite interactions¹¹. A similar strategy has been used to predict competition and mutualisms in bacteria¹². Alternatively, metagenomic approaches can be applied to investigate the functioning of the whole community as an integrated metabolism, analysing functions that are distributed among interacting species and identifying those taxa that carry out specific functions for the whole community. Such cooperative interactions could be more important than previously thought, determining community assembly and promoting the co-occurrence of taxa that exchange metabolites, particularly in nutritionally challenging habitats¹³. Cooperative interactions do not necessarily need physical contact, and therefore metagenomics of samples including important co-occurring taxa, together with relevant Single Amplified Genomes (SAGs) (Figure 2), can provide evidence that cells are interacting.

Figure 2. Single Cell Genomics workflow designed for this project.

1.6. References

1 Falkowski, P. G. et al. Science 320, 1034-1039, (2008).

2 DeLong, E. F. Nature 459, 200-206, (2009).

3 Logares, R. et al. J Microbiol Methods 91, 106-113, (2012).

4 Raes, J. & Bork, P. Nat Rev Microbiol 6, 693-699, (2008).

5 Faust, K. & Raes, J. Nat Rev Microbiol 10, 538-550, (2012).

6 Lidicker, W. Z. A. Bioscience 29, 475-477, (1979).

7 Leadbetter, E. R. & Poindexter, J. S., Vol. 2 (Springer-Verlag, 1986).

8 DeLong, E. F. & Karl, D. M. Nature 437, 336-342, (2005).

9 Montoya, J. M. et al. Nature 442, 259-264, (2006).

10 Pimm, S. L., (Princeton University Press, 1982).

11 Borenstein, E. & Feldman, M. W. J Comput Biol 16, 191-200, (2009).

12 Freilich, S. et al. Nature communications 2, 589, (2011).

13 Zelezniak, A. et al. Proc Natl Acad Sci U S A 112, 6449-6454, (2015).