The Prospects of Artificial Endosymbioses

Ryan Kerney. Zakiya Whatley, Sarah Rivera. David Hewitt. American Scientist. Volume 105, Issue 1. January/February 2017.

In the engineering of biological systems, it can be said unequivocally that art imitates nature. Nearly all efforts to control human health, the environment, and agriculture involve the appropriation of evolutionary processes. These processes typically originate through incremental changes in the genome that are sustained and promoted through natural selection in descendant lineages. Recombinant DNA technology and more recently genome editing help us imitate these genome-level changes in engineered systems. However, the dramatic evolutionary innovations that are attributed to singular beneficial endosymbioses, in which a mutualist microbial cell inhabits a host’s cell, are also worthy of imitation. For example, researchers are studying how to engineer endosymbiotic bacteria to control mosquito-borne viral diseases, tweak nitrogen-fixing microbes to help crop plants, and treat macular degeneration, just to name a few projects that are under way.

Endosymbioses have arisen independently many times in nature. They are essential for many plants, which use them to take up vital nutrients or defend against herbivores; in insects, their presence or absence may determine gender or population structure, and they are often needed for specialized diets; they allow many lineages to manufacture their own food through photosynthesis or chemosynthesis; and within our domain of life (eukaryotes), they are the origin of the organelles that manufacture energy storage molecules through photosynthesis (chloroplasts) and convert this energy for use in our cells (mitochondria). Indeed, chloroplasts and mitochondria are examples of just how successful and game-changing endosymbionts can be.

The power of endosymbioses to lead to innovations has not escaped the attention of modern bioengineers. The establishment of a novel symbiont in an otherwise naive host has the potential to radically alter the host cell physiology in many ways, without directly affecting the host genome. These approaches have a range of applications, which include public health, agriculture, medicine, and basic research. As biologists better understand these relationships, the potential grows for people to move endosymbionts from one organism to another to transfer or establish novel benefits in a new association.

While genetically modified organisms (GMOs) are often highly controversial, the prospects of co-opting symbiotic relationships are apparently more ethically palatable. Pedro Gundel and colleagues coined the term symbiotically modified organisms (SMOs) in 2013 to describe artificial fungal endosymbionts (endophytes) in grasses, even though by that time the approach had already expanded to other applications. To date, SMOs have not encountered the same ethical scrutiny as genetic modifications, even though naturally occurring endosymbioses are known to have had large and sweeping effects on ecology and evolution in the past. Although we are not arguing against these new applications here, we believe that it is useful to put the ethical contrasts between SMOs and GMOs into perspective.

Surprisingly, the study of artificial endosymbioses has a long history. However, an appreciation for the implications of early research from the 1930s has emerged only recently.

A Pioneering Artificial Endosymbiosis

Biologists still know little about how new endosymbioses become established, a point that intrigued the great invertebrate zoologists Ralph and Mildred Buschbaum when they set up a pioneering artificial symbiosis in 1934. The Buschbaums knew that many protozoans, jellyfish, corals, and flatworms were able to survive without feeding because of their algal symbionts, which make food through photosynthesis. In some instances, such as in giant clams, these algae live inside specialized tissues, but not within the cells, of the host. However, the majority of these unique invertebrates harbor intracellular algae (often belonging to the genera Symbiodinium or Chlorella, or to “blue-green” cyanobacteria). These photosynthetic cells reside inside cells of their host and provide simple sugars, lipids, or amino acids that are generated through photosynthesis. The algae in turn benefit from the host’s carbon dioxide and nitrogenous wastes, and from the intracellular refugia, making these interactions mutualistic endosymbioses. The Buschbaums were also aware that similar algal associations were not known for vertebrates. So they decided to form an artificial symbiosis in the lab with cultured vertebrate tissue expiants and the unicellular green alga Chlorella.

The Buschbaums’ co-cultures of Chlorella with chicken and amphibian cells revealed an apparent benefit to both partners. Although the exchange of metabolites was not measured, the Buschbaums did describe in their 1937 paper in Physiological Zoology a “marked” effect on the growth and “health” of both partners, and that result was later quantified by growth measurements. The co-cultured algae were greener and more abundant than controls, and the mixed vertebrate tissue cultures remained healthy for twice as long. The Buschbaums also noted something entirely unexpected: Embryonic chicken fibroblasts, which are cells derived from connective tissue, “occasionally took up” the algal cells in culture. This adoption created not only an “artificial symbiosis” but an artificial endosymbiosis in cells that otherwise have no business acquiring foreign microbes. These fibroblasts had reduced fat stores and “appeared to be much healthier” than controls. This uniquely engineered endosymbiosis appeared to replicate both the reciprocal benefits as well as the intimate cellular associations found in marine invertebrates and microbial protists (for example, Paramecium). However, their experimentally derived endosymbiosis was a first for vertebrate cells.

The Buschbaums had stumbled upon a discovery that was not fully appreciated until many decades later. Little is known even today about the rules of engagement that establish cooperative, rather than parasitic, intracellular interactions. Even the Chlorella algae they used to form their artificial symbiosis can occasionally become parasitic in rare cases of sheep, dog, gazelle, or even human infections reported in the medical literature. There is only descriptive research on these rare parasitic Chlorella infections, and the mechanisms that cause this alga to become harmful are currently unknown.

But as the Buschbaums’ and others’ work on Chlorella demonstrates, the line between mutualist and parasite can be vague. The development of any endosymbiosis includes the trade-offs of burden and benefit, as well as the unique evolutionary trajectory of an endosymbiotic microbe. These merit close scrutiny as artificial endosymbioses are increasingly employed in many novel applications-and some are already in use, often under outdated regulations for preventing unintended environmental or health consequences.

Control of Mosquito-Borne Diseases

Some of the most encouraging research on artificial endosymbioses involves mosquitoes. The use of a bacterial endosymbiont called Wolbachia in novel mosquito hosts is rapidly being developed to control several infectious diseases, including dengue, Zika, Chikungunya and, to a lesser extent, malaria. There is already a tremendous amount of basic research on Wolbachia, primarily because they have unique abilities to spread through a host population by manipulating genders. This common symbiont of many species of insects and nematodes can be a remarkably well-tolerated “reproductive parasite” or even a mutualist, depending on the particular Wolbachia strain and its host. Wolbachia lives inside the cells of its hosts’ organs, often in the ovaries or testes, and it can be transmitted to all offspring of an infected female. The bacteria also often abort the development of offspring from infected males mating with uninfected females through a process called cytoplasmic incompatibility. Both of these effects promote the rapid spread of Wolbachia infection throughout a population.

Although Wolbachia infections have been found in other mosquito species, there is no native infection in the dengue-transmitting species, Aedes aegypti, or most malaria-transmitting Anopheles species. By microinjecting Wolbachia cells into mosquito embryos, Australian researcher Scott O’Neill and his team have produced artificial endosymbioses with multiple strains that are sustained as these maternallytransferred intracellular symbionts are passed down from one Aedes aegypti generation to the next. Similar successful SMOs have also been established with the related Aedes albopictus and the malaria host Anopheles stephensi (although with limited inhibition of the malaria parasite).

Australia has seen two successful releases of Wolbachia SMO mosquitoes to control the devastating viral disease dengue. These efforts began with multiple small-scale trials in 2011 and expanded to a city-wide trial in Townsville, Australia, in October 2014. This trial was followed by expansion to other Queensland sites. Wolbachia-infected mosquito projects are currently under way in Brazil, Colombia, Indonesia, and Vietnam, where this pathogen and related viruses present a more pressing problem. These artificial endosymbioses exploit the unique biology of Wolbachia and its relationship with the host insect.

Depending on the bacterial strain employed, the Wolbachia-A. aegypti artificial endosymbiosis results in a shortened life span for the female mosquitoes (and to a lesser extent for the male ones) or a decreased ability to harbor or transmit viruses such as dengue. Subsequent studies have shown that Wolbachia infections limit the transmission of the Zika and Chikungunya viruses as well. The virus control mechanisms are not well known but appear to occur inside the host cells. Endosymbiotic Wolbachia organisms rarely share this space with viruses in a mosquito or in other insect hosts. This is apparently due to resource competition for ions or lipids, as well as some intrinsic inhibition of viral replication through a currently unknown mechanism. Although there is some evidence for an up-regulation of the insect’s immune response when Wolbachia organisms are present, experimental models in the fruit fly Drosophila suggest that the bacteria take a more passive role in reducing viral load.

In Australian neighborhoods where Wolbachia-infected mosquitoes were released in 2011, the bacteria’s frequency in resampled mosquitoes was as high as 90 percent. Still, it is too early to know what effect widespread Wolbachia infection will have on the long-term rate of dengue transmission. Researchers are monitoring these initial releases closely in hopes that the Wolbachia-intected mosquitoes are not at a selective disadvantage to local uninfected populations and that their virus-suppressing effects continue over multiple generations in these novel wild hosts. Results to date have been encouraging, since no dengue outbreaks have occurred in these areas since the releases.

However, in advance of releasing Wolbachia-infected mosquitoes, O’Neill and his colleagues completed extensive testing to measure the ecological impacts of the mosquitoes with Wolbachia on regional food webs, as well as the their potential to transmit the bacteria to other arthropod or human hosts. Both were determined to be negligible. They also addressed community concerns through focus groups, thorough interviews with identified stakeholders, and telephone surveys of random community members. This community outreach and transparent ecological testing was essential for the adoption of the SMO dengue-control system. By all accounts these efforts were a thorough assessment of potential off-target impacts and an example of vital community engagement.

The acceptance and use of Wolbachia in mosquito vector control has several direct parallels with GMO approaches toward the same ends. However, the ethical considerations for each approach have been weighed differently in public opinion and regulatory oversight. Currently, both avenues are under exploration either to limit a mosquito’s ability to transfer a pathogen or to sterilize male mosquitoes and release them into a wild population.

Such control efforts are not without precedent. From the 1950s through the 1970s, the release of x-ray-sterilized males effectively controlled the screwworm, a devastating parasitic maggot in humans and livestock. Subsequent sterile male eradication programs have been used against multiple insect pests, including mosquitoes.

Genetic modifications of A. aegypti have led to the development and release of male mosquitoes that are effectively sterilized GMOs. This process requires a dominant lethal gene cluster in male mosquitoes. These male mosquitoes then pass on their lethal gene cluster to all of their offspring, so that none end up living. Because these labreared mosquitoes competitively breed with females once they are released into the wild, the high numbers of unviable offspring result in a profound crash in the population.

Oxitec is a company with successful mosquito-control campaigns in the Cayman Islands, Brazil, and Panama that use sterilized GMO males. With a small trial proposed in Florida, they have recently focused on public education and outreach. However, their reception hasn’t been universally positive. On November 8, 2016, residents of Key Haven in Monroe County cast referendum votes against the use of sterilized GMO A. aegypti to control dengue and Zika in their area. Even after the U.S. Food and Drug Administration (with support from the Centers for Disease Control and Prevention and the Environmental Protection Agency [EPA]) found no significant environmental effects of using genetically modified mosquitoes, there was a split vote. The Florida Keys Mosquito Control District of Monroe County recently approved the release of genetically modified mosquitoes despite local opposition in Key Haven, the initial proposed release site. The board has decided to continue with the release at currently undetermined sites in the Florida Keys.

A second approach to controlling Aedes mosquitoes uses artificial Wolbachia infections that mimic the sterile-male approach. In 2015 the company MosquitoMate from Kentucky test-released Wolbachia-infected male mosquitoes in Los Angeles. When they mate with an uninfected female, their offspring fail to develop, so they are effectively sterile. Florida appears to be more open to this SMO approach than it is to the release of Oxitec’s GMO mosquitoes. The EPA granted an extension of their Experimental Use Permit for Wolbachiaintected A. aegypti in September 2016 in Monroe County, and the mayor of Miami-Dade said that mosquito control had already contacted the University of Kentucky to conduct a trial of sterile SMOs as part of the state’s Zika response, according to the Miami Herald.

Use of Wolbachia offers an alternative to another promising but even more controversial genetic control technique. Mosquitoes can be genetically modified to become worse hosts for a human pathogen. However, spreading these modifications through wild populations is difficult. Establishing them in a population requires a gene drive that will create transmission rates higher than those of Mendelian inheritance. The most prominent recent advances in gene drive technology exploit a sequence-specific DNA targeting mechanism in bacteria called CRISPRCas9 (CRISPR stands for clustered regularly interspaced short palindromic repeats). This technology has revolutionized precision genome editing. In the CRISPR-Cas9 gene drive approach, the introduced DNA on one chromosome (say, from the father) contains the ingredients to manipulate DNA on the matching chromosome from the mother. Both the foreign transgenes that mediate CRISPR-Cas9 editing and those that affect the transmission of diseases are transferred from one parent’s chromosome to the other. This results in 100-percent transmission of the transgene, which eventually spreads through an entire population.

A recent National Academies of Sciences conference focused on ethical concerns regarding the use of gene drives to spread transgenes in wild populations. Although the mechanisms are different (and more specific), the ability of a gene drive to selfpropagate is similar to the ability of viral infections to spread. The National Academies of Sciences urged extreme caution in the development of these techniques for fear of off-target effects. For instance, it is unclear whether these self-propagating transgenes would have the ability to jump into new nonmosquito genomes, including potentially our own.

The Wolbachia approach to controlling A. aegypti transmission of viruses has similarities to, as well as dramatic differences from, the gene-drive approach. Maternal inheritance in Wolbachia organisms and their ability to essentially induce sterility when infected males mate with uninfected females means that they are similarly capable of spreading through a population extremely rapidly. Despite advances, gene-drive technology will likely be held up by ethical concerns for the foreseeable future, whereas Wolbachiabased mosquito control is already in use. In A. aegypti, Wolbachia infection is a comparable substitute for gene drivers to control transmission of these viruses. However, to date, similar Anopheles mosquito SMOs have exhibited limited success in reducing the transmission of malaria through Wolbachia infection. Unlike the Wolbachia SMO, the CRISPR-based gene drive approach to malaria control using antibody transgenes has great potential, although projects using this approach are currently also held up by ethical concerns.

Although the current excitement about the Wolbachia intervention is warranted, there remains room for caution. Wolbachia-containing mosquitoes are reproductively less fit than the wild-type mosquito and could be vulnerable to competition and prone to eventual loss in the wild population, in which case multiple releases might be required. In addition, the persistence of Wolbachia’s antiviral effects has not been evaluated through any longitudinal study. This persistence may be limited by the continued evolution of the Wolbachia bacteria, the virus pathogen, or the mosquito host. Because of Wolbachia’s maternal inheritance and ability to spread through a population, it is likely the most successfully infectious bacterium known. However, Wolbachia species have coevolved with their native hosts. Whether the bacteria maintain the same types of relationships with novel mosquito hosts remains to be seen.

Symbiotically Modified Crops

Controversy surrounding GMOs and their effects on human lives is most prominent in agriculture, where there is already widespread use of genetically modified crops, including soy, com, and cotton. Transgenes from distantly related organisms confer resistance to herbicides or diseases, generate insecticides, protect from drought, or-potentially- help provide micronutrients to improve human well-being. These approaches continue to be under intense public scrutiny, With many detractors concerned with off-target effects on ecosystems or on human health, as well as commercial concerns regarding intellectual property rights and agricultural industry consolidation. Given the ongoing arguments surrounding GMOs, a potential role in agronomic technology for SMOs could be a less resisted path to feeding the world safely and sustainably. But SMOs may help address public fears without affecting actual risk.

Naturally occurring bacterial and fungal endosymbioses are integral to the functioning of agricultural ecosystems and already provide opportunities for engineered systems or potential for transformative advances. One of the leading areas of research focuses on nitrogen-fixing bacteria that live inside the roots of legumes (and alder trees) and “fix” atmospheric nitrogen into biologically available forms of this essential nutrient, such as ammonia. A second group of endosymbionts, arbuscular mycorrhizae, are fungi that enter the plant root’s cell walls, where they can shift nutrients from soil to the plant, and in turn gain carbohydrates made by those plants. Finally, endophytes, which are fungi that live in leaves or other aboveground plant tissues, are also widespread and critical to plant ecology, with effects on water management and resistance to herbivory. All three categories of these symbionts are critical to ecosystem functioning across an astonishing breadth of biological communities, climatic regimes, and levels of human impact in rural, suburban, and urban habitats. All are currently used extensively in modern agricultural systems. These symbionts’ uses can supplement or even replace the role of genetic modifications that have potential for expanded capabilities, including genetic modifications to the symbiont or genetic and symbiotic modifications to one host.

Nitrogen-fixing bacteria can help to allay one of the most environmentally impactful practices of modern agriculture: the use of nitrogenous fertilizers, which can lead to runoff that causes algal blooms and oxygen depletion in aquatic ecosystems. Because nitrogen is already abundant in the atmosphere, but not in a form that most plants can use, nitrogen-fixing symbionts reduce the required amount of applied fertilizer by making the plentiful atmospheric nitrogen usable to their hosts. Nitrogen-fixing bacteria are already a major component of modern agronomic systems. Unlike Wolbachia, nitrogenfixing bacteria are not inherited but instead are acquired from the environment through the roots, thereby conferring nitrogen-fixing capability to alders and legumes, and are thus a major reason that leguminous cover plants are used extensively on crop fields.

Establishing nitrogen-fixing bacteria in novel hosts could significantly reduce nitrogenous fertilizer application and subsequent increased burden of nitrogen on adjacent and hydrologically linked ecosystems. However, novel host-symbiont pairing would likely require extensive genetic manipulations, because the establishment of intracellular, bacterial nitrogen-fixers requires complex signaling between both sym- biont and host. The factors involved, which include bacterial signaling proteins such as nodulation (Nod) factors that specifically activate host signaling pathways (called SYM pathways), are under intensive research and may provide inroads into nonlegume uptake of symbiotic bacterial nitrogen-fixers through genetic modifications to either the bacteria, the plant, or both. These plants would then be regulated under the same governmental oversight as other genetically engineered plants, thereby reducing the “non-GMO” benefit of those SMOs.

Several applied fungal endosymbioses are already commercially available in agriculture. Numerous strains and communities of mycorrhizae (some kinds live inside plant cell walls; others live outside them) are available commercially for use in residential and agricultural use, with different products marketed for specific applications (for example, lawns, pastures, or ecological restoration). Like nitrogen-fixing bacteria, mycorrhizae are also acquired from the environment. Their artificial symbioses are established by inoculation- that is, the microbial organism is applied directly to the surrounding soil or the plant. However, the efficacy of controlling mycorrhizae varies between strain and application. Not all mycorrhizal interactions are equivalent (for example, some can reduce plant biomass). The inoculum source and its associated local adaptation both play roles in how well mycorrhizae affect the desired plant growth.

Mycorrhizae and nitrogen-fixing bacteria both trigger the host cell’s conserved symbiotic signaling (SYM) pathway, which results in periodic calcium concentration increases in the nucleus and eventually symbiont acquisition. This common host response suggests that the artificial acquisition of nitrogen-fixing bacteria may only require minor modifications in plants that harbor intracellular mycorrhizae. However, nitrogen-fixing bacteria also require specialized transport systems and a low-oxygen microhabitat that would likely be difficult to replicate and may come at a significant cost to the host plant.

These hurdles have led many researchers to focus on a strict GMO approach to acquired nitrogen fixation. Although that approach has its own unique set of challenges, it is arguably more precise and practical than the many modifications required for SMO acquisition. Although SMOs may pose risks and benefits similar to those of GMOs, artificial nitrogen fixation through SMOs likely will take longer to develop and will have less accuracy than a solely GMO-based approach.

Artificially introduced grass endophytes are already commercially available that can lead to more efficient water use in pasture grasses and to reduced herbivory on such grasses through the symbionts’ release of alkaloids, thus moderating the need for applied insecticides. Given that antiherbivory compounds may affect livestock, strains that are not toxic to sheep and cows have been developed for agricultural use. Endophytes are easy to apply-the inoculum just needs to be added to the plant stem. Unlike mycorrhizae or nitrogen-fixing bacteria, which are acquired from the environment each generation, the grass endophytes are passed down maternally once established. However, a downside to their use is that they can become parasitic and cause “choke” disease in the host plant.

GMOs may present a false dichotomy with SMOs. Many endosymbionts require genetic modification in their genome or in that of their target host prior to their use. For example, endophyte strains can be genetically engineered or selected for reduced toxicity to livestock. Similarly, the potential acquisition of nitrogen-fixing bacteria would require extensive genetic modifications to a novel host to successfully enable their profound and needed environmental benefits.

Some of the valid concerns about GMOs may well carry over to SMOs. Concerns about off-target toxicity of pesticides produced by a transgenic plant may also be applied to endophytes that produced anti-herbivory toxins in grass. Concerns about the “escape” of highly competitive genetically engineered plants that could then become weedy, such as those raised in response to development of glyphosateresistant bentgrass (Agrostis stolonifera, a commonly used turfgrass), could be applied to a grass imbued with a particularly adaptive endophyte. Concerns about the use of mycorrhizae and nitrogen-fixing bacteria with novel crop or pasture plants could include their potential to reduce biodiversity with minimally diverse agroecosystems, as these plants still would likely be planted in fields that contain few or only one species or strain.

Given the ubiquity of these symbiotic relationships in nature (relationships not brought about by direct human intervention), the general lack of genetic engineering involved, and the ease of their application, the use of SMOs can be seen as “natural” in comparison with GMOs. However, “natural” does not imply “without risk,” because many of the concerns that are raised for GMOs may apply to SMOs as well.

Vertebrate Artificial Endosymbioses

Most efforts have focused on plants and invertebrates, because that’s where many known examples of endosymbiosis occur. However, following the work of the Buschbaums, artificial endosymbioses have recently been extended to vertebrates as well. Although endosymbioses appear in many branches of eukaryotic life (meaning all organisms that keep their DNA in a nucleus), there is only one known example of a mutualist naturally entering vertebrate cells to form a symbiotic interaction. Cells of the green alga Oophila amblystomatis enter tissues and cells of the spotted salamander Ambystoma maculatum during the amphibian’s development. The acquisition of this alga has several similarities with the uptake of “algal” Symbiodinium by coral hosts, although Oophila algae produce only a limited photosynthetic benefit. Perhaps because there are so few close natural analogs, artificial endosymbioses may not be as readily accepted in vertebrates as they are in agriculture.

Nevertheless, there have been several attempts at establishing an artificial endosymbiosis in vertebrates using novel partners, much as the Buschbaums did in their experiment. Such approaches in the laboratory are mostly focused on introducing various types of algae into cell cultures.

Few attempts were made immediately following the Buchsbaums’ experiments to replicate their “invigorating” artificial symbiosis. In the late 1970s Dennis Taylor from the University of Miami had similar success with co-cultures between photosynthetic marine flagellates and fish expiants or chicken-cell cultures. The topic has only recently been revisited by synthetic biologists working on experimental systems. These attempts have included genetically modified cyanobacteria (Synechococcus sp.). Similar unmodified Synechococcus organisms are capable of establishing an artificial endosymbiosis with paramecium hosts after their naturally occurring algal symbionts have been removed. In vertebrates, however, the genes from human pathogens that produce invasin and listeriolysin were introduced to Synechococcus to enable its cellular entry and to bypass intracellular digestion mechanisms. These transformed bacteria were shown to enter and live in human cells in culture, resulting in the first human cell-algae endosymbiosis.

Although this work is still exploratory, artificial endosymbioses have exciting potential for medical treatments. Fór example, putting algae cells into human eyes could stave off age-related macular degeneration, a leading cause of impaired vision in the United States. Following the work on Synechococcus, researchers discovered entry of the algae Nannochloris eukaryotum into human cells after screening II algal strains for their ability to enter cultured eye (retina) tissue. This entry reduces expression of vascular endothelial growth factor, which can treat many forms of macular degeneration, and also increases cell viability. Unlike Synechococcus (a bacterium), N. eukaryotum is a eukaryote, as its name implies. It was isolated from a saltwater tank in the former Yugoslavia in the early 1980s and has been maintained in culture for more than 30 years. Phycologists (algal scientists) were initially interested in its uniquely small size (2-5 micrometers) and simple structure. The “voluntary” ability of N. eukaryotum to enter human cells also reveals a latent potential of microbes to interact with vertebrate cells, including human cells, in unexpected ways. This association did not require genetic manipulations of the symbiont or host and suggests that more vertebrate endosymbioses may exist without our knowledge.

What other value is there in an artificial endosymbiosis of vertebrates, aside from the potential to treat macular degeneration? One futuristic possibility is in the field of “synthetic” meat production from cultured muscle precursor cells. Current attempts at cultured meat production require animal feedstocks. However, artificial photosynthefic symbionts could potentially lead to meat cultures that can generate their own food from sunlight, creating a guilt-free product that doesn’t require butchering of the feedstock and that would have a much lighter environmental footprint. Other “futuristic” possibilities include the delivery of metabolites that augment host-cell physiology through genetic manipulations of an endosymbiont itself. The creation of a microbial delivery service into diseased or aberrant cells could lead to targeted cellular therapeutics without germline manipulations or the use of modified viruses.

Additional possibilities outside of photosynthesis and drug delivery abound. For example, artificial endosymbiosis could allow grazing animals to eat plants that they currently cannot digest. The insect gut microbiome includes microbes that aid in wood digestion, sap feeding, and blood feeding, tasks often attributed to intracellular bacterial symbionts. Expanding the digestive abilities of livestock could also help sustainably repfirpose fallow lands and reduce the competition for land between grazing livestock and crops.

The Trade-Offs of Symbioses

Artificial endosymbioses have the potential to dramatically alter host cell physiology, organismal biology, and ecosystem functioning without directly manipulating the genome. This potential has already been realized or is rapidly advancing in numerous systems. This rapid adoption is occurring in part because public acceptance and regulatory oversight of these approaches are distinct from acceptance and oversight of genetic modifications through genome editing or use of recombinant DNA technology. This difference in reception is apparent in the far greater intensity of debates surrounding GMOs, compared with those surrounding SMOs. Modifying symbioses may appear to be more “natural” than use of those other technologies, which can imply to many a reduced level of risk or an increased level of quality. However, the line between “natural” and “artificial,” as with so many distinctions, becomes less clear as we understand more about the methods and outcomes of individual technologies.

SMOs can, arguably, result in more dramatic physiological changes to a host cell through a less controlled modification than occurs with genetic engineering. The transfer of an endosymbiont is a less precise manipulation than altering a genetic code, because an entire organism-not just a single gene or limited set of genes-is introduced. Most changes on a nucleotide level are, from an individual cell’s perspective, less consequential than accommodating a foreign microbe. There is no reason to expect a priori that a foreign endosymbiont will obey the same rules of interaction as a native association. The latter has invariably resulted from a co-evolutionary process that can often result in dramatic genomic integration. However in engineering, as in evolution, profound changes can occasionally be a “good” thing, regardless of precision. Again caution is warranted, and we need to evaluate individual technologies. SMOs should not be seen as necessarily more or less risky than GMOs in any sweeping categorization.

Of key importance is the expectation that artificial symbionts will have consistent effects on their hosts. Basic research on symbiotic associations has paid tremendous attention to the establishment and maintenance of mutualistic associations. Reciprocal exchanges can often result in “cheaters” infiltrating transient mutualistic associations. Meanwhile, permament endosymbiotic associations have recently been described as establishing a “symbiont prison” for an intracellular microbe that becomes dependent on its host. The latter may be preferable in SMOs for maintaining a consistent, and controlled, intracellular relationship.

The one guarantee of artificial endosymbioses is that they will continue to evolve with or without our intervention. Maintaining an endosymbiotic mutualism requires more than the accidental cellular fusion discovered by the Buschbaums, because the reciprocal costs and benefits need to be both established and maintained. As Charles Darwin wrote in The Origin of Species by Means of Natural Selection: “Natural selection cannot possibly produce any modification in any one species exclusively for the good of another species.” Instead SMOs will always require, to varying extents, some degree of compromise.

Gray Area Between Parasitic and Mutualistic Relationships

Symbiotic relationships change over time, and biologists will need to understand and manage that change if people want to put such relationships to deliberate use. One of the biggest challenges is that parasitic relationships develop into mutualist ones and vice versa, so that relationships that at first seem useful can end up countering the human intent. Longitudinal studies of symbiotic and parasitic interactions reveal some of these transitions. For instance, the classic work in the late 1970s of Kwang Jeon from the University of Tennessee demonstrated a transition from parasite to obligate mutualist in bacteria infecting single-celled amoebas over multiple generations. Although the selective pressure for the transition from parasite to mutualist is not clear, the amoeba hosts not only began to tolerate their bacterial parasites, but also eventually required their presence for survival.

The opposite path, from mutualist to parasite, has been correlated with changing circumstances in a different system. The symbiotic alga Symbiodinium microadriaticum, whose presence is required for the survival of some jellyfish, has been shown to transition from mutualist to parasite depending on the mode of symbiont acquisition. A Symbiodinium alga that is passed down from one generation to the next ends up being beneficial, but when algal strains are transmitted between host jellyfish of the same species, they can potentially be parasitic. Similarly, biologist John Klironomos at the University of Guelph showed that different mycorrhizal plant-fungus combinations could yield increases or decreases of aboveground plant biomass, depending on which pair of partners was in play. The relative benefits of the endosymbiotic partner change through manipulations of their co-culture conditions, through hostsymbiont specificity, or over subsequent generations. Therefore, maintaining artificial endosymbioses over time may be problematic, because both strict resource codependence on metabolic byproducts and inheritance of the endosymbiont from parent to offspring are likely required for persistent beneficial associations.

So how can a mutual exchange relationship be maintained? The consideration of new innovations should focus on the features of consistently beneficial endosymbionts (such as mitochondria and chloroplasts). Intracellular microbes that consistently benefit their hosts tend to be “prisoners” of the host cell microenvironment. This relationship may be maintained through ensuring transmission from one generation to the next, a tight metabolic integration, or potentially the transfer of DNA from the endosymbiont to the host with associated genome reductions in the former. Replicating these features may become a useful tool in ensuring stable interactions in engineered systems.

However, replicating these features in any system would not be trivial and would likely require extensive genetic manipulations. These could be done by knocking out essential genes from the microbial symbiont (such as an endophyte or a nitrogen-fixing bacterium) and inserting those genes into the host genome (along with components that would target them to the microbe). There are several fascinating examples of convergent gene transfer in sap-feeding insects. Similar engineered transfers could tether the microbe’s metabolic needs to the host genome’s survival and success.

Initial Successes of Organelle Transfer

Many artificial endosymbioses require the transfer of a mutualistic microbe from one species to another. Yet biologists still do not fully understand how and when these transfers can be made without causing unintended dysfunctions. Nevertheless, even the transfer of organelles with endosymbiotic origins, such as mitochondria, has been surprisingly promising. Although they evolved from endosymbionts that became integrated into cells of some eukaryotes, mitochondria and chloroplasts are considered organelles rather than endosymbionts because of their extensive genomic and physiological integration with the host cell. Both of these endosymbiont-derived organelles retain their own highly reduced circular bacterial genomes. Their DNA encodes for several proteins involved in organelle replication, metabolic processes, and the production of organelle-specific ribosomes. However, most of the organelle’s original DNA has been transferred to the host’s genome in the cell nucleus.

The persisting organelle genome has a tight integration with its host cell, making organelles unlikely candidates for simple swapping experiments. But that is just what medical researchers and even clinicians are doing. This sort of exchange, combined with in vitro fertilization, could prevent mothers with mitochondrial genetic disorders from passing them on to their children.

Inborn genetic errors in mitochondria can be detected in a mother before she has children and potentially corrected through a nuclear transplantation technique that is similar to cloning but occurs before fertilization. A donor provides their mitochondria through eggs that have had their nucleus removed. The mother then provides a “pro-nucleus” from her own unfertilized egg, which is microinjected into the donor’s egg prior to in vitro fertilization. The resulting embryo, and eventual individual, has a genetic composition from three parents: the nuclear DNA from the sperm and egg nucleus donor and the mitochondrial DNA of their egg donor. The first baby bom through this experimental therapy was recently announced. Still, potential effects of unmatched mitochondrial and maternal genomes remain uncertain and raise concerns about imprecise integration between the mitochondria and nucleus.

Surprisingly, the initial acquisition of the bacteria that gave rise to mitochondria, as well as that of the cyanobacteria that gave rise to chloroplasts, each occurred only once in the history of eukaryotic life (with the arguable second origin of the photosynthetic organelles known as plastids in the amoeba Paulinella chromatophora). Despite many subsequent associations of eukaryotic, plastid-bearing endosymbionts, and occasional bacterial endosymbionts, true organelle acquisition has been exceedingly rare. Modifying organelles or establishing new organelles in novel hosts may similarly be exceedingly difficult because of the extent of genomic integration required for these associations. Initial organelle acquisitions did not occur in modem eukaryotes; rather, it took place between 1.0 and 1.8 billion years ago (depending on the analysis) with the extraordinarily different organisms that were our distant ancestors. However, to date, transplants of these organelles between close relatives appear to be tolerated with surprising efficiency.