Expectation is the root of all heartache: Plant species identity and the rhizosphere microbiome

January 30th, 2020

Noah Fierer and Corinne Walsh

Like nearly every other organism (but not all 1), plants harbor thriving microbial populations. In particular, plant roots are surrounded by diverse microbial communities in the rhizosphere (qualitatively defined as soil found in close proximity to roots). We know that the presence of roots changes the amounts and types of microbes found in their surrounding soil. Likewise, we know that these rhizosphere microbes can have important direct and indirect effects on plant health, soil fertility, and nutrient cycling. However, rhizosphere microbial communities can be highly variable in composition – functionally, taxonomically, and phylogenetically – and remain difficult to predict. What factors are responsible for this variation in rhizosphere microbial communities? This question has been the focus of many studies and it will continue to preoccupy scientists in the coming years, especially as the manipulation of rhizosphere communities continues to be hyped as a means to maximize crop production 2.

One dominant paradigm is that plant species identity is a key factor influencing what bacteria are found in the rhizosphere. In other words, different plant species, even different plant species growing in the same soil, should harbor distinct rhizosphere bacterial communities. There are numerous reasons why we would expect this paradigm to be robust. First, we know that distinct plant species can have distinct effects on soil conditions (e.g. soil pH and carbon pools) and we would expect these plant species-effects on soil environmental conditions to have corresponding impacts on the rhizosphere microbiome. Second, we have known for at least a century 3 that plants often form symbiotic associations with particular soil microbial taxa and many microbial pathogens are host-specific. For these reasons (and more) it is reasonable to expect that the rhizosphere microbiome should also be shaped in a host-specific manner. Knowing the identity of the plants growing in a given soil should allow us to predict what types of microbial communities we would see in the soil associated with that plant. This assumption underlies the design and implementation of many studies.

A closer look at the evidence suggests that this paradigm is not actually that robust. Instead, it may be more accurate to assume at the onset of any study that plant species identity has only a minor effect on rhizosphere microbial communities (at least minor compared to the many other factors that can influence the structure of rhizosphere microbial communities). If we look at some recent studies that have addressed this question for rhizosphere bacteria (e.g. refs 4-8) – we see that the percent of variance in rhizosphere bacterial communities explained by host identity is surprisingly small – usually 5-20%. Note that these studies were all ‘common garden’ experiments where multiple plant species were grown in the same soil (holding soil type constant). The proportion of variance explained by plant identity is typically far smaller when different soil types are studied (as typical in more “natural” field studies). The variance in rhizosphere bacterial communities across members of a given plant species grown in different soils is typically larger than the variance observed between different plant species grown in the same soil (e.g. refs 7,9,10,11). Only a small fraction of the variance in rhizosphere bacterial communities can be explained by the plant identity and this amount of variance explained is typically low or nonexistent when comparing across different soil types. It is not correct to assume that distinct plant species necessarily harbor distinct rhizosphere communities.

Why is this the case? Why are the effects of plant identity on the overall structure of the rhizosphere microbiome often so subtle?  There are multiple possible explanations and these explanations are not mutually exclusive nor exhaustive:

Microbial tourists”: Many of the bacteria found in the rhizospheres of different plant species could be just ‘tourists’ that are not intimately associated with their respective plants. The soil-dwelling ‘tourists’ dominate all plant species growing in a given soil even if they are not closely associated with that plant. We can apply a strained analogy and pretend that the Grand Plaza in Venice and the Roman Coliseum are different host plants. If you visit these sites in mid-summer, you will likely find hordes of tourists and a casual observer may mistakenly assume that the residents of Venice have a lot of overlap with the residents of Rome. The large numbers of tourists who visit the two sites (or two plant species) may make it difficult to distinguish between the actual residents and casual visitors.

“If all you have is a hammer, everything looks like a nail”: Most recent cultivation-independent studies of bacteria in the rhizosphere rely on 16S rRNA gene sequencing. We know that bacterial taxa which share similar, or identical, 16S rRNA gene sequences can have distinct traits and may interact with plant hosts in different ways (see Pseudomonas and many other examples from the plant pathology literature). The ‘function’ of bacteria in the rhizosphere cannot necessarily be predicted from bacterial taxonomy or phylogeny. Thus, it is possible that if we had higher resolution taxonomic/phylogenetic information (or more information on the particular traits of individual strains) we might be able to more readily identify host-specific effects on the rhizosphere microbiome that aren’t apparent when using a relatively blunt tool like 16S rRNA gene sequencing.

“Maybe Hubbell was right”: When a plant root is growing through soil, it is exposed to large numbers of potential microbial colonizers. Potential rates of microbial dispersal into the rhizosphere are likely very high. The bacteria that end up in the rhizosphere may not actually be those that are best adapted to living in the rhizosphere of a given species – but rather a stochastic or random subset of potential colonizers that were in the right place at the right time. This would be somewhat equivalent to Hubbell’s ‘neutral theory’ 12 – the communities that assemble in the rhizosphere are a product of what taxa can immigrate into the rhizosphere, not necessarily those taxa which are best suited to the particular conditions found in the rhizosphere of a given plant species.

“Microbes don’t care about plant taxonomy”: No microbe carries around an Audubon Field Guide to plants. Rather, microbes respond to, and are affected by, specific plant traits – most prominently, the amounts and types of root exudates released into soil. We assume that the microbially-relevant plant traits, including root exudation profiles, are distinct across different species, but that may not always be the case 13. If a given species has a high degree of intra-specific variation in root exudation profiles (for example) there may be a corresponding high degree of intra-specific variation in rhizosphere communities. Even if different plant species have distinct root exudates or other root traits, maybe the microbes just don’t care. To a microbe, whether a plant releases high concentrations of amino versus carboxylic acids may not matter – the same microbes may do equally well regardless of the carbon substrates in question. Plant traits that may seem important to us may be irrelevant to microbes.

Regardless of the underlying explanation, the effects of host species on the rhizosphere microbiome are often more subtle than we might expect. This has a number of potential implications for the study of rhizosphere microbiomes:

– Effect size is generally more important than statistical significance14. Just because there is a significant effect of plant species identity on rhizosphere communities does not necessarily mean that the effect is large or even important. Instead of focusing on why a host species effect is statistically significant, it may often to be more fruitful to try to explain why the host effect is not larger.

– Besides just quantifying the importance of plant species identity on overall bacterial community composition in the rhizosphere, it is arguably more useful and effective to focus on which specific bacterial taxa have high plant species specificity (and which don’t) and why (or when) the degree of host species specificity varies across taxa.

– We rarely measure the plant traits that are probably most relevant to the rhizosphere microbiome. Some of the commonly measured root traits (e.g. root length, architecture, tissue nutrient concentrations) may be largely irrelevant to microbes. Instead, traits like root exudate chemistry, nutrient uptake/release, or plant immune system regulation are likely more important (though clearly these traits can be very difficult to measure).

– Given that there are a number of companies focused on manipulating rhizosphere microbiomes to improve agricultural productivity and/or sustainability (e.g. AgBiome, LocusAg, Indigo Agriculture) – the weak host species influence may suggest that other factors (like particular soil abiotic conditions) may be more important to consider than plant species if we want to optimize the persistence and effectiveness of microbial inocula.

Instead of automatically assuming that distinct plant species harbor distinct rhizosphere communities, we should, instead, start with the hypothesis that rhizosphere communities are not strongly host specific. Re-considering our expectations will change how we study, report, and attempt to manipulate rhizosphere communities.

1          Hammer, T. J., Sanders, J. G. & Fierer, N. Not all animals need a microbiome. FEMS Microbiol. Lett. 366 (2019).

2          Busby, P. E. et al. Research priorities for harnessing plant microbiomes in sustainable agriculture. PLoS Biol. 15, e2001793 (2017).

3          Waksman, S. Principles of Soil Microbiology. (The Williams & Wilkins Company, 1927).

4          Fitzpatrick, C. R. et al. Assembly and ecological function of the root microbiome across angiosperm plant species. PNAS 115, E1157-E1165 (2018).

5          Hannula, S. E. et al. Time after Time: Temporal Variation in the Effects of Grass and Forb Species on Soil Bacterial and Fungal Communities. mBio 10, e02635-02619 (2019).

6          Leff, J. W. et al. Predicting the structure of soil communities from plant community taxonomy, phylogeny, and traits. ISME Journal 12, 1794-1805 (2018).

7          Thiergart, T. et al. Root microbiota assembly and adaptive differentiation among European Arabidopsis populations. Nature Ecol. Evol. 4, 122-131 (2020).

8          Schlaeppi, K., Dombrowski, N., Oter, R. G., Ver Loren van Themaat, E. & Schulze-Lefert, P. Quantitative divergence of the bacterial root microbiota in Arabidopsis thaliana relatives. PNAS 111, 585-592 (2014).

9          Vieira, S. et al. Drivers of the composition of active rhizosphere bacterial communities in temperate grasslands. ISME Journal 14, 463-475 (2020).

10        Yeoh, Y. K. et al. Evolutionary conservation of a core root microbiome across plant phyla along a tropical soil chronosequence. Nature Communications 8, 215 (2017).

11        Bonito, G. et al. Plant host and soil origin influence fungal and bacterial assemblages in the roots of woody plants. Molecular Ecology 23, 3356-3370 (2014).

12        Hubbell, S. The Unified Neutral Theory of Biodiversity and Biogeography.  (Princeton University Press, 2001).

13        Herz, K. et al. Linking root exudates to functional plant traits. PLoS One 13, e0204128 (2018).

14        Halsey, L. G. The reign of the p-value is over: what alternative analyses could we employ to fill the power vacuum? Biology Letters 15, 20190174 (2019).

One response to “Expectation is the root of all heartache: Plant species identity and the rhizosphere microbiome”

  1. Hey Noah Fierer and Corinne Walsh.
    A productive article but still there are numbers of questions remained and unanswered. For example, The plants still not fully capable to understand who is who and what is what. The plant species still couldn’t differentiate between FOE and FRIENDS. Although some of the latest studies reported that plant species recruit beneficial microbes instead of pathogens. For example, Zipfel published one of the articles in NATURE plant signaling in symbiosis and immunity where he stated that the plant badly failed to identify the BM and even provoke his immunity against them while the BM TRY TO COLONIZE plants’ roots.

    Let’s discuss one thing more and I would be grateful if I get a reply and guidance regarding this issue.
    How to determine the significant differences between total and active microbes in soil?

    If suppose the soil shows magnificent suppressiveness against Phytonematodes, what would be the possible and radical microbiological techniques (For biotic and abiotic) to use to ascertain the latent of the suppressive world? Soil microbial community play a crucial role in soil and plant health is Axiomatic factor, but the mystery is How to determine the significant differences between total and active microbes dwell over there. Let me explain here. DNA as sole evidence for the existence of a microbiota and you identify all the OTUs in that particular soil sample but the identification and existence of these OTUs doesn’t mean that all these microbes are active metabolically and additionally you can also find DNA from dying cells or spores or cell-free DNA but it does not necessarily indicate microbial life and an active microbiota in the sample. Transcriptomics could potentially address this issue but there are also some redundant and some not all of them criticize the mRNA based identification technique for the activeness of microbial community.
    Let’s agree for a while with Chu et al. (2017) who used Propidium monoazide (PMA) a dye which intercalates only into double-stranded DNA, preventing it from being amplified by PCR
    to remove free DNA from dead microbes prior to 16S rRNA gene amplicon sequencing, but Papp K, et al 2018 suggest that that RNA-based method to measure metabolic activity do not work equally well for all microbiome types.