Root interactions with soil Microbial Communities

and Processes

Christine V. Hawkes, Kristen DeAngelis, and Mary K. Firestone

I. Introduction

A common definition of soil is “the surface layer of earth, supporting plant life.” (Webster’s). In fact, most of the volume of the upper weathering layer of the earth’s crust has been influenced by plant roots at one time or another and hence by standard definition, most of soil would be or would have been at some time considered rhizosphere soil. Here we will focus on soil that is in active, current communion with living plant roots. However, the fact that a large proportion of surface soil was directly impacted by plant roots and associated microbes last year, ten years ago, or 100 years ago provides a potentially valuable context for discussion of soil microbial communities and processes generally (also see Richter et al., this volume).

Rhizosphere soil effectively forms a boundary layer between roots and the surrounding soil Because roots and soil act as both sources and sinks for a diverse range of compounds, this boundary layer of soil mediates large fluxes of solution and gas-phase nutrient (and non-nutrient) compounds (Belnap et al. 2003). From the microbial perspective then, rhizosphere soil is both a crossroads and a marketplace. The physical extent of the active rhizosphere zone is not easily defined, but at any time is expected to extend only a few millimeters from the root surface and to differ based on the process or characteristic of interest.

Plant roots grow into and through an extraordinary array of “indigenous” soil microorganisms. The phylogenetic and functional characteristics of the community that develops in concert with the plant root is thus framed by the background, bulk soil community. While we have been able to photograph relatively intact rhizosphere communities for some time now (Figure 1), understanding who these organisms are and what they are doing has been a longstanding challenge. This is an exciting time in rhizosphere microbial ecology. The development of new methods for studying the intact rhizosphere is opening up yet another black box. In this chapter, we discuss recent advances in rhizosphere microbial ecology, the impacts of rhizosphere microbial communities on nutrient cycling, and the importance of rhizosphere processes at larger scales.

II. The Composition of Rhizosphere Microbial communities

A. Microbial Populations and Communities in the Rhizosphere

Plant species can be important in determining the structure of rhizosphere bacterial and fungal communities (e.g. Stephan et al. 2000), with both positive and negative effects on different microbial groups. Within plant species, microbial communities can be affected by plant genotype (Smith et al. 1999), plant nutrient status (Yang and Crowley 2000), pathogen infection (Yang et al. 2001), and mycorrhizal infection (Johnson and Gehring, this volume). Within root systems, microbial communities can even differ among root zones (Yang and Crowley 2000) and at different distances from the root surface as rhizosphere soil grades into bulk soil (Marilley and Aragno 1999). The largest numbers of bacteria in the rhizosphere have been reported to occur in the zone of root elongation (Jaeger et al. 1999).

Studying organisms in the rhizosphere, and more generally in soil, is not a straightforward task. A complex community of bacteria may exist at the scale of a soil aggregate, a biofilm, or a section of root surface where boundaries can be difficult to delineate (Belnap et al. 2003). Physically removing microbes from soil is also non-trivial, particularly from intact rhizosphere soil. The recent development and popularity of molecular techniques to identify soil organisms has allowed us to move beyond the small subset of culturable soil organisms and begin defining populations and communities of microbes belowground. It is increasingly common to characterize complex microbial communities genotypically using the small subunit 16S ribosomal DNA gene (16S rDNA), a region that is very highly conserved, essential, subject to low homologous gene transfer, and a good reflection of overall phylogenetic relatedness. A collection of 16S genes can be analyzed partially, as with the fingerprinting methods T-RFLP and DGGE, or in detail by sequencing entire populations or communities in clone libraries. Using these methods, we have begun to understand how population and community ecology concepts apply to rhizosphere microbes.

Most population studies have focused on organisms that can be manipulated in agricultural settings either for biocontrol or for increased plant growth, including species of symbiotic nitrogen fixers (Carelli et al. 2000), plant growth promoting rhizobacteria (Bevivino et al. 1998), deleterious rhizosphere bacteria (Nehl et al. 1997), pathogens (Khan and Khan 2002), and bacteriophage (Ashelford 2003). Population-level studies are also common for rhizosphere bacteria useful for bioremediation. For example, Dalmastri et al. (2003) recently reported high genotypic and phenotypic diversity of a Burkholderia cepacia complex population in maize rhizosphere, potentially important in explaining the diverse ecological roles of these bacteria in biocontrol, bioremediation, and human illness.

Because the effects of microbes in the rhizosphere are often synergistic, understanding them at the community level is perhaps most ecologically meaningful. Microbial community characterization is often limited to a subset of the rhizosphere community, such as plant growth promoting bacteria (Dalmastri et al. 2003), pseudomonads (Misko and Germida 2002), nitrifiers (Priha et al. 1999), or mycorrhizal fungi (Johnson and Gehring, this volume). Alternatively, entire communities can be described. Microbial community characterizations have taken place most often in economically important agricultural species, primarily corn, but also alfalfa, avocado, barley, beet, canola, lettuce, pea, potato, rye, soybean, tomato, and wheat (Table 1). In a small number of cases the focus is on plants in natural communities (Priha et al. 1999, Kuske et al. 2002).

In past studies, researchers using culture-based methods have generally reported dominance of Gram-negative bacteria. Results from molecular-based characterizations are, however, more variable, with different groups of dominant microbes in the rhizospheres of individual plant species (Table 1). In a meta-analysis of published bacterial 16S rDNA community characterization from rhizospheres of 14 plant species, we discovered that bacteria from rhizosphere soil in fact span the entire tree of life (Figure 2). This analysis was based on rhizosphere soils from nine herbaceous dicots, two woody dicots and three grasses. Bacteria from 35 different taxonomic orders were reported in the rhizosphere. Based on prior results from culture-experiments, we expected to find the Proteobacteria and Actinobacteria well-represented, which was indeed the case. Proteobacteria dominated the rhizosphere in 16 of 19 studies (Table 1). Within the Proteobacteria, patterns were variable, but most often members of the γ-Proteobacteria were dominant. Gram-positive bacteria and the Cytophaga-Flavobacterium-Bacteroides (CFB) group followed the Proteobacteria in abundance. Most of the α-Proteobacteria were unclassified at the level of order, which suggests that there is potentially more sequence and functional diversity in the rhizosphere than was revealed in this analysis. A few unexpected bacteria were found including thermophiles and deinococcus; it is not clear whether these were indigenous soil bacteria or whether these sequences were miscategorized or erroneously sequenced.

Across plant groups, there was a great deal of overlap in the broad taxonomic divisions comprising rhizosphere microbial communities in this analysis (Figure 3). The herbaceous dicots exhibited the greatest microbial richness with representatives in 26 orders of bacteria, followed by woody dicots (22 orders) and grasses (20 orders); richness was unrelated to the number of sequences reported for each group. Compositional differences in the rhizosphere microbial community were also evident among the three groups. Relative to the dicot herbs, the woody plants had fewer organisms from the CFBgroup, Actinobacteria, and Firmicutes, and more Acidobacteria, unclassified β-Proteobacteria, Rhodospirillales, Geobacter, and most orders of α-Proteobacteria. The woody rhizospheres also harbored the only representatives from termite groups and several groups with no cultured representatives including TM6, OP10, and Gemmatimonadetes. Only two woody plant species were included in this analysis, Persea americana(avocado) and Pinus contorta (lodgepole pine), with the vast majority of sequences contributed by the pine. Pines are well known for their associations with ectomycorrhizal fungi, which may influence the composition of the bacterial community. Very few differences between dicot herbs and grasses could be seen at this coarse taxonomic scale, though they did exhibit slightly different distributions within the Proteobacteria. One study in corn also looked for and found Archaea in the rhizosphere (Chelius and Triplett 2001) and two studies (alfalfa and corn) reported Frankia in the rhizosphere (Chelius and Triplett 2001, Tesfaye et al. 1999).

Within herbaceous dicots, you might expect to see a difference in the rhizosphere microbial community of those plants that can and cannot fix nitrogen, given their different requirements for nitrogen from soil. A comparison between the rhizosphere of nitrogen-fixers (Medicago, Phaseolus, and Trifolium) and other dicot herbs (Beta, Brassica, Dendranthema, Fragaria and Solanum) revealed striking differences in the diversity of microbes associated with roots of these two groups (Figure 4). Rhizospheres of nitrogen-fixing plants supported a greater richness of bacteria compared to the non-fixers, with nearly double the number of monophyletic groups of bacteria from our analysis. These included the presence of - and -Proteobacteria, Nitrosomonas, Planctomycetes, Deinicoccus-Thermus, Sulfobacillus, and Chloroflexi. Because nitrogen-fixing plants rely less on nitrogen from soil microbial mineralization of organic nitrogen, root exudation may be altered and microbial communities may be selected based on characteristics other than rapid growth on labile root carbon.

While this is far from a complete picture of the diversity of the rhizosphere, it demonstrates that the rhizosphere is potentially capable of hosting an array of microbes far more diverse than what has been reported with other methods. It also suggests that, at coarse taxonomic scales, there is some degree of commonality in the bacterial components of rhizosphere communities of many plants and at the same time there is some degree of specificity in the selection of rhizosphere microbial communities. As more DNA-based characterizations of rhizosphere microbial communities become available, we can continue to increase our understanding of rhizosphere microbial communities and their controllers.

Does microbial diversity per se matter in the rhizosphere? Diversity is primarily important in terms of the specific composition of the community present and the amount of functional redundancy included. If multiple pathways for the same process are provided by the community composition, then increased microbial diversity may buffer microbial community structure and function from disturbance (Girvan et al. 2005). The rhizosphere is characterized by large environmental fluctuations (see section III A below), which may promote high diversity in the rhizosphere microbial community by maintaining high niche diversity. Thus, microbial community diversity may be important for broad functional continuity in the rhizosphere where disturbances occur in the form of daily environmental fluctuations (see section III A below) and in this way has potential for positive feedbacks to the plant (see section IV A below).

Community characterization is not always genotypic in nature, but may occur at different scales ranging from functional diversity to broader taxons to simple abundance. Functional diversity is a common measure of microbial community composition and may be more relevant to ecosystem function than taxonomic diversity. Indicators of functional diversity are those that measure the type, abundance, activity, and rate of microbial substrate use. The most common method is the sole-carbon-source utilization profile (Campbell et al. 1997). Functional diversity can also be estimated by measuring functional genes that play a role in ecosystem processes (Prosser 2002). This is commonly used for those processes in which a limited number of genes are involved, such as nitrification and denitrification.

In our 16S rDNA analysis of the rhizosphere microbial community, we can relate the phylogenetic diversity to function only through conventional interpretations (Table 2). Functional groups relevant to the rhizosphere include nitrogen cycling bacteria, anaerobic respirators, and pathogens, all of which have been reported in the sequence dataset analyzed here. Nitrospira, a genus of microbes important in nitrification, were virtually unreported in these studies, perhaps either as a result of sampling bias or other methodological constraints, or because this division of nitrifier (that has its ammonia monooxygenase gene unconstrained by internal membranes) is somehow less well-suited to the rhizosphere environment. Methanotrophs and iron oxidizers were also present. Enterics and Xanthomonads (mostly -Proteobacteria) represent the majority of known plant pathogens, and these groups were well-represented in this sampling.

The phylogenetic and functional composition of rhizosphere microbial communities is the net result of the plant interacting with the indigenous soil community. While we have some knowledge of how the rhizosphere soil environment differs from that of surrounding bulk soil (see below), how the aggregate rhizosphere environmental characteristics select/inhibit specific free-living bacteria and fungi is largely unknown. The specificity of plant-microbial interactions and the environmental characteristics of the rhizosphere are addressed in the following sections.

B. Specificity of Root-Microbial Interaction

Root-microbial interactions encompass a range of specificity from “highly-evolved” symbioses (legume-rhizobium) to less-specific associations (arbuscular mycorrhizae). The degree of specificity and coevolution of free-living rhizosphere heterotrophs is, however, quite unclear. As discussed above, host plant genotype can affect the composition of the rhizosphere microbial community and this community has the potential to affect plant growth and survival (Nehl et al. 1997). Thus the potential for coevolution exists. Furthermore, colonization of the rhizosphere environment may involve a complex array of microbial behaviors that require the development of some host specificity. A model biological control organism, Pseudomonas fluorescens, was put through the promoter-trapping technique IVET (in vivo expression technology) to determine what genes it needs in order to colonize the sugar beet rhizosphere. Twenty genes were identified as having a significant increase in transcription, of which one quarter were involved in nutrient acquisition (organic acid metabolism and xylanase), three were related to oxidative stress, one was a copper-inducible regulator and one was a component of the type-III secretion system (Rainey 1999).

Apparent symbioses are the most likely to develop host-specificity. Specificity in legume-rhizobia relationships, for example, is determined by plasmids which contain the genes responsible for both nodulation and nitrogen-fixation, and can be exchanged among strains (Hedges and Messens 1990). Some legume-rhizobia associations are clearly defined, with a single rhizobium species limited to a single plant genus or small group of genera (Hedges and Messens 1990, González-Andrés and Ortiz 1999), though individual plants may host several genetic strains of the same microbial species (Carelli et al. 2000) and some rhizobium species are promiscuous. Thrall et al. (2000) posit that rare rhizobia species may be more host-specific than widespread ones.

Non-symbiotic associations may also be specific to plant species and this has been demonstrated in a few specific cases. Bacterial antagonists of the soilborne fungal pathogen, Verticillum dahliae, varied significantly among four host plant species (Berg et al. 2002). Plant genotype in hybrid tomato plants accounted for 38% of the variability in the plant growth promoting rhizosphere bacteria, Bacillus cereus,mediating resistance to the pathogen, Pythium torulosum, (Smith et al. 1999). An examination of replicate plant species in our analysis of rhizosphere microbial communities, however, reveals some repeatable and some unique associations in both Medicago sativa (alfalfa; Figure 5A) and Zea mays (corn; Figure 5B). In the case of alfalfa and a cultivar of alfalfa, the cultivar had greater microbial diversity with 19 compared to seven of the monophyletic microbial groups in our analysis represented. Moreover, when alfalfa was grown in two agricultural soils with different management histories, the alfalfa rhizosphere was dominated in one soil by α-Proteobacteria (particularly rhizobia) and in the other by Bacteroidetes (Miethling et al. 2003). Microbial community composition in the rhizosphere of corn and two corn cultivars has a good deal more overlap, with only two microbial groups unique to one cultivar, Z. mays Bosphore, which was grown in the same soil as Z. mays KX844 but at a different time. Though this small sample is hardly conclusive, it furthers the case that microbial communities in the rhizosphere are primarily non-specific and are selected through a combination of the available bulk soil microbial pool, plant species, and environmental conditions.

III. Characteristics of rhizophere soil that impact microbial community composition

The physical, chemical, and biological environment that rhizosphere soil provides for microbial growth can differ substantially from that of nearby bulk soils. The differences in rhizosphere and bulk soil encompass virtually every environmental determinant that is critical to soil microbial activity and survival. Moreover, the biota of the rhizosphere environment differs substantially from that of bulk soils both in abundance, composition, and trophic interactions (Johnson and Gehring, Moore et al., this volume). The compositional and functional characteristics of the rhizosphere microbial community are thus determined by the integrated environmental determinants operating on the bulk soil microbial community together with the biotic interactions occurring in the rhizosphere habitat.

A. Physical and Chemical Characteristics of Rhizosphere Soil

Microbes in the rhizosphere are subject to an environment in which the supply of water, oxygen, and nutrients is strongly influenced by plant activity. An actively transpiring plant removes huge quantities of water from the soil. Depending in large part on the rate of water supply from the surrounding soil to the rhizosphere, the water potential in rhizosphere soil can be more than 1 MPa lower and much more variable than in the surrounding soil (Papendick and Campbell 1975). During the daytime, rhizosphere soil is commonly measurably drier than the surrounding bulk soil.