Correspondence: Leo van Overbeek, Wageningen University and Research Centre, Plant Sciences Group, Droevendaalsesteeg 1, 6708PB Wageningen, The Netherlands. Tel.: +31 317 480 606; fax: +31 317 418 094; e-mail: email@example.com
The rhizosphere environment selects a particular microbial community that arises from the one present in bulk soil due to the release of particular compounds in exudates and different opportunities for microbial colonization. During plant–microorganism coevolution, microbial functions supporting plant health and productivity have developed, of which most are described in cultured plant-associated bacteria. This review discusses the state of the art concerning the ecology of the hitherto-uncultured bacteria of the rhizosphere environment, focusing on Acidobacteria, Verrucomicrobia and Planctomycetes. Furthermore, a strategy is proposed to recover bacterial isolates from these taxa from the rhizosphere environment.
The rhizosphere, defined here as the volume of soil over which plant roots exert an influence (Hiltner, 1904), differs from bulk soil due to the biophysicochemical processes that occur as a consequence of root growth, water and nutrient uptake, respiration and rhizodeposition (Hinsinger, 1998). The rhizosphere extends from the surface of the plant root to a position in soil that depends on the diffusion rate of exudates and the roots' biochemistry and development (Hinsinger et al., 2005). This extension also varies with the plant type and microbial community composition (Huisman, 1982; Watt et al., 2006b). Besides, as the rhizosphere is known to contain strong gradients of compounds from the root surface into the soil, any rhizosphere sample will inevitably average out the effects of such gradients. In fact, the further away from the root the soil is sampled, the lower the rhizosphere effect on the sample will be. Thus, plant roots create selective pressures that influence the local microbial communities, resulting in effects on their abundance and composition. These selective pressures not only depend on biological processes but also on abiotic ones, such as temperature and water content. For instance, seasonal and daily temperature changes have been found to affect microbial activities (Turpault et al., 2007; Gaumont-Guay et al., 2008) and community composition (Pandey et al., 2001). Water content is also important as it directly influences the microbiota in soils due to the high correlation of microbial activity and soil moisture (Krivtsov et al., 2007). Different studies demonstrated that soil structure, as well as different granulometric fractions determine the presence of different bacterial populations (Sessitsch et al., 2001; Kotani-Tanoi et al., 2007); these factors might influence the microbial communities present in the rhizosphere depending on the soil where the root develops. Also, pH and CO2/O2 tensions will influence the rhizosphere microbial community functional diversity and activity (Jossi et al., 2006; Nelson & Mele, 2006).
The organisms thriving in the rhizosphere encompass a range of different taxa, including prokaryotic and eukaryotic microorganisms. Most abundant among these groups are the bacteria and fungi. Usually, the bacterial fraction ranges from 109 to 1010 cells g−1 of soil and c. 106 cells mm−3 of rhizosphere biofilm in most cases (Watt et al., 2006a). The description of the bacterial taxa present in the rhizosphere will be discussed later.
Key ecological functions can be ascribed to the different bacterial populations that thrive in the rhizosphere environment. For instance, bacteria with nitrogen-fixing and phosphorus-dissolving activities provide nitrogenous compounds and available phosphorus to plant roots, thus increasing the growth of plants, for instance in organic agriculture (Canbolat et al., 2006; Hameeda et al., 2008). Also, the production of particular phytohormones by rhizosphere bacteria can enhance plant growth (Lee & Song, 2007). The activity of the enzyme 1-aminocyclopropane-1-carboxylate deaminase, in particular plant-associated bacteria such as Burkholderia phytofirmans, has important implications for rhizosphere functioning as well (Hardoim et al., 2008). In addition, the rhizosphere microbiota is important, as a wide range of bacteria isolated from this environment have already shown to act as biocontrol agents of plant pathogens, for instance, those producing antimicrobial compounds or eliciting plant defence reactions (Sfalanga et al., 1999; Földes et al., 2000; Jiménez-Esquilín & Roane, 2005; Flores-Vargas & O'Hara, 2006; Romanenko et al., 2007).
The ‘great plate count anomaly’ states that 95–99% of the microbial community present in the environment is not readily accessible by traditional culture techniques (Nichols, 2007). This fraction of the microbial community has been termed as uncultured microbiota. Given the fact that a greater fraction of the total soil microbiota may be culturable than thought before, here we will address this fraction as the hitherto-unculturable bacteria. Most often, the biogeochemical transformations observed in soil have been ascribed to well-characterized culturable bacteria, simply because their ecological relevance has driven a plethora of studies based on culturable bacteria. The large gap between what we know about the culturable microbial diversity vs. the hitherto-unculturable members of these communities is illustrated in Fig. 1. The fraction of culturable members with regard to the total representatives differs for each prokaryotic group. This gap also demonstrates our lack of knowledge of the putative functions hidden in the hitherto-uncultured microbial fractions of particular groups that are involved in rhizospheric microbial–microbial and microbial–plant interactions and pinpoints the need to invest in this basic area of research. One key question is whether members of the uncultured majority are actually involved in key rhizosphere processes. Recently, using stable isotope probing-RNA analysis, a broad new range of inhabitants of the rhizosphere was found that were active under the experimental conditions, but did not have their specific role in the system identified (Vandenkoornhuyse et al., 2007). Moreover, the current ‘omics’ techniques (defined here as integrative studies of biological systems, including genomics, transcriptomics, proteomics and metabolomics) offer a good perspective in describing function or potential function in environmental studies. For instance, proteomics-based studies have already revealed the abundance of particular proteins in the rhizosphere, such as glycoside hydrolases, trypsin/protease inhibitors, plastocyanin-like domains, copper–zinc superoxide dismutases and plant basic secretory proteins (Kiely et al., 2006). Also, metagenomic studies of the rhizospheres of plants growing in acid mine drainage (AMD) revealed several novel microbial genes that are possibly involved in heavy metal resistance to their bacterial hosts (Mirete et al., 2007). Using a metabolomics approach, it was shown that the microbial populations in the rhizosphere were able to metabolize uncommon nutrients exuded by plants (Narasimhan et al., 2003).
Even though the omics techniques offer great advances in our capabilities to unravel the identity of genes present in the rhizosphere microbiota, there still are technological hurdles that hamper a thorough analysis. For instance, full genomes can only be assembled from dominant species in natural habitats, and only ‘mosaic’, instead of single-species genomes, may be assembled due to the presumed occurrence of closely resembling species in the same habitat (Tyson et al., 2004; Venter et al., 2004). In rhizosphere research, time-course data are of utmost importance due to the drastic changes occurring in associated microbial communities during plant growth. Thus, in cases where shotgun sequencing is used, the sequence information will relate only to a momentary snap shot of the dominant sequences present in the DNA extract, and information on community development in the rhizosphere is often lacking. Also, given the heterogeneity occurring in all natural environments, the interpretation of metagenomics data will be complicated due to the expected variations within residing communities (Tyson et al., 2004; Tringe et al., 2005). Moreover, it is questionable what knowledge of the genome of particular species populations actually means, given the expected genomic heterogeneity within a species, where genomic rearrangements, deletions or insertions may abound (Tringe et al., 2005). Metagenomics will thus provide, at best, information at the level of the prevalence of genes and will depend on the availability of information about the functions of such genes in cultured species (DeLong, 2005; Schirmer et al., 2005; DeLong et al., 2006). The increase in the number of genes with as-yet-unknown functions is keeping pace with the amount of metagenomic information that is becoming available over time. As long as no attempts are made to unravel the function of the many ‘unidentified’ or ‘hypothetical’ proteins commonly found in metagenomic databases, progress in the elucidation of the roles of uncultured microorganisms in ecosystem functioning will be hampered.
In this review, the following questions are posed:
1What are the key interactions among soil, plant and the local microbial communities?
2Which are the dominant hitherto-unculturable bacteria in the rhizosphere environment?
3Can rhizosphere-relevant functions of these organisms in their natural habitat be revealed by the actual knowledge in the literature?
4How to culture the hitherto-unculturable bacteria from rhizosphere environment?
5How to reveal the basic ecology of the hitherto-unculturable populations in the rhizosphere?
Soils, plant roots and microbial communities – an intricate triplet
Plant–soil–microbial community systems have evolved ever since plants started to colonize the terrestrial environment, around 700 million years ago. The first interactions between soil microbial communities and plant root systems may not have been beneficial to plants, as they may have mainly consisted of attacks to the roots by microorganisms, especially bacteria (Phillips et al., 2003). However, plants colonizing terrestrial environments, being ‘bathed’ in microorganisms, have also initiated evolutionarily useful interactions with microorganisms, for instance, to optimize growth and reproduction. Such – initially primitive – interactions are thought to have resulted in the evolution of all currently known mutually beneficial relationships (Phillips et al., 2003). Classical examples of such interactions are the intricate associations of both rhizobia and mycorrhizal fungi with plant roots. Fossilized remains of primordial root systems also provide evidence that epiphytic and endophytic microbial populations already existed on/in the primitive land plants, although the nature of these associations has not been elucidated (Phillips et al., 2003).
Here we postulate that, in particular highly selected cases, the coevolution of plants and microorganisms in soil has also resulted in firm and dedicated relationships between particular players in the rhizosphere. For one, the belowground competition for soil nutrients between plants certainly has exerted selective pressure that has driven the emergence of plants with wide variations in root structure and function, adapted to perform well under the different conditions in soil. As a consequence, the plant-associated microbial communities may have evolved to the actual scenery of positive, negative and neutral microorganism–microorganism and microorganism–plant interactions. It is known that current-day plant roots and their associated microorganisms are highly affected by soil conditions, which determine the distribution, density and depth of roots in various soils (Passioura, 1991; Jackson et al., 1996; Stewart et al., 1999; Jobbágy & Jackson, 2000).
Thus, firstly, rhizosphere microbial community structure and diversity, in different soil textural types, are strongly determined by the types of plant roots (Garbeva et al., 2004b; Kotani-Tanoi et al., 2007). It is known that during root growth, local soil characteristics, for instance, gas composition and water flow, are increasingly affected by the roots (Colmer, 2003; Lipiec et al., 2007; Benedict & Frelich, 2008; Whalley et al., 2008). In addition, carbonaceous compounds are often abundantly released from roots into the rhizosphere. These factors, in turn, will influence the root-associated microbial communities. Moreover, effects in the opposite way also occur, for example the modulation of root architecture by microorganisms that are locally present, for instance, via the production of phytohormones (Belimov et al., 2009). Some of the relevant interactions between plants and microorganisms, as well as among different microorganisms in the rhizosphere, are summarized in Fig. 2.
Many transformations in nutrient cycles that commonly occur in the rhizosphere are still unresolved, whereas others are attributed to microbial activities. Examples of microbial activities in nutrient cycles are, for instance, nitrogen fixation and demineralization (Canbolat et al., 2006) and solubilization of phosphorus (Canbolat et al., 2006) and carbohydrates (Kohler et al., 2006). Plants and other soil microorganisms profit from these transformations occurring in the rhizosphere. Examples of other interactions commonly occurring in the rhizosphere are commensalism, for example by creation of new niches for microorganisms via the secretion of exopolysaccharides (Kaci et al., 2005; Haggag, 2007) or via the production of phytohormones (Lee & Song, 2007), antagonism via production of secondary metabolites with antibiotic activity (Garbeva et al., 2004b; Berg et al., 2005; Costa et al., 2007) and syntrophism, for example via degradation of toxic compounds like recalcitrant hydrocarbons and herbicides (Biryukova et al., 2007; Vaishampayan et al., 2007). Often, a multitude of microbial species is responsible for particular conversions in the rhizosphere and hence a particular degree of functional redundancy may exist for important processes in the rhizosphere. For instance, several different microbial groups are known to be involved in ammonia oxidation in the rhizosphere (Nicolaisen et al., 2004; Herrmann et al., 2008).
However, all of these studies have so far ignored the role of the major part of the hitherto-unculturable soil microbiota. Exploring the hitherto-unculturable bacteria by improvement of culturability presumably will reveal a range of novel functions in the rhizosphere, including functions involved in enhancing rhizosphere competence. An examination of the bacterial groups that are most commonly found in rhizosphere environments and our understanding about their presumed roles in these environments will be discussed further.
Dominant hitherto-uncultured bacteria in the rhizosphere
Soil bacteria with comparable metabolic capacities will respond in a similar fashion to the emergence of different plants, provided that (nutritional) conditions are similar. However, it is still unclear as to what extent commonality exists in the microbial communities that associate with plant roots across plant genera, species or cultivars. This is mostly due to the fact that little attempts were made to synthesize data available in the literature into a systematic evaluation on the relationships between microbial communities, plants, soil types and geographical locations. Given this limitation, we performed such a limited study on data in the literature to indicate the most dominant bacterial groups in the rhizosphere using the following procedure:
1Collation of an inventory of culture-independent studies that addresses the rhizosphere bacterial diversity across a broad array of plant types.
2Identification of the most dominant phyla in the rhizosphere as found in each study.
3Singling out the most dominant groups as the ‘most commonly found’ taxa in the rhizosphere by frequency of detection across all studies.
Within these phyla, the origin of the 16S rRNA gene sequences, i.e. whether these were obtained from isolates or uncultured organisms, was taken into account; it was thus found that the contribution of isolates to the total number of sequences (redundant or not) per bacterial group was variable. For Proteobacteria, the taxonomical information available from isolates was about the same as that from uncultured organisms. However, for Actinobacteria, the available taxonomical information was even higher for isolates, reflecting better adaptation of this group for growth in a pure culture. For the other five groups, by far the highest amount of phylogenetic information was derived from uncultured organisms, i.e. most organisms occurring in these groups are regarded as as-yet-uncultured or plainly uncultured bacteria. Among these groups, the only knowledge with respect to rhizospheric Acidobacteria, Verrucomicrobia and Planctomycetes was obtained on the basis of culture-independent studies. Hence, our understanding of these taxa in the rhizosphere is still limited, as it concerns mainly their 16S rRNA gene fluctuations in the rhizosphere (Sanguin et al., 2006; Zul et al., 2007).
From the three groups, the Acidobacteria and Verrucomicrobia have a low number of species recognized in the literature. To date, the Acidobacteria contain three early-described species, i.e. Acidobacterium capsulatum, Geothrix fermentans and Holophaga foetida (Garrity et al., 2005), next to four recently recognized cultured species (Edaphobacter aggregans, Edaphobacter modestus, Chloracidobacterium thermophilum and Terriglobus roseus (Bryant et al., 2007; Eichorst et al., 2007; Koch et al., 2008). The Verrucomicrobia show a similar picture, with only 10 defined species, i.e. Verrucomicrobium spinosum, four Prosthecobacter spp., Opitutus terrae, Rubritalea marina and three Xiphinematobacter sp. (Ward-Rainey et al., 1995; Garrity et al., 2005). Although some representatives of the Acidobacteria and Verrucomicrobia have now been cultured from soil (Janssen et al., 2002; Sait et al., 2002; Stevenson et al., 2004; Davis et al., 2005), no isolates have as yet been obtained from the rhizosphere. Isolation directly from the rhizosphere is clearly needed to provide supportive data that these isolates are really rhizosphere-relevant organisms. Their availability would allow an assessment of their influences on plant growth, as well as of their interaction with plants and other soil microorganisms. Also, their involvement in important biogeochemical conversions, degradation of particular compounds and mobilization of nutrients would be facilitated. The members of the Acidobacteria and Verrucomicrobia that have previously been isolated from soil were slow growing, aerobic and heterotrophic bacteria (Janssen et al., 2002; Sait et al., 2002; Stevenson et al., 2004; Davis et al., 2005). Such characteristics are broadly distributed among rhizosphere isolates, which indicates that such isolates may thrive in the rhizosphere environment.
Ecology of the hitherto-unculturable bacteria in the rhizosphere
If particular niches exist for the hitherto-unculturable bacteria in the rhizosphere, it is logical to assume that selective forces present within such niches act on these rhizosphere populations. Such a selection might be nutritional, dependent on the availability of oxygen or other electron acceptors, or otherwise. To our knowledge, no conclusive data have extensively shown the factors that influence the fate and behaviour of these populations in the rhizosphere. Therefore, the following points should be addressed to elucidate the roles of these organisms in the rhizosphere: (1) their general occurrence and numerical dominance, (2) the ecological conditions driving these populations, (3) in situ (local) metabolic activities and (4) their functional roles in the soil–plant ecosystem.
When comparing bulk and rhizosphere soil, the general occurrence and numerical abundances of the Acidobacteria, Verrucomicrobia and Planctomycetes are often variable, as demonstrated in culture-independent studies based on 16S rRNA gene detection (Dunbar et al., 1999; Kuske et al., 2002; Gremion et al., 2003; Filion et al., 2004; Stafford et al., 2005; Sanguin et al., 2006; De Cárcer et al., 2007; Singh et al., 2007; Zul et al., 2007; Kielak et al., 2008). Specifically, a higher prevalence of 16S rRNA gene clones affiliated with particular groups of Acidobacteria, especially those of groups 1, 2 and 3, was found in rhizospheres of lodgepole pine (Chow et al., 2002), Proteacea sp. (Stafford et al., 2005) and grass (Chow et al., 2002; Stafford et al., 2005; Singh et al., 2007) as compared with corresponding bulk soils. Also, a higher number of 16S rRNA gene clones of Verrucomicrobia was observed in rhizospheres of lodgepole pine, particularly subdivisions 2, 3 and 4 (Chow et al., 2002), Thlaspi caerulescens, subdivision 2 (Gremion et al., 2003) and Proteacea sp., subdivision 3 (Stafford et al., 2005) in comparison with the corresponding bulk soils. Besides, Planctomycetes representatives were found in higher numbers in the rhizospheres of T. caerulescens than in the bulk soil, with 16S rRNA gene clones affiliating only with the Nostocoida limilula III cluster (Gremion et al., 2003) and Proteacea sp. (Stafford et al., 2005) (Table 1). In contrast, a study performed with different plants demonstrated higher numbers of Acidobacteria in a 16S rRNA gene clone library made from bulk soil DNA than in the ones made from rhizosphere DNA, although no information was given about the dominant groups in both soil compartments (Kielak et al., 2008). Also, the rhizosphere of maize plants showed lower numbers of Acidobacteria, Verrucomicrobia and Planctomycetes than the corresponding bulk soil, and again no information was given about the dominant subdivisions (Sanguin et al., 2006). In T. caerulescens (Gremion et al., 2003), different Acidobacteria groups were found in bulk and rhizosphere soil, although the number of detected 16S rRNA genes was the same in both soil compartments. In this study, more clones were found to be affiliated with Acidobacteria subdivision 6 in the rhizosphere, whereas in the bulk soil more clones were found to be affiliated with Acidobacteria subdivision 1. Furthermore, the diversities of the Acidobacteria and Verrucomicrobia in the rhizosphere of Lolium perenne L. were large. Such diversities even approximated those of the total Proteobacteria (Singh et al., 2007).
Table 1. Abundance of Acidobacteria, Verrucomicrobia and Planctomycetes and presumed factors influencing community sizes in the rhizosphere
Major observations made
Higher abundance of the hitherto- uncultured groups found at
Presumed factors influencing the Acidobacteria, Verrucomicrobia and Planctomycetes communities
Applied approaches for studying the bacterial community composition in described studies:
A, 16S rRNA gene clone library; B, terminal restriction length polymorphism; C, amplified ribosomal DNA restriction analysis; D, 16S rRNA gene-based taxonomic microarray; E, PCR–thermal gradient gel electrophoresis; F, PCR–denaturating gradient gel electrophoresis; G, quantitative PCR; PCB, polychlorinated biphenyls; ND, not determined.
Lodgepole pine (Pinus contorta)
Acidobacteria and Verrucomicrobia
• No significant effects of soil disturbance or geographical location • Plant roots exerted a strong selective pressure on Acidobacteria, Verrucomicrobia and Planctomycetes representatives
• Representatives of the division of Acidobacteria were higher in number and diversity in the rhizosphere of healthy plants than of plants with black spruce •Verrucomicrobia representatives were only found in the rhizosphere of diseased black spruce plants
• Rhizosphere effect increased numbers of Acidobacteria, Verrucomicrobia and Planctomycetes member in rhizosphere soil when compared with the respective bulk soil • The number of Acidobacteria, Verrucomicrobia and Plancomycetes also differed among root tip, root hairs and mature root
The numerical abundance of members of the Acidobacteria, Verrucomicrobia and Planctomycetes is not only dependent on differences between bulk and rhizosphere soil but also changes are incited upon root development. In a microcosm experiment performed with wild oat roots, members of the Acidobacteria, Verrucomicrobia and Planctomycetes occurred at a higher number, as demonstrated by the higher number of 16S rRNA gene copies in the rhizosphere of adult plants than in the bulk soil (DeAngelis et al., 2009). Besides, there was a greater abundance of Acidobacteria 16S rRNA genes at the surface of mature roots than at root hairs and tips; the Planctomycetes showed higher 16S rRNA gene abundance at the surface of mature roots and root tips, than at the root hairs, and the Verrucomicrobia, although having a lower abundance at the root tips, had greater abundance at the surface of mature roots and root hairs than in the corresponding bulk soil (DeAngelis et al., 2009).
Clearly, the high diversity of acidobacterial and verrucomicrobial species indicates that different bacterial populations are favoured by different selective forces exerted in the rhizosphere. At the microhabitat level, such forces can be different across space and time, thus allowing the emergence of diverse populations across each phylum. An important question here is which selective forces influence the acidobacterial, verrucomicrobial and Planctomycetales communities in the rhizosphere. On the basis of incidental data, this question has been difficult to answer. Thus, no consensus was found about which ecological factors would select for these organisms in the rhizosphere. Recent studies demonstrated that different factors select distinct members of these groups (Table 1). For instance, plant species (Chow et al., 2002; Gremion et al., 2003; Sanguin et al., 2006), soil type (Singh et al., 2007), field history (Kielak et al., 2008) and different abiotic factors (Stafford et al., 2005; Hao et al., 2008) have all been implicated as selective factors that differentially affect these groups (Table 1).
Moreover, another key question is whether we can actually show to what extent hitherto-uncultured organisms, for example members of the Acidobacteria, are part of the metabolically active community (derived from the higher abundances in RNA vs. DNA extracts) in the rhizosphere and how this is temporally organized. In the metal-hyperaccumulating plant T. caerulescens, a major discrepancy in abundance was found between the metabolically active and the total Acidobacteria (Gremion et al., 2003). In this study, it was indicated that only a subset of the Acidobacteria might be active in the rhizosphere, whereas the majority may be inactive or dormant. On the other hand, in a study using a culture-independent approach to search for 16S rRNA gene sequences in DNA and RNA extracts from the chestnut tree (Castanea crenata) rhizosphere (Lee et al., 2008), Acidobacteria were found to be not only numerically dominant but also metabolically active, implying that these Acidobacteria may play a key role in this habitat. Additionally, in a metagenomic study conducted on the rhizosphere of Erica andevalensis adapted to AMD, it was demonstrated that in this acid environment, Acidobacteria of subdivision 1 were metabolically active (Mirete et al., 2007). The authors also found that four of 13 clones with nickel resistance genes most likely were from representatives of Acidobacteria subdivision 1 (Mirete et al., 2007). However, it is unclear whether these genes were actually expressed in the environment and whether they play key roles in the heavy metal resistance of E. andevalensis.
Methods to improve bacterial and archaeal culturability
Recently, several successes in the culturing of the hitherto-uncultured bacteria were reported (Table 2). Key factors that presumably influence bacterial colony formation on solid media were found to be the type of growth medium, the incubation conditions (temperature, CO2 concentration and time of incubation) and several factors that are known to reduce oxidative stress (e.g. the presence of catalase or pyruvate) (Stevenson et al., 2004; Van Overbeek et al., 2004; Sangwan et al., 2005; Eichorst et al., 2007). It is a well-known fact that medium composition will determine which fraction of the microbial community is recoverable from the environment. Often, the medium itself limits bacterial growth due to the presence of excessive amounts of nutrients that do not reflect environmental conditions (Janssen et al., 2002; Sait et al., 2002). This phenomenon, known as ‘substrate-accelerated death’, has first been observed when organisms growing under oligotrophic conditions are transferred to high substrate concentrations (Postgate & Hunter, 1964; Straskrabová, 1983). It is a key issue, given the roughly low substrate concentrations in the rhizosphere compartment. Therefore, to overcome this limitation, novel media are designed to match conditions in the environment in which some of the key facets (e.g. amount of nutrients, nutrient composition, the presence of trace elements and pH) are adjusted for optimized bacterial growth. For instance, the use of soil-extract agar medium (Pochon & Tardieu, 1962; Hamaki et al., 2005) or of dilute standard media, both aiming to reduce nutrient levels in order to avoid substrate-accelerated death, has allowed the isolation of bacterial groups that had been regarded as uncultured only a short while ago (Joseph et al., 2003; Davis et al., 2005). Also, the presence of fast-growing bacteria that inhibit colony formation of slower growing ones, especially on nutrient-rich media, affects the recovery of hard-to-culture organisms. Media with low nutrient levels, in combination with long incubation periods at relatively low temperatures, will allow colony formation by bacteria that have low growth rates. The approach has been successful in revealing a range of novel organisms in various recent studies (Janssen et al., 2002; Rappe et al., 2002; Sait et al., 2002; Davis et al., 2005; Miteva & Brenchley, 2005; Sangwan et al., 2005; Eichorst et al., 2007). Hence, this allowed the cultivation from soil of a range of novel organisms such as bacteria belonging to the following groups: Chloroflexi, Planctomycetes, Acidobacteria, Verrucomicrobia and some Archaea representatives (Table 2). However, such approaches have seldom been used with rhizosphere samples, and only one reference was found reporting cultivation of hard-to-culture microorganisms from rhizosphere samples (Simon et al., 2005).
Table 2. Factors hampering bacterial growth upon isolation from natural environments and proposed approaches to improve culturability
Factors that limit growth of the hitherto-uncultured bacteria
Proposed approaches to improve culturability
Phylogeny of novel isolates that were obtained
Environments from where novel isolates were obtained
Inability to grow at high nutrient concentrations and overgrowth by faster growing species
• Reduce nutrient availability in growth medium • Apply a longer incubation time
Besides, the inadvertent presence of growth-inhibitory compounds in agar media may prevent the formation of colonies by particular bacteria. The use of gellan gum as an alternative solidifying agent was shown to allow the cultivation of many hitherto-unculturable bacteria (Tamaki et al., 2005), as the agar might actually be inhibitory to some groups of microorganisms, for example, those from freshwater. Also, incubation of samples (in this case from marine sediments) in diffusion chambers, which allows dilution of inhibitory compounds produced by bacteria during growth, permitted the growth of novel bacterial groups (Bollmann et al., 2007) (Table 2). Another reason why many bacteria cannot be cultured is the obligate growth of bacteria in consortia (Sakai et al., 2007). Thus, the growth of individual species in these consortia relies on the growth of others, for example when bacteria feed on compounds produced by others (Simon et al., 2005) or require growth factors produced by other bacteria (Bollmann et al., 2007). Therefore, it will be relevant to screen for unknown species in clone libraries made from bacterial mixtures, for example, by plating undiluted rhizosphere samples.
Towards the as-yet-uncultured microbial populations in the rhizosphere and their basic ecology
The recovery of the broadest spectrum of species diversity from the rhizosphere will offer a new insight into hitherto unculturable populations and their functions, as demonstrated with the recovery of novel nickel resistance genes from heather (E. andevalensis) rhizosphere metagenome (Mirete et al., 2007). A major focus should be placed on those bacterial phyla that are presumably abundant in the rhizosphere and that have so far only been described via culture-independent approaches, i.e. the Acidobacteria, Verrucomicrobia and Planctomycetes. Several studies have indicated strategies to improve culturability in natural ecosystems (Table 2), but only one has so far been addressed in the rhizosphere, in this case by improving the culturability of Archaea species (Simon et al., 2005).
Improving the cultivation of Acidobacteria and Verrucomicrobia from the rhizosphere
The rhizosphere clearly is a selective habitat for different groups of microorganisms in soil and this selectivity comes about as a result of a plethora of beneficial and stress conditions imposed on its inhabitants. Analysis of the factors that shape rhizosphere conditions might help to design methods for the recovery of a broad range of novel isolates. The following factors should be considered as relevant for the design of such methods: (1) the provision of the (solid) substrate for bacterial growth; (2) the nutritional status of the rhizosphere, which is often oligotrophic, although pulse-wise receiving nutrients; (3) the high partial CO2 and low partial O2 pressures (due to bacterial and root respiration); (4) the temporary and fluctuating lowering of the pH and (5) the presence of signalling and other (eventually toxic) molecules of root and bacterial origin. Given the fact that no single medium or technique is likely to solve the culturability issue all at once, here, a polyphasic approach is proposed, which should apply parallel cultivation systems in an attempt to yield as many hitherto-uncultured bacteria from the rhizosphere as possible. Key to such a polyphasic approach will be:
1Use of low-nutrient media in combination with long incubation periods to avoid substrate-induced death and overgrowth of fast growers, thus supporting the growth of slower ones.
2Inclusion of incubation at elevated carbon dioxide tension levels.
3Amendment with agents in the growth medium that protect cells from reactive oxygen species produced during their metabolism or present as a result of autoclaving.
4Amendment of media with rhizosphere extract to mimic the rhizosphere environment, with an emphasis on the presence of available nutrients and signalling molecules for growth.
5Parallel use of solid and liquid media – the rhizosphere offers both a solid substrate (root surface and soil matrix) and an aqueous phase (water film surrounding soil particles and water-filled soil pores) for bacterial growth; thus, the use of a combination of solid and liquid media will support the outgrowth of microorganisms with sessile and benthic lifestyles.
However, the successful isolation of the hitherto-uncultured bacteria will be no a priori guarantee that key bacteria with defined roles in the rhizosphere will be found. Indeed, so far, the efforts to increase the culturability of microorganisms from natural habitats have recovered a broad range of novel bacteria (see references in Table 2). However, their involvement in key ecosystem processes is often not clearly understood, which is mainly due to the lack of further in-depth studies concerning their ecological roles. Hence, we propose that – following their isolation – the novel taxa are thoroughly investigated with respect to their metabolism to indicate their ecological niches and functions in the soil–plant system. In such studies, the conditions reigning in the rhizosphere should be mimicked as accurately as possible and particular attention should be paid to the occurrence of signalling between plant roots and rhizosphere microorganisms or even between populations of rhizosphere microorganisms. Functions can be assigned to genes or operons in newly recovered isolates by knockout mutations and testing derived mutants in realistic plant–soil settings. Following the fate of knockout mutants upon introduction into the rhizosphere should reveal their fitness, activities and/or interactions with other species. Here, the omics tools are a welcome addition to the toolbox that is currently available for the molecular ecologist. From a genomics-based analysis of potential function, an analysis of transcriptomics and/or proteomics under different (soil, rhizosphere) conditions can quickly indicate as to what extent particular functions are operational under rhizosphere conditions.
Rhizosphere communities are important for key plant processes such as nutrient acquisition, protection against soil-borne diseases and plant development, as the rhizosphere is the transitional zone in between bulk soil and plant roots. The key functions of rhizosphere bacterial communities, for instance nitrogen fixation (Canbolat et al., 2006), phosphate solubilization (Kohler et al., 2006) and plant health protection (Garbeva et al., 2004a), have been described, mainly for well-characterized culturable species.
In this review, we show that among the groups of bacteria highest in density in the rhizosphere environment, the Acidobacteria, Verrucomicrobia and Planctomycetes are the ones whose functions remain to be unveiled. So far, these three groups have not been cultivated from this environment. Therefore, most of the information that emerged on these taxa in the rhizosphere was from measurements on the 16S rRNA gene level. The only exception is a metagenomics study conducted in the rhizosphere of plants adapted to AMD (Mirete et al., 2007). However, in this study, not all modes of action of the novel genes found could be elucidated nor was it shown that these genes were actually active in the rhizosphere environment.
One way to circumvent the hurdles of culture-independent techniques (Van Elsas et al., 2008) is the development of approaches that improve culturability, here, with a special focus on the rhizosphere environment. Several studies already aimed at the isolation of novel species from bulk soil, with variable successes in the recovery of Acidobacteria and Verrucomicrobia. Although these two groups have shown their recalcitrance to cultivation from the rhizosphere environment, it is likely to presume that they can be isolated using approaches directed to recover the widest possible range of bacteria. We propose the use of a polyphasic approach comprising low-nutrient solid and liquid media with the inclusion of oxidative stress-protective agents and/or rhizosphere extract, and long incubation periods at elevated carbon dioxide concentrations. The exploration of hitherto-uncultured organisms – following their growth to purity – will clearly simplify an assessment of their interactions with other organisms, which cannot be achieved via culture-independent approaches.
Moreover, there is a lack of studies that merge culture-dependent with advanced culture-independent studies. Such a combination of approaches will greatly increase our analytical power in order to answer one of the ultimate questions in microbial ecology –‘Which are the ecological niches, putative roles, functions and modes of interactions of the hitherto-uncultured microbiota?’
This work was supported by the Netherlands Genomics Initiative (Ecogenomics program) and the Dutch Ministry of Agriculture, Nature and Food quality (research program on sustainable agriculture).