Review

The global ocean microbiome

See allHide authors and affiliations

Science  11 Dec 2015:
Vol. 350, Issue 6266, aac8455
DOI: 10.1126/science.aac8455

The ocean microbial system

The vast translucent oceans are teeming with microscopic life that drives significant life processes and elemental cycling on Earth. Yet how climate change will affect the functioning of this microbiome is not well understood. Moran reviews progress and the transformative discoveries made recently in marine microbiology that have led environmental, plant, animal, and even human microbiome research.

Science, this issue p. 10.1126/science.aac8455

Structured Abstract

BACKGROUND

Oceanographers began studying the ocean microbiome in earnest over four decades ago, when it was recognized that microbes are responsible for nearly all of the energy flux in this largest and most dilute biological system on Earth. Much has been learned about the microbes that play key roles in every marine element cycle, but much is still unknown about the factors regulating their activity. Although the number of marine microbes per liter of seawater reaches into the billions, their small size means that, statistically, each microbe is separated by 100 to 200 body lengths from its closest neighbors. Yet recognition of microscale structuring of both microbial communities and marine organic matter suggests that the ocean microbiome does not operate as stand-alone cells in a watery soup.

ADVANCES

Several decades of 16S ribosomal RNA gene analysis has revealed distinct and recurring bacterial communities in the ocean. More recent characterizations of marine archaea, protists, and viruses are filling out the taxonomic inventories of the ocean microbiome and showing that membership is predictable over seasons, ocean depth, and organic matter features. The retrieval of proteorhodopsin—a gene that allows cells to harvest energy from sunlight without complex photosynthetic machinery—from an uncultured ocean microbe marked the first exciting discovery from the use of “meta-omics” methodologies in the ocean. Now these techniques are the central tools for converting inventories of organisms and functions into explicit linkages between the two. Substantial progress has been made toward unraveling how and where microbes participate in ocean biogeochemical processes, as well as toward recognizing new categories of nonpredatory microbial alliances that operate based on the exchange of nitrogen, vitamins, hormones, and antibiotics.

Several characteristics of the ocean microbiome distinguish it from microbiomes on or in animals, plants, and soils. First, the primary producers that fuel the ocean are exclusively microbial and thus are a part of the microbiome. This is the case for photosynthesis in the surface ocean and for chemosynthesis carried out in deeper waters. The ocean microbiome is responsible for half of all primary production occurring on Earth. Second, trophic categories are particularly difficult to assign in the ocean microbiome, with no clear division of organisms into canonical autotrophic and heterotrophic roles. Proteorhodopsin, anoxygenic phototrophy, and chemolithotrophic energy acquisition from inorganic compounds create trophic mayhem among members of the ocean microbiome. Having multiple strategies for meeting metabolic requirements may be an advantage in this chemically dilute and physically dynamic environment. Last, heterogeneity in the structure of seawater organic matter has become a foundational concept for the ocean microbiome because it aligns with differences in microbial attributes. Bacteria and archaea that live singly in seawater differ from those that intermingle on the various marine polymer networks and organic surfaces in terms of phylogenetic affinity, metabolism, and capabilities for motility, chemotaxis, and defense. Single bacteria and archaea are numerically dominant in terms of cells, genes, and transcripts, but those clustered near surfaces have higher per-cell rates of metabolism and growth. The importance of material exchanges and signaling networks between neighboring cells in the ocean, as well as the consequences spatial arrangements impose on biogeochemical processing, are not yet understood.

OUTLOOK

Earth’s changing climate is predicted to decrease carbon fixation by microbial primary producers, favor smaller picophytoplankton over larger nano- and microphytoplankton, and impose stress on photosynthetic microbes that form calcium carbonate shells. The structure of phytoplankton communities, in turn, has implications for the abundance and composition of organic substrates for heterotrophic microbes, as well as for dictating which trophic strategies will be under selection in the future ocean. Taking stock of the ocean microbiome in terms of cells, genes, transcripts, and proteins now has a long tradition in oceanography. Linking these stocks with the regulation of critical ecosystem functions is the next challenge. One key step in this process is the identification of the molecules that pass between microbes as substrates, nutrients, signaling molecules, and defensive compounds; these are the “currencies” of ocean microbiome function.

Sampling the ocean microbiome.

(Left) New instruments such as the Environmental Sample Processor (ESP) (Monterey Bay Aquarium Research Institute) autonomously sample ocean microbes and environmental conditions while deployed at sea. (Right) Microbial cells preserved in metal pucks inside the ESP are removed for gene expression analysis.

PHOTOS: (LEFT) J. MOTARD-CÔTÉ; (RIGHT) A. BURNS

Abstract

The microbiome of the largest environment on Earth has been gradually revealing its secrets over four decades of study. Despite the dispersed nature of substrates and the transience of surfaces, marine microbes drive essential transformations in all global elemental cycles. Much has been learned about the microbes that carry out key biogeochemical processes, but there are still plenty of ambiguities about the factors important in regulating activity, including the role of microbial interactions. Identifying the molecular “currencies” exchanged within the microbial community will provide key information on microbiome function and its vulnerability to environmental change.

The ocean microbiome is a highly dilute microbial system that covers the majority of Earth’s surface and extends an average of 3600 m down to the seafloor. As one of the first microbiomes to be studied, the diversity and distribution of its members is now becoming familiar. Key questions remain, however, regarding how functional capabilities are distributed and what biotic and environmental factors control their rates.

A consistent link is emerging between ocean temperature and both the composition (1, 2) and productivity (3) of microbes inhabiting surface seawater. Stratification of the ocean into more discrete layers and consequent reduction in the mixing of nutrient-rich deeper waters to the surface may be one key mechanism behind this link (3). Decreases in ocean pH, a direct chemical outcome of increased absorption of CO2 from the atmosphere, affect the energetic costs to microbes of obtaining inorganic carbon for photosynthesis and the availability of CaCO3 for forming calcareous shells (4). To expand our evolving knowledge of the ocean microbiome, recent work has been directed toward explaining how the resident microbes will be affected by the predicted impacts of climate change on the ocean ecosystem.

In 1974, the idea that microorganisms are the major consumers of energy in the sea was formally articulated (5). Marine microbes <60 μm in diameter had gradually been recognized as responsible for nearly all of the primary production and respiration occurring in the ocean. This “new paradigm” proposed four decades ago marked the first conceptualization of the ocean microbiome, a community of microorganisms inhabiting 71% of Earth’s surface (6) and taking part in every one of its major elemental cycles.

Despite population numbers reaching into the billions per liter of seawater, marine bacteria in today’s ocean are separated from their closest neighbor by an average of 200 body lengths, due to their small size. The larger but less abundant microbial eukaryotes are separated by an average of ~100 body lengths (7). Early views of the ocean microbiome were therefore premised on the idea that interactions between members were indirect, except in predator-prey relationships. Microbes added and removed organic and inorganic materials from the pool of nonliving matter in seawater, and their interactions were mediated largely through the compounds in this pool. Yet the groundwork for recognizing the presence of structured microbial communities, even in dilute ocean waters, had already been laid (8). Bacteria were shown to exhibit chemotaxis to phytoplankton exudates (9) and to gain energy benefits when clustered around patches of concentrated organic matter (10). These were early hints that the ocean microbiome does not operate as stand-alone cells in a watery soup.

The organic matter continuum

The microscale structure of organic materials in seawater that is relevant at the microbial scale, although daunting in its complexity, is now better understood. At the low end of the size range, seawater harbors a pool of dissolved organic molecules released from plankton or leaked during predatory interactions. There are tens of thousands of different structures of these dissolved compounds (11), an overwhelming diversity whose origin and fate are only beginning to be understood (12, 13). Intermediate structures include three-dimensional polymer networks, or “microgels” (14), that can serve as loci for microbial clustering (15), followed by “macrogels” (larger polymer networks, often dominated by polysaccharides). At the upper end of the size range are eukaryotic plankton cells, both alive and dying. Although eukaryotic microbes typically are not heavily colonized by bacteria or archaea when healthy (16), the larger cells create zones of concentrated dissolved organic matter extending from their surfaces—these zones are referred to as “phycospheres” (9) (Fig. 1). Fecal pellets and other remnants of multicellular organisms also contribute to the organic matter matrix.

Fig. 1 Microscale structure of the ocean microbiome.

The ocean microbiome is dispersed across a dilute matrix of organic matter, populated by bacterial, archaeal, and eukaryotic cells using complex trophic strategies and fueled by microbially fixed carbon. Molecules, gels, and a living diatom represent the organic matter matrix of seawater. Dashed lines indicate predatory or nonpredatory interactions between microbiome members.

The heterogeneity in the structure of seawater organic matter is a foundational concept for the ocean microbiome because it aligns with differences in composition, metabolism, and genome content of the member microbes. Bacteria and archaea that live singly in the ocean passively encounter small molecules and efficiently scavenge them at low concentrations (17). Those that intermingle with the various types of polymer networks and organic surfaces have better access to substrates but more neighbors to contend with compared to the free-living cells (Fig. 1). Along with differences in phylogenetic affiliation (18) and metabolic capabilities (19, 20), these bacteria and archaea, compared to their free-living counterparts, have genes biased toward capabilities such as signal transduction, defense, anaerobic metabolism, and carbohydrate processing (21, 22) and use motility and chemotaxis to find substrate-rich plumes and filaments associated with the organic matrix (23, 24). In some ways it is surprising that these differences consistently emerge, despite the fact that bacteria and archaea are quite coarsely assigned their place in the organic matter continuum based on passage through a filter: Those passing through a pore of a few micrometers in diameter (typically 2.0 μm) are considered free-living, whereas those retained are considered matrix-associated. Eukaryotic plankton are also divided across this size-based boundary. Small eukaryotes (picoeukaryotes) are collected with the free-living bacteria and archaea by virtue of their small size (25). Larger eukaryotic microbes (nano- and microeukaryotes) are collected with matrix-associated bacteria and archaea.

How do inventories and activities compare between microbes associated with different organic matter features of the ocean microbiome? As a general rule of thumb, free-living bacteria and archaea account for more cells (~80%) and more genes and transcripts per volume of seawater (26) than those that are matrix-associated (Fig. 2). Yet per-cell rates of metabolism and growth (27, 28) coincident with per-cell inventories of genes and transcripts (26, 29) are higher in the organic matrix-associated bacteria and archaea. Cyanobacteria and picoeukaryotes, minute enough to pass through a 2.0-μm filter pore, are the smallest photosynthetic organisms of the ocean. In seawater, picoeukaryotes are typically an order of magnitude fewer in number than cyanobacteria, but they contribute more biomass per cell (30) and can produce nearly equivalent numbers of transcripts (Fig. 2). The larger eukaryotes are generally most abundant where nutrient levels are high, such as in coastal oceans or upwelling regions, and can sometimes rival heterotrophic bacteria in their contribution of microbial genes and transcripts to seawater (Fig. 2).

Fig. 2 Organisms and “omics.”

(A) Typical inventory of microbial cells and (B) microbial genes and transcripts per liter of ocean water (26, 29).

The assemblage

Early studies typically used the word “assemblage” to refer to microbes in seawater; this term indicates that encounters might be too infrequent and random to be considered a structured ecological community. Eventually, distinct and predictable microbial communities were identified in the ocean, enabled by new methodologies that did not depend on cultivation. In the first application of a 16S ribosomal RNA (rRNA) gene amplicon surveying to a natural ecosystem, three previously unidentified bacterial lineages were discovered among just 12 sequences amplified from Sargasso Sea surface waters (31). The SAR11 lineage was uncovered in this pioneering survey and is now recognized as the most abundant bacterial group in the world’s oceans (32, 33). Several decades of 16S rRNA gene analysis have made us quite knowledgeable about the identity, relative abundance, and patterns of occurrence of marine bacteria and archaea (2, 34, 35). Most marine microbial communities are composed of few dominant species and a long tail of rare ones (36), and species are divided into coexisting populations that imply fine-scale dissection of available resources (37, 38). Latitudinal patterns in species richness similar to those for macroorganisms have been observed (1), and community composition is predictable over time (39), depth (40), and in relation to the structure and composition of seawater organic matter (18, 4143).

More recently, attention has turned to the taxonomic characterization of protists [recognized in 1974 as important in ocean microbial communities (1)], viruses [joining the ocean microbiome in 1989 (44, 45)], and archaea [joining in 1992 (46, 47)]. A recent comprehensive accounting of ocean microbes by the Tara Oceans expedition found 37,000 bacterial and archaeal species [based on 16 million metagenomic 16S rRNA genes (2)], 100,000 protist groups [based on 580 million amplified 18S rRNA genes (48)], and 5500 viral populations [based on 2 million metagenomic sequences and considering only double-stranded DNA bacterial and archaeal viruses (49)]. If each sample in this survey is representative of a 0.25-km3 homogeneous parcel of seawater (35), we now have a snapshot in time of a 26-millionth of the ocean microbiome.

Functions in the ocean

A gene retrieved from an uncultured coastal ocean bacterium in 1999 challenged perceptions about energy acquisition by the ocean microbiome (50). This gene encodes a rhodopsin protein—belonging to a family previously known only from microbes living in extreme environments—that allows cells without photosynthetic machinery to harvest energy through sunlight-driven proton pumping. The discovery overturned a fundamental distinction between microbes obtaining energy from sunlight versus from organic matter and was a “tip-of-the-iceberg” moment that hinted at other notable microbial physiologies yet to be discovered. Before this study, the main strategy for linking microbes to their roles in the ocean was through “guilt by association,” with functional roles of organisms known only by their 16S rRNA gene sequences inferred from the physiology of their cultured relatives (if they existed). But the discrepancies in evolutionary history between the highly conserved phylogenetic marker genes and the highly diverse accessory genes (more easily shaped by forces of selection, drift, and horizontal transfer) made this an unsatisfying approach. Further, it confined progress in functional understanding to physiologies known from a narrow suite of cultured microbes.

The proteorhodopsin discovery emerged from the idea to clone random fragments of bacterial DNA from seawater (51) and then sequence the regions flanking 16S rRNA genes (50). Eventually, capabilities for this proteorhodopsin-based energy transduction were found in >40% of bacterial cells in the surface ocean (52), representing all major heterotrophic bacterial lineages (53), as well as in marine archaea (54), microbial eukaryotes (55, 56), and even viruses (57). The fact that such a fundamental energy-acquisition mechanism had gone unnoticed for decades was equally sobering and exhilarating.

Metagenomics and other “meta-omics” methodologies thus became central tools for turning disjointed inventories of organisms and functions into explicit linkages between them. For example, our understanding of nitrogen controls on marine primary production was advanced when a widespread unicellular N2-fixing cyanobacterium, known only by its nitrogenase gene sequence, was successfully hunted down (58). Metagenomic sequencing of an enrichment of these elusive UCYN-A cells revealed a surprising cyanobacterial genome lacking genes for photosystem II and carbon fixation and therefore represented a novel category of ocean microbial function that potentially decoupled carbon and nitrogen fixation (58). Similarly, characterization of the marine bacteria responsible for metabolizing a sulfur-containing phytoplankton osmolyte (dimethylsulfoniopropionate) began with gene discovery efforts in cultured marine strains (5961), followed by metagenomic data mining (52, 59) and then autonomous ocean sensing (62). Regulation of these bacterial genes affects the primary source of biogenic sulfur emissions to the atmosphere (63) and influences sulfur cycling on a global scale (64, 65).

The Global Ocean Sampling data set released in 2004 consists of 1 gigabyte of metagenomic sequence data from the North Atlantic Ocean (66); the Tara Oceans data set released in 2015 consists of 7200 gigabytes of metagenomic data from multiple locations and depths across the globe (2). Our knowledge of marine microbial function has been rapidly propelled by the availability of these and other marine meta-omic data sets (67, 68). They provide windows into the multitude of organisms, genes, and transcripts of the ocean microbiome that run the global elemental cycles, buffer atmospheric CO2 concentrations, and are on the front line of Earth’s changing climate (3, 69).

Neighborhood associations

The assemblage perspective of early researchers has given way to a growing understanding of the diverse mechanisms by which ocean microbes interact. Even though microbial cells occupy only ~0.0001% of seawater volume (70), predatory activities are surprisingly efficient. Between 20 and 50% of bacterial cells are killed each day in the ocean by viral infection (71), and a similar amount are consumed by protist grazers (72). Microbial predators and prey are engaged in sophisticated evolutionary arms races; for example, cyanobacterial viruses carry copies of photosynthesis genes, originally captured from their hosts, that ensure a sufficient number of light-harvesting proteins during later stages of infection (73). Whole new categories of nonpredatory microbial alliances are also being recognized. These include bacteria that release a hormone promoting phytoplankton cell division (74); phytoplankton that synthesize novel organic molecules that only certain bacterial species can use as substrates (75); nitrogen-fixing cyanobacteria that are endosymbiotic (76) or episymbiotic (77) with eukaryotic hosts; ciliates that borrow functional chloroplasts from their otherwise-digested prey (78); microbes that rely on neighboring cells for essential vitamins and defensive enzymes because they have lost their own capability for synthesis (79, 80); and bacteria that attack neighboring microbes through the production of algicides (81), antibacterial compounds (82), or both (83). At the level of community assembly, there is evidence of patterns in microbial composition and function over time frames of tidal cycles, days, and seasons (40, 84, 85). The limitation of relying on filters to distinguish microbial roles in the ocean microbiome is now being complemented by methods that track individual cells moving through seawater (86), quantify metabolite exchanges between two microbes (87), and sequence parasite or symbiont genomes inside a single host microbial cell (88). Our appreciation of the richness of microbial interactions in the ocean and their relevance to microbial function is in the very early stages.

An unusual microbiome

Several combined features of the ocean microbiome distinguish it from those on or in animals, plants, and soils. First, the organisms responsible for fueling the system through primary production are exclusively microbial and, thus, part of the microbiome rather than external to it. This is the case for photosynthesis orchestrated by bacteria and eukaryotes in the surface ocean and also for chemosynthesis carried out by archaea and bacteria (89). The ocean microbiome is responsible for half of all primary production occurring on Earth (90).

Trophic categories are particularly difficult to assign in the ocean microbiome (Fig. 3). The division of organisms into strict autotrophic and heterotrophic roles began to unravel with the realization that proteorhodopsin allowed otherwise heterotrophic microbes to harvest light energy (91). It continued with the discovery of anoxygenic phototrophic bacteria in the ocean that acquire energy from sunlight, using bacteriochlorophyll a and associated photosynthetic reaction centers. However, unlike photosynthetic microbes, these bacteria are unable to fix CO2 (9294). Other ocean microbes are able to obtain energy from the oxidation of both inorganic compounds, such as carbon monoxide and sulfide, and organic compounds (95). One hypothesis to explain the choreographed trophic mayhem of the ocean microbiome is that variability in organic matter and light, as microbes are mixed into and out of surface waters and sink to deep ocean layers, creates an advantage for organisms with numerous metabolic mechanisms. Another is that very low concentrations and supply rates of the individual organic resources in oligotrophic seawater select for microbes that can take advantage of many different energy and carbon sources.

Fig. 3 Trophic categories of the ocean microbiome.

The mechanisms by which microbes of oxygenated seawater obtain energy and carbon are grouped in unconventional ways in the trophic hodgepodge of the ocean microbiome (50, 55, 93, 95, 107, 108).

Finally, because the majority of microbes in the ocean microbiome are free-living, only those cells populating the organic matter matrix experience physical interactions at the scale typical of most other microbiomes (Fig. 1). Questions regarding how close microbes need to be to establish networks of signaling and material exchange (9698) and what consequences spatial arrangements impose on biogeochemical processing (24, 99, 100) are still awaiting answers.

Earth’s changing climate will affect characteristics of the ocean microbiome. Strengthening of stratification under a warming climate is widely predicted to decrease carbon fixation rates of the ocean’s microbial primary producers (67) and to favor picophytoplankton over larger nano- and microphytoplankton because the former are better competitors for nutrients (101, 102). Members of the phytoplankton community that form calcium carbonate shells are anticipated to be negatively affected by increasing seawater pH (103). In turn, phytoplankton composition has implications for the abundance and composition of the polymer networks and organic surfaces that are the substrates for heterotrophic microbes (104). Picophytoplankton may be of insufficient size to create phycospheres detectable by chemosensory mechanisms (105), and their dominance would favor free-living over matrix-associated bacteria and archaea (104). Certainly, the intricate trophic schemes of marine microorganisms will be subject to redesign in the future ocean, with changes in surface seawater temperature (95), CO2 concentrations (96), O2 concentrations (95), nutrient regimes (68), and light availability (97) shifting the physiological and ecological payoffs of a microbe’s life history strategy.

Conclusions

The ocean microbiome fits the definition of an ecological community of microorganisms that share an environment (106). As is true for all microbiomes, it is considerably easier to inventory microbial parts (cells, genes, transcripts, proteins, regulatory molecules) than to link the parts with their functions in the community or to understand the give-and-take relationships with a shifting environment. Identifying the molecules—whether as substrates, nutrients, signaling molecules, or defensive compounds—that pass between microbes is an emerging area of microbiome research that will focus attention on the “currencies” of microbial activity. Ultimately, the flux of materials, both among microbes and between microbes and their surroundings, constitutes microbiome function and determines its response to external changes.

References and Notes

  1. Body length calculations assume 1 × 109 bacterial cells per liter of 0.5-μm average diameter and 1 × 106 microbial eukaryote cells per liter of 100-μm average diameter.
  2. The percent of seawater volume occupied by microbial biomass was estimated by assuming average volumes of 0.1, 2, and 550 μm3 for marine heterotrophic bacteria, cyanobacteria, and protists, respectively. Average cell numbers are assumed to be 1 × 109, 1 × 107, and 1 × 106 cells per liter, respectively.
  3. Acknowledgments: J. T. Hollibaugh, K. Ross, M. Landa, A. Burns, and B. Nowinski provided helpful discussions and C. English assisted with graphics. This work was supported by NSF grants OCE-1356010 and OCE-1342694 and, in part, by the Gordon and Betty Moore Foundation through grant 538.01.
View Abstract

Navigate This Article