Perception and modification of plant flavonoid signals by rhizosphere microorganisms


  • Liz J. Shaw,

    Corresponding author
    1. Department of Environmental and Geographical Sciences, Manchester Metropolitan University, John Dalton Building, Chester Street, M1 5GD, UK.
    Search for more papers by this author
    • Present address: Department of Soil Science, School of Human and Environmental Sciences, University of Reading, Whitenights, Reading, RG6 6DW, UK.

  • Phil Morris,

    1. Institute of Grassland and Environmental Research, Plas Gogerddan, Aberystwyth, SY23 3EB, UK.
    Search for more papers by this author
  • John E. Hooker

    1. Department of Environmental and Geographical Sciences, Manchester Metropolitan University, John Dalton Building, Chester Street, M1 5GD, UK.
    Search for more papers by this author

*E-mail; Tel. (+44) 118 3786971; Fax (+44) 118 3786660.


Flavonoids are a diverse class of polyphenolic compounds that are produced as a result of plant secondary metabolism. They are known to play a multifunctional role in rhizospheric plant-microbe and plant–plant communication. Most familiar is their function as a signal in initiation of the legume-rhizobia symbiosis, but, flavonoids may also be signals in the establishment of arbuscular mycorrhizal symbiosis and are known agents in plant defence and in allelopathic interactions. Flavonoid perception by, and impact on, their microbial targets (e.g. rhizobia, plant pathogens) is relatively well characterized. However, potential impacts on ‘non-target’ rhizosphere inhabitants (‘non-target’ is used to distinguish those microorganisms not conventionally known as targets) have not been thoroughly investigated. Thus, this review first summarizes the conventional roles of flavonoids as nod gene inducers, phytoalexins and allelochemicals before exploring questions concerning ‘non-target’ impacts. We hypothesize that flavonoids act to shape rhizosphere microbial community structure because they represent a potential source of carbon and toxicity and that they impact on rhizosphere function, for example, by accelerating the biodegradation of xenobiotics. We also examine the reverse question, ‘how do rhizosphere microbial communities impact on flavonoid signals?’ The presence of microorganisms undoubtedly influences the quality and quantity of flavonoids present in the rhizosphere, both through modification of root exudation patterns and microbial catabolism of exudates. Microbial alteration and attenuation of flavonoid signals may have ecological consequences for below-ground plant-microbe and plant–plant interaction. We have a lack of knowledge concerning the composition, concentration and bioavailability of flavonoids actually experienced by microbes in an intact rhizosphere, but this may be addressed through advances in microspectroscopic and biosensor techniques. Through the use of plant mutants defective in flavonoid biosynthesis, we may also start to address the question of the significance of flavonoids in shaping rhizosphere community structure and function.


The rhizosphere is a unique zone around a root characterized by complex biological interactions, many involving microbes. Influences of the plant root that define the rhizosphere are numerous but perhaps chief among these is the input of organic carbon compounds to the soil through exudation by active roots or root senescence and sloughing; the process of rhizodeposition. For mycorrhizal plants, mycorrhizal colonization is the rule rather than the exception in natural ecosystems (Smith and Read, 1997); the term mycorrhizosphere is more accurately used to describe the zone of soil influenced by both the colonized root and the extraradical hyphae of the mycorrhizal fungus (Johansson et al., 2004). (Mycor)rhizosphere microorganisms play a critical role in cycling of rhizodeposit carbon and in interacting biogeochemical functions (e.g. denitrification (Cheneby et al., 2004; Mounier et al., 2004) and have important plant growth promotion functions (e.g. biocontrol, hormone production, bioremediation) (Dobbelaere et al., 2003; Kuiper et al., 2004). Thus, understanding the drivers of rhizosphere microbial communities to allow rational manipulation or engineering of the rhizosphere for beneficial function is an important biotechnological aim.

Molecular studies have demonstrated reduced microbial diversity in the rhizosphere as compared with bulk soil and that the rhizosphere selects for a specific subset of genotypes from the bulk soil, and this effect is plant species specific and highly reproducible (Marilley et al., 1998; Kowalchuk et al., 2002). The reasons for the reduction in diversity in the rhizosphere and the plant–microbe interactions that determine host specificity of microbial community composition are not understood, but the general assumption is that roots, through rhizodeposition, are able to regulate the soil microbial community in their immediate surroundings (Walker et al., 2003).

Rhizodeposits are chemically diverse, with components ranging from simple sugar and organic acid monomers to polymeric lignocellulose containing root debris (Uren, 2001; Walker et al., 2003). The diversity of rhizodeposition not only sustains multitrophic rhizosphere foodwebs (Phillips et al., 2003), but also mediates chemical communications which include signal traffic between roots of competing plants and between roots and beneficial or detrimental rhizosphere-dwelling microorganisms. The existence of such diverse below-ground communications has been likened to an ‘information superhighway’ (Bais et al., 2004), and conversations between beneficial rhizobacteria and plant roots as a ‘love parade beneath our feet’ (Somers et al., 2004).

We have a reasonable understanding of how some root-derived chemical signals are perceived by some of their target rhizosphere microorganisms. A good example is the molecular integration of legume flavonoid signals by compatible rhizobia during the initiation of nitrogen fixing symbiosis (see Initiation and maintenance of symbiosis). Nevertheless, it is likely that we are ignorant of many other impacts of rhizodeposit signals on rhizosphere ecology. To begin with, a full inventory of the diversity of rhizodeposition has not been carried out, so all the possible compounds that could be acting as chemical signals are not known. Secondly, while signal perception by target microorganisms may be quite well characterized (as in the rhizobia example), potential impacts on so-called ‘non-target’ rhizosphere inhabitants (we use the term ‘non-target’ to distinguish those microorganisms not conventially known as signal targets) have not been thoroughly investigated. Thus, questions such as, do ‘non-target’ species ‘eaves-drop’ on rhizodeposit-mediated plant-microbe and plant–plant conversations, have not been addressed. If ‘eaves-dropping’ does take place, how are the chemical signals integrated by these ‘non-target’ microorganisms and what is the impact on rhizosphere community structure and function? Conversely, how does rhizosphere microbial ecology impact on chemical conversations? Is there potential for microbial attenuation of rhizodeposit messages and modification of their meaning? This review will critically explore these ‘non-target’ questions by focusing on the flavonoid class of plant metabolites. It will focus on flavonoids because, out of the myriad of chemicals produced via root exudation, flavonoids are known to play a multifunctional role in the rhizosphere. One of the most familiar roles, as alluded to above, is as signals in the establishment of symbiosis. In addition, flavonoids are also agents in plant defence against pathogens and in allelopathic interactions. Accordingly, flavonoids as nod gene inducers/inhibitors, phytoalexins/phytoanticipins, and allelochemicals and their effects on target organisms will be discussed. In subsequent sections, the emphasis will be more speculative; we will discuss how flavonoid signals and agents may impact on, be perceived by and modified by ‘non-target’ rhizosphere microorganisms.


Flavonoids are a diverse class of natural compounds produced as a result of plant secondary metabolism. They are polyaromatic compounds with a 15-carbon skeleton and can be divided in to subclasses depending on their structure. Figure 1 summarizes the biosynthetic relationships between the major flavonoid classes. Flavonoids are products of the central phenylpropanoid pathway; the first committed step to the flavonoid branch is provided by the action of chalcone synthase (CHS, Fig. 1) which catalyses the condensation of 4-coumaroyl CoA (a phenylpropanoid pathway product) and three molecules of malonyl CoA to form a chalcone flavonoid precursor. The generic flavonoid class is ubiquitous in higher plants, but specific structures may be peculiar to certain plant families. For example, the isoflavonoid skeleton is restricted to the Papilionoidae subfamily of the Leguminosae (Dixon et al., 2002). After biosynthesis, flavonoids are generally stored in a sugar-conjugated form (glycosides, e.g. genistin in Fig. 1) in plant vacuoles (Aoki et al., 2000).

Figure 1.

Summarized flavonoid branch of the phenylpropanoid biosynthetic pathway leading to flavonoid and isoflavonoid structures (for more details see Dixon and Steele, 1999; Aoki et al., 2000; Dixon et al., 2002). The enzyme abbreviations are: CHS, chalcone synthase; CHI, chalcone isomerase; IFS, isoflavone synthase; IOMT, isoflavone O-methyltransferase; I2 ′H, isoflavone 2′ hydroxylase; IFR, isoflavone reductase; VR, vestitone reductase; DMID, 7,2′-dihydroxy, 4′-methoxyisoflavanol dehydratase; GT, glycosyltransferase; FS I & II, flavone synthase I and II; F3βH, flavanone 3β hydroxylase; F3′H, flavonoid 3′-hydroxylase; F3′5′H, flavonoid 3′,5′ hydroxylase; ANS, anthocyanidin synthase. Note the isoflavonoid branch, of which the first biosynthetic step is catalysed by IFS, is structurally characterized by having the phenyl group attached to C3 rather than C2. Adapted from Dixon and Steele (1999); Dixon et al. (2002); Bowles et al. (2006).

Conventional roles and integration by target organisms

Initiation and maintenance of symbiosis.  In legumes, flavonoids are key signals in initiation of nodule formation in the nitrogen fixing symbiosis through acting as inducing agents of rhizobial nodulation and nodulation-related genes (Broughton et al., 2000; Cooper, 2004). Legume hosts exude flavonoids continuously, but concentrations in the rhizosphere increase significantly in the presence of compatible rhizobial strains (Schmidt et al., 1994; Zuanazzi et al., 1998). Flavonoid structures interact with rhizobial NodD proteins to activate transcription of nodulation genes encoding biosynthesis of lipo-chito-oligosaccharide Nod factors which elicit deformation of plant root hairs and assist rhizobial entry via infection threads. Successful infection thread development depends probably on rhizobial production of extracellular polysaccharides and proteins, the secretion of which may also be induced by flavonoid structures (Broughton et al., 2000). The specificity of signalling interactions involved have been reviewed in detail (e.g. Broughton et al., 2000; 2003). A wide variety of flavonoids (chalcones, flavanones, isoflavones, flavonols; Fig. 1) have been shown to have nod gene inducing activity in different legume/rhizobia interactions (Aoki et al., 2000). There is also evidence that flavonoids are involved in regulation of nodulation in actinorrhizal associations; where nitrogen-fixing nodules are formed as a result of colonization by an actinomycete, Frankia (Benoit and Berry, 1997; Hughes et al., 1999; Hocher et al., 2006).

Increasingly, it is being appreciated that flavonoids also form part of a cluster of signals that are exchanged between arbuscular mycorrhizal fungi (AMF) and their host plants at all stages of the symbiosis: presymbiotic, colonization and symbiotic. Numerous studies (reviewed by Larose et al., 2002) have shown an effect of flavonoids on spore germination, growth of hyphae and the extent of colonization. For example, presymbiotically, the flavonol quercetin stimulates AMF spore germ tube growth and hyphal branching (Gianinazzi-Pearson et al., 1989; Tsai and Phillips, 1991; Becard et al., 1992) and recent research has demonstrated how the effects of flavonoids on spore germination and other presymbiotic growth of AMF can be compound and genus specific (Scervino et al., 2005a). Addition of isoflavonoids (formononetin and biochanin A) to white clover roots has also been shown to stimulate colonization by a Glomus sp. (Siqueira et al., 1991). Similarly, exogenous treatment of tomato roots with flavonoids altered the number of penetration structures and colonization by AMF, and the response was flavonoid and AMF specific (Scervino et al., 2005b). However, despite this evidence for the role of flavonoids in the stimulation of AMF growth, other research suggests they may not be absolutely essential for hyphal growth from spores (Becard et al., 1995).

A comparative study of flavonoids in Trifolium repens has shown that infection with a mycorrhizal symbiont (Glomus intraradices) can significantly alter the composition of flavonoids accumulated in roots (Ponce et al., 2004), and, in studies with Medicago sativa, the roots began to accumulate flavonoids before colonization took place, indicating elicitation by an AMF-derived signal (Volpin et al., 1994; Larose et al., 2002). Following this, patterns of flavonoid accumulation by Glomus mosseae-colonized M. sativa roots (Larose et al., 2002) and Glomus versiforme-colonized Medicago truncatula roots (Harrison and Dixon, 1993) were altered compared with non-colonized roots throughout the time-course of colonization: from formation of appressoria and intraradical arbuscules to degradation of fungal structures. Alongside this metabolite evidence, transcript analysis has demonstrated changes in the level of expression of genes of the flavonoid biosynthetic pathway (e.g. chalcone synthase, isoflavone reductase; Fig. 1) both during the initial period of contact between AMF and plant host (Bonanomi et al., 2001; Liu et al., 2003) and during colonization (Harrison and Dixon, 1993). Another interesting finding is that the patterns of flavonoid accumulation also depend upon the species and genus of AMF colonist, indicating AMF specificity (Larose et al., 2002). As flavonoids also have roles as compounds in plant defence (see Defence against plant pathogens), it is suggested that this alteration of flavonoid profiles in response to AMF colonization may be a result of initiation of a general plant defence response (which is later suppressed) (Volpin et al., 1994). However, disaggregation of the precise involvement of flavonoids in different stages is difficult. The molecular basis of signalling in AM symbiosis needs to be elucidated in order to establish whether altered flavonoid profiles in AM-plants are merely a side-effect of AMF colonization or whether they are a key signal in initiation and maintenance of the symbiosis.

Defence against plant pathogens.  Phytoalexins and phytoanticipins can be defined as low molecular weight antimicrobial compounds that are synthesized by plants. A division can be made between the phytoalexins that are formed de novo by a plant in response to pathogen attack and phytoanticipins that are preformed (often chemically identical compounds) and stored in plant cells (Dakora and Phillips, 1996). A significant role of the isoflavonoid class of flavonoids as phytoalexins and phytoanticipins in disease response, in particular in legumes, has been postulated because of their broad spectrum in vitro antimicrobial activity (Dixon et al., 2002) and correlative evidence that isoflavonoids are present in plant tissue, or accumulate in infected tissues in response to pathogen attack, at concentrations shown to be antimicrobial in vitro (Dakora and Phillips, 1996; Aoki et al., 2000). Simple isoflavone compounds, such as daidzein, glycitein and formononetin glycosides are accumulated constitutively by many legume species (Dakora and Phillips, 1996) and the corresponding aglycones are inhibitory to growth of microbial pathogens (VanEtten, 1976; Kramer et al., 1984) and thus can be classified as phytoanticipins. Perhaps the best known isoflavonoid legume phytoalexins belong to the pterocarpan class (Aoki et al., 2000); for example, medicarpin (Fig. 1). Pterocarpans have proven microbial toxicity (Blount et al., 1992; Lozovaya et al., 2004) and are produced in roots either constitutively (Lopez-Meyer and Paiva, 2002) or after induction by pathogens (Lozovaya et al., 2004) or endogenous elicitors (Armero et al., 2001). The mode of antibiotic activity of pterocarpans may involve a non-specific action involving disruption of membrane structural integrity (Weinstein and Albersheim, 1983) or a site-specific inhibition of enzymes involved in electron transport (Boydston et al., 1983).

Allelopathy.  Allelopathy is derived from the Greek allelon‘of each other’ and pathos‘to suffer’. It is a term applied to describe the chemical inhibition of one plant species by another. The chemical inhibitors, referred to as allelochemicals, may be produced in above ground plant tissues and reach the rhizosphere soil through leaching or they may be deposited directly in root exudates (Weir et al., 2004). Perhaps the best-known allelochemical is sorgoleone, a lipophilic benzoquinone produced by Sorghum bicolour. Sorgoleone acts as an allelochemical by inhibition of photosystem II (Weir et al., 2004).

Of relevance here are the (+) and (–) isomers of catechin (Fig. 1) and 7,8-benzoflavone, flavonoids that have suggested roles in allelopathic interactions. (+/–)-Catechin is secreted by Centaurea maculosa (spotted knapweed) roots but (–)-catechin specifically inhibits seed germination and is a potent herbicide (Bais et al., 2002). Although the allelochemical mechanism of (–)-catechin has not been as thoroughly investigated as for sorgoleone, it is suggested that (–)-catechin inhibits seed germination by disrupting mitochondrial respiration (Weir et al., 2004) eliciting the generation of reactive oxygen species in susceptible plants leading to cell death (Bais et al., 2003). It is thought that secretion of (–)-catechin contributes to the invasive behaviour of C. maculosa in N. America (Bais et al., 2002; 2003; Fitter, 2003). However, other recent research indicates that catechin may not be as persistent in soils as previously thought, which suggests that its ecological role in the invasiveness of C. maculosa may have been overestimated (Blair et al., 2005). Allelochemistry may also contribute to the invasive behaviour of Acroptilon repens (Russian knapweed) which produces rhizotoxic 7,8-benzoflavone in its root exudates (Stermitz et al., 2003).

Potential for flavonoid–‘non-target’ microbe interaction in the rhizosphere

Evidence suggests that flavonoids are synthesized in the cytosol (Lopez-Meyer and Paiva, 2002), but are stored in vacuoles (Mackenbrock et al., 1992). However, in addition to accumulation in organized plant tissues, they also find their way to the external rhizosphere through exudation (Armero et al., 2001). The potential effects of exudation of pterocarpan phytoalexins on root interactions with non-pathogenic (i.e. ‘non-target’) microbes have been noted (Lozovaya et al., 2004). Not only do isoflavonoids have likely significance in the rhizosphere from root exudates, but also from sloughed root border cells that are released in large quantity (Hawes et al., 2003). It can also be hypothesized that, flavonoids and their glycosylated conjugates stored in vacuoles will eventually also be released to the rhizosphere during root senescence.

Soil concentrations of up to 390 μg g−1 of racemic catechin have been reported for areas invaded by C. maculosa (Bais et al., 2002). One other study has noted, qualitatively, the presence of 4′,7-dihydroxyflavanone, 4′,7-dihydroxyflavone and medicarpin glycoside in alfalfa rhizosphere soil (Phillips et al., 1997). In general, however, information regarding the actual concentrations of individual flavonoids in intact rhizosphere soil, likely to be experienced by rhizosphere microorganisms, is lacking. Alfalfa root concentrations of 98, 168 and 340 nmol g−1 fresh weight have been reported for daidzein, formononetin and formononetin malonyl glucoside respectively [extrapolated from data in Harrison and Dixon (1993) and Larose and colleagues (2002)]. These values fall within the lower range of those reported for other chemical classes (e.g. 15–19 700 nmol g−1 for the organic acid citrate; Jones, 1998) with implicated roles in rhizosphere processes. However, to our knowledge, there are no data on what proportion of flavonoids stored in roots is deposited to rhizosphere soil and at what rate. Therefore, it also is difficult to gain an idea from the literature what fraction of total rhizodeposition (a summary of rates is given in Toal and colleagues, 2000) is represented by flavonoids. Quantitative physiochemical property data (e.g. aqueous solubilities, octanol-water partition coefficients) is also difficult to find, but the glycoside forms are more water soluble than the free aglycone. For example, genistein (Fig. 1) is described as, ‘practically insoluble in water’, whereas genistin, the glucoside (Fig. 1), is ‘sparingly soluble in water’ (O'Neil et al., 1996). Thus, conjugated forms are expected to be less adsorbed to the soil matrix, more mobile and therefore more bioavailable than the free aglycone form. However, existence in the conjugated form is likely to be short-lived, because hydrolysis by glycosidase enzymes of both plant and microbial origin has been shown to occur shortly after release in exudate (Hartwig and Phillips, 1991), although, one study reports the detection of formononetin-7-O-glycoside in rhizosphere soil which may suggest that some glycosides have unexpected stability (Leon-Barrios et al., 1993). The fact that legumes growing in soil become nodulated with rhizobial symbionts can be taken as empirical evidence that flavonoids (a key signal in initiation of nodulation), whether in free or conjugated form, are bioavailable to microorganisms in the rhizosphere and that there is great potential for microbe–flavonoid interaction. The speculation here is with regard to the extent of the influence of flavonoid signals, i.e. whether aglycone and conjugated flavonoids have sufficient mobility to diffuse and be bioavailable in the outer extremities of the rhizosphere, or whether their impact is confined to the near-rhizosphere and rhizoplane. Much more data are required concerning flavonoid bioavailability and soil spatial and temporal distribution in relation to plant root architecture and rhizosphere microorganisms.

Impact of flavonoid signals on ‘non-target’ microbial communities

Selection for catabolic and resistance phenotypes

Flavonoids represent, to those rhizosphere microorganisms in possession of appropriate catabolic enzymes, a carbon rich source. Several studies (summarized in Table 1) have quantified and characterized aerobic flavonoid biodegradation for a number of bacterial species. From examination of those studies which attempt to clarify the pathway, a common flavonoid biodegradative route can be identified (reviewed in detail by Cooper, 2004). The upper biodegradative pathway for quercetin identified in plant growth promoting Pseudomonas putida PML2 (Pillai and Swarup, 2002) is shown as an example (Fig. 2). In the initial step there is a double dehydroxylation to produce naringenin, then the pathway proceeds via multiple C-ring fission, and transient chalcone formation, to produce conserved A and B ring hydroxylated aromatics, in the present example, phloroglucinol and 3,4-dihydroxycinnamic acid respectively. Generically, the structure of conserved A and B ring intermediates depends on initial flavonoid structure. For A ring products, flavonoids with an OH substitution at carbon 5 tend to yield resorcinol, whereas those with OH at both C5 and C7 yield phloroglucinol (Cooper, 2004). Phenylacetic acid, substituted cinnamic acids and protocatechuic acid have also been reported as B ring products (Rao and Cooper, 1994; Pillai and Swarup, 2002). Mono-aromatic A and B ring products most likely serve as substrates for, and are mineralized by, the β-ketoadipate pathway which is widely distributed in soil bacteria (Harwood and Parales, 1996). Thus, any bacterium that can initiate flavonoid biodegradation as far as the production of A and B ring products (i.e. the upper pathway, as shown in Fig. 2) is likely to possess the lower pathway enabling it to utilize flavonoid carbon for energy and growth. Given that flavonoid compounds are present in the rhizosphere (see Potential for flavonoid–‘non-target’ microbe interaction in the rhizosphere), we suggest the ability to exploit the flavonoid resource will have selective value in plant–microbe interactions. To our knowledge, the contribution of flavonoid catabolism to rhizosphere competence has not been widely investigated, but, Guntli and colleagues (1999) report an interesting example where the ability to catabolize another class of (non-flavonoid) plant secondary metabolites, the calystegines, confers a competitive advantage to rhizosphere colonization by Sinorhizobium meliloti Rm41.

Table 1.  Aerobic flavonoid biodegradation by bacteria: a summary of the literature.
Bacterial species/strainFlavonoidSummary of findingsLiterature reference
Acinetobacter calcaoceticus
MTC 127
(+)-CatechinUsed as a sole carbon source and mineralized via protocatechuic acid (PCA) and phloroglucinol carboxylic acid (PGCA) intermediates.Arunachalam et al. (2003)
Bradyrhizobium japonicumCatechinCleaved to produce PGCA and PCA. Mono-aromatics were further metabolized via ortho-ring cleavage. Degradative pathway was inducible.Hopper and Mahadevan (1997)
Actinobacterial strain isolated
from roots of Cicer arietium
FormononetinThe strain was also able to assimilate daidzein, quercetin, luteolin and various other isoflavonoids and flavonoids hydroxy-substituted at the 3-, 5-, 7-, 3′- or 4′-positionBarz (1970)
Rhizobium loti NZP2042 and
LC22, Bradyrhizobium sp.
CC814s and CC829
QuercetinCatabolized to produce PCA and phloroglucinolRao and colleagues (1991)
Bradyrhizobium japonicum
USDA 110spc4, Rhizobium
fredii HH103, Rhizobium sp.
Daidzein and
Biodegradation proceeded via closed C ring
modification followed by C ring fission and production of A (phloroglucinol, PGCA, resorcinol) and B (p-coumaric acid, p-hydroxybenzioc acid) ring products
Rao and Cooper (1995)
Various rhizobial strainsNaringenin, quercetin,
luteolin, genistein,
daidzein, apigenin
Biodegradation proceeded via C ring cleavage and
production of A and B ring products
Rao and Cooper (1994)
Pseudomonas putida
QuercetinBiodegradation via A ring cleavageRao and Cooper (1994)
tumefaciens C6-6
NaringeninNon-specific ring fission, no conserved A or B ring productsRao and Cooper (1994)
Pseudomonas putida PML2QuercetinUtilization as a sole carbon source via dehydroxylation to naringenin and formation of A (phloroglucinol) and B (3,4-dihydroxycinnamic acid) intermediatesPillai and Swarup (2002)
Figure 2.

Quercetin degradation pathway in P. putida strain PML2 (Pillai and Swarup, 2002).

In addition to the conventional antimicrobial roles of pterocarpan phytoalexins (see Defence against plant pathogens), other flavonoid structures (e.g. catechin, Veluri et al., 2004) have reported microbial toxicity and may be present in exudates in toxic concentrations (Bais et al., 2002). Dakora and Phillips (1996) highlight the probability that ‘non-target’ beneficial bacteria and fungi encounter toxic levels of flavonoids in the rhizosphere. Given that the rhizosphere is actually quite a chemically stressful microbial environment due to the presence of toxic concentrations of flavonoids, it follows that there will be selection for microbes possessing mechanisms to detoxify, or avoid the toxicity of, the rhizosphere environment.

Catabolism may be a potential detoxification mechanism, and several plant pathogenic fungi employ enzymatic degradation to non-toxic products (vanEtten et al., 2001), however, many root associated microbes have evolved an inducible resistance mechanism that probably does not depend on decomposition but on permeability changes in the outer membrane, or active exclusion (Parniske et al., 1991; Palumbo et al., 1998; Gonzalez-Pasayo and Martinez-Romero, 2000; Burse et al., 2004). For example, Rhizobium etli possesses a multidrug efflux pump, inducible by flavonoids, that confers enhanced resistance to bean pterocarpan phytoalexins (e.g. phaseollin) (Gonzalez-Pasayo and Martinez-Romero, 2000). Agrobacterium tumefaciens also possess an isoflavonoid-inducible isoflavonoid efflux pump which contributes significantly to its rhizosphere competitiveness (Palumbo et al., 1998).

From a competition perspective it would be advantageous for rhizosphere competent microbes to possess mechanisms to combat potentially toxic plant-secreted compounds and to exploit them as carbon sources. From the literature reviewed above, it is evident that flavonoid catabolism and resistance are traits reported for microorganisms already known to be rhizosphere-competent (i.e. rhizobia, agrobacteria, plant growth-promoting pseudomonads). Experiments should test whether the ability to biodegrade flavonoids and resist their toxicity are properties more commonly possessed by microbes inhabiting the rhizosphere and rhizoplane than those inhabiting the bulk soil.

Other roles for flavonoids in rhizosphere biology

Acceleration of xenobiotic biodegradation.  Several authors have noted that the chemical structure of many plant secondary metabolites, including flavonoids, have similarity to those of xenobiotic compounds (Fig. 3) (Gilbert and Crowley, 1997; Siciliano and Germida, 1998; Dunning Hotopp and Hausinger, 2001; Shaw and Burns, 2003; Singer et al., 2003). The similarity between natural product and anthropogenic structures may have been key to the evolution of xenobiotic catabolic pathways (Singer et al., 2003; 2004) and may explain the accelerated kinetics of xenobiotic biodegradation frequently recorded in rhizosphere compared with bulk soil (reviewed by Shaw and Burns, 2003). As discussed (see Selection for catabolic and resistance phenotypes), flavonoid structures can be biodegraded via a series of steps, and intermediates funnel into the β-ketoadipate pathway. The biodegradation of aromatic xenobiotic structures also proceeds in the same way (Pazos et al., 2003). Perhaps the mechanisms behind the accelerated biodegradation of xenobitoic structures in the rhizosphere arise from positive interactions between xenobiotic and flavonoid catabolic pathways. It can be predicted that structural analogues will accelerate xenobiotic degradation, either by being utilized directly as an additional growth substrate by biodegradative microorganisms, or, by acting as an inducer or cometabolite in biodegradative pathways. Experimental evidence in support of the former comes from a study with polychlorinated biphenyls (PCBs). Leigh and colleagues (2002) have shown for example that mulberry (Morus rubra) root flavones can be used as sole carbon sources by the PCB-degrading bacterium Burkholderia sp. LB400. The outcome of experiments with the herbicide, 2,4-dichlorophenoxyacetic acid (2,4-D) and legume rhizodeposits have also suggested a role for flavonoids, or more specifically, isoflavonoids, in 2,4-D mineralization enhancement (Shaw and Burns, 2004; 2005). In the 2,4-D example, the speculated mechanism of flavonoid-enhanced 2,4-D mineralization is that products produced by flavonoid biodegradation acted to induce the 2,4-D biodegradation in root associated bacteria. More research is required to establish, unequivocally, the role of flavonoids in rhizosphere-accelerated pollution biodegradation.

Figure 3.

Rhizodeposits as xenobiotic structural analogues.

Enhancement of root colonization and growth rate.  Exogenously applied flavonoids (naringenin and daidzein) promote lateral root crack (LRC) colonization of Brassica napus, Triticum aestivum, Arabidopsis thaliana and Oryza sativa by Azorhizobium caulinodans and Herbaspirillum seropedicae (Gough et al., 1997; Webster et al., 1998; O'Callaghan et al., 2000; Jain and Gupta, 2003). The mechanisms of flavonoid stimulation of root colonization have not been fully characterized, but it is thought that the effect does not arise from the use of flavonoids as growth substrates (Gough et al., 1997; Webster et al., 1998) or flavonoid enhancement of growth rate (Gough et al., 1997), although flavonoid structures have been shown to enhance the growth rates of bacterial strains (Hartwig et al., 1991; Jain and Nainawatee, 1999). Azorhizobium caulinodans stimulation of LRC colonization additionally did not depend on flavonoid activation of NodD proteins (Webster et al., 1998). However, Webster and colleagues (1998) do not rule out the possibility that flavonoids could be inducing expression of other bacterial genes involved in root colonization. This point stimulates a broader question concerning whether or not flavonoids act as regulators of expression of genes in ‘non-target’ microbes involved in (as yet undiscovered) non-symbiotic rhizosphere processes.

Impacts of rhizosphere community structure on flavonoid-mediated communications: modification of exudation patterns and attenuation

Instead of examining how flavonoids impact microorganisms, the reverse question can be asked; ‘how do rhizosphere microbial communities impact on flavonoid signals?’ The presence of microorganisms in the rhizosphere undoubtedly influences the quality and quantity of rhizosphere flavonoids (reviewed by Cooper, 2004). This may either be through modification of root exudation patterns, or, via microbial catabolism of exuded flavonoids.

Modification of patterns of exudation

Many studies have shown that the presence of microbes or microbial components/products (e.g. Nod factors, cell wall components) elicit changes in expression of enzymes involved in the plant phenylpropanoid biosynthetic pathway. Sometimes, enhanced expression of plant biosynthetic genes has been correlated with qualitative and quantitative changes in exudation or end product accumulation in roots. Rhizobial impacts on legume flavonoid accumulation and exudation have been reviewed thoroughly by Cooper (2004) and we have discussed previously instances where AMF and plant pathogens elicit the synthesis, accumulation and exudation of flavonoids. Associative and non-symbiotic plant growth promoting bacteria may also elicit changed patterns of flavonoid exudation (Burdman et al., 1996).

Microbial catabolism

The flavonoid catabolic capability of microbial species known to be inhabitants of the rhizosphere has been demonstrated in liquid culture (see Selection for catabolic and resistance phenotypes). It therefore follows that the potential exists for microbial degradation of flavonoids present in rhizosphere soil. However, the fate of flavonoids subject to the concert of catabolic activities likely to be present in the rhizosphere has scarcely been investigated. One exception is the study by Ozan and colleagues (1997), who measured the disappearance of the isoflavonoids formononetin and biochanin A in a soil (pH 7.5) : sand (1:1) mixture, and found ∼40% and 80% biodegradation, respectively, over a 15-day period.

We argue that the likely microbial catabolism of flavonoids in rhizosphere soil will impact on the potency of the original signal produced by the plant root. At first, exuded flavonoid glycosides can be hydrolysed to more potent nod-gene inducing aglycones by rhizboial activity (Hartwig and Phillips, 1991). Once present in the aglycone form, new flavonoid structures may be produced during biodegradation of a parent flavonoid (e.g. naringenin and chalcone intermediates are produced during quercetin biodegradation (Fig. 2), before C-ring cleavage destroys the flavonoid motif altogether). Rao and Cooper (1995) have highlighted the potential consequences of microbial transformations of pre-existing flavonoid pools, namely the production of de novo flavonoids which are either nod gene inducers or repressors and may act to induce rhizobial resistance toward phytoalexins, or the formation of mono-cyclic hydroxy aromatic metabolites which could have implications for other rhizosphere interactions, for example competition for nodule occupancy, and chemotactic responses. Thus, microbial attenuation or alteration of flavonoid signals may be an important aspect of rhizosphere ecology and in the establishment of symbiosis. However, to our knowledge, the hypothesis that rhizosphere catabolism of flavonoids will result in an altered rate or extent of colonization by symbiotic rhizobia or mycorrhizal fungi has not been tested.

In pesticide microbiology, it has been observed frequently that a repeated treatment of a soil-applied pesticide is biodegraded at a faster rate and/or with a reduced lag phase than it was when first applied. This phenomenon has been termed accelerated or enhanced biodegradation (Alexander, 1994) and is probably due to an increase in the number of microorganisms capable of utilizing the pesticide as a source of carbon for growth in response to the initial application. If flavonoid biodegradation is growth-linked and exudation is high enough to support microbial growth, it follows that over time, flavonoid exudation should select for an enriched community of competent catabolic genotypes and any new input of flavonoids will be susceptible to accelerated biodegradation. The theoretical consequence of this may be that the concentration of flavonoid signals from roots of plants with an enriched rhizosphere population will be diminished rapidly through biodegradation, and their diffusion from the root surface will be reduced. Accelerated flavonoid biodegradation has not yet been demonstrated explicitly, but if it exists, it could have a number of interesting ecological consequences. Accordingly, we speculate that the rate of flavonoid biodegradation by rhizosphere populations will increase throughout the life of a plant (as the population becomes progressively enriched by continuous flavonoid exudation). Thus, enhanced biodegradation will modify the soil chemical environment for offspring, and facilitate coexistence of allelopathic plant species with their neighbours.

Questions and methodological answers

This review highlights deficiencies in knowledge: The lack of information concerning composition and concentrations of flavonoids in an intact, non-sterile rhizosphere, their bioavailability, and their spatial distribution as mapped onto root architecture; the unknown role of flavonoids as regulators of gene expression in non-target microorganisms and in shaping rhizosphere microbial community diversity and function at the community level. We believe that through methodological advances, it is now possible to start filling some of these knowledge gaps.

Raman spectroscopic imaging, applied at microscopic and macroscopic scales, offers potential to be used to gain spatial information regarding the distributions of flavonoids in the rhizosphere. As it is not necessary to apply external stains or other labels for chemical or microbial detection (Petry et al., 2003), the technique has particular potential for application to non-disturbed intact systems. Urlaub and colleagues (1998) used microscopic Fourier Transform Raman spectrometry to localize alkaloid plant secondary metabolites in various parts of plant root cross sections and micro-Raman has also been successfully employed to detect and identify single bacterial cells on fused silica surfaces (Rosch et al., 2005). Thus, the methodological foundations which would enable superimposition of flavonoid maps onto bacterial distributions in the rhizosphere exist, although analysis in intact soil systems will require careful optimization as sample fluorescence, low concentrations and susceptibility to thermal decomposition under the Raman excitation laser may obscure or destroy the Raman spectrum (Petry et al., 2003). Another spectroscopic technique, infrared microscopy, also has ecological applications and Raab and Vogel (2004) have described a ‘rhizobox-IR spectromicroscopy’ system to examine the spatial distribution of exudates in the rhizosphere of mungbean.

Complementary to the spectroscopic mapping, biosensor technology could be exploited to address questions concerning flavonoid bioavailability. Bacterial biosensors consisting of fusions between reporter genes and inducible genes have been used previously to report on the bioavailability of exudate components in the rhizosphere, for example, galactosides (Bringhurst et al., 2001), sucrose and tryptophan (Jaeger et al., 1999). Probably the best known flavonoid-inducible loci are the rhizobial nod genes where compatible flavonoids interact with NodD regulatory proteins to activate transcription of nod genes. Numerous reporter fusions have been constructed in order to study nod gene expression (e.g. (Mulligan and Long, 1985). Our initial thoughts were that these ready-made biosensors may have potential to be used to quantify the bioavailability of flavonoids in soil and rhizosphere. However, further assessment reveals that complex and multiple mechanisms operate in regulating nod gene expression (Loh and Stacey, 2003) and interaction between flavonoids may inhibit induction (Peters and Long, 1988) whereas non-flavonoid compounds may also act as inducers (Phillips et al., 1992). Therefore, flavonoid biosensors need to be carefully designed and characterized, perhaps exploiting (as yet undiscovered) other flavonoid-inducible loci, in order to gain unambiguous information. The availability of complete genome sequences now for several soil and rhizobacteria (Kunst et al., 1997; Capela et al., 2001; Puhler et al., 2004; Paulsen et al., 2005) allows the study of differential gene expression using microarrays and could be applied to identify genes induced in response to flavonoids; both to facilitate biosensor design, and to answer the question of the role flavonoids play in regulation of gene expression in non-target microbes.

Regarding the impact of flavonoids at the rhizosphere community level, the simplest way to address this question would be a model exudates experiment. In this approach, which has been used effectively to demonstrate the involvement of exudates in determining bacterial community structure (Baudoin et al., 2003; Burgmann et al., 2005), synthetic solutions composed of simple exudate constituents (i.e. sugars, amino and organic acids) are supplemented with flavonoids and comparisons made with controls consisting of simple exudate constituents minus flavonoids. An artificial root system, such as the ones devised by Pearce and colleagues (1997) or Griffiths and colleagues (1999), whereby model exudates are supplied by radial diffusion from a cylindrical ultrafiltration membrane or glass-fibre wick could be used to retain greater spatial realism and reflect the likely continuous supply of flavonoids in exudates. However, the most elegant way to address questions involving the impact of flavonoid rhizodeposition would be the comparison of plant mutants defective in total flavonoid exudation, or defective in the production of a specific class of flavonoid(s) with a non-defective wild type.

In A. thaliana, mutants for genes required for the biosynthesis of flavonoids have been characterized (reviewed by Saslowsky et al., 2000). Transparent testa (tt) mutants have been identified on the basis of having partial or complete loss of pigment in seed coats. For example, an allelic series for the chalcone synthase (Fig. 1) locus, tt4, has been described. The main flavonoids produced by wild-type Arabidopsis are kaempferol and quercetin glycosides (Saslowsky et al., 2000). The tt4 mutants, generated by chemical or irradiation mutagenesis, either produced none, or greatly reduced concentrations of these flavonoids. In a pioneering rhizosphere bioremediation metabolomic study, Narasimhan and colleagues (2003) examined the survival of Pseudomonas putida PML2, a bacterium capable of degrading both flavonoids and PCBs, in the rhizosphere of wild-type Arabidopsis, a tt4 mutant and also ttg and tt8 (regulatory, over-producing) mutants. Pseudomonas putida PML2 better colonized the roots of the wild type and over producing mutants than the tt4 null mutant. In contrast, there was no difference in the extent of colonization by a flavonoid auxotrophic mutant of P. putida PML2.

Genes encoding proteins involved in the steps of the flavonoid biosynthetic pathway have been cloned and sequenced for a number of plant species (Dixon et al., 2002). Post-genomics techniques such as TILLING (McCallum et al., 2000) may be used to speed up the discovery of plants with mutations likely to result in an altered flavonoid phenotype that are not so easily identifiable by eye as the tt mutants. In TILLING [targeting induced local lesions in genomes (McCallum et al., 2000)] the gene of interest is polymerase chain reaction (PCR)-amplified from genomic DNA extracted from the M2 generation of a chemically mutagenized population, and lesions are discovered by high throughput screening of PCR products. TILLING is applicable to any species susceptible to chemical mutagenesis and for which sequence information is known, and can therefore be applied to identify flavonoid mutants in plant species additional to Arabidopsis, including those for which genomic resources are limited (Comai and Henikoff, 2006).

Through use of genetic engineering technologies (Dixon, 2005), it is also possible to engineer branches of the flavonoid pathway into species in which they are not naturally present. For example, isoflavone synthesis has been engineered in to non-legume plants that do not normally make isoflavonoids (Jung et al., 2000; Yu et al., 2000; Liu et al., 2002). On the other hand, metabolic engineering exogenous branches of the flavonoid pathway (e.g. the isoflavonoid branch) could result in modified accumulation of endogenous pathway products due to competition for common substrates (Liu et al., 2002). For example, it should be noted that in addition to reduced production of flavonoids, the tt4 mutants, discussed above, produced slightly elevated levels of sinapate esters relative to the wild type (Saslowsky et al., 2000). These examples illustrate that knocking out or adding a particular biosynthetic step may result in perturbations to other parts of phenylpropanoid, or wider, metabolism (Dixon, 2005). These side-effects need to be fully charaterized in potential mutants used for this type of work and should be borne in mind when interpreting results. Nevertheless, the existence of numerous well-characterized Arabidopsis mutants for structural and regulatory genes in the flavonoid pathway, the promise of discovery of mutants in other plant species, together with complementary targeted engineering methodologies have exciting potential to be used to ask questions regarding the role of flavonoids in rhizosphere ecology.