A new understanding of life is emerging today, one in which organisms are conceived of as multigenomic entities, comprising many species living together. We are genetic and physiological chimeras. We did not just evolve from bacteria, we have evolved with them: 90 % of the cells of our bodies are bacterial (Ley et al. 2006). Our largest microbial population (10–100,000 trillion) lives in our gut. These complex communities carry out many different functions for us: they boost our immunity, ward of pathogens, provide vitamins, and help in digestion; they promote differentiation of our animal cell tissue during development, and they affect our behavior (Gilbert et al. 2012). Symbiosis is at the core of all eukaryotic life. Mitochondria and chloroplasts are both vestiges of free-living bacteria that had become integrated within a host cell in a microbial world billions of years ago. The DNA in our cell nucleus is also chimeric.

The importance of microbial symbionts in the evolution and function of hosts has been and continues to be demonstrated and reinforced by advances in molecular methods. Nonetheless, the concept of the multigenomic organism remains a subject of dispute. Margulis (1991, p. 2) introduced the term “holobiont” simply to connote the contiguous associations of host and microbes. That term has come into more general usage in recent years to describe a host and its microbial communities, including all symbiotic viruses and microorganisms in an array of different, context-dependent relationships from beneficial to parasitic (Rosenberg and Zilber-Rosenberg 2015; Bordenstein and Theis 2015). The partner concept of the “hologenome,” to denote the sum of host and microbial genomes, has also been proposed (Zilber-Rosenberg and Rosenberg 2008, Rosenberg and Zilber-Rosenberg 2013).

The introduction of these terms has incited controversy and confusion. Moran and Sloan (2015), for example, have objected that the hologenome concept obscures more than it clarifies for two main reasons: (1) to be biologically meaningful necessarily implies that the host and its microbial community are a coevolved unit, and form “the primary unit of selection,” which is only sometimes true and often difficult to demonstrate; (2) it is “the wrong approach to start with the assumption that associated organisms have evolved to function as a cooperative unit.” Yet, those who support the hologenome and holobiont concepts contend that these arguments are misdirected because the terms were not introduced to connote solely a coevolved cooperative unit and need not be so restricted (Theis et al. 2016). Natural selection operates at several levels of the holobiont, from single genes to microbes and host, as well the holobiont itself. Conflict, coevolution, selection, and drift would occur in the hologenome just as they do in the genome.

Recent discussions of the holobiont concept are somewhat reminiscent of a debate over the term “symbiosis” itself in the mid twentieth century. Did symbiosis mean mutualism, or did it include parasitism as well? The term embraced a gradation of relationships when it was originally coined as “the living together of differently named organisms” (De Bary 1879), but it soon was also used as a synonym for mutualism. Some argued that if symbiosis implied mutual benefit, which was difficult if not impossible to demonstrate, then the term should be expunged from biology (Thorton 1951). Others argued that its meaning was not in strict reference to mutualism, but rather in the functional bonds of associates and in a broader concept of the organism as a functional whole. As Frederick Gregory put it (1951), “The value of the concept of symbiosis resides in the widening of the concept of organism to include heterogeneous systems overriding the limitations of genetic uniformity, and the supplementation of the concept of structural unity by that of a functional unity or functional field.”

A symbiotic view of the organism is indeed not new to biology. Such ideas have persisted on the margins for more than a century, in constant tension with the aims and concepts of classical biology and neo-Darwinism (Sapp 1994). The importance of symbiosis in evolution would come to the fore with the development of molecular phylogenetic methods for studying bacterial diversity and cell origins (Sapp 2009). Molecular methods have also enabled the exploration of our microbiomes and revealed their importance in our physiology and development.

Symbiosis in Conflict

Suggestions that symbiosis is at the foundation of all eukaryotic life first arose in the late nineteenth century. Lichens were shown to be “dual organisms,” composites of fungi and algae, in 1867. There were also startling reports, beginning in the mid 1880s, that mycorrhizal fungi in the roots of forest trees helped trees with water, mineral and nutrient uptake, and that plants feed them with carbohydrates. Bacteria in the root nodules of legumes were understood to fix atmospheric nitrogen and make it available to their hosts. What biologists had called “animal chlorophyll” in coral, hydra, sponges and protists were also shown to be algae living inside cells.

Based on these examples and cytological evidence that mitochondria and chloroplasts were self-reproducing bodies in cells, several biologists in the late 19th and early 20th centuries postulated that hereditary symbiosis was a major source of evolutionary change. Paul Portier (1918, p. vii) at L’ Institut Océanographic de Monaco proposed that mitochondria had arisen from symbiosis in the remote past. “All living beings,” he said, “all animals from amoeba to man, and all plants are constituted by an association, the emboitement of two different beings.” Félix d’Hérelle at Yale, who had discovered “bacteriophage” in 1917, referred to relations between some viruses and bacteria as “microlichens” when he declared, “Symbiosis is in large measure responsible for evolution” (d’Herelle 1926, p.320). Ivan Wallin at the University of Colorado postulated that mitochondria were bacterial symbionts, that hereditary symbiosis was the source of new genes, and that the establishment of intimate microsymbiotic complexes was “the fundamental factor or the cardinal principle involved in the origin of species” (Wallin 1927, p. 8).

Such proposals were generally met with silence. As famed cell biologist E.B. Wilson (1925, p.739) remarked, they were “too fantastic for present mention in polite biological society.” There were several reasons why this was so. The notion that bacteria played a beneficial role in the tissue of animals confronted a basic tenet of the germ theory of disease, according to which bacteria were “the enemy of man.” Wallin (1927, p. 8) noted the problem: “It is a rather startling proposal that bacteria, the organisms which are popularly associated with disease, may represent the fundamental causative factor in the origin of species. Evidence of the constructive activities of bacteria has been at hand for many years, but popular conceptions of bacteria have been colored chiefly by their destructive activities as represented in disease”.

The relatively few reports of the beneficial effects of microbes could not compete with the triumphs of germ theory: specific microbes were at the basis of smallpox, cholera, tuberculosis, bubonic plague and many other diseases that have dramatically affected human history. In some cases, reports of putative cooperation between microbes and plants were reinterpreted, and the microbes were depicted as thieves stealing the “rightful inheritance” of higher organisms. The notion of microbial symbiosis conflicted with immunological conceptions of “self,” whereby individuals were equipped with weaponry designed to fight off such “infections”. The individual organism was well defined and protected against “invaders”.

When microbes were found to be involved in an experimental result in genetics and physiology they were typically considered to be “contamination,” a source of error and confusion. Classical geneticists rejected the few known examples of hereditary symbiosis as pseudo inheritance, and deemed them to be unimportant for heredity and evolution. Harvard geneticist E.M. East (1934, 409) put it this way:

Thus there are several types of phenomena where there is direct transfer of alien matter capable of producing morphological changes. It is not to be supposed that modern biologists will cite such instances, when recognized, as examples of heredity. But since an earlier generation of students used them, before their cause was discovered, to support the inheritance of acquired characteristics, it is well to be cautious in citing similar, though less obvious, cases as examples of non-Mendelian heredity.

Symbiosis, hereditary or not, was generally considered to be an exceptional, rare phenomenon, not part of normal life. Reports of symbiosis and the “communal life” between species were also overshadowed by the emphasis on conflict and competition in nature, a view of life “red in tooth and claw”, which, it had long been argued, reflected human society and some theorists’ views of human social progress. In so much as cases of symbiosis implied mutualism, they were associated with socialist views of mutual aid, and held up in opposition to Social Darwinism.

The role of symbiosis in evolution also contradicted the central tenets of Darwinian evolutionary biology. From the moment the term symbiosis was coined, it was considered to be a saltational mode of evolutionary change- in contrast to the gradualist evolutionary theory of Darwin and his followers. As De Bary (1879) commented: “Whatever importance one wants to attach to natural selection for the gradual transformation of species, it is desirable to see yet another field opening up to experiment.” Neo-Darwinian theory is based on the principle that natural selection acts on small genetic differences among individuals of an interbreeding population- that is, within species. Evolution by leaps was explicitly denied by the architects of the “modern synthesis”, as was the inheritance of acquired characteristics.

The neo-Darwinian theory of evolution that surfaced in the 1930 and 1940s, based on gene mutation and recombination within species, was a sterile conception of life without microbes. It emerged from within a two-kingdom world of plants and animals and was concerned with the origin of species over the past 500 million years. Microbes and their evolution over the previous 3 billion years were not included any more than were the origin of cells, and multicellular organisms in which microbial symbiosis was rooted.

Benevolent Virology

Those who studied symbiosis argued that the modest number of reported cases did not reflect its occurrence in nature but rather the extent to which it was investigated. A few thinkers heralded the changes that were to come in the concepts of both the organism and of “germs”. Bacteria were shown to have several means for transferring genes between “species”- by conjugation, by viral transduction and by transformation (Lederberg 1952). Even so, some leading bacterial geneticists of the 1950s contended that such lateral gene transfer would be a rare occurrence in nature, but not all did (see Sapp 2009).

To embrace a conception of infectious heredity in biology, Joshua Lederberg (1952, p.403) coined the word “plasmid” as “a generic term for any extrachromosomal hereditary determinant” regardless of its origin. Genetic particles shown to be infectious should be treated as part of the genetic system, he argued, and genetic entities not shown to be infectious should be treated as if they once were. The concept of heredity would have to be enlarged to embrace genes coming from the outside, and the concept of organism had to be enlarged to embrace symbionts. As Lederberg commented (1952, 425): “The cell or organism is not readily delimited in the presence of plasmids whose coordination may grade from the plasmagene to frank parasites.”

René Dubos (January 2, 1958) subsequently wrote to Lederberg about “benevolent virology” searching for examples of adaptive effects of viruses in mammals. “I am amusing myself,” he said, “trying to work up the thesis that the science of microbiology would have taken a very different course if it had used useful effects instead of harmful effects as criterion for selection of microorganisms.” Lederberg was certain that such cases must exist, but pointed to “the technical difficulties in detecting such adaptive effects.” “Except under extraordinary circumstances,” he replied to Dubos (January 7, 1958), “evolutionary pressure would all be in favour of the ubiquitous occurrence of such a virus and we would have no means of perceiving its effects without special techniques.”

When Dubos (1961, 204) emphasized the creative role of viral infections in bacteria, and joined them with examples of microbial symbiosis, he chided his colleagues for maintaining themselves “as poor cousins in the mansion of pathology,” and prophesied: “There will soon develop a new science of cellular organization, indeed perhaps a new biologic philosophy. The time has come to supplement the century old philosophy of the germ theory of disease with another chapter concerned with the germ theory of morphogenesis and differentiation.” The importance of viruses in mammalian evolution and of horizontal gene transfer in bacterial evolution would become apparent several decades later with the development of molecular phylogenetic methods.

The Molecular Phylogenetics Revolution

Interest in the role of hereditary symbiosis in cell evolution was rekindled when DNA and ribosomes were discovered in mitochondria and chloroplasts in the early 1960s (Margulis 1970). But that evidence was not decisive because these organelles were shown to be genetically well-integrated into the cell system. At best they were vestiges of ancient symbioses, but one could not be certain. Debates over their exogenous or endogenous origins centered on the complexity of the stories told, and different accounts were assessed on the rather dull edge of Occam’s razor. All was a matter of speculation without direct phylogenetic evidence.

Bacteria lacked morphological complexity; they lacked a fossil record; their evolutionary relations could not be known by the methods of classical taxonomy used for plants and animals. Leading microbiologists insisted that the evolutionary history of the microbial world would always be unknowable, and therefore discussions of cell origins were doomed to be fruitless exercises in metascientific speculation. Famed microbiologist Roger Stanier (1971, p. 31) spoke for many when he addressed ideas about the symbiotic origin of organelles:

It might have happened thus, but we shall surely never know with certainty. Evolutionary speculation constitutes a kind of metascience, which has the same intellectual fascination for some biologists that metaphysical speculation possessed for some medieval scholastics. It can be considered a relatively harmless habit, like eating peanuts, unless it assumes the form of an obsession; then it becomes a vice.

Though such views were widespread, they were also soon out-of-date. Pauling and Zuckerkandl (1965) had already heralded the field they called “molecular evolution.” The sequences of amino acids and nucleic acids would serve as phylogenetic “traits” of a wholly different kind than the morphological traits of classical taxonomists. Evolutionary relatedness through molecular methods was quantifiable, and phylogenetics could reach deep into evolution’s past. During the 1970s, Carl Woese and collaborators at the University of Illinois began to construct a universal tree of life based on comparisons of the small subunit ribosomal RNA which was conceived of as “a universal chronometer” at the core of the genetic machinery of all organisms (Sapp 2009). Their research led to the taxonomic proposal of three domains: Bacteria, Archaea and Eukarya, representing the three primary lineages of life (Woese et al. 1990).

Molecular phylogenetic methods transformed cell origins from a field of speculation to one of empirical science. Evidence based on SSU rRNA comparisons showed irrevocably that mitochondria and chloroplasts were vestiges of ancient bacterial symbiosis: mitochondria were of alpha proteobacterial ancestry, chloroplasts of cyanobacterial descent. Centrioles lacked DNA and ribosomes; their ancestry remained unknown (Sapp 1998). The nucleus was also shown to be chimeric. It comprised informational genes (affecting replication, transcription, and translation) of Archaeal origin, and genes of Bacterial origin, and perhaps bacteria of still different ancestry- the chronocyte, now extinct (Hartman and Feerov 2002). The discovery of complex and large DNA mimiviruses supports the hypothesis that DNA viruses also may have been involved in the emergence of the eukaryotic nucleus (Bell 2001; Claverie 2006). The eukaryotic cell did not evolve in the classical Darwinian way by gene mutation and recombination within species, but rather through symbiotic mergers.

Nor do Bacteria (sensu lato) evolve in the classical Darwinian manner. Molecular phylogenetics data indicate that horizontal gene transfer between “species” is widespread in the bacterial world. Indeed, the universality of the genetic code may be due to horizontal transfer of genes among lineages at the dawn of life (Syvanen 1985). Genomic studies, beginning in the 1990s, have revealed that much of evolution is due to innovation sharing through the inheritance of acquired genes and genomes (Villarreal and Ryan 2011; Gontier 2015).

Viral infection has had an important role in human evolution as well. At least 8 % of the DNA in the cells of mammals has been shown to have an exogenous origin (Horie 2010). Though the function of many of those genes remains unknown, the effects of some have been profound. Genes of viral ancestry may have been involved in key events which led to the evolution of placental mammals from their egg-laying ancestors (Mi et al. 2000; Harmit 2012). Placental morphogenesis may have been facilitated in human and numerous mammalian families by the integration of viral DNA into the genome (Malik 2012).

The evidence for the evolutionary importance of the inheritance of acquired genes and genomes is overwhelming today. Although classical evolutionists had viewed hereditary symbiosis to be a curiosity, an evolutionary anomaly, the “quirky and incidental side of evolution” (Gould 1989, p. 310), it is also ubiquitous. In the vast and diverse world of heterotrophic protists, which feed on bacteria and microscopic algae, the adage “you are what you eat” may be taken literally (Gast, Sanders and Caron 2009). Some neo-Darwinian evolutionists had predicted that the transmission of symbionts through the egg would be rare in animals, and when it does occur, the microbes should be considered “encapsulated slaves” (Maynard Smith and Szathmáry 1999, p. 141). These assessments have also proven to be faulty.

Hereditary symbiosis is widespread among insects. The symbiosis between aphids and bacteria (Buchnera) is a 200 million year old relationship. The bacteria, inherited maternally through the egg cytoplasm, supply essential amino acids to the aphids, and other metabolites are transferred between the partners (Russell et al. 2013). Bacteria of the genus Wolbachia are inherited through the egg cytoplasm of 25–70 % of all insect species as well as nematodes in which they are essential for normal development and fertility (Taylor et al. 2005). In some cases, Wolbachia genes have been transferred to the host genome (Dunning et al. 2007). Far from being “slaves”, Wolbachia are master manipulators, affecting the reproduction of their insect hosts to aid in their own reproduction. They cause parthenogenesis, the production of female offspring without fertilization sperm; they prevent uninfected females from reproducing with Wolbachia-infected males; and they can cause feminization of infected male insects, transforming them into fully reproductive females (Warren et al. 2008). Males are an evolutionary dead end for Wolbachia. Hereditary symbionts cannot be presumed to be slaves anymore than they can be supposed to be coevolved mutualists.

Most symbiotic microbes are acquired from the environment during the lifetime of a plant or animal and they too have profound effects on the physiology and development of their hosts. In plants, fungi living in the tissue of leaves ward off pathogens and herbivores, and they act as vaccines to boost the plant’s immunity (Herre et al. 2007). The dynamic relations among plants and the microbes in their rhizospheres is at the core of one of the most perplexing questions in tropical biodiversity: how so many species of trees can coexist in a tropical rainforest such that a single hectare may hold 300 species (Mangan et al. 2010).

The great diversity in animal developmental patterns evolved first in the world’s oceans swarming with microbes. All stages of the life history of marine animals are adapted to microbes bathing in their tissues, and countless animal-bacteria associations have coevolved (McFall-Ngai 2002). Tropical corals are chimeric “plant-animals”. At least 60 % of their nutrition is supplied as carbohydrates by photosynthetic algae living in their cells, which they in turn provide with nitrogen. That symbiosis is broken with prolonged increase in sea-surface temperatures due to global warming. Corals “bleach”; they starve and die. When compounded with overfishing, pollution, ocean acidification, and disease, the world has been led to the brink of a massive extinction of tropical coral reefs (Knowlton and Jackson 2008). Those corals that have resisted diseases from pathogens and high temperature, however, have done so through changes in the populations of their algal and metabolically diverse bacterial symbionts (Reshef et al. 2006). This same evolutionary and ecological understanding of the organism as a symbiotic complex is critical to understanding our own well-being.

The Ecology of Health and Disease

Molecular phylogenetic methods have revealed a great diversity in the microbial populations in and on us. If the 20th century was “the century of the gene,” the 21st will surely be the century of the microbe and of symbiosis. The microbial world, which represents the greatest biomass on Earth, also has the greatest genetic and biochemical diversity (Bhat 2013). Yet, less than 1 % of microbes observable in nature can be grown as pure cultures in the laboratory and characterized by the methods of classical microbiology (Vartoukian et al. 2010). The same holds for the 100 s of bacterial “species” that live in an anaerobic environment in our gut.

Metagenomics has revitalized microbial ecology as the genomic structure and function of entire microbial communities can be characterized and sampled in their natural habitats without the need for culturing. Lederberg and McCray (2001) coined the term “microbiome” in 2001 to signify “the ecological community of commensal, symbiotic, and pathogenic microorganisms that literally share our body space” and which, he said, “have been all but ignored as determinants of health and disease.” Studies of our microbiomes have grown quickly as fast and cheap ways to sequence and analyze large genomes with “high-throughput sequencing methods” were developed during the first decade of this century (Gordon 2012).

Mammals begin to acquire bacterial communities the moment the amnion breaks and infants suckle and hug. Most microbes live in our gut, sometimes referred to erroneously as gut “flora”, a carryover from the days when bacteria were considered to be plants. Gut bacterial communities carry out many functions formerly thought to be the province of animal cells. They aid in food digestion, produce vitamins, and protect against pathogens. Studies of “germ-free” mice have further shown that the digestive system and the immune cells in the intestine do not develop properly without the microbiota of the gut (Lee and Mazmanian 2010).

Diet and host shape those microbial communities after birth. Newborn infants cannot digest some of the diverse carbohydrates in human milk; instead their nutritional function is to serve as prebiotics for certain gut bacteria (Marcobal and Sonnenburg 2012). Our appendix, long thought to be a vestigial organ of little use, like the muscles connected to our ears, may actually serve as a reservoir for gut bacteria, to be rapidly restocked when reserves are depleted after bouts of diarrhea (Bollinger et al. 2007). Our gut microbiota have systemic effects on liver function, and may well affect other organs Björkholm et al. 2009).

The delicate symbiotic relationships that confer benefits in many aspects of our lives can be broken by environmental changes in diet, infection and life style, thereby promoting disease. Diet and infection can increase and decrease susceptibility to cancer, for example, by changing the species composition of our bacterial communities (Schwabe and Jobin 2013). Various inflammatory diseases of the skin, mouth and digestive tract have been linked to changes in our microbial communities. Gut microbial community composition may predispose an individual to obesity (Turnbaugh and Gordon 2009). An imbalance in the species composition of our microbiomes has also been associated with type II diabetes (Quin et al. 2012).

Maintaining species diversity in our microbial ecological community is also fundamental to preventing asthma, the incidence of which is soaring. Traditionally, our respiratory tract was thought to be sterile, and asthma thought to be the result of a misdirected immune reaction. Not so today. Asthma is linked to a complex community of bacterial species living in our respiratory tract, and exposure to bacteria early in life helps to prevent it (Couzin-Frankel 2010; Huang et al. 2014). The data has come from diverse sources. Farm children are much less likely to get lung disease. Babies born by cesarean section are more likely to acquire asthma than those born vaginally, just as are children who have been treated with many doses of antibiotics. Children who develop asthma not only have different bacteria, they also have lower bacterial diversity. It is not simply the presence or absence of a certain species of bacteria that matters, but the composition of the entire microbial community. Our immune system is maintained through a balance in our bacterial communities. The one germ-one disease approach of classical germ theory must be supplemented with an ecological community-based approach to our microbiomes (Costello 2012a, b).

There is also evidence that our gut microbiota modulates our brain development and behavior. Changes in the gut bacterial community composition can affect motor control, anxiety and cognitive function (Herre et al. 2007). Autism has been linked to changes in the species composition of our gut communities. Children with autism have abnormal and less diverse communities of digestive bacteria. Harmful metabolites are produced that either function as neurotransmitters or effect neurotransmitter synthesis (Kang et al. 2013). Probiotic therapy may alleviate the problem in some cases (Hsiao et al. 2013).

The day has finally arrived when the germ theory of disease is complemented with the germ theory of morphogenesis and differentiation. And a new biologic philosophy emerges as the one genome-one organism conception yields to a symbiotic view of life. We are beginning to understand that nothing in evolution makes sense except in light of symbiosis. The concept of the symbiotic organism, which appeared now and then throughout the history of biology has been strengthened by new methods for investigating the microbial world. Evolution’s two-step process of divergence and integration is evidenced in the microbial communities that participate in the modulation of our physiology, development and behaviour. A dynamic, multispecies community-based conception of the organism emerges as genomics is transformed to symbiomics.