MODELING THE MICROBIOME: Using synthetic communities of microbes to colonize Arabidopsis plants grown in a sterile substrate—the botanical equivalent of germ-free mice—researchers can begin to understand how the microbiome affects plant health. © SIMON FRASER/SCIENCE SOURCE

A few years ago, as a postdoc in the lab of Paul Schulze-Lefert at the Max Planck Institute for Plant Breeding Research in Cologne, Germany, I used next-generation sequencing to study the bacterial communities that populate roots of the model plant Arabidopsis thaliana. Although scientists had known for many years that roots interact with a variety of microorganisms, the composition of these communities was still poorly understood. As our sequencing data began rolling in, I was stunned by the staggering taxonomic diversity of bacteria that a single, tiny root can host. Yet there was an order in this apparent chaos. Almost invariably, members of the phyla Actinobacteria, Bacteroidetes, and Proteobacteria were enriched, differentiating...

Subsequent studies by other labs supported our findings and posited Firmicutes as an additional dominant member of the plant microbiota. In addition to these bacterial groups, genomic surveys of plants have revealed certain fungal and eukaryotic microbes. And all of these groups of organisms are making themselves at home not just beneath the soil in and around plants’ roots, but in other tissues, such as leaves, as well.

This research immediately raised new questions: Why were certain microbes more abundant in roots and leaves? How did these microbial communities assemble? And most critically, how did they affect plant health?

Recently, in addition to genomic surveys of the microbes present in various plant tissues, researchers have begun to probe the functional consequences of these bacterial, fungal, and eukaryotic symbionts. A better understanding of the molecular dialog between plants and their microbiota could revolutionize agriculture. The world population is expected to reach 9.8 billion in 2050, more than 30 percent larger than at present. This will put enormous pressure on food production globally—pressure that won’t be relieved solely by the agrochemicals farmers currently use to increase yield and protect crops from pests and pathogens. To encourage a sustainable food source for humanity, radical changes in the crop production process are needed—changes that could come in the form of microbial manipulation.

The interface between plant roots and soil—a zone called the rhizosphere—and the root itself are sites of colonization for microbes capable of enhancing mineral uptake by the plant, of both actively synthesizing and modulating the plant’s synthesis of chemical compounds called phytohormones that modulate plant growth and development, and of protecting plants from soil-derived pests and pathogens. For these reasons, scientists are looking to manipulate the microbes populating this belowground habitat to sustainably increase crop production. And in my lab, we are looking at ancient varieties and wild relatives of crops as a source of insights into beneficial associations between plants and microbes that could be adapted for agricultural settings.

Surveying the plant microbiome

The roots of land plants thrive in soil, one of the richest and most diverse microbial reservoirs on Earth. It has been estimated that a single gram of soil contains thousands of different bacterial species, not to mention other microorganisms such as archaea, fungi, and protists. Perhaps not surprisingly, the establishment of interactions with the soil biota represented a milestone for plants’ adaptation to the terrestrial environment. Fossil evidence suggests that the first such interactions with fungal members of the microbiome occurred as early as ~400 million years ago.1

PLANTS’ MICROBIAL COMMUNITIES: Like animals, plants host communities of microbes that influence a wide variety of their biological processes. Recent surveys of the plant microbiome have begun to document which species are present—including not just bacteria, but fungi and microscopic eukaryotes as well—and how they affect the plant’s health and functioning.
See full infographic: WEB | PDF

Comparative studies indicate that soil characteristics such as nutrient and mineral availability are major determinants of the root microbiome. Just as digestive tract microbes interact with the food consumed by vertebrates, the root microbiome mediates the soil-based diet of plants. Also paralleling host/microbe interactions in the animal kingdom, individual members of the plant microbiome appear to be compartmentalized. I and other researchers working with Arabidopsis and with rice have identified at least three distinct microbiomes thriving at the root-soil interface: that in the rhizosphere; another one on the root surface, or rhizoplane; and a third one inside the root, an area known as the endosphere.2,3 In all three compartments, Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria dominate the bacterial communities in multiple plant species. The aboveground portions of plants such as leaves show similarly predictable microbial composition. (See illustration at left.)

While the categories of microbes that make up the plant microbiome are largely conserved, much variation exists in the species compositions of these communities across hosts. One key factor in determining how the microbiome is populated and maintained appears to be the plant’s release of organic compounds into the rhizosphere, a process known as rhizodeposition. The amount and composition of these organic deposits vary depending on plant species and developmental stage, but may account for up to 11 percent of net photosynthetically fixed carbon and 10 percent to 16 percent of total plant nitrogen.4 This process influences the chemical and physical composition of the rhizosphere and, in turn, provides signaling molecules and organic substrates for microbial growth.

ROOT BUGS: Plant roots and the interface between the roots and the soil—a zone called the rhizosphere—are home to diverse microbes that can affect mineral uptake by the plant.© BIOPHOTO ASSOCIATES/SCIENCE SOURCE

Another factor that likely shapes the composition of the plant microbiome is interaction between microbes. In 2016, Eric Kemen of the Max Planck Institute for Plant Breeding Research and colleagues surveyed the microbes thriving in and on wild Arabidopsis leaves at five natural sites in Germany sampled in different seasons. They then plotted correlations between the abundances of more than 90,000 pairs of microbial genera identified in their survey, revealing six “microbial hubs”—nodes with significantly more connections than other nodes within the network. These hubs were represented by the oomycete genus Albugo, the fungal genera Udeniomyces and Dioszegia, the bacterial genus Caulobacter, and two distinct members of the bacterial order Burkholderiales.5 Given the high degree of connectivity within the communities, it is likely that these microbial hubs play a disproportionate role in the microbiome, akin to that of keystone species in an ecosystem.

To validate this idea that certain species can drive the composition of the plant microbiome, Kemen’s team selected Albugo sp. and Dioszegia sp. as paradigmatic examples of microbial hubs. Albugo oomycetes are eukaryotic pathogens of Arabidopsis with an obligate biotrophic lifestyle—meaning that they cannot be cultured outside their host. Consistent with the central role of Albugo in the plant’s microbial community, Arabidopsis that had been artificially infected with Albugo laibachii and maintained in potting soil under controlled conditions displayed a bacterial microbiome composition that was less variable across plants than that of uninfected individuals. Conversely, differences between the bacterial microbiomes of three distinct Arabidopsis strains were amplified in the presence of A. laibachii infection. The fungal microbiome, however, was not significantly affected by the presence of A. laibachii and another Albugo species.

I was stunned by the staggering taxonomic diversity of bacteria that a single, tiny root can host.

Kemen’s team conducted a parallel set of experiments with Dioszegia sp., which—unlike Albugo sp.—are culturable under laboratory conditions, and six bacterial isolates from Arabidopsis leaves. The results confirmed that the presence of the fungal species can strongly inhibit the growth of Caulobacter—plants whose leaves were inoculated with Dioszegia sp. showed a 100-fold reduction in the number of colony-forming units of Caulobacter sp.—mirroring the significant negative correlation observed between these two groups of microbes in the network analysis.5

In 2017, Harvard University’s Roberto Kolter and colleagues demonstrated that such microbial interactions are not limited to Arabidopsis. The researchers developed a simplified version of the maize root microbiome, consisting of seven bacterial strains previously identified in sequencing surveys. By using a leave-one-out approach to colonizing naive maize plants, they demonstrated that removal of Enterobacter cloacae disrupts the composition of the microbial community, which became dominated by Curtobacterium pusillum, while the other five species had nearly disappeared. Interestingly, this effect was limited to plant colonization: when the seven strains of bacteria were monitored in a substrate that did not contain maize seedlings, the community’s composition was significantly different from the one retrieved from roots, and the regulatory role exerted by E. cloacae was not detected.6

These studies suggest that individual members of the microbiome can have a disproportionate role in assembling and stabilizing the community. Deciphering the interactions within and between the various taxa populating leaves and roots will be required to understand the regulation of the plant microbiome.

From composition to function

For years, researchers have observed that, despite the presence of pathogens and conditions favorable to infection, some regions produce plants that are less susceptible to disease than other areas. The soils in these areas, it turns out, support plant health via the microbiome.

Researchers are making strides in understanding the mechanisms underlying this support. In 2011, for example, a team led by Rodrigo Mendes, then at Wageningen University and Research Centre in the Netherlands, demonstrated that disease suppression was linked to the recruitment of a specific population of Pseudomonadaceae, a family of the phylum Proteobacteria. Using a PCR fingerprinting approach, the researchers discerned that this population could be grouped into ten haplotypes, which the team designated A to J. Of these, haplotypes A, B, and C represented some 90 percent of the isolated bacteria. When inoculated in soil, a representative strain of haplotype C suppressed the incidence of disease caused by the fungus Rhizoctonia solani on sugar beet roots, while, surprisingly, strains from haplotypes A or B did not.7

Similarly, in their study published last year, Kolter and colleagues found that maize plants inoculated with the seven selected bacterial strains showed significantly delayed development of Fusarium verticillioides, the causal agent of maize blight. This phenomenon was mediated by the specific strains chosen, and not by bacterial colonization per se, as seed treatment with a laboratory strain of Escherichia coli did not protect maize seedlings from pathogen development. Likewise, the seven strains together were required for the protective effect: inoculation with individual strains resulted in significantly less protection against F. verticilloides.

This method of combining sequencing data with microbial isolation is becoming a powerful tool to formulate testable hypotheses and gain novel insights into the function of the plant microbiome. Like Kolter, researchers are assembling microbial isolates into synthetic communities (SynComs) of known composition and testing their effects on host plants. This approach was once considered a daunting task, as only a very limited fraction—often less than 1 percent—of soil biota was considered culturable under laboratory conditions. But in 2015, Schulze-Lefert’s lab teamed up with Julia Vorholt’s group at ETH Zurich in Switzerland to investigate the proportion of Arabidopsis-associated bacteria that can be cultured, and found the 1 percent statistic to be a vast underestimate.

FUNGAL FINGERS: In addition to bacteria, the plant microbiome includes fungal species such as the Rhizoctonia solani shown here.© DENNIS KUNKEL MICROSCOPY/SCIENCE SOURCE

Comparing the taxonomic relationships among some 8,000 colony-forming microbes from leaves and roots of plants using cultivation-independent sequencing surveys of leaf and root microbiomes, the researchers demonstrated that more than 50 percent of the dominant members of the Arabidopsis microbiome can be cultured in vitro.8 Taking advantage of this finding, the team assembled SynComs representative of the microbiota of the Arabidopsis roots and leaves and tested the communities’ capacities to colonize these tissues on plants grown in a sterile substrate—the botanical equivalent of germ-free mice. These experiments revealed that, upon plant inoculation, root and leaf isolates form microbial communities resembling the natural microbiomes of those tissues, demonstrating that the SynCom approach accurately recapitulates the effects of a complete microbiota.8

Since then, numerous researchers have begun to develop SynComs to further explore the function of the plant microbiome. Earlier this year, for example, Jeff Dangl of the University of North Carolina at Chapel Hill and colleagues used the SynCom approach to explore the role of the root microbiome in phosphate uptake. In nature, less than 5 percent of the phosphorus content of soils is available to plants. To circumvent this limitation, farmers rely on the application of chemical fertilizers, but this approach is not sustainable in the long term. Thus, understanding how plants and their associated microbes can thrive under sufficient and limiting phosphorus supplies is a priority. There is a huge body of literature documenting the contribution of arbuscular mycorrhizal fungi to phosphorus uptake in plants, but the role of the bacterial microbiota remains mysterious.

Scientists are looking to manipulate soil microbes to sustainably increase crop production—and novel insights into the plant microbiome are now facilitating the development of such agricultural tactics.

In experiments with Arabidopsis, which does not engage in symbiotic relationships with mycorrhizal fungi, Dangl and his colleagues compared the microbiomes of wild-type plants with those of mutant lines that had impaired phosphate starvation responses (PSRs)—a set of morphological, physiological, biochemical, and transcriptional activities evolved by plants to cope with phosphorus deficiency. Using a SynCom represented by 35 taxonomically diverse bacterial isolates from Arabidopsis and related plants, the researchers demonstrated that wild-type plants and mutants, grown on agar plates, assemble distinct root communities when exposed to both low and high phosphorus concentrations. Remarkably, SynCom inoculation reduced accumulation of phosphorus when plants were grown under limited conditions but not when plants were grown in the presence of abundant phosphate, suggesting that bacteria and plants compete for the element.

By monitoring a core set of 193 marker genes, the team observed that SynCom inoculation greatly enhanced PSR-related transcription in wild-type plants. When the researchers transferred inoculated wild-type plants grown with limited phosphorus to plates with sufficient supplies, they observed a striking result: 20- to 40-fold increases in phosphorus concentration in the plant stem, as compared with mock-inoculated controls. Such a dramatic increase in phosphorus uptake was not detected in inoculated plants initially grown with sufficient phosphorus. Therefore, initial plant-bacteria competition for phosphorus might be part of an adaptive mechanism to maximize PSR in plants.9

Further investigation into the binding sites of transcription factors on Arabidopsis DNA revealed that PHR1, a master regulator of PSR, and its paralog PHL1 contribute to transcriptional regulation of plant immunity. In particular, phr1;phl1 mutant plants display enhanced activation of plant immunity genes in response to phosphate starvation and to SynCom inoculation, compared with wild-type plants. Together, these data suggest that the nutritional status of the host is a driver of microbiome composition; through master regulators of mineral starvation, plants can modulate immune responses, which could, in turn, shape microbiome composition. (See “Holding Their Ground,” The Scientist, February 2016.)

What’s next?

Characterizing the plant microbiome and its function could be applied in an agricultural setting, better equipping our crops to grow in resource-poor environments and to fight off dangerous pathogens. Indeed, the private sector has begun to invest in this approach. One strategy many companies are pursuing is a form of plant probiotic, which consists of preparations of beneficial microbes to be mixed with seeds at sowing and again once the seedlings germinate. Another approach is to use plant breeding to select for varieties that have enhanced symbiosis with the microbiota.

Many questions remain about the plant microbiome, however—not least of which is how thousands of years of cultivation have changed crops’ relationships with the soil biota. Using a cultivation-independent approach, my colleagues and I recently demonstrated that wild ancestors and modern varieties of barley (Hordeum vulgare) host distinct microbiotas.10 Likewise, Jos Raaijmakers of the Netherlands Institute of Ecology and colleagues last year identified a shift in the structure of the microbiome of modern and ancestral varieties of common bean (Phaseolus vulgaris); Bacteroidetes were more abundant in wild relatives, and their contribution to the community was progressively replaced by Actinobacteria and Alphaproteobacteria in the more domesticated plants.11

How do these differences translate to altered functionality of the microbiome? Thanks to the experience gained by Arabidopsis scientists, we are now in a position to address this question, and developing SynComs from crops will be an important step in the process.

Luckily, the field is motivated to do just that, as well as to define a road map to achieve the translational potential of the plant microbiome. In a few years, the plant microbiome manipulations may have moved from the lab to the field. 

Davide Bulgarelli is a principal investigator at the University of Dundee in the U.K. His research aims at understanding the structure, function, and host control of the microbiome thriving at the root-soil interface.


  1. P. Bonfante, A. Genre, “Plants and arbuscular mycorrhizal fungi: an evolutionary-developmental perspective,” Trends Plant Sci, 13:492-98, 2008.
  2. D. Bulgarelli et al., “Revealing structure and assembly cues for Arabidopsis root-inhabiting bacterial microbiota,” Nature, 488:91-95, 2012.
  3. J. Edwards et al., “Structure, variation, and assembly of the root-associated microbiomes of rice,” PNAS, 112:E911-E920, 2015.
  4. D.L. Jones et al., “Carbon flow in the rhizosphere: carbon trading at the soil-root interface,” Plant Soil, 321:5-33, 2009.
  5. M.T. Agler et al., “Microbial hub taxa link host and abiotic factors to plant microbiome variation,” PLOS Biol, 14:e1002352, 2016.
  6. B. Niu et al., “Simplified and representative bacterial community of maize roots,” PNAS, 114:E2450-E2459, 2017.
  7. R. Mendes et al., “Deciphering the rhizosphere microbiome for disease-suppressive bacteria,” Science, 332:1097-100, 2011.
  8. Y. Bai et al., “Functional overlap of the Arabidopsis leaf and root microbiota,” Nature, 528:364-69, 2015.
  9. G. Castrillo et al., “Root microbiota drive direct integration of phosphate stress and immunity,” Nature, 543:513-18, 2017.
  10. D. Bulgarelli et al., “Structure and function of the bacterial root microbiota in wild and domesticated barley,” Cell Host Microbe, 17:392-403, 2015.
  11. J.E. Pérez-Jaramillo et al., “Linking rhizosphere microbiome composition of wild and domesticated Phaseolus vulgaris to genotypic and root phenotypic traits,” ISME J, 11:2244-57, 2017.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?