Keywords

1 Marine Microbial Ecology: Opening the Black Box

1.1 Major Breakthroughs before the -Omics Revolution

As Sydney Brenner, Nobel Prize in Physiology or Medicine 2002 said, “Progress in science depends on new techniques, new discoveries and new ideas, probably in that order.” The field of marine microbial ecology has been no exception to this quote as it has seen major breakthroughs in the last 60 years following the application of new technologies. For example, the use of epifluorescence microscopy for the estimation of bacterial abundance unveiled that traditional plate counting methods were underestimating the real values by several orders of magnitude (Jannasch and Jones 1959) because most bacterial species do not form colonies on solid media (what was later coined as the “Great Plate Count Anomaly”, Staley and Konopka 1985). Likewise, epifluorescence microscopy was also crucial to uncover the presence of abundant and ubiquitous unicellular cyanobacteria in the surface ocean (Johnson and Sieburth 1979). Initial studies using radioisotopes to estimate respiration and primary production also revealed that much of the respiration, dissolved organic matter (DOM) processing, and primary production were carried out by picoplanktonic organisms (Li et al. 1983; Williams 1970, 1981). The application of techniques commonly used in the biomedical field like flow cytometry also constituted a tipping point in our understanding of marine microbial ecology. For example, it allowed the discovery of Prochlorococcus, the most abundant primary producer in the ocean (Chisholm et al. 1988), and enabled the delineation of two populations of heterotrophic bacteria based on their nucleic acid content (Li et al. 1995; Robertson and Button 1989). This separation in the DNA content was initially attributed to a difference in the physiological state with the high nucleic acid (HNA) cells representing the active fraction of the community (Lebaron et al. 2001; Servais et al. 2003). However, this view was later challenged when it was shown that low nucleic acid (LNA) cells were also important contributors to bacterial metabolism in the sea (Longnecker et al. 2005; Sherr et al. 2006), and that the nucleic acid content may partially reflect the genome size of the cells (Vila-Costa et al. 2012). The addition of flow cytometry to the microbial ecologist’s set of tools fueled a series of studies combining different stains probing cellular activity or growth (see Del Giorgio and Gasol 2008 for a review on this topic). These studies were crucial to open the black box of microbial communities unveiling the large heterogeneity in the physiological status of cells in the environment, which has an impact on their contribution to ecosystem function as we will expand on below.

The application of molecular tools like the 16S rRNA gene PCR amplification and sequencing from environmental DNA enabled access to the diversity of the “uncultured majority” (Amann et al. 1995; Rappé and Giovannoni 2003), which profoundly revolutionized the field of marine microbial ecology. These techniques opened a new venue to understand and classify marine microbial diversity and resulted in, among others, the discovery of the most abundant bacteria in the ocean, the SAR11 clade (Giovannoni et al. 1990), elusive for culturing using traditional methods. The rRNA approach also fueled the implementation of fluorescent in situ hybridization (FISH) methods to visualize and enumerate specific phylogenetic groups in natural samples. The combination of FISH with radioisotopic methods such as microautoradiography (MAR) was pivotal in unraveling niche partitioning in regard to the use of organic substrates between the different components of bacterial communities (Alonso-Sáez et al. 2012a; Cottrell and Kirchman 2000). Likewise, the estimation of the net growth rates of microorganisms belonging to various phylogenetic groups by using FISH has provided evidence for the important role that numerically unremarkable microbes can play in ecosystem functioning (Kirchman 2016). The advent of high-throughput sequencing approaches represented another revolution in microbial ecology that shed light on the composition of marine microbial communities, the relative abundance of their components, and the discovery of the “rare biosphere” (Sogin et al. 2006). The latter comprises low abundant taxa that act as reservoir for most phylogenetic and functional diversity (Pedrós-Alió 2012).

1.2 It Is Not Always Black and White: The Discovery of Photoheterotrophs

One of the most important findings in microbial oceanography in the last decades was the discovery of photoheterotrophs as an important component of the bacterioplankton. This discovery challenged the classic portrayal of bacterioplankton composed of photoautotrophic microorganisms as primary producers and of chemoheterotrophic microorganisms as consumers. Early genomic analyses of natural occurring marine bacterioplankton reported that an uncultured bacterium harbored a gene coding for proteorhodopsin (PR), a light-dependent proton pump, unveiling a new type of phototrophy in the ocean (Béjà et al. 2000). That same year, using infrared fluorometry Kolber et al. (2000) detected high signals of bacteriochlorophyll a (BChla) in the surface oligotrophic ocean suggesting that aerobic anoxygenic phototrophic (AAP) bacteria are a substantial component of the marine microbiome. These two reports represented the beginning of a change of paradigm in the field of marine microbial ecology that demanded to consider the direct effects of light on heterotrophic processes and, consequently, to rethink the models of organic carbon fluxes in the ocean.

Metagenomics and other molecular approaches have revealed an unsuspected large diversity among PR and AAP-containing bacteria and have shown that the genes responsible for photoheterotrophy are widespread among the most abundant microbial taxa in the surface ocean (DeLong and Béjà 2010; Koblížek 2015). Proton-pumping rhodopsins are found in marine Proteobacteria (including the SAR11 clade), Bacteroidetes, and Euryarchaeota (Pinhassi et al. 2016). These rhodopsins consist of only one opsin protein with a covalently bound pigment (retinal) and this simple structure has favored their lateral gene transfer and spreading among distant taxa, even across the domains Bacteria and Archaea (Frigaard et al. 2006). Contrarily, the machinery for light harvesting and energy synthesis in AAP bacteria consists of several pigments and proteins. Thus, even though the genes involved in aerobic anoxygenic photosynthesis have also been laterally transferred in marine bacteria, these events of gene transfer has been constrained to the Proteobacteria, mostly to the Alpha- and Gammaproteobacteria. Besides the size of coding operons, rhodopsins and bacteriochlorophyll systems display differences in the cost of biosynthesis and benefits of phototrophy (Kirchman and Hanson 2013). While the broad occurrence of marine photoheterotrophs has been well described, less is known about their physiology and ecology. Some axenic cultures of AAP and PR-containing bacteria have shown that they can use light to grow more efficiently (i.e., Gómez-Consarnau et al. 2007; Hauruseu and Koblížek 2012) and this has also been confirmed for natural populations of AAPs (Ferrera et al. 2017). In contrast, other PR-containing isolates do not grow better with light (González et al. 2008), and it has been reported that proteorhodopsins can also promote survival during starvation (Gómez-Consarnau et al. 2010). In addition to proton-pumping rhodopsins, other types of rhodopsins have been discovered in marine bacteria and archaea including chloride- and sodium-pumping rhodopsins (see Pinhassi et al. 2016), xanthorhodopsins (proton pumps associated with a particular carotenoid molecule named salinixanthin, Balashov et al. 2005), or sensory rhodopsins such as the newly discovered heliorhodopsins (Pushkarev et al. 2018). Further, rhodopsins, including heliorhodopsins, have been reported in marine viruses and single-celled eukaryotes (Needham et al. 2019). In short, following the discovery of photoheterotrophs it seems now clear that, in the marine environment, photoheterotrophy is not just an exception but probably the rule. In fact, there is growing evidence that bacteriochlorophyll- and rhodopsin-based dual phototrophy may have evolved in nature and is awaiting discovery (Zeng et al. 2020).

1.3 Are all Microorganisms Equally Active in the Ocean?

When single cell approaches started to be used in the late 70s (e.g., Hoppe 1976) it became evident that the activity of marine microbes is not homogeneous but on the contrary is tremendously heterogeneous. Within a given microbial community cells can be dead, injured, dormant (i.e., in a non-growth state), active but constrained by the availability of a certain resource, slow- or fast growing. Any of these physiological states will have strong implications for the role that such cells play in the environment. Yet, one of the major challenges that microbiologists face is how to define the physiological state of a microorganism.

A variety of methods has been developed in order to differentiate active- from non-active cells, probing cellular division, membrane integrity, respiratory activity, substrate uptake, or protein synthesis (see Del Giorgio and Gasol 2008; Sebastián and Gasol 2019; Singer et al. 2017 for reviews on this topic). These methods target different processes and differ in resolution, and thus the delineation of active cells is not always consistent (see Del Giorgio and Gasol 2008 for a comparison of methods). Therefore, the categorization within the various physiological states is purely operational. For example, using the tetrazolium salt 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) as an indicator of activity Del Giorgio and Scarborough (1995) found that the percentage of active cells represented only 5% of the cells in the oligotrophic ocean, which is consistent with the microautoradiography (MAR) values that were obtained using 3H-thymidine as substrate (Longnecker et al. 2010). However, when these authors used 33P-phosphate instead of thymidine the percentage of active cells increased up to 50%. This indicates that the categorization of active cells using substrate uptake methods like MAR strongly relies on the substrate of choice. Despite the variability found in the proportion of active cells, these single cell techniques have been highly informative, particularly when coupled with other approaches that allow taxonomic characterization such as fluorescence in situ hybridization (MAR-FISH). These techniques have shown that some members of the community are more active than others and that there is heterogeneity in the levels of activity even within a given population (e.g., Alonso-Sáez et al. 2007, 2012a; Cottrell and Kirchman 2000).

The delineation of dormant cells is also important because it has been hypothesized that these cells constitute a seed bank of taxonomic and functional diversity that guarantees the persistence and long-term maintenance of the diversity and function of the community (Lennon and Jones 2011). Yet, we still lack a method to assess whether or not a cell is dormant. Some attempts have been made to characterize the active and inactive members of bacterial and archaeal communities through the joint analysis of ribosomal RNA and DNA of individual taxa, using RNA:DNA ratios below 1 as a threshold to delineate “inactivity” or dormancy (Bowsher et al. 2019; Campbell et al. 2011; Jones and Lennon 2010; Kearns et al. 2016). However, the delineation of the inactive members of a community is sometimes problematic (Steven et al. 2017) because some taxa accumulate ribosomes during dormancy in order to be able to respond swiftly to favorable conditions (see Blazewicz et al. 2013 for a review on this topic). Despite these caveats, the sequencing of ribosomal RNA has provided valuable information about marine taxa that have or lack the potential for protein synthesis and how this changes over space or time (Campbell and Kirchman 2013; Campbell et al. 2009, 2011; Ghiglione et al. 2009; Hugoni et al. 2013; Hunt et al. 2013; Zhang et al. 2014).

Labeling with the thymidine substitute 5-bromo-2′-deoxyuridine (BrdU) has also been used to define the active members of the community (Galand et al. 2013; Pernthaler et al. 2002). This and other approaches that link activity and identity, such as stable isotope probing, have been utilized to identify the major players in a certain biogeochemical process or follow the uptake of specific compounds (Bryson et al. 2017; Mou et al. 2008; Nelson and Carlson 2012; Orsi et al. 2016; Taubert et al. 2017). Other methods that are used in marine microbial studies but are still in their infancy include Bioorthogonal non-canonical amino acid tagging (BONCAT) (Couradeau et al. 2019; Hatzenpichler et al. 2016; Leizeaga et al. 2017; Samo et al. 2014) and Raman micro-spectroscopy (Berry et al. 2015; Huang et al. 2007; Lorenz et al. 2017). These methods could have a tremendous potential to address the physiological status of individual cells as will be expanded on below.

Altogether these studies have shed light on microbial processes in the ocean and identified the major players driving them. Yet, accurate knowledge is lacking of how many cells are active, how dynamic is the transition between activity and inactivity of individual cells, and what are the factors driving these transitions. Furthermore, most studies are temporally and geographically restricted and are particularly focused on the sunlit ocean. When inexpensive techniques like BONCAT will be broadly applied by the research community these gaps of knowledge may be filled. This is crucial to understand microbial function and to predict how it is going to evolve in future scenarios.

2 The Marine Microbiome over Space and Time

2.1 The Beginning of the Global Exploration of the Marine Microbiome

Microbial biogeography relies on the description of how microbial communities are distributed in space, in the vertical as well as in the horizontal dimension. While multiple studies that were restricted to particular biogeographical areas provided hints on the spatial distribution of marine microorganisms (e.g., Agogué et al. 2011; DeLong et al. 2006; Pham et al. 2008; Pommier et al. 2010) it was not until the global oceanographic circumnavigations took place (Fig. 8.1) when the exploration of the worldwide distribution of marine microorganisms became feasible. The pioneering large-scale survey was the Global Ocean Sampling Expedition (GOS) that was launched in 2003 in the Sargasso Sea. This survey continued as a several years expedition across the globe although most of the data were generated between 2004 and 2006 from a transect running from the North Atlantic to the South Pacific through the Panama Canal. The GOS expedition indeed represents the first approximation to the microbial diversity of the global surface ocean (Rusch et al. 2007; Venter et al. 2004). The GOS unveiled ~1300 different 16S rRNA gene sequences in surface seawater samples from the Sargasso Sea (Venter et al. 2004) and thousands of gene families in the surface ocean (Rusch et al. 2007). Contemporary to the GOS, the International Census of Marine Microbes (ICoMM) was launched in 2005 aiming at using standardized sampling and data analysis procedures to study microbial diversity in a multitude of marine habitats including pelagic and benthic systems using 454-pyrosequencing of the 16S rRNA gene (Amaral-Zettler et al. 2010). At that time, the ICoMM represented the most comprehensive diversity picture of global ocean bacterial and archaeal communities and reported remarkable horizontal and vertical large-scale segregation showing contrasting patterns for the different explored habitats (pelagic, benthic, and anoxic ecosystems, or vents, Zinger et al. 2011). This study showed that alike yet remote habitats can harbor similar communities in the pelagic realm supporting the Baas-Becking and Beijerinck hypothesis that “everything is everywhere, but the environment selects” (Baas Becking 1934; Beijerinck 1913). Hence, marine planktonic bacteria display an unlimited potential for dispersal while abiotic environmental filtering is responsible for their different distributions in the global ocean. Furthermore, it was the ICoMM initiative that allowed the discovery of the abovementioned “rare biosphere” (Pedrós-Alió 2012; Sogin et al. 2006).

Fig. 8.1
figure 1

The most important global marine circumnavigations for the exploration of marine microbiomes and when they took place. Figure courtesy of Dr. Marta Royo-Llonch from SHOOK Studio

Full size image

After the GOS and the ICoMM initiatives the main surveys with focus on the worldwide exploration of the marine microbiome have been the Tara Oceans Expedition (2009–2013) (Karsenti et al. 2011), the Malaspina 2010 Expedition (Duarte 2015), the Ocean Sampling Day, which is a simultaneous sampling campaign of the world’s coasts on the summer solstice that has been carried out since 2014 (Kopf et al. 2015; Tragin and Vaulot 2018), the BioGEOTRACES (Biller et al. 2018), the Bio-GO-SHIP programs (Larkin et al. 2021) (Fig. 8.1), and other initiatives that are in the planning phase. Between 2009 and 2013, the Tara Oceans Expedition sampled the global ocean from surface waters to the mesopelagic layer. It used standardized sampling procedures to obtain seven size fractions of planktonic diversity, from viruses to small metazoans, and comprised a large sequencing effort (>30 Gb/sample) (Pesant et al. 2015). One of the main legacies of this expedition was the generation of the Ocean Microbial Reference Gene Catalog (OMR-GC) containing a total of >40 million-non-redundant genes of the global marine microbiome of which over 80% of the sequences were new (Sunagawa et al. 2015). This catalog was updated with the OMR-GCv2 integrating the metagenomes of the Arctic region and the metatranscriptomes of the global ocean (Salazar et al. 2019). The multidisciplinary and extensive effort of the Tara Oceans consortium has resulted in many important outcomes (see Sunagawa et al. 2020 for a review) such as the genetic repertoire of bacteria and archaea (Sunagawa et al. 2015) including more than 500 bacterial and archaeal metagenome assembled genomes (MAGs) from the polar Arctic Ocean (Royo-Llonch et al. 2021) and eukaryotes (Carradec et al. 2018) in the sunlit and mesopelagic global ocean, the diversity of eukaryotic plankton (De Vargas et al. 2015) and viruses (Brum et al. 2015; Gregory et al. 2019), and the planktonic interactions occurring in the photic ocean (Lima-Mendez et al. 2015). Other relevant achievements from the Tara Oceans Expedition were studies on the potential contribution of unexpected components such as bacteria, archaea, and viruses to carbon export in the nutrient-depleted oligotrophic ocean (Guidi et al. 2016), the potential impact of ocean warming on community composition and gene expression (Salazar et al. 2019), or the discovery of latitudinal gradients of diversity for most planktonic groups (Ibarbalz et al. 2019). Furthermore, data from the Tara Oceans Expedition confirmed that the geographic distance only plays a subordinate role in determining the taxonomic and functional microbial community composition in the photic open ocean whereas environmental selection seems to be an important driver of microbial biogeography (Sunagawa et al. 2015).

Between 2010 and 2011, the Malaspina 2010 Circumnavigation Expedition (Duarte 2015) sampled the marine microbiome in the tropical and subtropical oceans from the surface down to bathypelagic waters (~4000 m depth). This expedition showed that shifts towards communities enriched in rare taxa in the sunlit ocean reflect environmental transitions (Ruiz-González et al. 2019) and it explored the role of dispersion on planktonic and micro-nektonic organisms (Villarino et al. 2018). The Malaspina Expedition also contributed with an assessment of the diversity and biogeography of deep-sea pelagic bacteria and archaea (Salazar et al. 2016) as well as provided an account of the diversity of heterotrophic protists in the deep ocean, particularly unveiling the special relevance of fungal taxa (Pernice et al. 2015). This expedition also unraveled that the particle-association lifestyle is a phylogenetically conserved trait in bathypelagic microorganisms (Salazar et al. 2015), and provided a metabolic characterization of the deep ocean microbiome based on Metagenome Assembled Genomes (Acinas et al. 2021). Nonetheless, the Malaspina -omics datasets have not yet been fully exploited and new inputs are expected in the coming years, such as those derived from ongoing detailed analyses of vertical profiles and of specific water masses or insights in the microbiome associated with the deep scattering layer in the ocean.

The datasets of the BioGEOTRACES and Bio-GO-SHIP programs have recently been made available (Biller et al. 2018; Larkin et al. 2021) and these will surely push further our knowledge of the marine microbiome. The BioGEOTRACES initiative spans 610 metagenomes collected from diverse regions of the Pacific and Atlantic oceans (Biller et al. 2018). It also adds a temporal dimension to the marine microbiome by providing metagenomes collected every month during two years at the stations HOT (North Pacific) and BATS (North Atlantic), which are part of long-term time series programs (Karl and Church 2014; Steinberg et al. 2001) and for which a suite of physicochemical and biological data are available. The Bio-GO-SHIP program aims at providing high-resolution spatiotemporal sampling of the marine microbiome in order to link microbial traits with ecosystem function and biochemical fluxes. This program has released 720 globally distributed surface ocean metagenomes from samples collected every 4–6 h representing a median distance between sampling stations of only 26 km (Larkin et al. 2021) and thus providing a much higher spatial resolution than other global expeditions such as Tara Oceans (~700 km) or bioGEOTRACES (~200 km).

Overall, these global expeditions together with other local and regional studies have unveiled some general microbial patterns across horizontal and vertical scales in the ocean. These general trends include among others: (i) the key concept that microbial communities are formed by a few abundant taxa and a long tail of low abundant taxa (the rare biosphere) (Sogin et al. 2006), (ii) that temperature is one of the main drivers to predict the taxonomic and functional gene composition of microbial communities in epipelagic waters of the open ocean (Sunagawa et al. 2015), (iii) that phosphorus exerts a strong selective pressure in the surface ocean (Coleman and Chisholm 2010; Grote et al. 2012), (iv) that the latitudinal gradient of diversity observed in macroorganisms, meaning that species richness increases in latitudes closer to the equator and mid-latitudes, also applies to most marine planktonic microorganisms (Amend et al. 2013; Ibarbalz et al. 2019; Sul et al. 2013), (v) that the microbial community composition of contrasting polar zones, even though being distant from each other, is more similar than that of microbial communities of temperate or tropical latitudes (Cao et al. 2020; Ghiglione et al. 2012; Royo-Llonch et al. 2021; Sul et al. 2013), (vi) a vertical segregation between photic and aphotic microbial communities (Amaral-Zettler et al. 2010; DeLong et al. 2006; Sunagawa et al. 2015), and (vii) the existence of a vertical connectivity between surface and deep ocean communities (Cram et al. 2015; Mestre et al. 2018; Parada and Fuhrman 2017; Ruiz-González et al. 2020).

Despite the relevant information generated by all these studies, the myriad of microbial processes ocurring in the ocean are far from understood. For instance, there is an increasing recognition of the role of sub-mesoscale hydrographic features in key processes such as the carbon pump (Boyd et al. 2019; Resplandy et al. 2019) or the dispersion of marine microbes, pollutants, and microplastics, but studies on the changes of the marine microbiome at fine spatial scales are still missing. Current initiatives looking at this heterogeneity like the EXPORTS program (https://oceanexports.org/about.html) and further implementation of genomic sensors (Scholin et al. 2017) will help to address this issue. Similarly, the aim of the recently launched AtlantECO EU project (https://www.atlanteco.eu) is a better understanding and integration of the marine microbiome in the context of ocean circulation and the presence of pollutants, e.g. plastics. AtlantECO also pursues to assess the role of the marine microbiome in driving the dynamics of the Atlantic ecosystem at basin and regional scales. The joint effort of all those initiatives will help to boost our knowledge of the marine microbiome.

2.2 Seasonality and Temporal Dynamics of Marine Microbial Communities

Besides studies aiming at understanding how microbial communities vary across space, increasing efforts are being invested towards exploring how they change over time. The establishment of microbial observatories across the globe has allowed the monitoring of microbial communities over time from short- to long-term scales (see reviews by Bunse and Pinhassi 2017; Buttigieg et al. 2018). Defining seasonality is essential to understand how microbes react to changes in environmental conditions or perturbations. Time series of marine microbiomes also allow addressing fundamental ecological questions such as which patterns of biodiversity are present in an ecosystem, how are these patterns governed, how stable and predictable are microbial communities, how do species interact or what is the ecological niche of a given taxon. Seasonality had been observed in phytoplankton blooms but it was only after the molecular revolution that it could be investigated in bacterioplankton communities. The pioneer application of fingerprinting methods and clone libraries to samples collected from marine microbial observatories over 1–2 year periods elucidated community shifts over seasons and demonstrated the existence of temporal niches for specific organisms (Alonso-Sáez et al. 2007; Brown et al. 2005; Ghiglione et al. 2005). Nevertheless, time series of many consecutive years are required to test the robustness of seasonal patterns. Such long-term time series with large sampling efforts have been undertaken in oceanic and coastal monitoring stations; the San Pedro Ocean Time Series (SPOT) and the Hawaii Ocean Time Series (HOT) in the Pacific Ocean, the Bermuda Atlantic Time Series (BATS) in the Atlantic Ocean, the Western Channel Observatory in the English Channel, or the Service d’Observation du Laboratoire Arago (SOLA Station; Banyuls-sur-Mer, France) and the Blanes Bay Microbial Observatory (BBMO) in the Mediterranean Sea are some examples of such long-term programs (for a detailed list, see Bunse and Pinhassi 2017; Buttigieg et al. 2018).

Over a decade ago the application of high-resolution molecular fingerprinting methods over monthly samples from SPOT revealed remarkably repeatable and predictable seasonal patterns in the distribution and abundance of microbial taxa (Fuhrman et al. 2006). High-throughput sequencing of bacterioplankton communities confirmed this observation at higher resolution in different locations (Cram et al. 2015; Eiler et al. 2011; Fuhrman et al. 2015; Gilbert et al. 2012; Lambert et al. 2019) and unveiled that also the rare members of the bacterioplankton showed seasonality (Alonso-Sáez et al. 2015). Moreover, long-term time series captured the occasional blooming of some rare members of the community which became dominant when the conditions were favorable (Gilbert et al. 2012). Likewise, microbial eukaryotes showed recurrent seasonal patterns (Giner et al. 2019; Lambert et al. 2019). In addition, changes in community composition were accompanied by repeatable shifts in alphadiversity (Gilbert et al. 2012; Giner et al. 2019). Besides, the seasonal patterns of relevant phylogenetic groups (Díez-Vives et al. 2019; Salter et al. 2015; Vergin et al. 2013) or of certain functional groups (AAPs, Auladell et al. 2019) have been studied, revealing remarkably similar patterns to those of the whole communities (Fig. 8.2).

Fig. 8.2
figure 2

Time-decay plot showing the recurrence of aerobic anoxygenic phototrophic communities in the Blanes Bay Microbial Observatory (NW Mediterranean). Bray–Curtis similarity between samples is plotted against the time lag between each comparison. Blue dots represent mean values for each time lag and gray vertical bars the standard error (background gray dots show each comparison). A linear regression with 95% confidence intervals is shown (modified from Auladell et al. 2019)

Full size image

Current methodologies allow to investigate the dynamics of individual taxa, unveiling that closely related populations can represent distinct ecotypes that temporally occupy different niches (Auladell et al. 2019, 2021; Chafee et al. 2018). Further, analysis of finely resolved taxonomic units in combination with high frequency sampling over multiple years has shown that regardless of interannual variation in phytoplankton blooms recurrent modules of co-varying microbes exist (Chafee et al. 2018). Beyond long-term studies, high frequency sampling over a phytoplankton bloom has shown that biological interactions among bacteria, archaea, and eukaryotic microorganisms may play critical roles in controlling plankton diversity and dynamics (Needham and Fuhrman 2016), contradicting the traditional view that blooms are mainly controlled by physical- and chemical processes. Daily sampling over periods of several months has also revealed that microbial plankton is organized in clearly defined but ephemeral communities whose turnover is rapid, mirroring environmental variability (Martin-Platero et al. 2018). On shorter time scales, metatranscriptomics has shown evidence of diel transcriptional oscillations of both phototrophic and chemotrophic microorganisms (Ottesen et al. 2013, 2014) as well as of viruses (Aylward et al. 2017; Kolody et al. 2019).

Altogether, long time series have provided evidence for seasonal and interannual recurrence of some microbial taxa, highly resolved time series have demonstrated that communities fluctuate on a daily and monthly scale along with changes in environmental conditions, and sampling on an hourly scale has unveiled diel periodicity of gene transcription. These studies demonstrate that the temporal scale of sampling is directly linked to the scale of temporal variability that we are able to capture. Highly resolved sampling over long periods will be eased by the development of automated samplers. This, in combination with the decreasing analytical and computing costs, will provide further insights into the stability and reproducibility of the short-term changes over longer periods of time. Moreover, increasing efforts on obtaining data from high and low latitudes will help defining a more global picture of the temporal dynamics of marine microorganisms. All these data will be useful to feed model-based analyses aimed at better predicting the temporal dynamics of marine microbial communities in future scenarios.

3 Approaches to Link Taxonomy and Function of Marine Bacteria and Archaea

3.1 The Genome-Centric Approaches: Single Amplified Genomes (SAGs) and Metagenome Assembled Genomes (MAGs)

Whereas direct analysis of metagenomes can provide a community overview, other strategies have been developed to either access individual environmental genomes without the need for cultivation (Single Amplified Genomes, SAGs) or group the community’s metagenomic information into meaningful genomic units reflective of a population of close taxa as in Metagenome Assembled Genomes (MAGs). Moreover, properly assigning function to taxonomy, which has been an essential goal in the microbial ecology of uncultured microorganisms, requires the genetic information to be considered in a genomic context.

3.1.1 Single-Amplified Genomes (SAGs)

SAGs are generated from the direct amplification of DNA from previously sorted individual cells, its sequencing and assembly (Fig. 8.3). SAGs may represent environmental genomes sequenced from the most fundamental units of life (Blainey 2013; Stepanauskas 2012; Sieracki 2007; Woyke et al. 2009). Individual cells are selected from a sample by fluorescence-activated cell sorting (FACS) or microfluidics. Subsequently, the cells undergo lysis and whole genome amplification through various methodologies: degenerate oligonucleotide-primed PCR, multiple displacement amplification (MDA), or WGA-X, an MDA method that utilizes a thermostable mutant of the phi29 polymerase (Stepanauskas et al. 2017). This is followed by sequencing and genome assembly. SAGs retrieves all the DNA molecules of a cell, unveiling microbial intimate interactions in their natural environment otherwise overlooked like infections, symbioses, and predation (Castillo et al. 2019; Labonté et al. 2015; Martinez-Garcia et al. 2014; Roux et al. 2014; Yoon et al. 2011). Because SAGs circumvents the taxonomic binning used in metagenome assembly, it improves the understanding of microevolutionary processes in the environment (Kashtan et al. 2014), providing as well unique reference genome datasets from uncultured microbes. Indeed, the analysis of 2715 partial SAGs from the tropical and subtropical euphotic ocean enabled the functional and taxonomic annotation of about 80% of metagenomic reads from diverse oceanographic cruises and marine stations (Pachiadaki et al. 2019).

Fig. 8.3
figure 3

Simplified workflow for SAGs generation from seawater samples. After sample collection cells are sorted in a flow cytometer. Individual cells are then lyzed and nucleic acids undergo Multiple Displacement Amplification (MDA). After that, genomes are sequenced and assembled resulting in SAGs of variable quality, from low to high coverage. Figure courtesy of Dr. Marta Royo-Llonch from SHOOK Studio (https://www.instagram.com/shookstudio/?hl=en)

Full size image

Some of the valuable lessons learnt from SAGs studies include: (i) genome streamlining is a prevalent feature in the oligotrophic surface ocean (Swan et al. 2013), (ii) the extensive microdiversity and co-existence of hundreds of genomes within Prochlorococcus populations (Kashtan et al. 2014), (iii) the reconstruction of novel uncultured bacterial species (Royo-Llonch et al. 2020), (iv) the ubiquity of light harvesting and secondary metabolite biosynthetic pathways across microbial lineages (Pachiadaki et al. 2019), (v) the impact of chemolithoautotrophic microorganisms such as the SAR324 clade and the Gammaproteobacteria ARCTIC96BD-19 (Swan et al. 2011), (vi) the overlooked role of the nitrite-oxidizing bacteria (Pachiadaki et al. 2017), or (vii) the adaptation to anoxic niches such as Oxygen Minimum Zones (OMZ) of unique SAR11 lineages with capacity for nitrate respiration (Tsementzi et al. 2016).

3.1.2 Metagenome Assembled Genomes (MAGs)

Metagenomic reads can be assembled into contigs and later binned into the so-called Metagenome Assembled Genomes (Fig. 8.4). MAGs are composite genomes of closely related populations from natural communities. The first attempts to reconstruct genomes from metagenomic DNA sequences of environmental communities started in the early 2000s (Martín et al. 2006; Tyson et al. 2004) but more reliable methods and larger-scale results emerged during the last decade (Albertsen et al. 2013; Alneberg et al. 2014; Parks et al. 2017; Sharon and Banfield 2013; Wrighton et al. 2012). Nowadays, thousands of genomes have been reconstructed from marine metagenomes both from discrete sampling events and from global circumnavigations (Acinas et al. 2021; Delmont and Eren 2018; Delmont et al. 2018; Royo-Llonch et al. 2021; Tully et al. 2017). Just recently, a large-scale study used 10,450 metagenomes sampled from a variety of habitats, including marine environments, and recovered 52,515 medium- and high-quality MAGs, which constitute the Genomes from Earth’s Microbiomes (GEM) catalog, of which 8578 represent marine microbial MAGs (Nayfach et al. 2021).

Fig. 8.4
figure 4

Simplified workflow for MAGs generation from seawater samples. Community DNA is extracted from seawater samples and undergoes high-throughput sequencing, which yields large amounts of metagenomic reads. Reads are then assembled into contigs which are later binned to form Metagenome Assembled Genomes (MAGs) of varying coverage levels. Figure courtesy of Dr. Marta Royo-Llonch from SHOOK Studio (https://www.instagram.com/shookstudio/?hl=en)

Full size image

There has been an increase in the development of tools for estimating genome completion and contamination (Eren et al. 2015; Parks et al. 2015) improving the approaches to retrieve completed (circularized curated, no gaps) Metagenome Assembled Genomes (CMAGs) (Chen et al. 2020) and guidelines on genome quality standards and complementary analyses for the correct deposition in public databases (Bowers et al. 2017; Konstantinidis et al. 2017). However, a community consensus on the pipeline for MAG reconstruction is missing since different strategies in each step of the process (e.g., assemblies, binning, or annotation) are still used.

The reconstruction of bacterial and archaeal MAGs has provided insights into the ecology and evolution of marine microbial taxa unveiling, among other things: (i) the prevalence of diazotrophs in the surface ocean belonging to the Proteobacteria and Planctomycetes (Delmont et al. 2018), (ii) the potential for primary productivity in a globally distributed AAP bacterium (Graham et al. 2018), (iii) the metabolic diversity within Marine Group II Euryarchaea (Tully 2019), (iv) the biogeography and evolutionary processes within the SAR11 clade (Delmont et al. 2019), (v) the widespread potential for mixotrophy found in the genomes of uncultured bacteria and archaea in the bathypelagic ocean (Acinas et al. 2021) or (vi) transcriptional patters of unique bacteria and archaea polar Arctic MAGs (Royo-Llonch et al. 2021).

Both SAGs and MAGs have been successfully combined to derive conclusions from organisms and their ecosystems and to improve the binning quality of single cell assemblies or metagenomes. Finally, the emergence of new pipelines for analysing bacterial and archaeal individual genomes, metagenome assembled genomes, and single-amplified genomes such as MetaSanity (Neely et al. 2020) will facilitate genome quality evaluation, phylogenetic and functional annotation through a variety of integrated programs. It is clear that these two genome-centric approaches will enhance the understanding of functional and evolutionary processes of the prevalently uncultured marine microbes and will as well serve as a starting point for future experimental validation processes of their potential metabolisms, for in situ taxonomic quantification using CARD-FISH, or for providing essential information in order to design strategies for isolation of key microorganisms into culture.

3.2 The Relevance of Culturing Marine Bacteria in the -Omics Era

Despite the fact that gene- and genome-centric approaches have allowed the description of marine microbial diversity from diverse habitats at an unprecedented scale, as overviewed in this chapter, isolates are still a necessary and complementary resource of knowledge. Isolating bacteria and archaea in the laboratory is a fundamental requirement to investigate their physiology under different scenarios, to test ecological hypotheses raised from metagenomics and genome-centric studies, to have access to their complete genomes, to assess the function of novel genes (Muller et al. 2013), to interpret multi-species interactions in co-culture experiments (Stomp et al. 2004, 2008), or to investigate evolutionary principles and population dynamics in long-term monitoring efforts (Good et al. 2017; Rosenzweig et al. 1994). At the same time, genomes from isolates are important to update and improve the existing databases that are needed for the correct annotation of sequencing data (Giovannoni and Stingl 2007; Gutleben et al. 2017). Moreover, isolation is still the only current option for the official procedures for classification and characterization of novel prokaryotic species (Parker et al. 2019) although new efforts to create guidelines for nomenclature of uncultured microorganisms are being developed (Murray et al. 2020). Finally, the short generation times and the nearly 4 billion years of evolution of marine microorganisms have resulted in an enormous biodiversity and a plethora of metabolic pathways and thus, having access to pure cultures, represents an excellent opportunity for biotechnology research (Luna 2015), including bioremediation of polluted ecosystems.

There is little overlap between taxa retrieved by molecular techniques and those retrieved by isolation (Crespo et al. 2016; Lekunberri et al. 2014). This is mainly due to the fact that molecular techniques usually recover the abundant bacteria present in a given environment, while cultures often retrieve those taxa that belong to the rare biosphere (Pedrós-Alió 2012; Sogin et al. 2006). Isolation is thus still essential to decipher the full spectra of diversity of the marine ecosystem (Sanz-Sáez et al. 2020). New culture-dependent techniques have been developed to expand the range of bacteria that can be cultured like microfluidics (Boitard et al. 2015; Ma et al. 2014), culturing chips (Gao et al. 2013; Hesselman et al. 2012; Ingham et al. 2007), manipulation of single cells (Ben-Dov et al. 2009; Park et al. 2011), high-throughput culturing techniques termed “culturomics” (Giovannoni and Stingl 2007; Lagier et al. 2012), or culturing following large-scale dilution to extinction (Henson et al. 2020). A common feature of all these new techniques is that they are based on the same principles of recreating the nutrient conditions of natural environments and overcoming the tendency of rapidly growing cells to outcompete species that reproduce more slowly. The use of these methodologies allowed the isolation of previously uncultured groups such as the alphaproteobacterium Candidatus Pelagibacter ubique within the SAR11 clade (Morris et al. 2002; Rappé et al. 2002), or the chemolithotrophic ammonia-oxidizing archaeon Nitrosopumilus maritimus (Könneke et al. 2005). The existence of these and other cultures has been essential to test hypotheses derived from genomic data and confirm the role of some proteins in the cell’s physiology. For example, experiments with isolates have unveiled light-stimulated growth in some bacteria harboring proteorhodopsin (Gómez-Consarnau et al. 2007), whereas in other bacteria this protein is involved in cell survival (Gómez-Consarnau et al. 2010) or in improving cell fitness (González et al. 2008; Steindler et al. 2011). Likewise, experiments with Candidatus Pelagibacter ubique have made possible to understand the growth requirements of bacteria with a limited genetic repertoire (Carini et al. 2013; Tripp 2013; Tripp et al. 2008). Altogether, these culturing depending approaches are fundamental to microbial ecologists to fully understand the ecology, function, and biotechnological potential of microorganisms in marine ecosystems.

3.3 Shedding Light on the Active Microbiome

Genome-centric approaches such as SAGs and MAGs have increased our understanding of the metabolic capabilities of uncultured bacteria and archaea and culturing efforts are increasing the array of model organisms to be used for carrying out physiological studies. However, the ultimate goal of microbial ecologists is to understand the activity and function of the different microbes in situ and how they are affected by changes in environmental conditions. Only then we will be able to have a predictive understanding of microbial processes in the ocean.

Different approaches have been used to unravel the role of key taxa driving microbial processes (Berry et al. 2015; Hall et al. 2011; Musat et al. 2012; Singer et al. 2017). Microautoradiography (MAR) coupled with fluorescence in situ hybridization (FISH) has been undoubtedly the most widely used technique to assess the groups of bacteria or archaea that are taking part in the uptake of a given substrate (Alonso-Sáez et al. 2012a; Cottrell and Kirchman 2000; Sintes and Herndl 2006), but it suffers from the poor taxonomic resolution of FISH. Stable isotope probing (SIP), which involves the incubation with a stable isotope-labeled substrate and the downstream analysis of heavy-isotope enriched cellular components such as DNA, RNA, or proteins has also been widely used (see Dumont and Murrell 2005; Musat et al. 2012 for reviews on this technique). Nano-scale secondary ion mass spectrometry (nanoSIMS) is a SIP-based technique that enables the quantification of stable isotopes with high spatial resolution. NanoSIMS has been crucial for unveiling the interactions between individual microbial cells and biochemical processes (see Mayali 2020 and references therein) and particularly relevant for the study of metabolic fluxes between symbionts and their hosts (Foster et al. 2011; Thompson et al. 2012). However, NanoSIMS is usually coupled with FISH and it is therefore also limited by its phylogenetic resolution. An alternative approach is applying nanoSIMS to an isotope-labeled whole community RNA hybridized to a phylogenetic microarray. This method is called Chip-SIP and it increases the taxonomic resolution to the level of individual taxa (Mayali et al. 2012; Bryson et al. 2017).

Given the short half live of mRNA, metatranscriptomics provides information about the gene expression profiling in near-real time conditions (Moran et al. 2013). This approach has been pivotal to elucidate daily transcriptional oscillations in heterotrophic bacterioplankton (Ottesen et al. 2014) and in viral genes in host cell assemblages (Aylward et al. 2017; Kolody et al. 2019). Besides, degradation pathways may be identified based on shifts in the proportion of certain transcripts upon substrate additions (Li et al. 2014; McCarren et al. 2010; Mou et al. 2011; Vila-Costa et al. 2010), and the fluctuations of transcripts in the mRNA pool are highly informative for how cells sense shifts in environmental conditions and the machinery involved in the response to these shifts (Moran et al. 2013). Nevertheless, high sequencing depth is required to detect transcripts of functional genes that are not involved in core metabolic pathways or transcripts of rare microbes. Metaproteomics is an emergent field in ocean studies (see Saito et al. 2019 for a review), and it has been useful to shed light on transport functions and microbial nutrient utilization (Morris et al. 2010; Sowell et al. 2009), nutrient stresses (Saito et al. 2014) as well as on changes on substrate utilization by microbial communities in the water column (Bergauer et al. 2018). However, metaproteomics is limited to well described proteins produced by the abundant fraction of microbial communities (Saito et al. 2019). Targeted meta-omics with SIP (Chen and Murrell 2010; Coyotzi et al. 2016; Grob et al. 2015) or targeted Single Cell Genomics with fluorescently labeled substrates (Doud et al. 2020; Martinez-Garcia et al. 2012) is a powerful alternative to circumvent the limitation of metatranscriptomics and metaproteomics to only the abundant fraction of the community. They allow the detection of rare taxa participating in a given biochemical process and facilitate the identification of the different enzymes involved. Similarly, other approaches like Bioorthogonal non-canonical amino acid tagging (BONCAT) or Raman micro-spectroscopy (see Hatzenpichler et al. 2020 and references therein) present great prospect for targeted meta-omics. BONCAT is a sensitive technique that uses a synthetic amino acid that upon incorporation can be fluorescently detected via copper-catalyzed alkyne–azide click chemistry (Dieterich et al. 2006). It has been applied to environmental samples to identify protein-synthesizing cells and the taxonomic identification of these cells has been performed by FISH (Hatzenpichler et al. 2014; Sebastián et al. 2019) or by 16S rRNA tagging after fluorescence-activated cell sorting (Couradeau et al. 2019; Hatzenpichler et al. 2016; Reichart et al. 2020). BONCAT has also the potential to study short-term proteomic responses of bacteria (Bagert et al. 2016) and opens new avenues of research to study the proteins involved in relevant biochemical processes. Likewise, Raman activated cell sorting after incubation with heavy water or other isotope-labeled substrates (Berry et al. 2015; Lee et al. 2019; Wang et al. 2013; Zhang et al. 2015) can be used for subsequent single cell genomics or mini-metagenomics of selected populations (Yu et al. 2017).

Although single cell RNA-Seq has not made its way yet into environmental studies it has the potential to unravel transcriptional heterogeneity of individual bacterial cells (Imdahl et al. 2020) or to discern different stages of viral–host interactions among a single population and monitor how the host metabolism is rewired during viral infection (Ku et al. 2020). All these emerging techniques will surely contribute to a better understanding of microbial processes in the ocean and the key organisms involved, which is central to assess how changes in diversity in the future will impact global biogeochemical cycles.

4 What Have we Learnt from the Exploration of the Marine Microbiome?

4.1 The Unknown Marine Microbial Diversity

Current predictions indicate that the ocean may be home to ~1010 microbial species of which only ∼104 have been cultured (Locey and Lennon 2016). Moreover, the few cultured microbial species often represent rare members of microbial communities while the most abundant taxa remain largely elusive and begun only to be elucidated after the development of culturing-independent methodologies (Rappé and Giovannoni 2003). Certainly, although the “Great Plate Count Anomaly” was known at the time (Staley and Konopka 1985), the pioneer studies that investigated bacterioplankton diversity in the 90s led among others to the discovery of SAR11 and were groundbreaking because they brought to light that the most abundant organisms in the ocean were in fact unknown (Fuhrman et al. 1993; Giovannoni et al. 1990). Ever since, sequencing DNA from the ocean has continuously unveiled hitherto unknown microorganisms, which has expanded the tree of life to a great extent. Only a negligible fraction of the diversity detected in molecular surveys has been eventually cultured such as the SAR11 (Candidatus Pelagibacter ubique) (Rappé et al. 2002), while most of the observed diversity remains uncultured and therefore largely unknown (the so-called “microbial dark matter”). Indeed, most microbial isolates belong to only a few phyla (Hug et al. 2016) and most phyla do not have a cultured representative. Yet, metagenomics and single cell genomics have provided genetic information of many uncultured lineages including those represented by many of the first clones reported in the 90s. This has helped to get insights on the potential functional and ecological roles of these uncultured lineages (Parks et al. 2017; Rinke et al. 2013). For example, the SAR86, an abundant marine clade belonging to the Gammaproteobacteria and detected for the first time in clone libraries constructed almost 30 years ago (Britschgi and Giovannoni 1991) and for which no isolate exists, shares traits with SAR11 such as the presence of proteorhodopsin (Béjà et al. 2000) and metabolic streamlining, but also displays distinct carbon compound specialization that might possibly avoid competition with SAR11 (Dupont et al. 2012). Moreover, analyses of the SAR86 pangenome indicate that the clade is composed of different ecotypes with unique geographic distributions (Hoarfrost et al. 2020). The SAR406 clade, first reported by Gordon and Giovannoni (1996) and now referred to as the “candidate” phylum Marinimicrobia, represents a deeply branching lineage of bacteria that is abundant in the aphotic zone. Marinimicrobia possesses the potential to degrade complex carbohydrate compounds as well as performing nitrate reduction, dissimilatory nitrite reduction to ammonia, and sulfur reduction (Thrash et al. 2017; Wright et al. 2014). The SAR202, also common in the aphotic zone, appears to metabolize multiple organosulfur compounds, oxidize sulfite, and oxidize recalcitrant organic compounds and thus are predicted to play major roles in the sulfur and carbon cycles in the aphotic water column (Landry et al. 2017; Mehrshad et al. 2018; Saw et al. 2020).

In addition, SAGs and MAGs have unveiled the existence of many new “candidate” phyla (Rinke et al. 2013; Parks et al. 2017). These “candidate” phyla are often detected in marine environments other than seawater, for example, in hydrothermal vents, blue holes, marine sediments, or associated with marine animals (Dudek et al. 2017; He et al. 2020; Parks et al. 2017; Rinke et al. 2013). However, analyses of the Tara Oceans metagenomes have revealed the presence of diverse members of “candidate” archaeal and bacterial lineages in the pelagic realm (Delmont et al. 2018; Lannes et al. 2019; Parks et al. 2017; Royo-Llonch et al. 2021; Tully et al. 2018). Moreover, many unknown clades within the “known” phyla are constantly being discovered in the marine water column (Parks et al. 2017; Yilmaz et al. 2016) and the publication of the genomic catalog of Earth’s microbiomes has disclosed a breadth of phylogenetic diversity from multiple biomes including the marine aquatic biome (Nayfach et al. 2021), highlighting the ocean as a reservoir of hidden diversity.

4.2 Insights into New Metabolic Capacities of Uncultured Microorganisms

Although the wealth of genomic data of the “uncultured majority” in the ocean is exponentially increasing, there is still a long way to go to be able to interpret these data (Ferrera et al. 2015). Approximately 50% of the predicted genes detected in the ocean have an unknown function (Sunagawa et al. 2015) and the other half has an assigned putative function based on sequence homology to a gene from a distantly related cultured isolate for which the function has been experimentally demonstrated. Indeed, experiments with isolates have demonstrated that homologous genes yield different phenotypes in distinct microbes as is the case with proteorhodopsins mentioned earlier in this chapter. Despite this, homology searches in global metagenomic datasets have been pivotal to explore the potential relevance of different processes once the genes involved in a particular process are identified.

Like the discovery of proteorhodopsin changed the concept of phototrophy in the ocean, the discovery of ammonia oxidation genes in a genomic fragment belonging to Thaumarchaeota in the marine metagenome from the Sargasso Sea (Venter et al. 2004) changed our understanding of the global nitrogen cycle. Until then, nitrification was thought to be performed by a few low abundant bacterial genera. The abundant Thaumarchaeota were recognized as the organisms that exerted the primary control on ammonia oxidation in oligotrophic waters (Martens-Habbena et al. 2009; Wuchter et al. 2006). Genomic data of the “uncultured majority” was also fundamental to unveil other aspects of the nitrogen cycle such as that cyanate and urea may serve as potential substrates for nitrification (Alonso-Sáez et al. 2012b; Pachiadaki et al. 2017; Shi et al. 2011; Yakimov et al. 2011), which was subsequently tested by using cultures or single cell culture-independent approaches (Kitzinger et al. 2019, 2020; Palatinszky et al. 2015).

Likewise, our perspective of the life of microorganisms in the aphotic ocean has changed with the increase in genomic data from the deep-sea realm. In addition to heterotrophic and chemolithotrophic microorganisms, the deep ocean contains also mixotrophs that may play an important role in the carbon cycle (Acinas et al. 2021; Anantharaman et al. 2013; Pachiadaki et al. 2017; Sheik et al. 2014; Swan et al. 2011; Tang et al. 2016). These mixotrophic microorganisms may obtain energy from the oxidation of a plethora of compounds including ammonia, nitrite, CO, sulfur, hydrogen, and recalcitrant DOM (Acinas et al. 2021; Anantharaman et al. 2013; Landry et al. 2017; Martin-Cuadrado et al. 2009; Mehrshad et al. 2018; Pachiadaki et al. 2017; Sheik et al. 2014; Swan et al. 2011). Besides Thaumarchaeota (Reinthaler et al. 2010), other key players participating in carbon fixation in the dark ocean have been identified such as the ubiquitous SAR324 (Swan et al. 2011), the Thiomicrospirales (SUP05 cluster, Mattes et al. 2013), and Nitrospina (Pachiadaki et al. 2017). Moreover, a novel pathway for carbon fixation, the reductive glycine pathway, has just been described (Sánchez-Andrea et al. 2020), and the wealth of genomic data available allows the assessment of the distribution and potential relevance of this pathway in the ocean (Fig. 8.5).

Fig. 8.5
figure 5

Approaches to unveil new functions in uncultured microorganisms. (a) Novel genes are identified and functionally characterized in experiments with isolates from diverse environments (e.g., ocean, soil, gut). Search for these genes in global environmental -omics datasets provides information about the relevance and biogeography of the novel genes, their taxonomic distribution, and the conditions under which these genes are expressed. (b) Genes with unknown function can also be identified through functional metagenomics in which environmental DNA fragments are cloned into a plasmid and expressed in a surrogate host. Then, the phenotype can be screened (e.g., enzymatic activity assays, development of color in rhodopsin containing cells). After the function is identified its biogeography and global relevance can be assessed in the global -omics databases. Some items have been created with BioRender.com

Full size image

After the discovery that the cosmopolitan SAR11 clade produces methane in phosphorus deficient waters as a by-product of the decomposition of methylphosphonate (Carini et al. 2014), the role of the vast oligotrophic gyres as a source of methane to the atmosphere became clear. The genes coding for the enzymes of this pathway of methanogenesis is widespread in marine bacterial genomes (Villarreal-Chiu et al. 2012) and in marine metagenomes (Sosa et al. 2019). Similarly, a relevant strategy to cope with phosphorus stress was unveiled after the identification of a phospholipase C responsible for lipid remodeling in a soil bacterium (Zavaleta-Pastor et al. 2010). This led to the discovery that lipid substitution is a widespread strategy to decrease the phosphorus demand of heterotrophic bacteria in the vast phosphorus depleted waters of the ocean (Carini et al. 2015; Sebastián et al. 2016) similarly to what had been previously observed in phytoplankton (Van Mooy et al. 2009). The discovery of the strategy of lipid substitution to overcome phosphorus depletion shows the need to implement the flexible stoichiometry of planktonic cells in biogeochemical budgets in which a fixed stoichiometry is assumed to be the rule.

Strategies to deal with the vast number of genes with unknown function are constantly being developed. Some approaches involve the functional screening of metagenomic libraries (e.g., Colin et al. 2015; Pushkarev et al. 2018) while others rely on computational workflows that narrow down the potential role of genes based on the clustering of their coding sequence spaces and their contextualization with genomic and environmental information (Vanni et al. 2020). A large fraction of the unknown genes is phylogenetically conserved and may be relevant for niche adaptation (Vanni et al. 2020). Thus, identifying these unknown genes is crucial to gain a deeper insight of microbial processes in the ocean. Given that sequencing effort has been directedly linked to the increase in the rate of gene discovery (Duarte et al. 2020), it is expected that many new genes will be uncovered in the near future. This calls for ways to unveil the function of these “unknown” genes.

4.3 Delineation of Ecological Meaningful Units of Uncultured Microorganisms

Bacterial and archaeal populations, and by extension species, are fundamental units of ecology and evolution (VanInsberghe et al. 2020). Identifying genetically and ecologically congruent units remains a great challenge in microbial ecology because bacteria and archaea reproduce asexually. Their genomes are subjected to homologous recombination events (Fraser et al. 2007; Papke et al. 2007) and lateral gene transfer between similar or distant relatives frequently occurs (Doolittle and Papke 2006; Fernández-Gómez et al. 2012). Furthermore, even within delineated species, there is no homogeneity in the genetic content or in the total nucleotide composition of bacteria and archaea. This variability is known as microdiversity (Acinas et al. 2004; Fuhrman and Campbell 1998) and it can be seen as “bushy tips” of distinct sequence clusters in phylogenetic reconstructions (Cohan 2001; Giovannoni 2004), by clusters of gene orthologous groups from genomes (Thompson et al. 2019), or by the identification of gene-flow discontinuities derived from recent genetic exchanges (Arevalo et al. 2019). Microdiversity may persist thanks to forces like periodic selection (Cohan 2001) and homologous recombination (Fraser et al. 2007; Konstantinidis and DeLong 2008; Shapiro et al. 2012; Whitaker et al. 2005) in which gene gain and losses are often involved and may promote divergence between populations favoring the delineation of ecotypes (Cordero et al. 2012). The “ecotype concept” describes a collection of strains that show some ecological distinctiveness within its species. Ecotypes preserve nearly the full phenotypic and ecological potential of the species with slight changes in their genetic repertoire that enables them to exploit a slightly different ecological niche. However, microbial speciation occurs in a continuous spectrum in which microbial populations are in constant evolutionary tradeoffs between gene flow and natural selection (Shapiro and Polz 2015; Shapiro et al. 2012) and therefore the delineation of ecological meaningful microbial populations or units remains challenging.

Despite the “species concept” still remains highly controversial, a widely accepted view of microbial species is the “pangenome concept” that classifies the genetic repertoire of a species into the core genome and the flexible genome. The former includes the shared genes between all individuals categorized as the same species and the latter includes the gene pool that is partially shared or strain-specific (Mira et al. 2010; Tettelin et al. 2005). Thus, the genomes of multiple representatives of a species are needed to accurately define the genetic potential and size of the pangenome. Even though the core genome contains the essence of the species and is indispensable, it is the flexible genome that can confer selective advantages like niche adaptation, new host colonization, or antibiotic resistance, and contributes to the species diversity (Tettelin et al. 2005, 2008). The size of the pangenome is dynamic and pangenomes can be more open or close depending on the multitude of niches in which the species is able to live (Medini et al. 2005) (Fig. 8.6).

Fig. 8.6
figure 6

Visual representation of the pangenome concept for genomes of bacteria and archaea. The pangenome classifies de genetic repertoire of a species into the core genome, that includes all genes shared between all individuals categorized as the same species, and the flexible genome, which includes the gene pool that is partially shared or strain-specific. Depending on the genome size and genetic diversity of each species, the pangenome can be open or close (some examples from the literature are shown). Figure courtesy of Dr. Marta Royo-Llonch from SHOOK Studio (https://www.instagram.com/shookstudio/?hl=en)

Full size image

In marine ecosystems, the pangenome exploration was first performed with genomes from bacterial isolates including Prochlorococcus marinus (Kettler et al. 2007), Synechococcus (Dufresne et al. 2008), Shewanella (Konstantinidis et al. 2009), Alteromonas macleodii (López-Pérez and Rodriguez-Valera 2016), and Vibrio alginolyticus (Chibani et al. 2020) but generally using a limited number of genomes. A large-scale analysis of pangenomes using ~7000 high-quality cultured genomes from 155 phylogenetically diverse species belonging to ten phyla revealed the important role of environmental preferences and phylogeny in explaining the majority of variation of pangenome features across different species (Maistrenko et al. 2020). The advent of Single Cell Genomics, however, has enabled the study of the pangenome concept in uncultured microbial taxa (Fig. 8.6). Some pioneer studies have explored the pangenomes of uncultured Prochlorococcus (Kashtan et al. 2014; Thompson et al. 2019), SAR11 lineages (Grote et al. 2012; Haro-Moreno et al. 2020; Thrash et al. 2014; Tsementzi et al. 2016), or uncultured Bacteroidetes relatives of Kordia sp. (Royo-Llonch et al. 2020).

Further, the combination of pangenomes and fragment recruitment analyses on marine metagenomes has led to the “metapangenomes concept” that uses the delineation of single amino acid variants to explore the biogeography of distinct populations. This has been successfully applied on single cell genomes of Prochlorococcus (Delmont and Eren 2018) and the SAR11 clade (Delmont et al. 2019) (Fig. 8.6). Similarly, the “genomospecies concept” represents a species that can be differentiated from others based on the average nucleotide identity (ANI) value by comparative genomics, phylogenomic and fragment recruitment analyses (Haro-Moreno et al. 2020). Alternatively, the reverse ecology approach (Arevalo et al. 2019; Shapiro and Polz 2015) used on isolate’s genomes or single cell genomes relies on the identification of microbial populations as gene-flow units using network analyses. This approach enables the delineation of ecologically relevant populations and the identification of genes that have been under recent positive selection (Arevalo et al. 2019; Shapiro et al. 2012). All these alternatives will become more powerful when applied at large scale using uncultured genomes from different oceanic regions spanning a wider array of phylogenetic taxa, which will altogether enhance our knowledge on the evolution processes driving microbial speciation.

5 Future Perspectives

Despite last years’ integrative research has boosted the knowledge of the marine microbiome to an unprecedented scale, as overviewed in this chapter, it is still insufficient for understanding the functioning of the microbiome and how it will respond to climate change-driven alterations. Here, we discuss some of the issues that in our opinion will become increasingly important in marine microbiome research.

  • Large-scale sequencing and inputs from new technologies. Analyses of marine microbial communities from global circumnavigations have pointed out that only the “tip of the iceberg” has been unraveled and that even larger sequencing efforts are necessary to cover the immense genetic diversity of marine microbiomes. The generation of high-quality genomic references (using long read sequencing such as Nanopore) should be a priority to enhance the genome completeness and quality of SAGs and MAGs.

  • Large-scale single cell genomics studies. Extending the analyses of SAGs at large scale from a wide array of phylogenetic taxa combined with cutting-edge analyses of population genomics will enhance our knowledge of the evolutionary processes driving microbial speciation.

  • Increasing efforts in high-throughput culturomics. Large marine microbial culture collections will improve the microbial reference gene catalogs and will also be a fundamental tool to investigate the physiology of marine bacteria and archaea and to test physiological traits inferred from -omics data.

  • The scale matters. Choosing the right spatial and temporal scales is essential to fully comprehend microbial processes in the ocean and to understand the triggers driving changes in community composition and ecosystem function. Despite the increasing effort to move from local to global scales, analyses at the fine scale are still lacking, such as how sub-mesoscale oceanographic features drive changes in the composition and function of the marine microbiome. Other unique features like the deep scattering layer, which is a hotspot for microbial activity, deserves further investigation. Likewise, at the temporal scale, highly resolved sampling over long periods must be carried out to be able to investigate the stability and repeatability of short-term changes.

  • Expand environmental genomic datasets, particularly on the temporal dimension. Global expeditions have an incalculable value. Yet, they provide a snapshot of the marine microbiome at the moment of sampling. Repeated global expeditions at different times of the year in combination with time series from different latitudes will be necessary to have a complete view of the marine microbiome.

  • Beyond DNA sequencing. DNA-based approaches have been highly informative on the diversity, potential metabolic capabilities, and evolution of the uncultured majority, but more effort should be invested towards understanding how this genetic information translates into function. The extended use of other -omics approaches, such as metatranscriptomics, metaproteomics, and metabolomics, will help to identify which metabolic processes are actually occurring in the environment and their magnitude. Likewise, the application of next generation physiological approaches such as BONCAT and Raman will be pivotal to experimentally validate information inferred from -omics data. A multifaceted approach in microbiome research is needed to grasp the relevance that the observed diversity has for ecosystem functioning.