From protists to humans, all animals and plants are inhabited by microbial organisms. There is an increasing appreciation that these resident microbes influence the fitness of their plant and animal hosts, ultimately forming a metaorganism consisting of a uni- or multicellular host and a community of associated microorganisms. Research on host–microbe interactions has become an emerging cross-disciplinary field. In both vertebrates and invertebrates a complex microbiome confers immunological, metabolic and behavioural benefits; conversely, its disturbance can contribute to the development of disease states. However, the molecular and cellular mechanisms controlling the interactions within a metaorganism are poorly understood and many key interactions between the associated organisms remain unknown. In this perspective article, we outline some of the issues in interspecies interactions and in particular address the question of how metaorganisms react and adapt to inputs from extreme environments such as deserts, the intertidal zone, oligothrophic seas, and hydrothermal vents.
1. All organisms are metaorganisms
Discovering that individuals are not solitary, homogenous entities but consist of complex communities of many species that likely evolved during billions of years of coexistence led to the «metaorganism» concept (Rosenberg et al., 2007, Zilber-Rosenberg and Rosenberg, 2008, Bosch and McFall-Ngai, 2011, McFall-Ngai et al., 2013, Theis et al., 2016; for the terminology see Table 1) which considers the dynamic communities of microorganisms on epithelial surfaces as an integral part of the functionality of the respective organism itself. Consequently, microbes are considered to be an essential part of the animal phenotype influencing the fitness and thus ecologically important traits of their hosts (McFall-Ngai, 2002; O’Hara and Shanahan, 2006; McFall-Ngai, 2007; Fraune and Bosch, 2010; Gilbert et al., 2012; Bordenstein and Theis, 2015; Deines et al., 2017) (Fig. 1).
|Holobiont||A eukaryotic host with all its associated microbial partners. This multispecies assemblage includes viruses, phages, eubacteria, archaea, fungi and protozoa.|
|Hologenome||Genetic information encoded in the eukaryotic host and all of its associated partners. This collective genome forms the theoretical genetic repertoire of a holobiont (definition by Deines et al., 2017)|
|Metaorganism||Includes the function of a holobiont in a given environment. The function of a holobiont depends on (i) the presence and composition of the associated partners, framing the genetic potential of the holobiont, the hologenome; (ii) the activity, abundance and the transcriptionally active part of the genome of every single partner of the holobiont; (iii) this subsequently results in host–microbe and microbe–microbe interactions which must be kept in a state of homeostasis in order to maintain a stable holobiont. To emphasize this highly dynamic functional state (capacity) of a holobiont we will in the following use the term “metaorganism” (see also Deines et al., 2017).|
|Microbiome||Refers to an “ecological community of commensal, symbiotic, and pathogenic microorganisms within a given host” (Lederberg and McCray, 2001).|
The microbiome research of the last decades has revealed that in particular human and animal hosts can be associated with highly complex microbial consortia. The Human Microbiome Project in particular has shown the impact of host genetics, diet, or antibiotictreatment on the microbial diversity within the human gut (Lozupone et al., 2012, Goodrich et al., 2014, Blaser, 2016), while laboratory and preclinical tests suggest that several metabolic and neuronal disorders are linked to alterations in this diversity (Tremaroli and Bäckhed, 2012, Hsiao et al., 2013, Gilbert et al., 2016). Besides the human model, animal and plant models for host–microbe interactions exist that offer unique opportunities to address questions related to symbiosis and host–microbe interactions (Ruby, 2008, Bosch and McFall-Ngai, 2011, Kostic et al., 2013). Some years ago, Ruby (2008) already distinguished between artificial and natural model systems. Artificial laboratory models, such as flies (Drosophila melanogaster), worms (Caenorhabditis elegans), zebra fish or mice, allow the assessment of the influence of specific genetic or environmental factors on holobiont fitness. However, these laboratory models are highly derived and often lack the link to their original existence in nature. In contrast, natural models provide a relevant ecological and evolutionary framework; however, their experimental manipulation is typically limited and restricted to one host–one symbiont types of interactions (e.g., the squid Euprymna scolopesand the bacterium Vibrio fisheri), or to symbioses between animal hosts and low-diversity microbiota (e.g., Hydra sp. and some bacteria or Arabidopsis and the phyllosphere microbiome).
Extreme environments provide a particularly interesting natural context because of the specific challenges to maintain basic life functions under challenging conditions. Yet, due to the extremes in physico-chemical characteristics (temperature, pressure, salinity, light), but also the limited access and the problems of collection render experimental work with organisms from extreme environments difficult. In consequence, with a few noteworthy exceptions (e.g., deep-sea hydrothermal vents), we still know remarkably little about the biology of metaorganisms from these environments, while the available information from established experimental models is unlikely to be representative of these ‘extremophile’ metaorganisms. Additionally, the study of complex (“high-diversity”) consortia, as they frequently occur in nature, remains challenging because multiple interactions take place simultaneously, making it difficult to decipher the specific roles of each symbiont in the community. Methodologies such as the establishment of gnotobiotic, germfree, or axenic hosts are still limited in many natural models, and symbionts are frequently recalcitrant to cultivation. Owing to new deep-sequencing technologies, the extent of diversity and function in highly complex systems is thus just beginning to be unraveled (McFall-Ngai et al., 2013). These omics-driven approaches in combination with meaningful functional assays promise to reveal new insights into the functional and mechanistic underpinnings of metaorganism biology in natural and extreme environments. Current research is focused on understanding the general principles by which these complex host–microbe communities function and evolve: which selective forces drive the evolution of these interactions, i.e. how do the associated organisms influence each other’s fitness?
Life is found in almost all places on Earth under a multitude of different conditions, including extreme environments. These habitats are characterized by either extreme conditions in one (or more) abiotic parameter(s) or by parameters that greatly fluctuate during relatively short timespans. Some of the most challenging habitats are deserts (Ward, 2016), tidal zones with rapidly fluctuating physico-chemical parameters (Raffaelli and Hawkins, 1999), oligotrophicoceans (Miller and Wheeler, 2012), and the deep sea (Van Dover, 2000). Here we address the potential role of the associated microbial community in adapting a given host to extreme environments. We first describe the diverse ecological and evolutionary processes that shape the microbial community. We continue by discussing metabolic interactions and inter-species communication pathways in metaorganisms and their potential impact on living in extreme environments. We then highlight several examples of metaorganisms in extreme environments such as desert conditions, the intertidal zone, oligotrophic seas, and hydrothermal vents. We also throw some light on forgotten players such as viruses and archaea, and visit extreme conditions on terrestrial planets before summarizing how the associated microbiome can promote the host’s vigour and proliferation under extreme conditions.
2. Understanding the adaptation to extreme environments requires an understanding of the evolutionary and ecological processes that shape the associated microbial community
Adaptation to extreme environments represents a particular challenge for any organism. The involved processes are usually studied by characterizing the genetic changes spreading through populations in response to natural selection or drift. Genetic adaptation of eukaryotic species is usually slow because new variants require several host generations to achieve marked frequencies in populations. A faster response can occur through phenotypic plasticity, and/or the reversible adjustment of individual phenotypes to environmental conditions. Such plastic responses can also be inherited when involving epigenetic change, such as alterations in DNA methylation patterns. The fastest responses may be mediated by associated microbes due to their significantly shorter generation time (Fig. 2). In particular, changes in the presence/absence of microbial species and strains and also in their relative proportions (i.e., ecological processes) can result from novel abiotic or biotic constraints. These ecological changes are likely able to effectively promote rapid adjustments of the entire metorganism’s physiology and safeguard metabolic homeostasis of the host. Moreover, even single bacterial species can display extensive phenotypic variability/heterogeneity (Raj and van Oudenaarden, 2008) that can enhance resilience (Justice et al., 2008) to environmental changes and thus facilitate adaptation of the host.
Rapid adaptation can further occur through genetic changes at the level of individual bacterial cells. One particular phenomenon is phase variation (e.g., Van der Woude, 2011). Mediated mostly through intragenomic recombination processes (e.g., site-specific recombination, genomic rearrangements, etc.) and even epigenetic switches (e.g. Van der Woude, 2011, Li et al., 2016), phase variation comprises abrupt and distinct on/off-like phenotypic switches based on differential protein expression profiles in a reversible and irreversible manner. Moreover, bacterial adaptability is fueled by spontaneous mutations or intragenomic recombination (Harms et al., 2016).
Even though the proportion of beneficial mutations among neutral and detrimental mutations appears to be small (e.g., 2 × 10−9 per genome per replication for E. coli; Imhof and Schlötterer, 2001), the associated microbes are expected to adapt faster to changing environmental conditions than their hosts, due to their shorter generation times and their much greater population sizes (both relative to their hosts). The generation of genetic novelty might be facilitated by reversible mutator phenotypes within bacterial populations. Such mutators can arise spontaneously, often at higher frequencies than expected from the mutation/selection equilibrium, suggesting a selective advantage under specific conditions (Denamur and Matic, 2006), such as fluctuating environments (Travis and Travis, 2002). Furthermore, adaptation may also be enhanced by bacterial horizontal gene transfer (HGT), which is the exchange of genetic material, including mobile genetic elements (MGEs) such as plasmids, transposons, and phages. HGT events allow rapid access to genetic innovations of non-parental lineages and contribute to the dissemination of beneficial mutations, especially between closely related species (Aminov, 2011). MGEs serve as indispensible vehicles promoting recombination during HGT, and their acquisition can, depending on the context-dependent beneficial genes which are carried, promote rapid adaptation. HGT shows high prevalence in, for example, the human gut microbiome (Smillie et al., 2011, Brito et al., 2016).
The potential of microbial communities to promote host evolution is additionally influenced by population genetic processes. In particular, genetic drift and population bottleneck events result in a reduction of the population size of specific microbiome members and thereby affect the adaptation of the metaorganism in different ways. Neutral changes in the microbial community composition due to drift could facilitate shifts in the occupation of ecological niches by newly incoming species. Severe fluctuations of environmental conditions could potentially eliminate standing genetic variation and low-fitness genotype variants within the host’s microbiota. Thereby, species–species competition within the microbial community might be reduced and selective advantages conferred to better-adapted genotypes. Clonal interference, where lineages with different beneficial mutations coexist and compete in the population, is likely to positively affect the adaptability of the metaorganism by prolonging transient polymorphic population states with greater genetic diversity (Barroso-Batista et al., 2014).
In summary, understanding the adaptation to extreme environments requires an understanding of the diverse possible responses within the metaorganism (Fig. 1), especially the evolutionary and ecological processes that shape the associated microbial community (Fig. 2). The study of metaorganism evolution should thus be placed in a more general framework of species interactions where the unit of selection is defined by the lineages/populations of both host and microbes within the hierarchical structure of the metaorganism.
3. Metabolism and nutrient cycling in the metaorganism
Can microbes help hosts to perform well in extreme environments? The metabolic capacities encoded in metazoan genomes are in general limited to a narrow range of environmental conditions (Nursall, 1959; Pörtner, 2001). Yet, despite these limitations metazoan life is found even in the most extreme environments (Weber et al., 2007). Understanding how organisms manage to flourish in environments that exceed their own metabolic limitations requires the consideration of the entirety of metabolic processes of the metazoan host and its associated bacteria; in other words, the metaorganism rather than the individual (Bosch and McFall-Ngai, 2011) (Fig. 1). This is particularly apparent as the diversity of microbial metabolic processes by far exceeds those of their metazoan hosts. As a consequence, the metabolic capacity of metaorganisms is governed by the interactions of their individual members (McFall-Ngai et al., 2013). The emerging properties of the combined metabolism thus allow the metaorganism to flourish in environments otherwise uninhabitable for its individual members (McMullin et al., 2000). The nature of the associations supporting such metabolic interactions within metaorganisms can be highly diverse, ranging from parasitic (Kanaani and Ginsburg, 1989) through commensalistic (Hooper et al., 2001) to mutualistic (Douglas, 1998) associations, depending on whether nutrient limitation is increased, unaffected or attenuated by the symbiotic partners.
The importance of these host–microbe interactions for metaorganism performance and fitness is well documented by two illustrative examples:
The symbiosis between wood-feeding termites and their gut microbes. The evolutionary success of wood- and litter-feeding termites is based on their ability to efficiently digest lignocellulose although it is highly recalcitrant to enzymatic attack and has a low nitrogen content (Abe et al., 2000). An intimate symbiosis with archaeal, bacterial, and eukaryotic gut microbes facilitates the breakdown and utilization of this otherwise inaccessible food source (Brune, 2014, Waidele et al., 2017) (Fig. 3A). Furthermore, the low nitrogen content of this diet is compensated by dinitrogen (N2)-fixing bacteria which provide a critical source of organic nitrogen in an otherwise strictly nitrogen-limited system (French et al., 1976).
The symbiosis between tropical corals and their endosymbiotic algae. Tropical reef-building corals flourish in highly oligotrophic waters with low food and nutrient availability (Gove et al., 2015). The productivity of these metaorganisms is supported by the association between corals and their endosymbiotic dinoflagellate algae (Muscatine and Porter, 1977). In this symbiosis, the algal symbionts translocate the majority of photosynthetically fixed carbon to the host, which in return provides inorganic nutrients, i.e. nitrogen derived from its metabolism, to the symbiont (Tremblay et al., 2012, Rädecker et al., 2015) (Fig. 3B). Given that neither heterotrophic feeding nor nutrient uptake from seawater can sustain the nutritional requirements of these symbiotic partners on their own, the recycling of nutrients within this symbiosis is fundamental to ensuring the survival and growth of the coral metaorganism (Muscatine and Porter, 1977, Tanaka et al., 2018).
In both examples, the mutualistic symbiosis between a carbon-limited (coral host, termite gut bacteria) and a nitrogen-limited (termite host, endosymbiotic algae) organism enables these metaorganisms to overcome their nutrient limitations. At the same time, these symbiotic interactions do not occur in isolation, but are part of a much more complex metabolic interaction network within the metaorganism that ultimately provides sufficient nutrition to all its members. In this way, these metaorganisms could conquer new ecological niches, making them important ecosystem engineers of their respective environments (Jones et al., 2008).
Yet, these two types of nutrient-exchange symbioses represent two different strategies regarding their ecological maintenance and the regulation of nutrient transfer. In the case of the termites, the gut microbes process a food source which would otherwise be inaccessible for the host. Microbial biomass is excreted from the hindgut and re-ingested by the host, thereby completing the nutrient exchange between these partners (Brune, 2014). In this scenario, growth and fitness of the symbiotic partners are positively correlated, as the growth of one partner directly benefits the other (Shantz et al., 2015). In the case of corals, on the other hand, the nutrient exchange is facilitated by the constant nutrient-limited state of both symbiotic partners (Shantz et al., 2015). Here, the nitrogen derived from the host metabolism is insufficient to satisfy the nitrogen demand of the algal symbiont (Rädecker et al., 2015). Consequently, a large fraction of the photosynthetically fixed carbon cannot be converted into algal biomass as growth requires the availability of nitrogen (Cunning et al., 2017). The resulting accumulation and subsequent release of organic carbon is driving the translocation of carbon in this symbiosis (Cunning et al., 2017). At the same time, this carbon is insufficient to fully satisfy the carbon requirements of the coral host, resulting in the accumulation and release of inorganic nutrients and thereby completing the nutrient exchange to the benefit of both partners in this symbiosis (Rädecker et al., 2015, Cunning et al., 2017). Consequently, in this mode of regulation the disruption of the nutrient limitation of one symbiotic partner would pose a threat to its partner as it reduces the amount of nutrients available for translocation (Muller et al., 2009, Wooldridge, 2013). For example, Pogoreutz et al. (2017a) suggested that the stimulation of coral-associated N2-fixing bacteria may disrupt the nitrogen limitation of the algal symbionts, ultimately resulting in the retention of photosynthates and the disruption of this symbiosis. Microbial N2-fixation, therefore, may have potentially significant implications for coral resilience to environmental stress(Pogoreutz et al., 2017a, Pogoreutz et al., 2017b). As a consequence, growth and fitness may be negatively correlated between the two partners in this mode of regulation (Cunning and Baker, 2012, Shantz et al., 2015).
Despite these fundamental differences (i.e., positive vs. negative correlation of metaorganism performance or fitness between host and symbiont), both types of nutrient exchange symbioses may ultimately increase the productivity and fitness of the overall metaorganism under stable environmental conditions (Muller et al., 2009, Hill and Hill, 2012). Under changing environmental conditions, however, especially symbioses with a negatively correlated fitness of the partners may be highly vulnerable as changes in the fitness of one of the two partners may destabilize the overall balance of the delicate symbiotic system (Shantz et al., 2015). For corals, the fragility of such symbioses is tragically documented in the form of coral bleaching, i.e. the loss of algal endosymbionts due to the disruption of the coral–algae symbiosis that has been repeatedly observed in recent years on a global scale due to ocean warming (Hughes et al., 2017, Hughes et al., 2018).
While such instabilities of metabolic interactions within a metaorganism due to environmental changes may pose a serious threat to metaorganism functioning, rapid alterations in metabolic interactions may also aid in fast acclimatization/adaptation responses. Reshef et al. (2006) postulated that the environmental selection of optimal microbial communities within a metaorganism should result in highly adapted, beneficial microbiomes which may increase metaorganism fitness under these conditions. Indeed, such rapid microbiome restructuring can convey new metabolic traits which facilitate the metaorganism’s adaptation and acclimatization to environmental changes (Moran and Yun, 2015, Ziegler et al., 2017). In corals, for example, heat-induced symbiont shuffling or strong environmental selection selects for novel communities of algal symbionts with increased heat tolerance and altered nutrient cycling properties, thereby ultimately improving the heat tolerance of the overall coral metaorganism (Baker et al., 2013, Silverstein et al., 2014, Hume et al., 2016). Hence, by being able to change their metabolic profiles through microbiome restructuring rather than having to rely on mutation and recombination alone, metaorganisms can rapidly respond to changing environmental conditions (Rosenberg and Zilber-Rosenberg, 2016).
Microbiome flexibility, however, may proof detrimental to the metaorganism if microbial metabolic processes become increasingly important for metaorganism functioning, due to the potential loss of critical symbionts and associated microbial functions (Bennett and Moran, 2015). Hence, redundancy of important microbial functional groups and/or strong regulation of important symbionts are important factors for maintaining metaorganism structure and function during environmental changes (Allison and Martiny, 2008, Wittebolle et al., 2009). However, the increased reliance on symbionts to fulfill critical metabolic functions may be accompanied by high rates of horizontal gene transfer between all members of these metaorganisms (Nikoh and Nakabachi, 2009, Bhattacharya et al., 2016). Via this mechanism, important functional metabolic traits may be transferred from one symbiotic partner to another, thereby reducing the dependence on individual symbionts, and ultimately increasing metaorganism stability (Raina et al., 2013, Bhattacharya et al., 2016). In highly obligate nutrient exchange symbioses such gene transfer and deletion ultimately lead to the degeneration of symbiont genomes (Moran, 2002). In isolated evolutionary events, this may even have resulted in the formation of new cell organelles, such as mitochondria and plastids, thereby disrupting our traditional understanding of organism identity and boundaries (Keeling and Palmer, 2008, Theis et al., 2016).
Taken together, the metabolism of metaorganisms is an emerging feature of all its members combined. The rapid acquisition and the flexibility of metabolic traits has enabled metaorganisms to conquer even the most extreme environments and environmental niches, and may also assist them in a time of anthropogenically driven environmental change.
4. Living in extreme environments requires extensive inter-species communication
Since the beginning of metazoan evolution, the development of multicellular organisms has depended on the association with bacterial communities (King, 2004). These associations are mediated by communicating with each other through various chemical compounds. This interkingdom cell-to-cell signaling involves small molecules such as hormones that are produced by eukaryotes and hormone-like small molecules that are produced by bacteria (Pacheco and Sperandio, 2009).
Bacterial cell-to-cell signaling, referred to as quorum sensing (QS), allows bacteria to coordinate gene expression within a population. QS is based on the synthesis and perception of autoinducers (AI), which either diffuse through the cytoplasmic membrane or are actively transported and specifically detected by a certain receptor. When the AI binds to its corresponding receptor, the subsequent signal transduction activates the transcription of target genes, often including those encoding the respective AI synthase (autoregulation) (Fuqua et al., 1994). When the population density increases, the concentration of the signaling molecule exceeds a certain threshold (“quorum”), thus causing more autoinducers to be synthesized through the induction of AI synthase. Activation of the receptor changes the regulation of target genes, leading to synchronized transcription in the population (Bassler, 2002). Thus, cell density-dependent behaviors are coordinated (for review see Castillo, 2015), e.g., colonization, biofilm formation, virulence, motility, bioluminescence and pathogenicity (Landini et al., 2010, Castillo-Juárez et al., 2015).
In a complex microbial community, such as a metaorganism, a large number of chemical “languages” are spoken. A single bacterium has to selectively respond to certain “dialects”. As different bacterial species use QS to enhance their competitive advantages (Bassler, 2015), it is rational that competitors also have evolved mechanisms to interfere with the QS systems to prevail over the competition. The mechanisms causing the inactivation of QS systems are generally known as “quorum sensing interference” (QSI) or “quorum quenching” (QQ) (Kalia, 2015). The interference with QS can in general be achieved by inhibition of signaling molecule synthesis, inhibition of signal transport, inhibition of molecule–receptor interaction, modification or degradation of signaling molecules or by antagonistic small molecules (Grandclément et al., 2016, Weiland-Bräuer and Schmitz, 2017). However, recent evidence shows that QS is not restricted to the domain of bacteria, but also allows communication between bacteria and their hosts. The research field of “interkingdom signaling” is still in its infancy, but the increasing number of publications in this area demonstrates that microbe–host communication is in the spotlight. Prominent examples of microbe–host communication, which illustrate the enormous impact of bacteria–host interactions on the development and evolution of a metaorganism, will be presented in the following.
The first example of interkingdom signaling between bacteria and plants was found in the relationship between the marine bacterium Vibrio anguillarum and the green seaweedEnteromorpha (Joint et al., 2002). Biofilm-forming Vibrio cells release N-acyl homoserine lactones (AHLs) and attract zoospores, the motile reproductive stage of the seaweed, which subsequently settle to establish and develop in a certain habitat. The specific response of a plant to bacterial AHLs was first shown for the legumes Phaseolus vulgaris (Joseph and Phillips, 2003) and Medicago truncatula (Mathesius et al., 2003). Here, AHLs from both symbiotic (Sinorhizobium meliloti) and pathogenic (Pseudomonas aeruginosa) bacteria caused significant changes in the gene expression of the plants. The probably most-studied plant–bacteria interaction occurs in the red alga Delisea pulchra, which secretes brominated furanones to protect itself from fouling microorganisms. The algae release brominated furanones which inhibit multiple AHL-dependent processes including swarming motility in Serratia liquefaciens and bioluminescence in Vibrio species as well as AI-2-based QS in Vibrio species (Givskov et al., 1996, Nys et al., 2006). However, one of the best-studied interkingdom signaling mechanisms is found in the relationship between Rhizobium spp. and their symbiotic legume host. In this symbiosis, a complex exchange of signals between bacteria and plant leads to the successful formation of root nodules in which bacteria reside and fix atmospheric nitrogen (Sanchez-Contreras et al., 2007). In summary, plants have evolved multiple mechanisms to interpret bacterial QS signals and initiate attraction/defense responses in a tissue-specific manner, which are even signal-specific (Mathesius et al., 2003, Bauer and Mathesius, 2004).
During the last decades, scientists have realized that bacterial signals are also able to modulate signal transduction and immune responses of animal hosts (Telford et al., 1998, Kravchenko et al., 2008), and, conversely, that host hormones can modulate bacterial gene expression (Sperandio et al., 2003). The AI-3/epinephrine/norepinephrine signaling system is the prime example. The enteric pathogen Escherichia coli uses AI-3 produced by the microbial gastrointestinal (GI) community to activate the expression of its virulence genes, resulting in colon lesions. In enterohemorrhagic E. coli (EHEC), the eukaryotic hormones epinephrine and norepinephrine present in the GI tract also activate the expression of its virulence genes (Sperandio et al., 2003). Thus, EHEC “captures” the eukaryotic hormones to promote its colonization of the human colon mucosa causing colon lesions. Opioids such as endorphin and dynorphin are also known to be hijacked by pathogenic bacteria like P. aeruginosa. The bacteria recognize and use those opioids to enhance their virulence by increasing the production of their three QS systems, leading to persistent P. aeruginosacolonization in the lungs of cystic fibrosis patients (Zaborina et al., 2007). In addition, it was recently shown that mammalian epithelial cell cultures produce an AI-2 mimic that activates QS-controlled gene expression, e.g. in the enteric pathogen Salmonella typhimurium (Ismail et al., 2016). Further studies have found cytoplasmic factors in mammalian cells that degrade 3-oxo-C12-HSL, most likely to prevent interference with host cellular functions. Those paraoxonases (PONs) are involved in the detoxification of many organophosphatesand additionally act as lactonases, cleaving the lactone ring of AHLs to make them ineffective (Kiran et al., 2017).
Interkingdom communication also occurs in early-branching metazoans. In the freshwater cnidarian Hydra a new eukaryotic QQ mechanism was identified, which enables the polyp to specifically modify long-chain 3-oxo-HSLs by oxidoreductase activity (Pietschke et al., 2017). In addition, whole-genome sequencing of Hydra’s main bacterial colonizer Curvibacter sp. revealed the presence of LuxI/LuxR homologs. Functional characterization of the QS systems of Curvibacter sp. showed that both the host-modified and the non-modified AHL version were recognized by the same AHL receptor. Remarkably, even though both AHLs were recognized by the same receptor, gene expression profiles of the bacterium differed depending on the AHL version present. Investigating the impacts of the different QS signals on metaorganism assemblies in vivo elucidated that the host-modified signal promotes, while the non-modified signal represses, symbiont colonization. These results suggest that Hydra is able to regulate the behavior of its bacterial colonizer to promote metaorganism assembly and resilience (Pietschke et al., 2017).
To identify QQ activities of Aurelia aurita-associated microbes, over hundred bacteria were isolated from different life stages of the jellyfish, of which 25% showed QQ activities against AHL as well as AI-2-mediated QS using a recently established reporter system (Weiland-Bräuer et al., 2015, and unpublished data). Selected isolates have been further characterized to identify the QQ-conferring open reading frames (ORFs) (Weiland-Bräuer et al., 2015, and unpublished data); first, to gain insight into the bacteria–bacteria communication and its impact on host colonization; and second, to unravel possible protective effects for the host. In addition, several QQ-ORFs of the host, whose gene expression patterns are currently under investigation, were identified (Weiland-Bräuer et al., unpublished data).
Those examples clearly show that QS and QQ are common languages used in many metaorganisms. Thus, QS and QQ may represent a general language of communication that spans different domains of life and has evolved early during the co-evolution of bacteria and eukaryotic hosts. Future studies in this field will enhance our understanding of the signaling networks that have driven the co-evolution of prokaryotes and eukaryotes.
5. Living in the soil: rhizosphere microbes and plant stress tolerance
Soil-grown plants are immersed in a multitude of microbes and diverse beneficial microorganisms such as plant growth-promoting bacteria (PGPB) as well as plant growth-promoting fungi (PGPF). The root-associated microorganisms, generally denoted as the rhizosphere, receive a considerable fraction of photosynthetic products from their host plants. However, plants require significant quantities of nitrate, phosphate and other minerals from the soil to support their growth and development. These soil components are often only available in limited quantities or in inaccessible forms, but can be made available through rhizosphere microbes. While mycorrhizae provide phosphate and nitrate to plants, free-living or endophytic rhizobia can fix atmospheric nitrogen (Corradi and Bonfante, 2012, Geurts et al., 2012). Other beneficial microbes can confer enhanced resistance to biotic and abiotic stresses (Lugtenberg and Kamilova, 2009). The establishment of beneficial plant–microbe interactions requires the mutual recognition and a considerable orchestration of the responses at both the plant and the microbial side and in this way contributes to the establishment of the complex microbial communities that are shaped by soil- and plant-specific features. The interaction mechanisms are best understood in rhizobial and mycorrhizal symbioses which share a common plant signaling pathway that is activated by the distinct factors that are produced by each microbe (Corradi and Bonfante, 2012). The same signaling pathway also seems to be activated by certain beneficial bacteria, suggesting that different beneficial and pathogenic microbes initiate common plant signaling pathways. Evidence also indicates that beneficial and pathogenic microbes suppress the host defense system by a number of different strategies, including the production of effectors, exopolysaccharides, or phytohormones (Zamioudis and Pieterse, 2012).
Different bacterial families can improve the growth of vegetables and crops under abiotic stress conditions (Egamberdieva and Kucharova, 2009). Enhanced salt tolerance of Zea mays upon co-inoculation with Rhizobium and Pseudomonas is correlated with decreased electrolyte leakage and maintenance of leaf water contents (Bano and Fatima, 2009). Some microorganisms produce plant hormones, such as indole acetic acid and gibberellic acid, which induce increased root growth and thereby lead to enhanced uptake of nutrients (Egamberdieva and Kucharova, 2009). PGPBs can also induce systemic resistance to pathogens and prime the plant innate immune system to confer resistance to a broad spectrum of pathogens with a minimal impact on yield and growth (Van Hulten et al., 2006). A variety of PGPBs colonizes roots and has been shown to protect a large variety of plant species, including vegetables, crops, and even trees, against foliar diseases in greenhouseand field trials (Van Loon, 2007).
PGPF, such as mycorrhizal and endophytic fungi, can also interact with many plant species and thereby significantly enhance stress tolerance of the plants against a variety of conditions, including drought, heat, pathogens, herbivores, or limiting nutrients (Rodriguez et al., 2008). Interestingly, some plants are unable to withstand stress conditions in the absence of their associated microbes, as, for example, shown for the geothermal plant Dichanthelium lanuginosum which interacts with the fungus Curvularia protuberata. While C. protuberata confers heat tolerance to D. lanuginosum plants, neither the fungus nor its host plant can survive alone at elevated temperatures (Redman et al., 2002). Some PGPF can also have a dual nature, being beneficiary to their host plants, but pathogenic to non-host plants, as shown, for example, for Colletotrichum acutatum, which is a pathogenic ascomycete for strawberry, but beneficial when colonizing pepper, eggplant, bean, and tomato (Freeman et al., 2001). Depending on the PGPF, biotic stress tolerance appears to be conferred by different mechanisms. For example, a non-pathogenic Colletotrichumconfers disease resistance but does not activate defense in the host plant in the absence of a pathogen infection. Moreover, the defense response is confined to fungus-colonized tissues (Redman et al., 1999). In contrast, the fungus Piriformospora indica (Sebacinales) colonizes the roots of many plant species and confers disease resistance systemically (Stein et al., 2008).
Microbes seem to be able to ameliorate plant stress responses to abiotic environmental stresses by influencing not only physiological but also developmental processes of plants (De Zelicourt et al., 2013). This is best exemplified by the modification of the root systems of the host plants by various beneficial microbes (Egamberdieva and Kucharova, 2009). In contrast to plants, which have relatively stable genomes and long life cycles, microbes have short generation times and can evolve a high degree of genetic diversity on short time scales. These attributes allow for rapid changes in microbes and for the development of new compounds that can influence the biochemical regulation of key steps of the plant system. This enormous capacity for generating novel compounds from scratch has been seen in the incredible inventory of pathogen effectors (Toruño et al., 2016), and it cannot be excluded that beneficial microbes have also developed a battery of factors which help plants to rapidly respond to environmental changes to which the plant genomes are only slowly adapting.
6. Adaptation of plant metaorganisms to desert conditions
Ever since the colonization of land, plants have evolved mechanisms to settle and survive even in extreme habitats, such as deserts. In spite of the challenging conditions of desiccation and low nutrient availability, desert plants can also support a diversity of rhizosphere microorganisms. These observations pose the question how the growth of desert plants is possible under extreme conditions when most other plants from temperate zones quit long before. Besides specific morphological and physiological adaptations that are fixed in the plant genetic system, the discovery of specific microbial strains conveying heat, drought or salt tolerance suggests alternative explanations (De Zelicourt et al., 2013).
In particular, the ability of the desert PGPF P. indica to stimulate the growth of many plants and crops by promoting nitrate and phosphate uptake and at the same time conferring resistance against abiotic and biotic stresses (Waller et al., 2005, Stein et al., 2008) laid the ground for launching the DARWIN21 project with the aim to systematically search for beneficial rhizosphere microbes in different deserts and use them to develop an organic and sustainable agriculture. So far, the DARWIN21 project (http://www.darwin21.net/index.htm) has generated over 1500 culturable microbial strains from various deserts in the Middle East (Fig. 4). These strains are tested on a plant stress phenotyping platform for their capacity to confer salt, drought and heat tolerance on both model and crop plants, and selected elite strains are used for their performance with local crops in field trials. For certain beneficial microbes, transcriptome, proteome and metabolome analyses of both the microbes and the host plants have been conducted to better understand the communication, signaling and metabolic processes between the symbiotic partners. A genetic model system was established for the symbiotic interaction of several highly divergent bacterial genera with the genetic model plant Arabidopsis thaliana, and both reverse and forward genetic screens have been launched for both host plant and microbial partner. The search for new medicinally important chemical compounds complements the metabolic networkreconstruction from the fully sequenced microbial genomes and also opens the door for large-scale comparative genomics. Together with the above approaches, DARWIN21 is expected not only to help establish a future sustainable agricultural system, but also to develop new applications in medicine, chemical engineering and synthetic biology.
7. Invertebrates in the intertidal zone
The intertidal zone is a marine habitat that is characterized by periodical changes between high and low tide. The daily fluctuations in water level shape microhabitats with rapidly changing and extreme conditions with regard to light intensity, UV radiation, temperature, oxygen, and salinity. These environmental extremes exert high stress on intertidal organisms. For example, temperature oscillations of up to 20 °C daily (Fraune et al., 2016), salinities of up to 51‰ (Hand and Uhlinger, 1995), and direct air and sun exposure pose challenges to organisms living in this habitat. Therefore, predominately sessile species must have considerable physiological plasticity, i.e., a strong ability to acclimatize, paired with genetic adaptations to thrive in the intertidal zone. For instance, polychaetes and other organisms that live in high marsh estuarine environments (i.e., the upper intertidal zone) are typically characterized by reduced genetic connectivity between populations and local adaptations (Bilton et al., 2002, Virgilio et al., 2006). Thus, animals that colonize the intertidal zone are interesting study objects to probe for adaptation mechanisms to extreme environments. Currently, two animal groups of the Anthozoa, namely sea anemones with their representative Nematostella vectensis and stony corals are used to investigate the mechansisms of adaption to these highly variable environments.
Nematostella vectensis occurs in brackish habitats, particularly in tide pools of the high marsh along the Atlantic coast of the United States and Canada (Hand and Uhlinger, 1994, Reitzel et al., 2008). In its north–south distrubition Nematostella spans a thermocline of around 10 °C mean temperature over 10° latitude (Hand and Uhlinger, 1994, Reitzel et al., 2008), which has resulted in different thermal optima for growth and thermal tolerance in different populations (Reitzel et al., 2013a). These differences in optimal growth conditions and stress resistance are correlated with genetic differentiation between populations (Pearson et al., 2002, Darling et al., 2004, Reitzel et al., 2008). Comparisons of allelefrequencies support the hypothesis that some of this genetic variation is a result of natural selection (Sullivan et al., 2009, Reitzel et al., 2013b). Although there is significant genetic and phenotypic variation between populations of Nematostella, individuals from different populations are interfertile (Hand and Uhlinger, 1994, Reitzel et al., 2008), allowing the generation of hybrid lines and their further investigation in the laboratory.
Tropical and subtropical coral reefs extend from deep in the subtidal zone, where corals experience relatively stable environmental conditions, to the shallow intertidal zone, where they have developed mechanisms to cope with large environmental fluctuations. As illustrated by examples from American Samoa (Oliver and Palumbi, 2011), Florida (Kenkel et al., 2013), and Western Australia (Schoepf et al., 2015), corals from these highly variable environments are generally more heat tolerant. This increased heat tolerance is at least in part explained by (long-term) adaptation and (short-term) acclimatization processes in the coral host (Palumbi et al., 2014). Corals from highly variable environments are characterized by genomic adaptations (as detected by RNA-based single nucleotide polymorphisms) that may give them an advantage over their less exposed conspecifics under extreme environmental fluctuations (Bay and Palumbi, 2014, Palumbi et al., 2014). The higher acclimatization capacity of corals from highly variable environments may further be related to ‘front-loading’ of stress tolerance genes (Barshis et al., 2013) and a parallel muted transcriptional stress response (Bay and Palumbi, 2015). Generally, their capacity for acclimatization is determined by the high variability of the environment, which promotes more flexible corals with higher acclimatization capacity (Kenkel and Matz, 2016).
Bacteria and other microbial organisms are also a vital component of cnidarian holobionts (Rohwer et al., 2002, Fiore et al., 2010, Fraune et al., 2014, Torda et al., 2017). The anemone Nematostella is characterized by a stably associated bacterial community, which is highly dynamic but conserved in response to host development (Mortzfeld et al., 2016). Long-term acclimation to different environmental conditions results in reproducible shifts in the associated bacterial community, e.g. in an increase of rare bacterial species (Mortzfeld et al., 2016). In addition, polyps of natural populations of Nematostella reveal a strong correlation between host biogeography and bacterial diversity (Mortzfeld et al., 2016), which may represent a mechanism to buffer the impact of the environment on the holobiont (Rosenberg et al., 2007).
In shallow-growing corals, subaerial exposure during summer or noon low tides can lead to coral bleaching, which is the loss of the obligate endosymbiotic dinoflagellates(Symbiodinium spp.) from the coral host tissue (Brown et al., 1994). However, coral hosts in the intertidal zone can be associated with particularly tolerant Symbiodinium species (Oliver and Palumbi, 2010). But, in contrast to the coral host, Symbiodinium shows no transcriptional heat stress response (Barshis et al., 2014). Instead, post-transcriptional RNAediting has recently been implicated as an alternative mechanism of acclimatization in these coral endosymbionts (Liew et al., 2017). Much less is known about how environmental variability influences other coral-associated microbes, such as bacteria, and how environmental variability influences the interplay between the host and these microbes. Generally, coral-associated bacterial communities on reef flats can quickly change with tidal cycles, with younger colonies exhibiting changes faster than older colonies (possibly due to a more stable microbiome) (Sweet et al., 2017). Further, bacterial communities in highly variable reef flat environments are distinct from those in less extreme conditions, and bacterial community dynamics are linked to the heat tolerance of their coral host (Ziegler et al., 2017).
Fast changes in bacterial communities may thus represent a mechanism for cnidarian holobionts to adjust to rapidly changing environmental conditions in the tidal zone and possibly to environmental changes elsewhere (Ziegler et al., 2017). Potential microbial mechanisms to facilitate holobiont acclimatization include processes such as symbiontshuffling (proportional changes of microbiome members) and switching (loss or acquisition of microbes), mutations in bacterial genomes, and horizontal gene transfer (Buddemeier and Fautin, 1993, Reshef et al., 2006, McFall-Ngai et al., 2013, Fraune et al., 2016, Theis et al., 2016, Webster and Reusch, 2017). The level of flexibility in regard to metaorganism structure and associated acclimatization potential, however, may drastically differ between holobiont assemblages. Red Sea Pocillopora verrucosa corals, for instance, maintain structurally stable symbiotic algal assemblages and bacterial microbiomes even under adverse environmental conditions (Ziegler et al., 2015, Ziegler et al., 2016, Pogoreutz et al., 2018). In light of Ziegler et al. (2017) suggesting that coral microbiome adaptation aligns with increased stress resistance, the contrasting high stress susceptibility of the stable P. verrucosa holobiont may provide evidence that coral holobionts in fact harbor differing levels of microbiome flexibility with consequences for their ability to respond to environmental disturbance (Pogoreutz et al., 2018). Yet, it remains to be determined whether microbiome flexibility is the cause or consequence of physiological instability in a stressed holobiont (Grottoli et al., 2018). More research is needed to understand the potential roles and the contribution of bacterial communities to holobiont acclimatization and adaptation on short time scales (Torda et al., 2017). In particular, knowledge on acclimatization and adaptation mechanisms of organisms inhabiting extreme environments such as the intertidal zone may give insights into organismal responses to scenarios of a changing climate on a global scale.
8. Coral reefs in the oligotrophic seas
The tropical open ocean is characterized by warm, sunlit, and nutrient-poor (oligotrophic) waters, the latter factor ultimately limiting its primary productivity (Vitousek and Howarth, 1991). Nonetheless, highly productive shallow-water ecosystems such as coral reefs or seagrass beds flourish along its coastlines even in the absence of major nutrient input (Connell, 1978, Hatcher, 1990). The apparent contradiction of high ecosystem productivity in nutrient-poor tropical waters was first observed by Charles Darwin, and was hence coined the ‘Darwin Paradox’ (Darwin, 1842, Sammarco et al., 1999). Today we know that the high productivity of these ecosystems is enabled by functional traits of their key ecosystem engineer species, such as sponges, seagrasses, or reef-building (scleractinian) corals (Muscatine and Porter, 1977, Wild et al., 2004b, De Goeij et al., 2013). These engineers constitute complex metaorganisms (holobionts) comprising an animal or plant host living in intimate association with a diverse suite of prokaryotic and eukaryotic partners (Wilkinson and Fay, 1979, Rohwer et al., 2002, Thomas et al., 2016).
The critical importance of these holobionts for ecosystem health and functioning is best illustrated by the scale of coral reef ecosystems. Coral reefs can attain physical structures of hundreds to more than 2,000 kilometers in size, such as the Great Barrier Reef (Birkeland, 1997). The structural and functional basis of this ecological success is mainly provided by scleractinian corals, due to their calcium carbonate skeletons that build the structural backbone of reef ecosystems. Cnidarians are highly adapted to life in the tropical oligotrophic ocean and significantly contribute to the high primary production on reefs (Muscatine and Porter, 1977, Falkowski et al., 1984, Yellowlees et al., 2008). Primary production in the coral holobiont is mainly attributed to the coral–algal endosymbiosis, a mutualistic nutrient exchange relationship between the coral animal host and the algal endosymbiont Symbiodinium (Muscatine and Porter, 1977, Falkowski et al., 1984, Falkowski et al., 1993, Muscatine and Kaplan, 1994). In this intimate symbiosis, a heterotrophic multicellular organism is hosting a photoautotrophic partner, a rare holobiont constellationknown from only few other host–microbe systems, such as between Symbiodinium and other Cnidaria, giant clams (genus Tridacna), nudibranchs (Ziegler et al., 2014), and Foraminifera (Mies et al., 2017), or the Hydra viridis–Chlorella association (Muscatine et al., 1975, Muscatine and Neckelmann, 1981, Bosch, 2012, Habetha et al., 2003).
The unique constellation of the coral holobiont facilitates reef nutrient (re)cycling on all levels of biological organization: on the cellular level, nutrient cycling and recycling is mainly driven by the coral–algae symbiosis (Hatcher, 1990, Muscatine and Kaplan, 1994, Yellowlees et al., 2008, Rädecker et al., 2017). Here, the coral provides inorganic nitrogen compounds (along with other compounds such as carbon dioxide) from its catabolism to the algal symbionts, thereby tightly regulating their population density and growth (Muscatine and Porter, 1977, Rahav et al., 1989, Falkowski et al., 1993). Presumably elicited by a group of free amino acids provided by the coral, so-called host release factors, Symbiodinium in turn translocates most of its photosynthates (organic carbon, i.e., sugars and amino acids) to the coral host (Gates et al., 1995). On the holobiont level, bacteria (and other microbes) further contribute to nutrient cycling, helping maintain productivity (Lesser et al., 2007, Wegley et al., 2007, Raina et al., 2010, Cardini et al., 2015, Rädecker et al., 2015, Benavides et al., 2016). As a result of this productivity, the coral holobiont can afford the release of more than 50% of its organic carbon (i.e., mucus) into the water column (Wild et al., 2004b). Here, the coral-derived organic carbon enters the reef food web via the microbial and sponge loop (Wild et al., 2004b, Wild et al., 2004b, De Goeij et al., 2013, Rix et al., 2016). Briefly, the released coral mucus traps suspended particles and is subsequently mineralized in the coral reef sediments (Wild et al., 2004a, Wild et al., 2004b) or via the sponge loop (De Goeij et al., 2013, Rix et al., 2016), thereby reducing the loss of energy and nutrients from the reef. Coral (and sponge) holobionts thereby widely contribute to the recycling of nutrients via the (re)generation and transformation of (in)organic matter even on the ecosystem scale (Wild et al., 2004a, De Goeij et al., 2013, Cárdenas et al., 2015, Cardini et al., 2015, Cardini et al., 2016, Rix et al., 2016). Curiously, much of the energy released by scleractinian corals is rapidly channelled into higher trophic levels on pristine reefs, resulting in inverted, top-heavy biomass pyramids, i.e., high biomasses of apex predators and low biomasses of primary producers (DeMartini et al., 2008, Rohwer et al., 2010), which is unparalleled by any other ecosystem on Earth. Thereby, the high productivity of coral reefs extends well beyond its ecosystem boundaries even into the human realm, sustaining the livelihoods of some 500 billion people via the provisioning of food (reef fisheries), coastal protection, and income from tourism (Moberg and Folke, 1999).
The peculiar constellation of the coral holobiont not only allows for an efficient cycling of nutrients, but may also convey disease and stress resistance and drive adaptation via the rapid restructuring of microbial communities (Reshef et al., 2006, Rosenberg et al., 2007, Krediet et al., 2013, Torda et al., 2017, Ziegler et al., 2017). Consequently, microbial dysbiosis, i.e. the disruption of the coral–algae symbiosis (bleaching) or the coral bacterial community, can lead to holobiont breakdown (Pogoreutz et al., 2017a, Pogoreutz et al., 2018), its potentially devastating effects rapidly cascading on to the ecosystem level, ultimately resulting in the local or regional extirpation of coral reefs (Hughes et al., 2003, Hughes et al., 2017).
Scleractinian corals are complex and diverse metaorganisms shaping highly productive coral reef ecosystems in the oligotrophic tropical ocean. Thereby, corals and the reefs they shape share common drivers from the cellular to the ecosystem level, rendering the coral holobiont a micro-ecosystem – and the coral reef a holobiont of impressive spatial scales. Similarly, the implications of disrupted holobiont functioning, in particular bleaching, extend well beyond the individual colony level, threatening marine ecosystems on a global scale (Hughes et al., 2017, Hughes et al., 2018). Disentangling the complex interactions within the coral as well as the ‘ecosystem holobiont’ will therefore be critical to our understanding of their stress responses from the micro- to the macro-scale (Voolstra et al., 2015, Putnam et al., 2017).
By comparison, the symbioses between marine sponges (phylum Porifera) and associated microbial consortia are much less understood. Sponges harbor dense and diverse microbial communities extracellularly within the sponge matrix and are thus excellent examples of marine holobionts (Hentschel et al., 2012, Webster and Thomas, 2016). Sponges influence ecosystem functioning by modifying biotic and abiotic factors (Bell, 2008). For example, they provide habitat for a wide range of fauna and play an important role in benthic–pelagic coupling due to their impressive filtering capacity (De Goeij et al., 2013, Rix et al., 2016). Novel meta-omic approaches have begun to reveal the functions of the collective microbial community and individual symbiont groups (Thomas et al., 2010, Jahn et al., 2016, Lackner et al., 2017, Moitinho-Silva et al., 2017a, Slaby et al., 2017). One major finding is that many of the functional roles provided by sponges are indeed mediated by their associated microbes. Another finding is that sponges can generally be classified into two groups according to the abundance and density of microbes in their tissues. High microbial abundance (HMA) sponges harbor densities of microbes 2–4 orders of magnitude higher than low-microbial abundance (LMA) sponges (Moitinho-Silva et al., 2017b). Therefore, the experimental combination of HMA and LMA sponges can reveal the role and function of sponge microbes in the process under investigation. This natural variability across sponge holobionts and environments, together with the possibility of laboratory experiments, opens up the opportunity to address the dynamics of these complex symbiotic systems under climate change scenarios (Pita et al., 2016).
9. Hydrothermal vents
Since their discovery in 1977 during the exploration of the Galapagos Spreading Center, hydrothermal vents have been considered one of the most productive primary biomass producer habitats on Earth. These areas are typically located at depths greater than 1000 m, below the euphotic zone, and can be characterized by submarine springs that emanate from the sea floor and contain high concentrations of reduced energy sources – mainly hydrogen sulfide and methane. These nutrient-rich flows allow the establishment of successful invertebrate communities worldwide based on symbiosis interactions with microbial chemoautotrophic producers (Van Dover, 2000).
The invertebrate species found in hydrothermal vents belong to the phyla Mollusca, Annelida or Arthropoda. Mollusca is the most diverse group, with one worm species from the class Aplacophora, several clam and mussel species from the class Bivalvia and some snail and limpet species found within the Gastropoda group. As for the Annelida phylum, up to four worm species of Polychaeta could be described so far. This community also comprises one shrimp and one crab species from the Decapoda subgroup (Dubilier et al., 2008). One of the best-known examples of these associations is perhaps the symbiosis between the gutless giant tube worm of the genus Riftia and its sulfur-oxidizing bacteria (Cavanaugh et al., 1981). Riftia larvae are free-swimming organisms. When they reach the juvenile stage, they become sessile and acquire the bacteria from fresh water through the skin. After the infection, sharp developmental changes occur, giving rise to the development of the specialized trophosome tissue which harbors the symbionts (Nussbaumer et al., 2006). The larvae use their characteristic and highly vascularized red plume to exchange compounds with the water and thus provide the endosymbionts with nutrients (Minic and Hervé, 2004). Bathymodiolus mussels and Calyptogena clams are also capable of establishing associations with intracellular chemosynthetic bacteria that inhabit the bivalve gill epithelium. While so far only sulfur-oxidizing bacteria have been discovered in clams, some Bathymodiolus species, such as B. azoricus and B. broksii from the Gulf of Mexico, also harbor methanotrophic symbionts (Kuwahara et al., 2007, Ponnudurai et al., 2017). Another characteristic that differentiates both bivalves is the transmission mode. While it has been shown that in clams the bacteria are acquired vertically, in mussels this occurs horizontally and most likely by means of a continuous process during the whole lifetime of the host (Wentrup et al., 2014). Alviniconcha is one of the few examples of marine invertebrates that harbors thioautotrophic symbionts related to Epsilonproteobacteria inside their gills. These deep-sea snails are also believed to take up the endosymbionts from the environment (Suzuki et al., 2005).
Chemoautotrophy is the basis of all these symbiotic interactions. In this special type of metabolism, the energy required for the fixation of the inorganic carbon is obtained by oxidizing the reduced compounds that emanate from the hydrothermal vents. In this way, bacteria become the primary producers to support this biological community. The main two symbiont functional groups are thiotrophic and methanothrophic bacteria belonging to the Gammaproteobacteria clade, although in a few cases the thiotrophs could be identified as Epsilonproteobacteria (Nakagawa and Takai, 2008). Phylogenetic analyses revealed that sulfur-oxidizing endosymbiont clades are usually interspersed with other free-living Gammaproteobacteria, indicating that the establishment of these symbiont associations has evolved several times in a convergent manner (Petersen et al., 2012). In addition, different transmission modes may explain the acquisition of the symbionts by the host. While some species show vertical transmission from the parental host to the offspring, in many cases the symbionts are directly acquired from the sea water. Recent studies revealed a mixed transmission of the symbionts in the shallow-water Solemya velum symbiosis, suggesting that this genetic diversity may be related to the adaptation of the hosts to environmental changes (Russell et al., 2017). Also, the morphological plasticity of these interactions can vary among different host, from epibionts attached to the mouth appendages and gill chambers of vent shrimps to endobionts that can be associated either extracellularly – as in the case of gutless oligochaetes – or intracellularly as in most of the bivalves.
In conclusion, hydrothermal vents may be perceived as the oases of the deep sea. In this hostile environment, factors such as the absence of light and low oxygen concentration prevent the settlement of animal communities in the vast majority of areas. It is only the establishment of symbiotic associations with microbes that has permitted the development of animal diversity in the deep sea.
10. The forgotten players: viruses and archaea
In the last decade research has focused on the elucidation of host-associated bacterial communities. The development of new sequencing technologies, the availability of universal primers targeting bacterial communities and well established analysis pipelines facilitated the entry of many researchers from different disciplines into the field of host–bacterial interaction. During the last years, intensive microbiome research has uncovered the essential role(s) microorganisms play in regard to the physiology of their hosts (McFall-Ngai et al., 2013). While the use of eubacterial 16S rRNA gene primers targeting the V1–V2 or V3–V4 regions is well established in most research facilities, the use of different primer sets targeting other host-associated microorganisms, such as archaea, viruses, protozoans or fungi is still in its infancy. Thus, the majority of these studies has overlooked these microbiome members, most probably due to methodological challenges and underestimated significance.
Viruses are the most abundant entity in the world (Suttle, 2005) and are known to be important regulators within ecosystems (Brussaard, 2004, Suttle, 2016). This regulatory role of viruses can also be expected within metaorganisms. However, viruses as associated partners of metaorganisms cannot be compared to bacteria without prior exploration. The main reasons why viruses have been almost completely neglected are their high diversity and the fact that they are composed of dsDNA, ssDNA, dsRNA, ssRNA or a mixture of both DNA and RNA. In addition, viruses feature high mutation rates and thus are lacking any conserved genes that can be targeted by PCR (Sanjuán et al., 2010). Nevertheless, recent studies exploited large-scale sequencing datasets to identify viral sequences that were present as a sort of by-catch (Brüwer and Voolstra, 2017, Brüwer et al., 2017). Although such an approach comes with its own set of limitations and caveats, it provides a first insight into viral diversity and how it aligns with environmental or species differences.
Yet, to get an in-depth understanding of viral communities within metaorganisms, one cannot simply isolate DNA or RNA from host tissue, biopsies, surface swabs or fecal samples. Viral particles have to be isolated and separated from host tissue and purified from contaminating nucleic acid (Fig. 5). The separation of viruses from host tissue is not trivial as many host compounds can bind viral particles leading to aggregates (Ueda et al., 2013, Gerba and Betancourt, 2017) which may get lost during subsequent purification processes like filtration or low-speed centrifugation. The choice of the correct extraction buffer and the use of subsequent purification steps such as gradient density ultracentrifugation and digestion of contaminating host DNA and RNA by nucleases are labor intensive and prone to experimental biases. Giant viruses such as Mimiviridae and Pandoraviridae might get lost by filtration. Some virons are sensitive towards nuclease digestion at higher pH values, while some are unstable in extraction buffer, chloroform or cesium chloride (Wetz and Kucinski, 1991, Hema et al., 2010).
Another important point is that not all host-associated viruses are present in their encapsulated form, where viral nucleic acid is protected from enzymatic digestion by viral capsid proteins. During latent viral infections of eukaryotic hosts but also during the lysogenic state of temperate phages, viral sequences are integrated into the genome of their host or are present in the form of an episome in the cell. Most of these viruses are transcriptionally active and interact with the metaorganism. These viruses can only be identified by metagenomics and metatranscriptomics (Bikel et al., 2015). However, missing reference data or viral sequence data partially wrongly annotated as part of the host genome make it difficult to identify these viruses. Especially when working with non-model organisms with limited information about the host genome it is difficult to distinguish viral sequences. Finally, it has to be pointed out that the viral community of metaorganisms originates from the eukaryotic host and its associated eukaryotic and prokaryotic partners. In order to understand the function of viruses within metaorganisms the viral community has to be linked back to their specific interacting partners (viral host for replication) within the metaorganism.
Since 1977, archaea have been described as unique prokaryotic organisms that are neither bacteria nor eukaryotes (Woese and Fox, 1977). Carl Woese and colleagues (Woese and Fox, 1977, Woese, 1987, Woese et al., 1990) found explicit differences between the sequence identities of the 16S rRNA gene sequences and the RNA polymerases of archaea, bacteria and eukaryotes and thus proposed that Archaea form an evolutionarily independent domain. In the following years, more individual properties of archaea were observed. These include their unique chemical cell membrane composition that is composed of L-glycerol-ether lipids whose sidechains consist of isoprenoids (De Rosa et al., 1986, Koga et al., 1993, Koga and Morii, 2005, Koga and Morii, 2007). Moreover, the cell wall of archaea is highly diverse and includes not only surface layer proteins, but also pseudomurein as well as methanochondroitin (Kandler and König, 1998). In contrast to their overall morphological similarity to bacteria, archaea share several cellular characteristics with their eukaryotic counterparts, e.g. regarding their translation and transcription machinery (Cavicchioli, 2007). Recent studies even hypothesize that eukaryotic hosts evolved from an archaeal ancestor in which numerous components with eukaryote-specific features are found (Spang et al., 2015; Adam et al., 2017).
In the meantime, the diversity found within the domain Archaea has become incessantly higher and their phylogeny has been continuously updated – with the biggest steps achieved during the last five years. Currently, the domain is divided into ten phyla and two main superphyla (TACK and DPANN) (Zuo et al., 2015). Initially, archaea were only found in extreme environments such as places with very high (up to 113 °C) and very low (down to −20 °C) temperatures as well as in environments with very high salinities (Barns et al., 1996, DeLong, 1998, Valentine, 2007). However, recent studies have frequently discovered them under non-extreme conditions and in high numbers under mesophilic conditions, including in the gastrointestinal tract of insects and mammals (recently reviewed by Moissl-Eichinger et al., 2017). Within the eukaryotic ecosystem, particularly methanogenic archaea are highly flexible in forming syntrophic interactions with a broad range of bacteria and are thus crucially involved in fermentation processes (Samuel and Gordon, 2006, Samuel et al., 2007). Despite their metabolic capabilities, information on archaea’s cross-talk to their hosts, such as immunological response, is generally rare – most probably due to the fact that no archaeal pathogen has been described so far (Bang and Schmitz, 2015).
With respect to the human microbiome, Methanobrevibacter smithii, identified in stool samples 35 years ago, was the first human-associated archaeal species detected (Miller et al., 1982) and it is also the most abundant. It is known to comprise approximately 10% of all anaerobes within the human intestine and was found in nearly every human subject in a representative dataset (Bäckhed et al., 2005, Ley et al., 2005, Dridi et al., 2009). Additional methanogenic archaea cultured from stool samples include Methanobrevibacter oralis, Methanosphaera stadtmanae as well as Methanomassiliicoccus sp., yet their abundance varies within different studies (Dridi et al., 2009, Dridi et al., 2012). Due to the widespread usage of universal prokaryotic primers such as V1–V2 16S rDNA primers as the standard method, the archaeal diversity within the human microbiome is often not captured accurately (Eloe-Fadrosh et al., 2016), since their 16S rRNA gene is not necessarily amplified by the universal primers. However, due to the development and application of high-throughput sequencing analyses over the last years, the number of archaeal members detected in association with the human microbiota has rapidly increased, and currently numerous archaeal communities are reported at different human body sites (Dridi et al., 2011). In addition to the human intestine, archaeal strains have also been frequently found in the oral cavity (Methanobrevibacter sp. and Methanomassiliicoccus sp.) (Kulik et al., 2001, Vianna et al., 2006, Vianna et al., 2008, Horz et al., 2012) and on the skin (Thaumarchaeota) (Probst et al., 2013). Data on archaea associated with other human body sites are still generally rare (Belay et al., 1990, Sundquist et al., 2007); however, knowledge on this will certainly increase as a result of archaea-specific detection methods during high-throughput analyses as recently demonstrated (Koskinen et al., 2017).
The molecular cross-talk between archaea associated with eukaryotes has rarely been evaluated until now. So far, published information has exclusively focused on human immune cells or mice (Blais-Lecours et al., 2011, Blais-Lecours et al., 2014, Bang et al., 2014). Whereas experiments with human epithelial cells revealed high tolerance against the most common archaeal strains in the gut (M. smithii and M. stadtmanae), peripheral blood mononuclear cells (PMBCs) as well as monocyte-derived dendritic cells (moDCs) of healthy subjects responded with high pro-inflammatory cytokine release after exposure to M. stadtmanae (Blais-Lecours et al., 2011, Blais-Lecours et al., 2014, Bang et al., 2014). The response of human immune cells to M. stadtmanae was shown to be phagocytosis-dependent and resulted in the maturation of moDCs, thus suggesting a subsequent adaptive immune activation (Bang et al., 2014). Due to the phagocytosis-dependence a specific intracellular pattern recognition receptor (PRR) was hypothesized to sense archaeal cell components in an endocytic compartment. Indeed, most recently, M. stadtmanae as well its purified RNA were identified as potent stimuli for the human immune system, through recognition by TLR7 and TLR8 (Vierbuchen et al., 2017). Besides, evidence for archaea-specific monoclonal antibodies had already been obtained in earlier studies (Conway de Macario et al., 1982, Conway de Macario et al., 1983, Conway de Macario et al., 1984) and was confirmed in more recent studies (Yamabe et al., 2008, Blais-Lecours et al., 2011, Blais-Lecours et al., 2014). Interestingly, one of those studies could demonstrate significant enhancement of strain-specific serum IgGs for M. stadtmanae in patients that suffered from inflammatory bowel diseases and whose stool was positive for this strain (Blais-Lecours et al., 2014). On the other hand, related strain-specific serum IgG-levels against the more common strain M. smithii were found independently of the abundance of this strain detected in the corresponding stool samples (Blais-Lecours et al., 2014). As summarized in Fig. 6, these results suggest not only a potential involvement of M. stadtmanae in the establishment of inflammatory conditions involving the human gut, but also reveal strong evidence for differential cross-talk between various archaeal species and their (human) host as has been shown for numerous bacterial strains. Thus, it is conceivable that either more than one specific PRR is recognizing archaeal structures or that a few archaeal strains possess virulence factors, though there is no evidence for this to date.
Besides this immunologically driven cross-talk between archaea and their human host, several reports have demonstrated a high physiological impact of archaea during crucial fermentation processes in the gut. Particularly methanoarchaea are highly flexible in forming syntrophic interactions with a broad range of bacteria (primary and secondary fermenters) by using their products, like acetate, methanol, hydrogen and carbon dioxide (Thauer and Shima, 2006). A recent study showed that human-associated Methanomassiliicoccales strains are capable to degrade methylamines – known to potentially cause metabolic disorders – and thus might be an essential part of a healthy microbiota (Brugère et al., 2014). On the other hand, methanogenic archaea in the intestine produce methane as an end product, which has been shown to slow down the gut passage and support constipation as well as the development of obesity (Triantafyllou et al., 2014). Moreover, several studies proposed an indirect pathogenicity of archaea by promoting the growth of pathogenic bacteria (Conway de Macario and Macario, 2009). Recent studies also demonstrated a high immunological potential of the human gut-associated archaeon M. stadtmanae and proposed a role of this strain in the development of systemic intestinal inflammation (Blais-Lecours et al., 2011, Blais-Lecours et al., 2014, Bang et al., 2014, Bang and Schmitz, 2015, Vierbuchen et al., 2017). Thus, additional studies on archaea as part of metaorganisms are urgently needed in order to answer the many open questions on their overall impact on the health and disease of their eukaryotic hosts. In order to expand our current knowledge, future studies should use state-of-the-art methods that cover archaeal diversity in samples of interest, because only if we take a look at the whole microbiome, can we learn from it!
11. Searching for metaorganisms on terrestrial planets – a future laboratory for astrophysicists and biologists
The last few decades have taught us under what unexpectedly extreme conditions life can exist on Earth. In all likelihood, however, the next few years and decades will give us the opportunity to search for and to investigate life under even (much) more extreme conditions, namely in planetary systems outside of the Solar System, and perhaps even on planets or moons within the Solar System. The latter, however, is not the topic of this contribution.
As early as in the 18th century, Swedenborg (1734), Kant (1755) and Laplace (1796) put forward the concept that the solar system was formed out of a clump of interstellar material. Under the conservation of angular momentum, this clump formed not only the Sun itself, but also a gaseous disk around it, out of which a planetary system evolved. The fact that in our solar system the eight major planets (and many of its less massive members) orbit the sun with rather small orbital inclinations relative to a single plane, though at very different distances from the sun, is direct evidence of this concept. Since the second half of the 20th century, with the advent of powerful computing, this so-called nebular hypothesis of our planetary system’s origin can also be quantitatively modelled. These numerical simulations show that planets around stars are a normal outcome of star formation rather than a rare exception (for a state-of-the-art description of our current understanding of star and planet formation, see, e.g., Kippenhahn et al., 2012).
Finally, in the middle of the 1990s, the first planets around stars other than the Sun, so-called exoplanets, were found (Mayor and Queloz, 1995), and in the meantime several thousands of them have been confirmed, with more than one planet in many of the systems (for a daily updated list, see http://exoplanets.eu). As planets, they do not emit energy themselves but reflect radiation from their central star.1 The combination of the ensuing small luminosity and the apparent very close proximity to the star leads to a luminosity contrast between exoplanet and star that, as yet, makes a direct observation of exoplanets impossible in the vast majority of cases. There is, however, an indirect way (Fig. 7). Simply on statistical grounds, in some cases the orbital plane of an exoplanet is oriented in such a way that, from our point of view, the exoplanet, on its way around the star, moves in front of the latter. In that situation, the planet faces the star with its illuminated side and shows us the dark side, leading to the occultation of a small part of the star. The ensuing temporary small decline of the star’s brightness can thus be used as a technique for detecting exoplanets. In our context, however, these occultations also make an analysis of the planetary atmosphere(if present) possible. While the planet moves in front of the star, the main body of the planet not only occults part of the star. Along with it moves the exoplanet’s atmosphere. The light of the star will radiate through the transparent parts of that atmosphere. A comparison of spectra of the star with and without the planet and atmosphere in front of it will allow for the determination of spectral contributions originating from the atmosphere of the planet. The advantage of this technique is that a spatial resolution neither of the star nor of the planet is required. The difficulty is rather that the projected size of the planet in front of the star is very small; in the case of the Earth in front of the Sun, seen from another planetary system, it would be 8·10−5 of the star’s radiating area. This is even more true for a planet’s atmosphere; again in the Earth–Sun configuration with a characteristic thickness of the atmosphere of 100 km, it would be only 3·10−6 of the star’s radiating area. These numbers depend only on the relative sizes of the exoplanet, the star, and the atmosphere, not on the distance between the two, so long as it is a lot smaller than the distance from the observer, which is true in all possible cases (by many orders of magnitude). In the next few years, spectrographs will become available at all the telescopes which will make such transmission spectroscopy of planetary atmospheres routinely possible.
A second very favorable position occurs just before or immediately after the planet’s own occultation, i.e., when it moves behind the star (Fig. 7). In those moments the planet turns its entirely illuminated side towards us and allows for the spectroscopy of the stellar light reflected by its atmosphere. The contribution of that signal to the total light of the system star–planet depends on the reflectivity of the planet’s atmosphere, the relative sizes of planet and star, and the distance of the planet from the star. Knowledge of the uncontaminated stellar spectrum then allows for a derivation of the exoplanet’s atmospheric reflection spectrum.
Once either one or both of these approaches of the spectroscopic investigation of exoplanetary atmospheres become routinely available – and this will be the case very soon, in all likelihood – an Eldorado for biologists and astrophysicists will open up which can only be fully exploited by a close collaboration between the two fields.
Astrophysicists have already determined the orbits of several thousand exoplanets. From this, we can estimate the general conditions on the surfaces of these planets, such as the expected temperature ranges, for instance. And from this, we can make statements about the possible habitability of individual planets. Habitability of a planet, in this context, usually refers to the type of life known to us, and requires the continuous presence of liquid water. Quite a number of potentially habitable planets are known already, even exoplanetary systems with several habitable planets. Interestingly enough, the closest star, Proxima Centauri, has a planet in its habitable zone (Anglada-Escudé et al., 2016). It is important, however, to be aware that habitability is only a necessary condition for the occurrence of life on an exoplanet, and far from a sufficient one. Moreover, the definition of habitability is a rather loose one, as, for instance, the true temperature on a planet depends not only on the incoming radiation from the star, but also very much on local conditions like mass, composition of the atmosphere, rotation of the exoplanet, inclination of its rotational axis relative to the orbital plane about the star, magnetic field, presence of moon(s), etc., most of which are as yet unknown. But in the near future we will be able to routinely provide spectra of exoplanetary atmospheres and test the local environment.
What do astrophysicists need from biologists? In order to analyze the spectra and to extract the desired information from them, we need to know what we have to look out for. This is important as, at least at the beginning, the relevant signals will be close to the detection limit. So we need to learn about the tracers of possible (simple) forms of life under the conditions prevailing on the exoplanets.
What can astrophysicists provide to biologists? We astrophysicists can provide an almost unlimited variety of spectra from planets throughout the universe. These will represent a vast variety of environments, different from each other in all respects, and many, if not most of them, very different from what we know on Earth.
Together, by combining our expertise, we will be able to investigate life outside of our planet, if it exists at all. If so, we may discover where life exists and where it does not (if we recognize it as such). We may be able to determine under what other conditions (our form of) life can exist.
Initially, we will have to concentrate on the nearer systems, because of their higher brightness. Right at the beginning we will have to determine whether life exists on exoplanets at all. And at the beginning we will be restricted to the most prominent cases because of the week signal and the ensuing unfavorable signal-to-noise ratio. But it will only be a matter of time to resolve these issues.
12. Important areas for future research
The increasing realization that organisms exist only as metaorganisms has led to three important insights necessary for understanding metaorganism adapation in extreme environments. First, an in-depth understanding of the adaptation, physiology, and development of a given host species requires that it must not be studied in isolation. Second, the health and fitness of a metaorganism depend on its modularity that allows the change or exchange of parterns. This view is consistent with the “plasticity first” model of evolution, which states that plasticity can precede and facilitate evolutionary adaptation(Levis and Pfennig, 2016). Finally, the holobiont or metaorganism may be an important unit of evolutionary selection, a selection of “teams” containing many genomes and species. These unexpected insights require new research initiatives to systematically evaluate the critical position of microbes in the metaorganism’s adaptation. Despite the recent advances in our understanding of commensal colonization, competition, and impact on immune celldevelopment and differentiation (Bosch, 2014), many aspects of the host–microbe interaction remain unknown. Important areas of future research now include developing approaches to examine at a mechanistic level how a complex microbiota interacts as a spatially and temporally dynamic network.
Microbial symbionts represent a specific form of genetic inheritance, which has often been acquired vertically, either through the egg or from the maternal environment (Moran et al., 2008; Gilbert et al., 2010; Funkhouser and Bordenstein, 2013). Genetic variation in the symbionts can produce phenotypic variation of the metaorganism and may impact the ecological tolerance and distribution of the host. For instance, symbiotic bacteria in pea aphids provide selectable allelic variation (e.g., thermotolerance, color, and parasitoid resistance) that enables the host to better persist under different environmental conditions (Oliver et al., 2009; Tsuchida et al., 2010; Moran and Yun, 2015).
Overall, these findings suggest a possible evolutionary process whereby populations become increasingly reproductively isolated through the divergence of their microbiomes. Symbionts are also thought to be involved in major transitions in the history of life. The endosymbiotic theory of eukaryotic cell formation holds that the origin of eukaryotic life began through the merging of archaea and bacterial cells and genes (Margulis, 1981). Similarly, animal multicellularity might have emerged from the symbiosis of a choanoflagellate protist with a particular bacterial partner (e.g., Alegado et al., 2012). Another evolutionary transition, the origin of placental mammals, may have been promoted by symbiosis, namely, through the incorporation of retroviruses from other organisms. These retroviruses, which contain their own enhancer elements, appear to have allowed the rewiring of cell circuitry to produce the progesterone-responsive uterine decidual cell (Lynch et al., 2015). This particular role of beneficial microbes in eukaryotic development points to the complexity of genetic interrelations within the metaorganism species community. Taken together, symbionts are critical to normal development, fitness and natural selection. They help generate organs, they can produce selectable variant phenotypes, they can create the conditions for reproductive isolation, they are crucial for the fitness of their eukaryotic hosts under changing environmental condition, and they may be the facilitators of evolutionary transitions. Symbiotic relationships are the signature of life on earth, and evolutionary biology has to include the species community within the metaorganism as a central unit of selection. “Biology has entered a new era with the capacity to understand that an organism’s genetics and fitness are inclusive of its microbiome” (Brucker and Bordenstein, 2014). The past few years have seen a remarkable demonstration of the role of microbiota for multicellular host functions, but what is needed now is a sound understanding of the molecules and mechanisms driving this role and then capitalizing on this knowledge to improve health and decrease disease (Waldor et al., 2015).